diff --git a/.gitignore b/.gitignore index 42b83bf..0b6571c 100644 --- a/.gitignore +++ b/.gitignore @@ -3,4 +3,6 @@ tmp.bib .DS_Store _build/ -_deploy/ \ No newline at end of file +_deploy/ + +~$* \ No newline at end of file diff --git a/docs/lecture_03/Lecture_3.1.md b/docs/lecture_03/Lecture_3.1.md index 088996e..7df5c0d 100644 --- a/docs/lecture_03/Lecture_3.1.md +++ b/docs/lecture_03/Lecture_3.1.md @@ -1,95 +1,45 @@ --- -title: Mini-Lecture 3.1 -- Energy demands in energy systems modelling +title: Mini-Lecture 3.1 –- Residential Sectors in MUSE keywords: -- Energy demand -- Energy systems models +- Residential sector +- Sectors in MUSE authors: - Alexander J. M. Kell --- -To begin lecture 3, this mini-lecture provides an overview of energy demands within an energy system. We will cover differences in energy demands by sector, time and population classes. We will also begin to explore why these differences are important within energy models. Lecture 3 will take you through the basics for modelling energy demand in MUSE, the different options available to do so, and some specific examples +This mini-lecture introduces the concept of the residential sector -# Learning objectives - -- Learn what energy demands are in an energy modelling context -- Understand how demands can change based on different variables - -# Introduction - -Everyone needs energy for many different purposes. The form in which this energy should be delivered is dependent on the specific application. These demands for energy come from all sectors of society such as: - -- The residential sector (rural and urban) - - Cooking - - Heating - - Cooling - - Lighting - - Appliances -- Industry - - Chemical processes - - Steam production - - Heating -- Commerce - - Lighting - - Heating - - Cooling buildings - - Keeping products at low temperatures -- Transport - - Cars - - Trucks - - Buses - - Aviation - - Shipping - - Trains -- Agriculture - - Tractors - - Machinery - - Pumping water - -## Variations in daily energy demand - -These energy demands can vary on hourly, daily, weekly and monthly timescales. This mainly reflects the schedule of consumers' activities. For example, on a monthly timescale more cooling will be used in summer and more heating in winter. However, these energy demands can also vary by sector, as shown by Figure 3.1.1. - -![](assets/Figure_3.1.1.png){width=100%} - -**Figure 3.1.1:** Variations of energy demand by sector in a hypothetical example [@Taliotis2018]. - -Figure 3.1.1 shows us that the magnitude of demand varies by sector, with agricultural demand significantly lower than residential and commercial demand, in this example. The reason that the commercial and residential sectors consume more is because their activities are more energy intensive or they are simply larger. -We can also see that the daily profile of demand varies by sector. For example, in Figure 3.1.1 we can see that there is a clear evening peak in residential demand, whereas agricultural and industrial demand remains flat throughout the day. This is because agricultural and industrial demands are consistent throughout the day. This is likely because the industrial and agricultural sector operate constantly, whereas energy use in homes peaks in the evening when consumers use more electricity for cooking, lighting and appliances when they return from work or other business. - -## Sector specific demands - -The differences between sectors means that it can sometimes be important to model demands separately by each sector. This feature allows the models to consider the specific characteristics of each demand. +# Learning objectives -Within each of these sectors, the energy demand varies over time and across different types of consumers. For example, within the residential sector, demands can differ between rural and urban households, as shown in Figure 3.1.2. This can also be true between grid-connected and off-grid areas. Energy planners must ensure that energy demand is always met for all types of consumers. Therefore, it is important that the key characteristics of different demands are represented in energy models. +- Understand the role of the residential sector, its technologies and the main energy and societal challenges -![](assets/Figure_3.1.2.png){width=100%} +# Overview of the residential sector and its demands? -**Figure 3.1.2:** Variations of energy demand for the residential sector by population types [@Olaniyan2018] +Energy is used for many different reasons in the residential sector, as shown by Figure 3.1.1. This image shows the share of residential energy by service demand. We can see that energy is used for many different purposes, from heating and cooking to cleaning and ironing. This split of energy demand will vary across different countries. Figure 3.1.1 shows residential energy demand in Italy, which will differ to countries in Asia, for instance. This is largely dependent on different climates, levels of development and lifestyles. +![](assets/Figure_3.1.1.png){width=100%} -## Long-term variations in energy demands +**Figure 3.1.1:** Residential sector in Italy and the different demands [@en12112055]. (Note: DHW refers to Domestic Hot Water). -A major challenge in energy planning is that energy demands can change over time. This could be due to population growth or the creation of new industries. Figure 3.1.3 displays historical variations in energy demands. It is likely that these demands are correlated to changes in society. For example, increases in energy demand likely reflect increased industrial activity. For energy planning, we must also think about how energy demands are likely to change in the future. +The total magnitude of energy demand varies by country as a total value, but also as energy demand per capita. This is strongly dependent on the level of electricity access and availability of other fuels in the country. Residential activities can use different forms of energy. For example, cooking can be met by burning biomass, oil products, natural gas or electricity. The fuels used vary by country. -We can often forecast energy demand, such as with future projections as shown in Figure 3.1.3. These forecasts can be created using estimates of the key influencers of energy demand, such as population growth and economic activity. Future projections are often based on how energy demands have changed historically. +## Residential sector technologies -![](assets/Figure_3.1.3.png){width=100%} +Some of the key residential technologies include lamps, cooking stoves, heating and air conditioning systems, as well as other electrical appliances. Some of these technologies can only use one fuel, such as electrical appliances and air conditioning which rely on electricity. -**Figure 3.1.3:** Long-term energy consumption by source +However, in other cases multiple different fuels can be used for the same purpose. For example, heating. Heating can be met by burning biomass, natural gas, oil or electricity, for instance. These technologies have differing performance parameters. For example, electric stoves are usually much more efficient than biomass stoves. Different technological options also have different impacts on the environment and on human health. For example, the emissions from biomass can have detrimental impacts on human health, whereas electric stoves do not have emissions in the home. +It is possible to model these different options in MUSE, which allows us to gain insights into their environmental and cost implications. Modelling can allow us to model the entire system as a whole, understand the trade-offs between certain technologies and make decisions on which policies to implement. -## Capacity expansion planning +## Residential sector in MUSE -One of the key purposes of MUSE is for capacity expansion. Figure 3.1.4 displays this key issue which MUSE can address. Essentially, if total demand increases (green line) and existing system capacities are retired (blue line), how can we invest to meet the energy capacity needed to supply demand (red line)? +Within MUSE we can model different technology options. For instance, if we are to model an electric stove and a biomass stove we would have different inputs (CommIn.csv file). However, we would have the same output (CommOut.csv file) of cooking demand. We can also model an increase in efficiency of a technology by lowering the value in the CommIn.csv file. It is possible to change the efficiency over time using interpolation or a flat-forward extension as explained in mini-lecture 5.4. We can also consider the costs of investing in more energy efficient appliances by increasing the cost of these high efficiency appliances relative to the low efficiency appliances. By doing this, we can understand where and when investments in energy efficiency might be economic. -![](assets/Figure_3.1.4.png){width=100%} -**Figure 3.1.4:** Capacity expansion [@Taliotis2018] -You may notice that the red line is higher than the green line at all points. This is due to losses due to lower generating efficiencies. The gap between the red and blue lines demonstrates the required capacity expansion over time. MUSE enables us to plan such a capacity expansion whilst considering technical, economic and environmental constraints. +# Summary +In this lecture we have explored the residential sector. We considered the different demands that can reside within the residential sector and the different technologies that can be used to meet these demands. We also learnt of the difference in demands between countries and how we can model different technologies within MUSE. -# Summary -In this mini-lecture we covered the differences between energy demands in different population types, sectors and timescales. We learnt why it is important to model these differences in demand in energy systems models. We also explored how energy systems models can be used to meet a changing demand profile in the future. diff --git a/docs/lecture_03/Lecture_3.2.md b/docs/lecture_03/Lecture_3.2.md index c82aaac..2053bbb 100644 --- a/docs/lecture_03/Lecture_3.2.md +++ b/docs/lecture_03/Lecture_3.2.md @@ -1,57 +1,57 @@ --- -title: Mini-Lecture 3.2 -- Energy demands in modelling +title: Mini-Lecture 3.2 -- The transport sector in MUSE keywords: -- Energy demands -- Scenario analysis +- Transport sector +- Energy modelling authors: - Alexander J. M. Kell --- -Mini-lecture 3.2 outlines the general requirements for defining energy demands and how modelling different scenarios can help assess potential future energy demand. +This mini-lecture introduces the transport sector. We will explore the different demands and technologies within the transport sector and how we can model them within MUSE. # Learning objectives -- Understand how to define energy demands -- Understand why we need scenario analysis +- The main characteristics of the transport sector +- How these can be modelled within MUSE -# Introduction +# Overview of the transport sector and its demands -Within modelling we can break up the previously defined energy demands by sector. Electricity comes from the power sector and can be used to fulfil demand from each of the final service sectors. For example, the residential, commercial or industrial sector. +The transport sector is vital in the modern age. In the last few decades, the use of transport has increased significantly. This is as more people gain access to vehicles and develop lifestyles which rely on transport. -These sectors can have different electricity demands and needs and which can evolve over time as was seen in the last mini-lecture. We will now explore how these energy demands can be defined. +Figure 3.2.1 shows different modes of transport. As can be seen, road transport is the most used transport mode. We can also see that over 90% of fuel used in the EU transport sector is petroleum based. This is similar across the world. However, this creates challenges due to the unsustainability of fossil fuels. -## Defining energy demands +![](assets/Figure_3.2.1.jpg){width=100%} -When defining an energy demand for energy systems models, it is important to identify the following: +**Figure 3.2.1:** Transport modes and fuel share in the EU [@en13020432]. -- The energy carrier which the demand arises for. For example, electricity, gasoline for transportation or biomass for cooking. -- The sector the demand arises from. For example, residential (urban and/or rural, off- or on-grid), industrial or commercial. -- The average variability of the demand within a year. This is usually expressed using average demand profiles, which are explained in more detail later in this lecture. -- The current and expected future annual average demand. +Due to the unsustainability of fossil fuels, other solutions have been taken up with support from governments around the world. For example, cars, motorbikes and buses can be fuelled by electricity. Electric vehicles have seen large reductions in cost and improvements in performance. Electric vehicles could play an important role in overcoming the sector's challenges. -However, it is very difficult to predict future demand, and there will always be uncertainty in our predictions. Due to this it is important to model different scenarios. +It is possible to model the different technologies in MUSE, and observe competition between technologies based upon their technoeconomic parameters. -## Defining our own energy demand +## Emissions -As has just been seen, when we want to define our own energy demand, we need to identify a number of different features. Let's say, for example, that we want to define the demand for electricity in urban homes. To do this, we need to define: +The transport sector was estimated to be responsible for around 16% of global emissions in 2016 [@owidco2andothergreenhousegasemissions]. Thus, scenarios consistent with meeting global climate targets require transport sector emissions to decline rapidly. Therefore a rapid move towards sustainable technologies, such as electric vehicles is required. It is true, however, that some of the modes of transport are difficult to decarbonise. For example, it is difficult to decarbonise shipping and aviation technologies. This is because the energy density of lithium ion batteries and other technologies are lower than oil-based products. It is worth mentioning, however, that decarbonising transport is only useful if the energy sector increases its low-carbon electricity sources to supply the transport sector. -- The energy carrier for which the demand arises for. In this case it is electricity. -- The sector the demand arises from. In this example it is the residential sector, or the urban residential sector if you would like to be more specific. -- The average variability of the demand over the year. In this example we can look at daily and yearly electricity demand profiles for a residential urban area. This will tell us how the demand varies on a daily and seasonal scale. -- Current and predicted future demand. For this, we can look at an energy balance (covered in more detail later) to get data for the current and historical residential electricity demand. We can use these data as a baseline, and we could combine it with an estimate of population growth to create a future projection for the demand. +## Transport sector in MUSE -## Scenario analysis +Similar to the residential sector, we can define different technologies for the transport sector using technoeconomic parameters. For example, we can split road transport into three categories: -Within energy systems modelling, we must explore different possibilities of what could happen in the future. This is known as scenario analysis. We do this as the future is uncertain, particularly over the long-term horizon. We therefore might want to consider multiple scenarios to assess how demand could vary in the future. +- Cars +- Motorcycles +- Buses -For example, for different scenarios, key predictors of energy demand, such as population growth, economic development and energy policy can be varied across the scenarios. This would mean that each scenario has a different energy demand projection. +We can then split these three categories into their propulsion system. For instance: -Since we can not be certain of the scenario which will be the best predictor of the future, it is useful to model several scenarios and consider the implications of each of them to give useful insights for policymaking. This allows policy makers to assess which of the different policies and mixes suit their needs based upon likelihoods and risk tolerances. +- Electric vehicles +- Conventional vehicles -# Summary +We can source road transport data from national energy balances such as from the IEA, and divide this between cars, motorcycles and buses based on the split of transport by mode in the country. + +We can then run a MUSE model with the different parameters and see the effect of these different parameters on agent investment decisions. These parameters could be fuel prices, technology costs or performance parameters. We can also run the model with a carbon limit, which places a tax on carbon emissions, allowing us to work out how to pick a desirable policy depending on what we are trying to achieve. -Mini-lecture 3.2 provided an overview of energy demands, how we can define them and the details which make them up. We also explored how we can perform scenario analysis with energy demands, to understand what could happen in the future. +# Summary +In this mini-lecture we have considered the transport sector and how we can model this within MUSE. We discussed the emissions of the transport sector, and how different technologies can be used to reduce these emissions. diff --git a/docs/lecture_03/Lecture_3.3.md b/docs/lecture_03/Lecture_3.3.md index ff9993f..feb327a 100644 --- a/docs/lecture_03/Lecture_3.3.md +++ b/docs/lecture_03/Lecture_3.3.md @@ -1,62 +1,47 @@ --- -title: Mini-Lecture 3.3 -- Energy demand in MUSE +title: Mini-Lecture 3.3 -- The industrial and commercial sectors keywords: -- Energy demand -- MUSE +- Industrial sector +- Commercial sectors +- MUSE modelling authors: - Alexander J. M. Kell --- -## Short description +This mini-lecture reflects on -Following mini-lecture 3.2, this mini-lecture provides an insight into how to model service demand within MUSE. There are two possible methods to model service demand in MUSE, from user input and by correlation. In this mini-lecture we will learn what the difference is between these. +# Learning objectives -## Learning objectives +- The main characteristics of the industrial and commercial sectors +- How these can be modelled within MUSE -- Understand how to input exogenous service demand -- Understand what service demand by correlation is +# Overview of the industrial and commercial sectors -# Lecture content +Next, we will explore the industrial and commercial sectors and their respective energy demands. Figure 3.3.1. shows the energy consumption for different sectors, including industrial, by OECD (generally high-income countries) and non-OECD countries (generally low- and middle-income countries). It is evident that the industrial sector is responsible for a large share of energy consumption across the world. The industrial sector is forecast to rise in non-OECD countries significantly. We must also consider this growing expected demand in the modelling process and during policy design. -## Service Demand +![](assets/Figure_3.3.1.png){width=100%} -A service demand is a term used to describe the consumption of energy by human activity. This could be, for instance, energy for lighting or cooking in the residential sector, personal vehicles in the transportation sector or machine usage in the industrial sector. The service demand drives the entire energy system, and it influences the total amount of energy used, the location of use and the types of fuels used in the energy supply system. It also includes the characteristics of the end-use technologies that consume energy. +**Figure 3.3.1:** Energy consumption by sector, OECD and non-OECD [@world1020007]. -## Exogenous service demand +Energy is used in industry for a number of different purposes. For instance, heating and cooling, running machinery and chemical processes. These processes use a large variety of fuels and depend on the purpose, location and the technoeconomics. -Within MUSE we must set the energy demand exogenously. That means that the model does not calculate how much the service demand is. Effectively, this means that the user must make an assumption on how much electricity is consumed in, for example, the residential sector for a particular region in the model. +The commercial sector has a lower energy demand when compared to the industrial sector. This is because commercial processes, typically, are less energy intense and on smaller scales. This demand is often lighting, heating and to run office equipment and appliances. -We can change this per scenario, but these values will not change during a simulation run, even if the price for all fuels increases significantly, for instance. We are able to define the exogenous service demand by year, sector, region and timeslice. +## Industrial and commercial technologies -## Service demand by correlation. +Commercial activities use many different technologies which require energy inputs. For example, office electronics, lighting and heating systems. Many of these technologies use electricity. However, for some demands natural gas is used, for example for heating commercial buildings. -In the previous section we learnt about the exogenous service demand. That is, we can explicitly specify what the demand would be per year, sector, region and timeslice. However, it may be the case that we do not know what the electricity demand is per year, especially in the future. We may instead conclude that our electricity demand is a function of the GDP and population of a particular region, as previously discussed. +The industrial sector uses a wide range of technologies. This includes heavy machinery, boilers, heating and air conditioning. Again, a wide variety of fuels can be used for this. However, there exist a number of processes, such as steel manufacturing which requires very high temperatures. This is usually only done by burning fossil fuels, as it can be difficult to reach these high temperatures with electricity. -To accommodate such a scenario, MUSE enables us to choose a regression function that estimates service demands from GDP and population projections, which may be more predictable or have more accessible data in your case. A regression function is simply a mathematical model which fits a linear model to your data to predict what may happen in the future. +## Modelling industrial and commercial sectors in MUSE -## Sources for energy demand data +Similarly to the residential and transport sectors, we can use an energy balance [@iea_world_energy_balance] to estimate industry demands -- for instance, for industry heating demands. There are different technologies available for industrial heating. These can be grouped in a way that makes sense for your case study. However, as an example we can group these into high heat and low heat, which are modelled as separate demands. This is because generating very high temperatures requires different technologies and processes to generating low heat. -We can get publicly available energy balance data and/or demand projections from the following sources: +Again, we can group the technologies by their input fuel, such as biomass, coal, oil products or electricity with the `CommIn.csv` file. Through modelling with MUSE we can understand the emissions and economics of different technologies. -- International Energy Agency -- International Renewable Energy Agency -- United Nations Statistics -- Asia-Pacific Economic Cooperation +In addition, the commercial sector will have a different demand load profile to the residential sector. This is because, typically, the demand will follow office times for the specific region, whereas the residential sector will follow the inverse of the office schedule. -Energy balances tell us the amount that each energy commodity is used in a country or region in a given year. This is usually broken down by sector. +# Summary - -## Summary - -In this mini-lecture we introduced service demands, and the way we can input these into MUSE. The two ways we can input service demands are: -- Exogenous service demand -- Service demand by correlation - -We also learned where we can get energy data from for various countries. - -In the hands-on we will see how we can actually do this within MUSE. - - - -  +In this mini-lecture we explored the industrial and commercial sectors. We learnt the difference between these two sectors in terms of demand and the different types of technologies used in these sectors. We saw that demand for the industrial sector is expected to rise significantly in non-OECD countries. Finally, we learnt how we can model different technologies in MUSE. diff --git a/docs/lecture_03/Lecture_3.4.md b/docs/lecture_03/Lecture_3.4.md index 7a7e2e3..f6e1f97 100644 --- a/docs/lecture_03/Lecture_3.4.md +++ b/docs/lecture_03/Lecture_3.4.md @@ -1,69 +1,36 @@ --- -title: "Mini-Lecture 3.4 -- Demand examples and units" +title: Mini-Lecture 3.4 -- Sector coupling keywords: -- Infrastructure performance +- Preset sectors +- Service demand authors: - Alexander J. M. Kell --- -# Short description - -Mini-lecture 3.4 explains how we can use timeslices to approximate the real-world demand profile. We will look into the difference between power and energy. Finally, we will learn how to convert units to ensure we are consistent within MUSE. +In this mini-lecture we will investigate the role of electrification in different sectors, as well as find out what sector coupling is. # Learning objectives -- Understand how timeslices can be used in the context of demand -- Understand the difference between power and energy -- Know the units to use within MUSE and how to convert these - -# Demand profile - -Figure 3.1.5 shown an example demand profile for electricity that could be used in MUSE. In this demand profile there are 96 bars: one for each of the timeslices used in MUSE. These timeslices are split into 16 different sections – seasonal and into day and night. This is because there are four different seasons, which are split into day and night (twice). The demand profile is used to represent the proportion of demand occurring in each timeslice. - -![](assets/Figure_3.1.5.png){width=100%} - -**Figure 3.4.1:** Example demand profile for MUSE - -The chart shows us that electricity demand, in this example, is highest during the day in winter, while it is lowest during the night in spring. However, it is important to note that this is a simplification: in reality demand varies in the season and with each hour of the day. This simplification means that we model one representative day for each season, and we assume equal demand within days and nights of those seasons. - -Whilst this is a simplification, it allows us to consider the variation in demand across seasons and days without having an incredibly complex model structure. This reduces the amount of time required to run a full model relative to having timeslices for each hour and day of the year, as well as reducing the data input requirements. - -## Units - -We must ensure that during our data input process we are consistent with our units. Usually we will use the petajoules unit as this is the unit for energy for different sectors. If you were just modelling the power sector, you could use megawatt hours. +- Understand the importance of sector electrification +- Understand the need for sector coupling -## Power vs. Energy +# Sector electrification -When using energy modelling tools it is important to remember the difference between power and energy. Sometimes these terms are used interchangeably. However, there is an important difference between the two: +Electrification is becoming increasingly important in all sectors of the economy in order to achieve decarbonisation goals. As we saw earlier, electrification can be used to decarbonise the residential, transport, industrial and commercial sectors. However, some sectors are likely to be easier to electrify than other sectors. We have seen rapid progress with electric vehicles in parts of the transport sector, but sectors such as shipping and steel, which are harder to decarbonise, still have a way to go. -- Energy is the total amount of work done or the total capacity for doing work -- Power is the rate at which this energy is supplied or used. +However, different options exist for the decarbonisation of steel, for example. This can be done by retrofitting blast furnaces and adding carbon capture and storage (CCS) or scaling up hydrogen-based direct reduced iron. However, this will require innovation and further research on the key technologies, such as CCS. -Therefore, energy and power have different units. For example, energy is often measured in Joules, while power is often measured in Joules per Second (or Watts). +## Sector coupling -For example, providing the weight stays the same, lifting a weight requires the exact same amount of energy no matter how quickly we lift it. However, if we lift the weight more quickly, the power has increased. We used the same amount of energy, but over a shorter amount of time. +We have seen that we must decarbonise to meet global climate targets. However, this is not a straightforward process. A large reason for this is the inflexibility of intermittent renewable resources such as solar and wind technologies. One method of mitigating this variability and inflexibility is through sector coupling. Sector coupling is where we connect energy demands and processes across differing sectors and increase the efficiency and flexibility of energy use. This would allows us to use renewable energy for all sectors. -## Units for demand +One way this could be achieved is through power to gas conversion. When there is a high supply of renewable power, excess electricity could be used to produce hydrogen and methane. This would allow us to store this energy for later use across multiple sectors. This would enable sectors that are difficult to electrify to be based on renewable energy. -It is important that we convert our data from different sources to petajoules (PJ) when we include it in MUSE. - -Here are some example conversion factors: - -- 1 Petajoule (PJ) = 1000 Terajoules (TJ) -- 1 Petajoule (PJ) = 1,000,000 Gigajoules (GJ) -- 3.6 Petajoules (PJ) = 1 Terawatt hour (TWh) -- 0.0036 Petajoules (PJ) = 1 Gigawatt hour (GWh) - -We must ensure that we are consistent with the units we use within MUSE. +It is possible to model this sector coupling process within MUSE and to understand the tipping points which would make sector coupling possible. This could be based on the price and capacity of renewable energy, as well as the price of generating hydrogen or methane compared to the incumbent technologies. # Summary -In this lecture we have learnt the difference between power and energy. We have also learnt how to use timeslicing to speed up our model and reduce complexity. Finally, we learnt that we must use consistent units. - - - - - +In this lecture we have covered the importance of electrifying different sectors to reduce carbon emissions and meet some of the Sustainable Development Goals. We have also learnt of the importance of sector coupling to address hard to decarbonise sectors. diff --git a/docs/lecture_03/assets/Figure_3.1.1.png b/docs/lecture_03/assets/Figure_3.1.1.png index 486c983..80234e3 100644 Binary files a/docs/lecture_03/assets/Figure_3.1.1.png and b/docs/lecture_03/assets/Figure_3.1.1.png differ diff --git a/docs/lecture_06/assets/Figure_6.2.1.jpg b/docs/lecture_03/assets/Figure_3.2.1.jpg similarity index 100% rename from docs/lecture_06/assets/Figure_6.2.1.jpg rename to docs/lecture_03/assets/Figure_3.2.1.jpg diff --git a/docs/lecture_06/assets/Figure_6.3.1.png b/docs/lecture_03/assets/Figure_3.3.1.png similarity index 100% rename from docs/lecture_06/assets/Figure_6.3.1.png rename to docs/lecture_03/assets/Figure_3.3.1.png diff --git a/docs/lecture_03/bibliography.bib b/docs/lecture_03/bibliography.bib index 16a3e17..0d1365b 100644 --- a/docs/lecture_03/bibliography.bib +++ b/docs/lecture_03/bibliography.bib @@ -1,24 +1,62 @@ -@article{Taliotis2018, - abstract = {Defining final energy demands }, - author = {Taliotis, Constantinos and Gardumi, Francesco and Shivakumar, Abhishek and Sridharan, Vignesh and Ramos, Eunice and Beltramo, Agnese and Rogner, Holger and Howells, Mark}, - file = {:Users/alexanderkell/Downloads/Defining final energy demands in OSeMOSYS.pdf:pdf}, - keywords = {Demand,Energy,Energy System,Modelling,Osemosys}, - number = {January}, - title = {{Defining final energy demands in OSeMOSYS}}, - year = {2018} + +@article{en12112055, + author = {Mancini, Francesco and Lo Basso, Gianluigi and De Santoli, Livio}, + title = {Energy Use in Residential Buildings: Characterisation for Identifying Flexible Loads by Means of a Questionnaire Survey}, + journal = {Energies}, + volume = {12}, + year = {2019}, + number = {11}, + article-number = {2055}, + url = {https://www.mdpi.com/1996-1073/12/11/2055}, + issn = {1996-1073}, + abstract = {This work shows the outcomes of a research activity aimed at the energy characterization of residential users. Specifically, by data analysis related to the real energy consumption of sample buildings, the flexible loads amount has been identified so as to investigate on the opportunity to implement a demand/response (DR) program. The most meaningful input data have been collected by an on-line questionnaire created within an Excel spreadsheet allowing one to simulate and compare the calculations with the actual dwellings’ consumption; 412 questionnaires have been used as statistical sample and simulations have been performed based on single-zone dynamic model. Additionally, once the energy consumptions have been sorted by the different services, reference key performance indicators (KPIs) have been also calculated normalising those ones by people and house floor surface. From data analysis, it emerges how the Italian residential users are not very electrified. Furthermore, the flexible loads are low and, implementing minor maintenance interventions, the potential of flexibility can decrease up to 20%. For that reason, the current research can be further developed by investigating on suitable flexibility extensions as well as on the automation system requirements which is needed managing the flexible loads.}, + doi = {10.3390/en12112055} +} + + +@article{en13020432, + author = {Arens, Stefan and Schlüters, Sunke and Hanke, Benedikt and Maydell, Karsten von and Agert, Carsten}, + title = {Sustainable Residential Energy Supply: A Literature Review-Based Morphological Analysis}, + journal = {Energies}, + volume = {13}, + year = {2020}, + number = {2}, + article-number = {432}, + url = {https://www.mdpi.com/1996-1073/13/2/432}, + issn = {1996-1073}, + abstract = {The decarbonization of the energy system will bring substantial changes, from supranational regions to residential sites. This review investigates sustainable energy supply, applying a multi-sectoral approach from a residential site perspective, especially with focus on identifying crucial, plausible factors and their influence on the operation of the system. The traditionally separated mobility, heat, and electricity sectors are examined in more detail with regard to their decarbonization approaches. For every sector, available technologies, demand, and future perspectives are described. Furthermore, the benefits of cross-sectoral integration and technology coupling are examined, besides challenges to the electricity grid due to upcoming technologies, such as electric vehicles and heat pumps. Measures such as transport mode shift and improving building insulation can reduce the demand in their respective sector, although their impact remains uncertain. Moreover, flexibility measures such as Power to X or vehicle to grid couple the electricity sector to other sectors such as the mobility and heat sectors. Based on these findings, a morphological analysis is conducted. A morphological box is presented to summarize the major characteristics of the future residential energy system and investigate mutually incompatible pairs of factors. Lastly, the scenario space is further analyzed in terms of annual energy demand for a district.}, + doi = {10.3390/en13020432} +} + + +@article{world1020007, + author = {Mendoza, Daniel L. and Bianchi, Carlo and Thomas, Jermy and Ghaemi, Zahra}, + title = {Modeling County-Level Energy Demands for Commercial Buildings Due to Climate Variability with Prototype Building Simulations}, + journal = {World}, + volume = {1}, + year = {2020}, + number = {2}, + pages = {67--89}, + url = {https://www.mdpi.com/2673-4060/1/2/7}, + issn = {2673-4060}, + abstract = {The building sector accounts for nearly 40% of total primary energy consumption in the U.S. and E.U. and 20% of worldwide delivered energy consumption. Climate projections predict an increase of average annual temperatures between 1.1–5.4 °C by 2100. As urbanization is expected to continue increasing at a rapid pace, the energy consumption of buildings is likely to play a pivotal role in the overall energy budget. In this study, we used EnergyPlus building energy models to estimate the future energy demands of commercial buildings in Salt Lake County, Utah, USA, using locally-derived climate projections. We found significant variability in the energy demand profiles when simulating the study buildings under different climate scenarios, based on the energy standard the building was designed to meet, with reductions ranging from 10% to 60% in natural gas consumption for heating and increases ranging from 10% to 30% in electricity consumption for cooling. A case study, using projected 2040 building stock, showed a weighted average decrease in heating energy of 25% and an increase of 15% in cooling energy. We also found that building standards between ASHRAE 90.1-2004 and 90.1-2016 play a comparatively smaller role than variation in climate scenarios on the energy demand variability within building types. Our findings underscore the large range of potential future building energy consumption which depends on climatic conditions, as well as building types and standards.}, + doi = {10.3390/world1020007} +} + + + +@article{owidco2andothergreenhousegasemissions, + author = {Hannah Ritchie and Max Roser}, + title = {CO₂ and Greenhouse Gas Emissions}, + journal = {Our World in Data}, + year = {2020}, + note = {https://ourworldindata.org/co2-and-other-greenhouse-gas-emissions} } -@article{Olaniyan2018, - abstract = {Considering the challenge of accessing reliable household metering data in Nigeria, how can electricity consumption levels be determined? And how do disparities in electricity consumption patterns across the country affect the pursuit of sustainability, universal access and energy transition objectives? This study combined household-reported data on ownership of electrical appliances and energy expenditure with online sales records of household appliances to estimate current and future residential electricity demand in Nigeria, as well as the required generation capacity to achieve 100% electricity access, under various scenarios. Median residential electricity consumption was estimated at 18-27 kWh per capita but these estimates vary between the geographical zones with the North East and SouthWest representing extremes. Under a universal access scenario, the future electricity supply system would be expected to have installed generation capacity sufficient to meet the estimated residential demand of 85 TWh. To further understand the required infrastructure investment as a whole and the approaches that might be preferred in rural versus urban areas, the disaggregated, zone-by-zone and urban/rural data may offer more insight than a whole-of-country approach. The data obtained is useful for identifying specific transitions at the sub-national level that can minimize the required investment while maximizing households' energy access.}, - author = {Olaniyan, Kayode and McLellan, Benjamin C. and Ogata, Seiichi and Tezuka, Tetsuo}, - doi = {10.3390/su10051440}, - file = {:Users/alexanderkell/Downloads/sustainability-10-01440.pdf:pdf}, - isbn = {8180378314}, - issn = {20711050}, - journal = {Sustainability (Switzerland)}, - keywords = {Electricity access,Energy transition,Household survey,Nigeria,Sustainability}, - number = {5}, - title = {{Estimating residential electricity consumption in Nigeria to support energy transitions}}, - volume = {10}, - year = {2018} +@article{iea_world_energy_balance, + author = {IEA}, + title = {World Energy Balances: Overview}, + journal = {IEA}, + year = {2021}, + note = {https://www.iea.org/reports/world-energy-balances-overview} } diff --git a/docs/lecture_04/Lecture_4.1.md b/docs/lecture_04/Lecture_4.1.md index 156df1a..331fd77 100644 --- a/docs/lecture_04/Lecture_4.1.md +++ b/docs/lecture_04/Lecture_4.1.md @@ -1,37 +1,71 @@ --- -title: Mini-Lecture 4.1 -- Timeslicing in energy systems modelling +title: Mini-Lecture 4.1 -- Energy technologies keywords: -- Timeslices -- Energy modelling -- Energy demands +- Energy technologies +- Technoeconomic data authors: - Alexander J. M. Kell --- -This mini-lecture provides an overview of timeslicing in energy systems modelling. +This lecture will introduce the various technologies and how we can represent them within MUSE. We will also learn about the supply chains in which these technologies exist. Finally, we will learn about the key characteristics of the different technologies in the context of MUSE. # Learning objectives -- Learn why we use timeslices in energy systems models -- Understand the importance of representative days +- Understand the concepts of technologies and supply chains + +- Learn how to represent technologies in MUSE + +- Understand the key characteristics of technologies # Introduction -With energy systems models we must model how demand is met by supply. However, over the course of a year, or even over the course of 30 years we have large variations in demand and supply. For instance, the weather changes between years, seasons, and days. This all has an effect on the amount of energy that can be supplied by renewable energy sources such as solar and wind. +A technology in MUSE represents a process, or a group of processes, that: + +- Converts energy from one form into another. For example, the conversion of crude oil to oil products, oil products to electricity or electricity to heat. +- Transfers, transmits or distributes a form of energy, for example electricity transmission technologies. +- Supplies or produces a form of energy, for example oil imports or extraction, or a hydropower plant generating electricity. + +## Technology examples + +Now we will discuss specific technologies and their role in the energy system. + +Within the energy system there exists natural gas for the generation of electricity. However, we have to represent a technology which extracts natural gas in the system. We can call this technology "gas extraction", which outputs natural gas. This technology does not have any input fuel as it is a primary energy supply technology. + +A coal power plant, on the other hand, has an input of coal and an output commodity of electricity. This technology is an energy conversion technology and converts the energy in coal to electricity. + +Similarly, an oil power plant converts the energy in oil to electricity. It therefore has an input fuel of oil and an output commodity of electricity. + +It must be noted that some technologies can have more than one input or output fuel, such as a refinery with oil as the input fuel, producing both gasoline and heavy fuel oil as output fuels. + +## Parameters that define technologies + +There are three main groups of parameters that are used to define technologies. These can be seen in Figure 4.1.1 below. These include input commodities, which refer to the fuel supply to the technology. For instance, what is the input fuel, what is the price of this, and what is the availability? Crucially, it can also contain the greenhouse gas emissions associated with the fuel. + +Secondly, there is techno-economic and environmental characteristics of technologies. These include technology costs, efficiency, lifetime and availability. + +Finally, we need to define each technology's output commodity. This is the commodity which it produces, such as electricity from solar PV. Important data on output commodities includes their demand, impacts and when it is needed. + +![](assets/Figure_4.1.1.png){width=100%} + +**Figure 4.1.1:** Technology definitions by example parameters [@Taliotis2018] + + +## Representing technologies in MUSE -It is also true that this variation in demand has a large impact on the demand. In a particularly cold year, or on a particular cold day, energy demand may significantly increase as consumers use more energy for heating. The same may be true during a particularly warm period if people need energy for cooling systems. We therefore need to model this variability. +Since models are abstractions of reality, we can define technologies at different levels of abstraction depending on the nature of our energy model. Within MUSE, for instance, a single technology can represent a single power plant, or a group of similar power plants (for example, a technology could represent all coal power plants in a region if they had similar characteristics). The information provided can create a model with more or less granular data based upon the requirements of the user. It must be noted, that with increased granularity, an increase in computation time will be observed. -## Representative days +It is possible within MUSE to represent all power plants as a single technology. This is appropriate when technologies do not change significantly between power plants or extraction plants. -As you can probably imagine, matching supply and demand for every 30 minutes in a year is very costly in terms of computation time. If we must match supply and demand for every 30 minutes for 30 years (or more), we may end up with a very slow model in return for some gains in accuracy. +## Key characteristics of technologies -However, it may be the case that we do not need to model a year in such high detail. In most cases, for long-term energy systems models, we can reduce the amount of detail to significantly increase the speed of the model, without losing significant accuracy [@Kell2020]. +There are a number of different important technology characteristics that should be considered in capacity expansion planning. MUSE allows for several of these characteristics to be included. Such as: -A common approach is to model 4 days for each year. Each day corresponds to a season of the year and is split into 24 timeslices (which equates to a timeslice representing one hour). Therefore, we maintain the variability within a day, but also within seasons. We will lose some of the extremely hot or cold days, but that matters less when we're considering the long-term planning horizon. +- Variation in the availability, efficiency and costs of a technology over short and long timescales. For example, it may be the case that solar power reduces in costs over the next 30 years. If this happens, we would like to model this process and see the long-term effect on the market. +- MUSE can consider the limits on production by technology and capacity constraints. For example, there may only be a certain amount of hydro resources in a particular country, based on the number of rivers etc. It is important that MUSE takes this into account to ensure that the results are aligned to the reality in a region or country. +- Finally, the emissions associated with technologies can be captured. For example, we may want to reduce the carbon dioxide emissions of an entire system. This would allow us to compare scenarios and enable us to understand how we can reduce these emissions to reduce the impact of climate change. MUSE is also able to impose a limit on emissions through a constraint. -We do not always have to take into account entire days, to reduce the complexity further. For instance, we could have 8 days, but with only 2 timeslices (day and night). This will make the model run quickly, but may lose some detail. It is up to you, as the modeller, to find a sweet spot between accuracy and speed of computation. Various papers have been published to find this sweet spot, which you can look into in your own time [@Poncelet2017]. # Summary -In this mini-lecture we discovered why long-term energy models consider timeslices and representative days. Through this approach we are able to maintain high accuracy whilst also reducing computation time. +In this mini-lecture we have learned the importance of technologies within MUSE. We learnt that a technology can refer to a single power plant, to all coal power plants, for example. This is largely based on the requirements of individual case studies. We also learnt that technologies can also be processes, such as the extraction of natural gas. All of these different technologies come together to build an entire energy system, which MUSE is able to model. diff --git a/docs/lecture_04/Lecture_4.2.md b/docs/lecture_04/Lecture_4.2.md index a97bf69..2600b48 100644 --- a/docs/lecture_04/Lecture_4.2.md +++ b/docs/lecture_04/Lecture_4.2.md @@ -1,37 +1,66 @@ --- -title: Mini-Lecture 4.2 - Technologies by timeslice +title: Mini-Lecture 4.2 -- Technoeconomic characteristics keywords: -- Energy technologies -- Energy modelling -- Timeslices +- Technoeconomic data +- Parametrisation + authors: - Alexander J. M. Kell --- -In this mini-lecture we describe how different technologies can have different characteristics by timeslices. +This mini-lecture will describe the techno-economic data that defines technologies in MUSE. These technoeconomics are fundamental to the functioning of a good MUSE model. Most technologies can be characterised by their efficiencies, technoeconomics and inputs and outputs. This is because the technologies must be competitive against each other in an economic sense. # Learning objectives -- Understand the different characteristics of technologies by timeslice -- Understand how to characterise technologies by timeslice +- Understand the main technoeconomic parameters +- Understand how these parameters can impact investment decisions -# Introduction +# Technology costs -In the previous lecture we discovered the importance of timeslices. In this mini-lecture we will learn about how different technologies have different characteristics when it comes to timeslices, and how this can be modelled within MUSE. +In this mini-lecture we will describe the different techno-economic parameters that MUSE defines, primarily in the `Technoeconomic.csv` file found in the different sector folders. +Figure 4.2.1 displays the different cost types as defined in MUSE. The total costs are largely split into capital costs and annual costs. Capital costs, as shown by the figure, are the costs of depreciation, return on investment and other one-time fixed charges. This can include the initial costs of the technology such as construction. -# Technologies by timeslices +Then there are annual costs, which are split into variable and fixed costs. There is a distinction between these two types of costs, where fixed costs depend on the capacity of the power plant, whereas variable costs depend on the amount of energy output in a year. For instance, if a power plant does not output any electricity, it will not have to pay for fuel. However, it will still have to pay for salaries to look after the plant. -Different technologies and supply sectors have different characteristics when it comes to timeslices. For instance, solar photovoltaics do not produce any energy when it is dark (for instance, at night) and produce less in the winter. Wind, on the other hand, has a completely different profile and is largely dependent on geography. Therefore, it would make sense to provide a maximum output of the technologies at different times. For instance, it would be useful if the model limited solar output at night time in the form of a maximum utilization factor. Where utilization factor is the ratio of average amount of energy output to total possible output of an energy technology if it were to run 100% of time. +![](assets/Figure_4.2.1.png){width=100%} -However, it can be very difficult to turn off some technologies, such as a nuclear power plant. Nuclear power plants are expensive to turn on and can be unsafe if constantly varying their power. Also, their marginal cost, or the cost to produce 1MWh of electricity excluding capital costs, is usually much lower than other power plants such as gas or coal plants. It, therefore, makes sense that we place a minimum service factor, or minimum output allowed, on nuclear, to ensure their output does not fall below a certain level. +**Figure 4.2.1:** Cost types [@Taliotis2018] -Other technologies, however, such as gas power plants, can be turned on and off readily; therefore we can simply leave an average utilization factor for all the timeslices. +In MUSE, these are defined in the `cap_par`, `cap_exp`, `fix_par`, `fix_exp`, `var_par`, and `var_exp` variables where: -All of these features exist in MUSE, and during this lecture's hands-on, we will show you how to do this within MUSE. +-- `cap_par` is the capital costs, and `cap_exp` is the exponential component of this. Effectively, the `cap_exp` defines the reduction in cost due to economies of scale as the investment into this technology and its capacity increases. This should be a number between 0 and 1. +-- `fix_par` is the fixed costs, and `fix_exp` is the exponential component similar to the exponential component in `cap_exp`. +-- `var_par` is the fixed costs, and `var_exp` is the exponential component. -# Summary +The exponential component can be chosen from relevant data, but can often by difficult to find. In that case it is okay to use a number such as 1 or 0.95 as a rough indication. + +## Growth constraints + +As previously mentioned, it is important to place realistic constraints on the growth of technologies. For instance, there is only so much resource or land potential for renewable energy resources, such as offshore wind. If a country or region does not have any access to land offshore, the limit for offshore wind should be zero. On top of this, it may not be possible to grow and install technologies faster than a certain rate. For instance, there may not be enough resources, such as steel and labour, to double the capacity of wind in a certain country. + +The parameters which set these can be found in the `Technodata.csv` file and are called: + +- `MaxCapacityGrowth` +- `MaxCapacityAddition` +- `TotalCapacityLimit` + +## Other technoeconomic parameters + +Other technoeconomic parameters include the lifetime of a technology, scaling size and interest rate. A technology may become much more attractive if we are able to use it for a longer amount of time. For instance, the economics of nuclear power plants can be very sensitive to the length of time they can be used for due to their high capital costs. It is therefore important that we have good data on the lifetime of the plant. This is set by the `TechnicalLife` parameter. -In this mini-lecture we have explored the importance of characterising technologies not just by their economic data, but also by their physical characteristics. We discovered that different technologies have different outputs at different times, such as solar and wind. We also found out that nuclear power, for instance, must output a certain level to remain within a safety range. +The scaling size defines how small a single unit can be. For instance, a single nuclear power plant outputs a lot more energy than a single solar photovoltaic panel. This detail can be set by the `ScalingSize` parameter. + +The interest rate is the parameter which defines the discount rate. For instance, a technology may have a 2% return on investment, which may seem good. But it could also be possible to put the money required to build a technology into a high interest savings account and have a 4% investment. Thus the 2% return would actually reflect a loss relative to the rate of interest. This opportunity cost is the interest rate defined in the `InterestRate` parameter. + +## Inputs and outputs + +Finally, there are the input and output parameters. For a gas power plant, the input is gas and the end use is electricity. This can be set in the `Fuel` and `EndUse` parameters respectively. + + + +# Summary + In this mini-lecture we have discovered the main components which make up the Technodata sheet. We discovered the importance of properly defining the costs, lifetime and other characteristics which have a large impact on the final investment decisions. +  diff --git a/docs/lecture_04/Lecture_4.3.md b/docs/lecture_04/Lecture_4.3.md index 6ba9147..f10b2ef 100644 --- a/docs/lecture_04/Lecture_4.3.md +++ b/docs/lecture_04/Lecture_4.3.md @@ -1,39 +1,51 @@ --- -title: Mini-Lecture 4.3 - Different energy demands by timeslice +title: Mini-Lecture 4.3 -- Input and output commodities keywords: -- Energy demands -- Timeslice -- Energy modelling +- Technology efficiency +- Input commodities +- Output commodities authors: - Alexander J. M. Kell --- -This mini-lecture will continue exploring the importance of timeslices in energy modelling; however, it will have a particular focus on energy demands, and how these can change by timeslice and over the years. - -In the previous lecture we explored energy demands and timeslices. In this lecture we will have a brief recap of this, and explore how energy demand can be represented within MUSE. +In this mini-lecture we will learn about the input and output commodities within MUSE. Specifically we will learn what the `CommIn.csv` and `CommOut.csv` files do and how these relate to the energy system. # Learning objectives -- Understand how energy demand can change by timeslice -- Learn how energy demand is represented in MUSE +- To learn the importance of input and output commodities +- To learn how we can modify these commodities in MUSE -# Energy demand +# Introducing commodities -Energy demand can come in various forms. For instance, the demand we model can be for heating or cooling in the residential sector. It is the case that these demands have different characteristics. For instance, they may have different magnitudes and different technologies which serve these demands as well as they may be able to run at different times. +Input commodities are the commodities consumed by each technology. This could be coal for a coal power plant, uranium for a nuclear power plant or electricity for an electric heater. This is dependent on the technology, and some technologies can have multiple inputs. -Within MUSE, similarly to the supply sectors, we can model this time varying capability with timeslices. For instance, if we have 4 representative days which refer to the different seasons, we can model the high heating demand in winter and cooling demand in summer. On top of this we can vary these demands by time of day. +Output commodities are similar, but are the outputs of technologies. For example the output of any power plant will be electricity, and for heaters the output will be heat. Again, this is dependent on the technology, and some technologies can have multiple outputs such as combined heat and power plants. -To do this, we must edit the demand in the `preset/Residential2050Consumption.csv` sector. An example of which is shown in Figure 4.3.1. +The ratio between these two parameters is very important in MUSE and in energy modelling in general. This is because it defines the efficiency of the technology. For instance, if a coal power plant requires 1 PJ of energy stored in coal to output 0.8 PJ of electricity, the coal power plant has an efficiency of 0.8. The higher the efficiency the more economical the power plant is and the more competitive it will be when compared to different technologies. -![](assets/Figure_4.1.1.png){width=100%} +## Editing the CommIn and CommOut files -**Figure 4.3.1:** Example input for the preset sector. +Within MUSE there are two files which one should change to edit these parameters: the `CommIn.csv` and `CommOut.csv` files. These files are found within the sector folders of the case study. For instance, in the `power/CommIn.csv` or `gas/CommOut.csv` directories. -In this small example we see that there is only a demand for `heat` in the residential sector. However, this demand changes per timeslice (which are listed in the leftmost column). For instance, there is low demand for heat in timeslice 0 and a high demand for heat in timeslice 4. These timeslices refer to a single representative day, and therefore timeslice 4 has the highest demand for heat as it is in the late-evening, when people generally come home from work and turn on their radiators. +In this example we will look at the residential sectors `CommIn.csv` and `CommOut.csv` files. An example `CommIn.csv` file can be seen in the figure below: -In your models you can use datasets to disaggregate the demand into different types, or you can aggregate demand to include all gas or electricity utilised in the residential sector. This is largely dependent on the data available and the complexity of the model you would like. +![](assets/Figure_4.3.1.png){width=100%} -# Summary +**Figure 4.3.1:** CommIn file for the residential sector + +Here we see two technologies: `gasboiler` and `heatpump`. They are both in region R1 and we are specifying the characteristics for the year 2020. The `gasboiler` only requires gas, but requires 1.16 PJ, whereas the `heatpump` requires only 0.4 PJ to produce some energy. + +However, it is important to note that these figures are meaningless without the `CommOut.csv` file. We need to know how much energy does the 1.16 PJ of energy produce in the `gasboiler`? As can be seen in the figure below showing an example `CommOut.csv` file, it is convention to select an output of 1. That way we only have to vary the `CommIn.csv` to change the efficiencies consistently. + +![](assets/Figure_4.3.2.png){width=100%} -In this mini-lecture, we explored the importance of timeslicing for modelling demand in energy models. We also covered how this can be done within MUSE using the preset sector. +**Figure 4.3.1:** CommOut file for the residential sector + +Therefore, we can now conclude that the `heatpump` is much more efficient than the `gasboiler` as only 0.4 PJ are required to output 1 PJ of heat. If we divide 1 by 0.4, we get the efficiency of the `heatpump`, where 1/0.4= 2.5. Notice that the `gasboiler` also outputs carbon dioxide. It is important to take these emissions into account to have a complete understanding of the energy system. MUSE calculates these emissions endogenously. + + + +# Summary +This mini-lecture has explored the input and output commodities in MUSE. We have learnt that the `CommIn.csv` and `CommOut.csv` files relate to efficiencies when brought together in a ratio. +  diff --git a/docs/lecture_04/Lecture_4.4.md b/docs/lecture_04/Lecture_4.4.md index 53eab1f..5e34529 100644 --- a/docs/lecture_04/Lecture_4.4.md +++ b/docs/lecture_04/Lecture_4.4.md @@ -1,33 +1,51 @@ --- -title: Mini-Lecture 4.4 -- Timeslicing and climate policy +title: Mini-Lecture 4.4 -- Interpolation and future years keywords: -- Climate policy -- Timeslicing +- Interpolation +- Energy technologies authors: - Alexander J. M. Kell --- -This mini-lecture explores the relevance of timeslicing to climate policy. We will explore how different timeslicing can affect modelling results, why it is important to consider realistic timeslicing and how these can affect policy decisions. +MUSE is flexible in its approach. It requires inputs for at least the base year, but does not necessarily need more than that to project forward. In this mini-lecture we will cover how MUSE deals with missing data and how to model future years + # Learning objectives -- Understand the impact of timeslicing on modelling outputs -- Learn how timeslicing can affect policy decisions +- Learn how to model costs in multiple years +- Understand how MUSE deals with missing data +- Understand interpolation -# Timeslicing and policy +# Introduction -Timeslicing is a core component of an energy systems model as we have previously discussed. If one were to use an inappropriate number of timeslices in an energy systems model, it is likely that this would have major implications on the model outputs. +Within the input sheets you may have noticed the `Time` column. In the default example this is set to 2020. However, what happens beyond these years if we do not specify a cost, for example? Also, what happens in 2030 if we only specify a cost in 2020 and 2040? -Let's look at an example: if we were to model solar panels with an average capacity factor for the entire time horizon of the model this would assume that the solar panels can be used at night and could displace other technologies, such as gas turbines. However, in reality, solar panels contribute to the grid during the day and produce nothing at night. Therefore, we need some sort of flexibility in the system to ramp up after the sun sets. This needs to be modelled explicitly within MUSE, so to allow gas (or other technologies) to fill this gap in supply. +Within MUSE, we make some assumptions. We assume that if there are no costs input into a model beyond a certain year, that the costs remain the same. This is known as a flat-forward extension. If, for example, we input costs in 2020 and 2040, we will interpolate the values in between these years linearly. -If we take this conclusion further, it is possible to see scenarios where the intermittency of solar and wind are not modelled, and therefore we observe scenarios with a majority in solar or wind. With current technologies this is not possible, and this therefore underscores the importance of timeslicing. +An example of this is, say that the capital costs for a gas boiler is set to be 4 for a gas boiler in 2020 and 2 in 2040. We have not explicitly defined 2025, 2030 or 2035. Based on linear interpolation, MUSE will assume a value of 2.5 for 2025, 3 for 2030 (halfway between the year 2020 and 2040) and 3.5 for 2035. -If we do not use accurate timeslicing then the model outputs can skew resulting policy, and so due care must be taken for sourcing data from different geographies. +It must be noted, however, that MUSE does not allow a user to just update a single technology. For instance, if we want to specify the technology costs in 2035 for a coal power plant, we must also define the technology costs for every other technology in 2035 – although this cost need not be changed from the original value. We also do not need to define every year, however, as interpolation and a flat-forward extension can still be used. -# Summary +## Practical example + +The figure below shows a snippet of the technodata file for the residential sector. We can see that we have data parametrising the technologies in 2020. + +![](assets/Figure_4.4.1.png){width=100%} + +**Figure 4.4.1:** Technodata for residential sector + +Let's say that we want to update the capital costs (`cap_par`) for heat pumps in 2040, but do not want to update the prices for gasboilers. This is how we do it: -In this lecture we have looked into the implications of different timeslicing decisions made when creating an energy systems model. We learnt that if we do not get this right, the investments made could be skewed and unrealistic. +![](assets/Figure_4.4.2.png){width=100%} + +**Figure 4.4.2:** Updated technodata for residential sector + +Notice that we need separate rows for both `heatpump` and `gasboiler` even though we are only making a change in the `heatpump` capital cost. If we do not do this we will encounter an error. In between 2020 and 2040 we will get interpolation. + + +# Summary +In this mini-lecture we learned how to update costs in the time domain, and the assumptions MUSE makes if we do not give costs for every year. Namely, flat-forward extension and interpolation. We also learnt how to practically input these values in MUSE with the `Technodata.csv` file. diff --git a/docs/lecture_04/assets/Figure_4.3.1.png b/docs/lecture_04/assets/Figure_4.3.1.png new file mode 100644 index 0000000..33eac5d Binary files /dev/null and b/docs/lecture_04/assets/Figure_4.3.1.png differ diff --git a/docs/lecture_05/assets/Figure_5.3.2.png b/docs/lecture_04/assets/Figure_4.3.2.png similarity index 100% rename from docs/lecture_05/assets/Figure_5.3.2.png rename to docs/lecture_04/assets/Figure_4.3.2.png diff --git a/docs/lecture_05/assets/Figure_5.4.1.png b/docs/lecture_04/assets/Figure_4.4.1.png similarity index 100% rename from docs/lecture_05/assets/Figure_5.4.1.png rename to docs/lecture_04/assets/Figure_4.4.1.png diff --git a/docs/lecture_05/assets/Figure_5.4.2.png b/docs/lecture_04/assets/Figure_4.4.2.png similarity index 100% rename from docs/lecture_05/assets/Figure_5.4.2.png rename to docs/lecture_04/assets/Figure_4.4.2.png diff --git a/docs/lecture_05/assets/Picture_5.1.1.png b/docs/lecture_04/assets/Picture_4.1.1.png similarity index 100% rename from docs/lecture_05/assets/Picture_5.1.1.png rename to docs/lecture_04/assets/Picture_4.1.1.png diff --git a/docs/lecture_05/assets/Picture_5.2.1.png b/docs/lecture_04/assets/Picture_4.2.1.png similarity index 100% rename from docs/lecture_05/assets/Picture_5.2.1.png rename to docs/lecture_04/assets/Picture_4.2.1.png diff --git a/docs/lecture_04/bibliography.bib b/docs/lecture_04/bibliography.bib index a0fb1e6..f706272 100644 --- a/docs/lecture_04/bibliography.bib +++ b/docs/lecture_04/bibliography.bib @@ -1,22 +1,9 @@ -@article{Kell2020, - author = {Kell, Alexander J. M. and Forshaw, Matthew and McGough, A. Stephen}, - isbn = {9781450366717}, - journal = {The Eleventh ACM International Conference on Future Energy Systems (e-Energy'20)}, - keywords = {agent-based modelling,en-,energy market simulation,ergy models,genetic algorithm,long-term,op-,policy,simulation,validation}, - mendeley-groups = {Energy Modelling}, - title = {{Long-Term Electricity Market Agent Based Model Validation using Genetic Algorithm based Optimization}}, - year = {2020} -} -@article{Poncelet2017, - abstract = {Due to computational restrictions, energy-system optimization models (ESOMs) and generation expansion planning models (GEPMs) frequently represent intra-annual variations in demand and supply by using the data of a limited number of representative historical days. The vast majority of the current approaches to select a representative set of days relies on either simple heuristics or clustering algorithms and comparison of different approaches is restricted to different clustering algorithms. This paper contributes by: (i) proposing criteria and metrics for evaluating representativeness, (ii) providing a novel optimization-based approach to select a representative set of days and (iii) evaluating and comparing the developed approach to multiple approaches available from the literature. The developed optimization-based approach is shown to achieve more accurate results than the approaches available from the literature. As a consequence, by applying this approach to select a representative set of days, the accuracy of ESOMs/GEPMs can be improved without increasing the computational cost. The main disadvantage is that the approach is computationally costly and requires an implementation effort.}, - author = {Poncelet, Kris and Hoschle, Hanspeter and Delarue, Erik and Virag, Ana and Drhaeseleer, William}, - issn = {08858950}, - journal = {IEEE Transactions on Power Systems}, - keywords = {Energy-system planning,generation expansion planning,power system economics,power system modeling,wind energy integration}, - mendeley-groups = {Electricity Market Simulations/Selecting Representative RES Days}, - number = {3}, - pages = {1936--1948}, - title = {{Selecting representative days for capturing the implications of integrating intermittent renewables in generation expansion planning problems}}, - volume = {32}, - year = {2017} +@article{Taliotis2018, + abstract = {Defining final energy demands }, + author = {Taliotis, Constantinos and Gardumi, Francesco and Shivakumar, Abhishek and Sridharan, Vignesh and Ramos, Eunice and Beltramo, Agnese and Rogner, Holger and Howells, Mark}, + file = {:Users/alexanderkell/Downloads/Defining final energy demands in OSeMOSYS.pdf:pdf}, + keywords = {Demand,Energy,Energy System,Modelling,Osemosys}, + number = {January}, + title = {{Defining final energy demands in OSeMOSYS}}, + year = {2018} } diff --git a/docs/lecture_05/Lecture_5.1.md b/docs/lecture_05/Lecture_5.1.md index d4a2479..38964f5 100644 --- a/docs/lecture_05/Lecture_5.1.md +++ b/docs/lecture_05/Lecture_5.1.md @@ -1,71 +1,27 @@ --- -title: Mini-Lecture 5.1 -- Energy technologies +title: Mini-Lecture 5.1 -- Agents in energy systems models keywords: -- Energy technologies -- Technoeconomic data +- Agent-based model authors: - Alexander J. M. Kell --- -This lecture will introduce the various technologies and how we can represent them within MUSE. We will also learn about the supply chains in which these technologies exist. Finally, we will learn about the key characteristics of the different technologies in the context of MUSE. +In this mini-lecture we will describe the importance of agents within MUSE and also within an energy systems modelling context. # Learning objectives -- Understand the concepts of technologies and supply chains +- Understand why agents are important in an energy modelling context +- Understand how we can characterise these agents within MUSE -- Learn how to represent technologies in MUSE +# Agents overview -- Understand the key characteristics of technologies +Within real-life energy systems there are many different objectives that investors or consumers have. These objectives may differ by sector, by investor type or by proportions of the population. For instance, a certain percentage of the population may be willing to be spend more money on heating their homes than others. -# Introduction - -A technology in MUSE represents a process, or a group of processes, that: - -- Converts energy from one form into another. For example, the conversion of crude oil to oil products, oil products to electricity or electricity to heat. -- Transfers, transmits or distributes a form of energy, for example electricity transmission technologies. -- Supplies or produces a form of energy, for example oil imports or extraction, or a hydropower plant generating electricity. - -## Technology examples - -Now we will discuss specific technologies and their role in the energy system. - -Within the energy system there exists natural gas for the generation of electricity. However, we have to represent a technology which extracts natural gas in the system. We can call this technology "gas extraction", which outputs natural gas. This technology does not have any input fuel as it is a primary energy supply technology. - -A coal power plant, on the other hand, has an input of coal and an output commodity of electricity. This technology is an energy conversion technology and converts the energy in coal to electricity. - -Similarly, an oil power plant converts the energy in oil to electricity. It therefore has an input fuel of oil and an output commodity of electricity. - -It must be noted that some technologies can have more than one input or output fuel, such as a refinery with oil as the input fuel, producing both gasoline and heavy fuel oil as output fuels. - -## Parameters that define technologies - -There are three main groups of parameters that are used to define technologies. These can be seen in Figure 5.1.1 below. These include input commodities, which refer to the fuel supply to the technology. For instance, what is the input fuel, what is the price of this, and what is the availability? Crucially, it can also contain the greenhouse gas emissions associated with the fuel. - -Secondly, there is techno-economic and environmental characteristics of technologies. These include technology costs, efficiency, lifetime and availability. - -Finally, we need to define each technology's output commodity. This is the commodity which it produces, such as electricity from solar PV. Important data on output commodities includes their demand, impacts and when it is needed. - -![](assets/Figure_5.1.1.png){width=100%} - -**Figure 5.1.1:** Technology definitions by example parameters [@Taliotis2018] - - -## Representing technologies in MUSE - -Since models are abstractions of reality, we can define technologies at different levels of abstraction depending on the nature of our energy model. Within MUSE, for instance, a single technology can represent a single power plant, or a group of similar power plants (for example, a technology could represent all coal power plants in a region if they had similar characteristics). The information provided can create a model with more or less granular data based upon the requirements of the user. It must be noted, that with increased granularity, an increase in computation time will be observed. - -It is possible within MUSE to represent all power plants as a single technology. This is appropriate when technologies do not change significantly between power plants or extraction plants. - -## Key characteristics of technologies - -There are a number of different important technology characteristics that should be considered in capacity expansion planning. MUSE allows for several of these characteristics to be included. Such as: - -- Variation in the availability, efficiency and costs of a technology over short and long timescales. For example, it may be the case that solar power reduces in costs over the next 30 years. If this happens, we would like to model this process and see the long-term effect on the market. -- MUSE can consider the limits on production by technology and capacity constraints. For example, there may only be a certain amount of hydro resources in a particular country, based on the number of rivers etc. It is important that MUSE takes this into account to ensure that the results are aligned to the reality in a region or country. -- Finally, the emissions associated with technologies can be captured. For example, we may want to reduce the carbon dioxide emissions of an entire system. This would allow us to compare scenarios and enable us to understand how we can reduce these emissions to reduce the impact of climate change. MUSE is also able to impose a limit on emissions through a constraint. +It is straightforward to specify these objectives and characteristics within MUSE. For instance, you may want to split a population based upon their geospatial and economic characteristics. This could be done by, for example, splitting a population into rural and urban categories. That would provide us with two groups. However, it is possible to go further, and we may want to split the rural and urban groups into different socioeconomic demographics, such as disposable income. +Say for example, we only split the population into rural and urban. We can specify these groups as two agents within MUSE. Once we have specified the two agents, we would have to give them characteristics which differentiate them from the each other and define the proportion of the population that they make up. It must be noted, at this stage, that we do not need to have a separate agent for each individual or entity. It is perfectly fine to group and aggregate similar individuals or agents. # Summary -In this mini-lecture we have learned the importance of technologies within MUSE. We learnt that a technology can refer to a single power plant, to all coal power plants, for example. This is largely based on the requirements of individual case studies. We also learnt that technologies can also be processes, such as the extraction of natural gas. All of these different technologies come together to build an entire energy system, which MUSE is able to model. +In this mini-lecture we understood the concept of agents and how they relate to an energy modelling context. We briefly understood how we can translate these concepts into MUSE. Urban populations might have greater energy needs or rural populations may not have access to the same energy sources. Giving the model a bit more detail will allow you to make sure that the model is both more accurate, and that its projections take into account different parts of society. In the hands-on we will learn how to add a new agent. diff --git a/docs/lecture_05/Lecture_5.2.md b/docs/lecture_05/Lecture_5.2.md index aad9c9f..a43f826 100644 --- a/docs/lecture_05/Lecture_5.2.md +++ b/docs/lecture_05/Lecture_5.2.md @@ -1,66 +1,38 @@ --- -title: Mini-Lecture 5.2 -- Technoeconomic characteristics +title: Mini-Lecture 5.2 -- How to relate agent representations to the real world keywords: -- Technoeconomic data -- Parametrisation - +- Agent-based modelling +- Characterisation authors: - Alexander J. M. Kell --- -This mini-lecture will describe the techno-economic data that defines technologies in MUSE. These technoeconomics are fundamental to the functioning of a good MUSE model. Most technologies can be characterised by their efficiencies, technoeconomics and inputs and outputs. This is because the technologies must be competitive against each other in an economic sense. +In this mini-lecture we will introduce some methods to translate socioeconomic data into MUSE with a quantitative approach. # Learning objectives -- Understand the main technoeconomic parameters -- Understand how these parameters can impact investment decisions - -# Technology costs - -In this mini-lecture we will describe the different techno-economic parameters that MUSE defines, primarily in the `Technoeconomic.csv` file found in the different sector folders. - -Figure 5.2.1 displays the different cost types as defined in MUSE. The total costs are largely split into capital costs and annual costs. Capital costs, as shown by the figure, are the costs of depreciation, return on investment and other one-time fixed charges. This can include the initial costs of the technology such as construction. - -Then there are annual costs, which are split into variable and fixed costs. There is a distinction between these two types of costs, where fixed costs depend on the capacity of the power plant, whereas variable costs depend on the amount of energy output in a year. For instance, if a power plant does not output any electricity, it will not have to pay for fuel. However, it will still have to pay for salaries to look after the plant. - -![](assets/Figure_5.2.1.png){width=100%} - -**Figure 5.2.1:** Cost types [@Taliotis2018] - -In MUSE, these are defined in the `cap_par`, `cap_exp`, `fix_par`, `fix_exp`, `var_par`, and `var_exp` variables where: - --- `cap_par` is the capital costs, and `cap_exp` is the exponential component of this. Effectively, the `cap_exp` defines the reduction in cost due to economies of scale as the investment into this technology and its capacity increases. This should be a number between 0 and 1. --- `fix_par` is the fixed costs, and `fix_exp` is the exponential component similar to the exponential component in `cap_exp`. --- `var_par` is the fixed costs, and `var_exp` is the exponential component. - -The exponential component can be chosen from relevant data, but can often by difficult to find. In that case it is okay to use a number such as 1 or 0.95 as a rough indication. - -## Growth constraints - -As previously mentioned, it is important to place realistic constraints on the growth of technologies. For instance, there is only so much resource or land potential for renewable energy resources, such as offshore wind. If a country or region does not have any access to land offshore, the limit for offshore wind should be zero. On top of this, it may not be possible to grow and install technologies faster than a certain rate. For instance, there may not be enough resources, such as steel and labour, to double the capacity of wind in a certain country. - -The parameters which set these can be found in the `Technodata.csv` file and are called: - -- `MaxCapacityGrowth` -- `MaxCapacityAddition` -- `TotalCapacityLimit` - -## Other technoeconomic parameters +- Discuss surveys and socioeconomic data and how these can relate to MUSE +- Discover ways that surveys can be used in quantitative modelling -Other technoeconomic parameters include the lifetime of a technology, scaling size and interest rate. A technology may become much more attractive if we are able to use it for a longer amount of time. For instance, the economics of nuclear power plants can be very sensitive to the length of time they can be used for due to their high capital costs. It is therefore important that we have good data on the lifetime of the plant. This is set by the `TechnicalLife` parameter. +# Qualitative representation in agent-based models -The scaling size defines how small a single unit can be. For instance, a single nuclear power plant outputs a lot more energy than a single solar photovoltaic panel. This detail can be set by the `ScalingSize` parameter. +Through the use of qualitative data, such as using qualitative surveys, it is possible to gain greater insight into the different characteristics of consumers or investors. One example of how this can be done was by Moya et al. (2020). In this paper the authors explore fuel-switching investment in the long-term energy transitions of India's industry sector. They inform the modelled agents through a questionnaire that was carried out to inform MUSE. -The interest rate is the parameter which defines the discount rate. For instance, a technology may have a 2% return on investment, which may seem good. But it could also be possible to put the money required to build a technology into a high interest savings account and have a 4% investment. Thus the 2% return would actually reflect a loss relative to the rate of interest. This opportunity cost is the interest rate defined in the `InterestRate` parameter. +Some of the types of questions asked in the questionnaire to industrial companies are listed below: -## Inputs and outputs +- Geographical location +- Financial details +- Investment plans +- Type of fuels used +- Willingness to switch fuels -Finally, there are the input and output parameters. For a gas power plant, the input is gas and the end use is electricity. This can be set in the `Fuel` and `EndUse` parameters respectively. +Once these data have been collected, they can be used to find similar groups of investors and to start characterising the agents. For instance, if from the data it is clear that geographical location is an important consideration, the decision could be made to group companies by geographical region and form an agent on this basis. If the more important consideration is the investment plans, then a group can be made there. +This approach is a more than efficient method of better understanding the characteristics of agents of a system, and it can help to inform a better modelling process. The work by Moya et al. ([@Moya2020]) finds that the results represent the unique heterogeneity of fuel-switching industrial investors with distinct investment goals and limited foresight on costs. In other words, the survey results have an impact on the outcome of the energy system over the long-term. # Summary - In this mini-lecture we have discovered the main components which make up the Technodata sheet. We discovered the importance of properly defining the costs, lifetime and other characteristics which have a large impact on the final investment decisions. +In this mini-lecture we explored how surveys can be used to inform agents within MUSE. We also discovered how these results can affect the modelling outcomes of energy systems.   diff --git a/docs/lecture_05/Lecture_5.3.md b/docs/lecture_05/Lecture_5.3.md index 84079c6..bcb3bcd 100644 --- a/docs/lecture_05/Lecture_5.3.md +++ b/docs/lecture_05/Lecture_5.3.md @@ -1,51 +1,38 @@ --- -title: Mini-Lecture 5.3 -- Input and output commodities +title: Mini-Lecture 5.3 -- Agents by sector keywords: -- Technology efficiency -- Input commodities -- Output commodities +- Sectors +- Agent differentiation +- Key agent parameters authors: - Alexander J. M. Kell --- -In this mini-lecture we will learn about the input and output commodities within MUSE. Specifically we will learn what the `CommIn.csv` and `CommOut.csv` files do and how these relate to the energy system. +In this mini-lecture we will cover how agents and their characteristics can differ between sectors. We will also investigate the similarities between agents and sectors and consider the key parameters that make up agents. # Learning objectives -- To learn the importance of input and output commodities -- To learn how we can modify these commodities in MUSE +- Understand the differences between agents of different sectors +- Understand the key parameters that differentiate agents -# Introducing commodities -Input commodities are the commodities consumed by each technology. This could be coal for a coal power plant, uranium for a nuclear power plant or electricity for an electric heater. This is dependent on the technology, and some technologies can have multiple inputs. +# Agent parameters -Output commodities are similar, but are the outputs of technologies. For example the output of any power plant will be electricity, and for heaters the output will be heat. Again, this is dependent on the technology, and some technologies can have multiple outputs such as combined heat and power plants. +Different sectors may mean having agents with different characteristics. For instance, within the residential sector socioeconomic data can be used to characterise the agents. We could use wealth to characterise our agents in different geographic locations. For example we could place a constraint on the `Budget` parameter for residential users, and split these agents into different proportions. For example, we could prohibit 70% of residential users from spending more than a certain amount on heating which could affect their technology choice. The other 30% of users would form an agent that was not constricted in this way, and thus their choices may end up being differet in the model. -The ratio between these two parameters is very important in MUSE and in energy modelling in general. This is because it defines the efficiency of the technology. For instance, if a coal power plant requires 1 PJ of energy stored in coal to output 0.8 PJ of electricity, the coal power plant has an efficiency of 0.8. The higher the efficiency the more economical the power plant is and the more competitive it will be when compared to different technologies. - -## Editing the CommIn and CommOut files - -Within MUSE there are two files which one should change to edit these parameters: the `CommIn.csv` and `CommOut.csv` files. These files are found within the sector folders of the case study. For instance, in the `power/CommIn.csv` or `gas/CommOut.csv` directories. - -In this example we will look at the residential sectors `CommIn.csv` and `CommOut.csv` files. An example `CommIn.csv` file can be seen in the figure below: +Another way we could classify residential agents is through the `Maturity` parameter. This would limit investments in novel technologies until the specified technology had a certain market share. This could be informed by the innovation adoption lifecycle, as shown by Figure 5.3.1. Where, for example, innovators make up 2.5% of the population but have no `Maturity` constraints. As we work our way up the curve from innovators to laggards, this `Maturity` constraint increases. ![](assets/Figure_5.3.1.png){width=100%} -**Figure 5.3.1:** CommIn file for the residential sector - -Here we see two technologies: `gasboiler` and `heatpump`. They are both in region R1 and we are specifying the characteristics for the year 2020. The `gasboiler` only requires gas, but requires 1.16 PJ, whereas the `heatpump` requires only 0.4 PJ to produce some energy. - -However, it is important to note that these figures are meaningless without the `CommOut.csv` file. We need to know how much energy does the 1.16 PJ of energy produce in the `gasboiler`? As can be seen in the figure below showing an example `CommOut.csv` file, it is convention to select an output of 1. That way we only have to vary the `CommIn.csv` to change the efficiencies consistently. - -![](assets/Figure_5.3.2.png){width=100%} +**Figure 5.3.1:** Innovation adoption lifecycle -**Figure 5.3.1:** CommOut file for the residential sector +# Sectors -Therefore, we can now conclude that the `heatpump` is much more efficient than the `gasboiler` as only 0.4 PJ are required to output 1 PJ of heat. If we divide 1 by 0.4, we get the efficiency of the `heatpump`, where 1/0.4= 2.5. Notice that the `gasboiler` also outputs carbon dioxide. It is important to take these emissions into account to have a complete understanding of the energy system. MUSE calculates these emissions endogenously. +In this mini-lecture we have focused on the residential sector and seen the way we can characterise agents. Although these characteristics may not directly translate to the power sector, in some cases investors in the power sector can have similar characteristics. For instance, some companies are larger, and are more willing to invest their capital, reflecting a larger `Budget` parameter. Others may be less willing to invest in new technologies. The differing objectives of agents will often be the reason behind differences with other agents. For instance, some agents may only want to minimise their costs, whereas others may want to reduce their capital expenditure. It is easy to change these characteristics within MUSE to create diverse energy scenarios. # Summary -This mini-lecture has explored the input and output commodities in MUSE. We have learnt that the `CommIn.csv` and `CommOut.csv` files relate to efficiencies when brought together in a ratio. +In this mini-lecture we covered the differences between agents and the different parameters that can be used to inform these differences. We saw how the `Maturity` constraint maps to the innovation adoption lifecycle and how the `Budget` parameter can be informed by socioeconomic characteristics. These parameters lead to a large amount of possible scenarios that can be tested and run.   diff --git a/docs/lecture_05/Lecture_5.4.md b/docs/lecture_05/Lecture_5.4.md index 6e1cf13..aff51d1 100644 --- a/docs/lecture_05/Lecture_5.4.md +++ b/docs/lecture_05/Lecture_5.4.md @@ -1,52 +1,33 @@ --- -title: Mini-Lecture 5.4 -- Interpolation and future years +title: Mini-Lecture 5.4 -- Agent parameters keywords: -- Interpolation -- Energy technologies +- Agent parameters +- MUSE authors: - Alexander J. M. Kell --- -MUSE is flexible in its approach. It requires inputs for at least the base year, but does not necessarily need more than that to project forward. In this mini-lecture we will cover how MUSE deals with missing data and how to model future years - +This mini-lecture explores all the major parameters that can define agents within MUSE. # Learning objectives -- Learn how to model costs in multiple years -- Understand how MUSE deals with missing data -- Understand interpolation - -# Introduction - -Within the input sheets you may have noticed the `Time` column. In the default example this is set to 2020. However, what happens beyond these years if we do not specify a cost, for example? Also, what happens in 2030 if we only specify a cost in 2020 and 2040? - -Within MUSE, we make some assumptions. We assume that if there are no costs input into a model beyond a certain year, that the costs remain the same. This is known as a flat-forward extension. If, for example, we input costs in 2020 and 2040, we will interpolate the values in between these years linearly. - -An example of this is, say that the capital costs for a gas boiler is set to be 4 for a gas boiler in 2020 and 2 in 2040. We have not explicitly defined 2025, 2030 or 2035. Based on linear interpolation, MUSE will assume a value of 2.5 for 2025, 3 for 2030 (halfway between the year 2020 and 2040) and 3.5 for 2035. - -It must be noted, however, that MUSE does not allow a user to just update a single technology. For instance, if we want to specify the technology costs in 2035 for a coal power plant, we must also define the technology costs for every other technology in 2035 – although this cost need not be changed from the original value. We also do not need to define every year, however, as interpolation and a flat-forward extension can still be used. +- Understand the different agent parameters and their role within MUSE -## Practical example +# Overview agent parameters -The figure below shows a snippet of the technodata file for the residential sector. We can see that we have data parametrising the technologies in 2020. +Within MUSE each agent can have their own objectives. MUSE is flexible enough to allow for up to 3 objectives, which can be summed together at various weightings. To input these objectives into MUSE one would use the `Objective1`, `Objective2` and/or `Objective3` parameters and select an objective such as `comfort`, `lifetime_levelized_cost_of_energy` or `fixed_costs`. -![](assets/Figure_5.4.1.png){width=100%} +Then we would select the weight of each of the objectives using the `ObjData1`, `ObjData2`, `ObjData3` inputs. For example, if we had 3 objectives, we could make the objective of `Objective1` dominant by setting `ObjData1` to 0.5. This would mean it would make up 50% of the final objective. -**Figure 5.4.1:** Technodata for residential sector +We can edit the `SearchRule` to reduce the space of technologies that those agents are likely to consider. For example, we could fill this with `same_fuels`, or `same_enduse`. -Let's say that we want to update the capital costs (`cap_par`) for heat pumps in 2040, but do not want to update the prices for gasboilers. This is how we do it: - -![](assets/Figure_5.4.2.png){width=100%} - -**Figure 5.4.2:** Updated technodata for residential sector - -Notice that we need separate rows for both `heatpump` and `gasboiler` even though we are only making a change in the `heatpump` capital cost. If we do not do this we will encounter an error. In between 2020 and 2040 we will get interpolation. +The rest of the parameters include the parameters discussed in the previous lecture: +- `MaturityThreshold` +- `Budget` # Summary -In this mini-lecture we learned how to update costs in the time domain, and the assumptions MUSE makes if we do not give costs for every year. Namely, flat-forward extension and interpolation. We also learnt how to practically input these values in MUSE with the `Technodata.csv` file. - - +In this mini-lecture we discovered the main parameters that are used by agents within MUSE. For a full breakdown of the parameters please refer to the MUSE documentation that can be found online. diff --git a/docs/lecture_05/assets/Figure_5.3.1.png b/docs/lecture_05/assets/Figure_5.3.1.png index 33eac5d..04d9d66 100644 Binary files a/docs/lecture_05/assets/Figure_5.3.1.png and b/docs/lecture_05/assets/Figure_5.3.1.png differ diff --git a/docs/lecture_05/bibliography.bib b/docs/lecture_05/bibliography.bib index f706272..4fd8773 100644 --- a/docs/lecture_05/bibliography.bib +++ b/docs/lecture_05/bibliography.bib @@ -1,9 +1,15 @@ -@article{Taliotis2018, - abstract = {Defining final energy demands }, - author = {Taliotis, Constantinos and Gardumi, Francesco and Shivakumar, Abhishek and Sridharan, Vignesh and Ramos, Eunice and Beltramo, Agnese and Rogner, Holger and Howells, Mark}, - file = {:Users/alexanderkell/Downloads/Defining final energy demands in OSeMOSYS.pdf:pdf}, - keywords = {Demand,Energy,Energy System,Modelling,Osemosys}, - number = {January}, - title = {{Defining final energy demands in OSeMOSYS}}, - year = {2018} +@article{Moya2020, + abstract = {This paper presents the formulation and application of a novel agent-based integrated assessment approach to model the attributes, objectives and decision-making process of investors in a long-term energy transition in India's iron and steel sector. It takes empirical data from an on-site survey of 108 operating plants in Maharashtra to formulate objectives and decision-making metrics for the agent-based model and simulates possible future portfolio mixes. The studied decision drivers were capital costs, operating costs (including fuel consumption), a combination of capital and operating costs, and net present value. Where investors used a weighted combination of capital cost and operating costs, a natural gas uptake of $\sim$12PJ was obtained and the highest cumulative emissions reduction was obtained, 2 Mt CO2 in the period from 2020 to 2050. Conversely if net present value alone is used, cumulative emissions reduction in the same period was lower, 1.6 Mt CO2, and the cumulative uptake of natural gas was equal to 15PJ. Results show how the differing upfront investment cost of the technology options could cause prevalence of high-carbon fuels, particularly heavy fuel oil, in the final mix. Results also represent the unique heterogeneity of fuel-switching industrial investors with distinct investment goals and limited foresight on costs. The perception of high capital expenditures for decarbonisation represents a significant barrier to the energy transition in industry and should be addressed via effective policy making (e.g. carbon policy/price).}, + author = {Moya, Diego and Budinis, Sara and Giarola, Sara and Hawkes, Adam}, + doi = {10.1016/j.apenergy.2020.115295}, + file = {:Users/alexanderkell/Downloads/1-s2.0-S0306261920308072-main-2.pdf:pdf}, + issn = {03062619}, + journal = {Applied Energy}, + keywords = {Agent-based,Decarbonisation,Energy survey,Energy systems modelling,Investment metrics,Iron and steel}, + pages = {115295}, + publisher = {Elsevier}, + title = {{Agent-based scenarios comparison for assessing fuel-switching investment in long-term energy transitions of the India's industry sector}}, + url = {https://doi.org/10.1016/j.apenergy.2020.115295}, + volume = {274}, + year = {2020} } diff --git a/docs/lecture_06/Lecture_6.1.md b/docs/lecture_06/Lecture_6.1.md index f2cc6f7..d8ef9e7 100644 --- a/docs/lecture_06/Lecture_6.1.md +++ b/docs/lecture_06/Lecture_6.1.md @@ -1,45 +1,29 @@ --- -title: Mini-Lecture 6.1 –- Residential Sectors in MUSE +title: Mini-Lecture 6.1 -- Introduction to regions and aggregation keywords: -- Residential sector -- Sectors in MUSE +- Regions +- MUSE authors: - Alexander J. M. Kell --- -This mini-lecture introduces the concept of the residential sector - +This mini-lecture provides an overview of different regions within energy systems models and how these can be represented within MUSE. # Learning objectives -- Understand the role of the residential sector, its technologies and the main energy and societal challenges - -# Overview of the residential sector and its demands? - -Energy is used for many different reasons in the residential sector, as shown by Figure 6.1.1. This image shows the share of residential energy by service demand. We can see that energy is used for many different purposes, from heating and cooking to cleaning and ironing. This split of energy demand will vary across different countries. Figure 6.1.1 shows residential energy demand in Italy, which will differ to countries in Asia, for instance. This is largely dependent on different climates, levels of development and lifestyles. - -![](assets/Figure_6.1.1.png){width=100%} - -**Figure 6.1.1:** Residential sector in Italy and the different demands [@en12112055]. (Note: DHW refers to Domestic Hot Water). +- When to aggregate data into different regions -The total magnitude of energy demand varies by country as a total value, but also as energy demand per capita. This is strongly dependent on the level of electricity access and availability of other fuels in the country. Residential activities can use different forms of energy. For example, cooking can be met by burning biomass, oil products, natural gas or electricity. The fuels used vary by country. +# Aggregation -## Residential sector technologies +Regions within energy models play an important role. We often want to aggregate technoeconomic data from multiple regions into one. For example, the UK is made up of many different counties with different energy demands and supply. However, it could be the case that we do not have comprehensive data for each of these counties. We may, however, have plentiful data for the UK as a whole, or even for England, Scotland, Northern Ireland and Wales. We can therefore aggregate these data and make assumptions about the geographical locations of supply and demand. -Some of the key residential technologies include lamps, cooking stoves, heating and air conditioning systems, as well as other electrical appliances. Some of these technologies can only use one fuel, such as electrical appliances and air conditioning which rely on electricity. +This is an example of aggregation and can make the modelling process more straightforward, whilst losing a small amount of accuracy. This is because we do not need to model each individual power plant, demand centre or end-use sector. This means we can use aggregated data which are often easier to access. -However, in other cases multiple different fuels can be used for the same purpose. For example, heating. Heating can be met by burning biomass, natural gas, oil or electricity, for instance. These technologies have differing performance parameters. For example, electric stoves are usually much more efficient than biomass stoves. Different technological options also have different impacts on the environment and on human health. For example, the emissions from biomass can have detrimental impacts on human health, whereas electric stoves do not have emissions in the home. - -It is possible to model these different options in MUSE, which allows us to gain insights into their environmental and cost implications. Modelling can allow us to model the entire system as a whole, understand the trade-offs between certain technologies and make decisions on which policies to implement. - -## Residential sector in MUSE - -Within MUSE we can model different technology options. For instance, if we are to model an electric stove and a biomass stove we would have different inputs (CommIn.csv file). However, we would have the same output (CommOut.csv file) of cooking demand. We can also model an increase in efficiency of a technology by lowering the value in the CommIn.csv file. It is possible to change the efficiency over time using interpolation or a flat-forward extension as explained in mini-lecture 5.4. We can also consider the costs of investing in more energy efficient appliances by increasing the cost of these high efficiency appliances relative to the low efficiency appliances. By doing this, we can understand where and when investments in energy efficiency might be economic. +We can also aggregate multiple countries into regions. For example, we can merge the European continent together. This would be especially useful if we are considering a global model. However, it must be noted that we would lose significant detail by aggregating up to a supranational level. It is up to you, the model user, to consider the trade-offs between aggregation and disaggregation. For example, if you only wanted to model a single country, it would be possible to have a single region. However, if you had good access to data at the local level, you could disaggregate the data further. It does not matter whether the region is a single country, a number of counties or at a supranational level. The regions depend on your case study and the data you have access to. # Summary -In this lecture we have explored the residential sector. We considered the different demands that can reside within the residential sector and the different technologies that can be used to meet these demands. We also learnt of the difference in demands between countries and how we can model different technologies within MUSE. - +In this mini-lecture we learnt about the trade-offs between aggregation and disaggregation when defining regions. We learnt that the more aggregated the model, the less granular data are required. This can be helpful in cases where the data are not available at a local level, but available at a national level. diff --git a/docs/lecture_06/Lecture_6.2.md b/docs/lecture_06/Lecture_6.2.md index 697331d..7107df3 100644 --- a/docs/lecture_06/Lecture_6.2.md +++ b/docs/lecture_06/Lecture_6.2.md @@ -1,57 +1,32 @@ --- -title: Mini-Lecture 6.2 -- The transport sector in MUSE +title: Mini-Lecture 6.2 -- Disaggregation of regional data keywords: -- Transport sector -- Energy modelling +- Disaggregation +- Regions authors: - Alexander J. M. Kell --- -This mini-lecture introduces the transport sector. We will explore the different demands and technologies within the transport sector and how we can model them within MUSE. +This mini-lecture introduces the concept of disaggregation of regions in further detail. # Learning objectives -- The main characteristics of the transport sector -- How these can be modelled within MUSE +- When to disaggregate regional data in MUSE and energy systems models -# Overview of the transport sector and its demands -The transport sector is vital in the modern age. In the last few decades, the use of transport has increased significantly. This is as more people gain access to vehicles and develop lifestyles which rely on transport. +# Disaggregation -Figure 6.2.1 shows different modes of transport. As can be seen, road transport is the most used transport mode. We can also see that over 90% of fuel used in the EU transport sector is petroleum based. This is similar across the world. However, this creates challenges due to the unsustainability of fossil fuels. +Disaggregation of regions can often be a good way of gaining a deeper understanding of the interactions between regions. For example, if you have a lot of technoeconomic data on the locations of supply and demand, then it may make sense to disaggregate regions. This will also allow the modeller to understand where there may be issues within a specific region or country. -![](assets/Figure_6.2.1.jpg){width=100%} +An interesting example of this would be for the Southeast Asia region. Laos has a good amount of hydropower availability, whereas Thailand has more solar and wind resources. If we modelled the Southeast Asia region as a single region in MUSE, we would lose information on the potential for trade between these two countries. -**Figure 6.2.1:** Transport modes and fuel share in the EU [@en13020432]. +It is also interesting to see energy flows between regions within a country, similar to the Southeast Asian example. For example, if a country has a large demand centre in the south of the country, but large energy resources in the north, it could be interesting to disaggregate this country into those two nodes. -Due to the unsustainability of fossil fuels, other solutions have been taken up with support from governments around the world. For example, cars, motorbikes and buses can be fuelled by electricity. Electric vehicles have seen large reductions in cost and improvements in performance. Electric vehicles could play an important role in overcoming the sector's challenges. - -It is possible to model the different technologies in MUSE, and observe competition between technologies based upon their technoeconomic parameters. - -## Emissions - -The transport sector was estimated to be responsible for around 16% of global emissions in 2016 [@owidco2andothergreenhousegasemissions]. Thus, scenarios consistent with meeting global climate targets require transport sector emissions to decline rapidly. Therefore a rapid move towards sustainable technologies, such as electric vehicles is required. It is true, however, that some of the modes of transport are difficult to decarbonise. For example, it is difficult to decarbonise shipping and aviation technologies. This is because the energy density of lithium ion batteries and other technologies are lower than oil-based products. It is worth mentioning, however, that decarbonising transport is only useful if the energy sector increases its low-carbon electricity sources to supply the transport sector. - -## Transport sector in MUSE - -Similar to the residential sector, we can define different technologies for the transport sector using technoeconomic parameters. For example, we can split road transport into three categories: - -- Cars -- Motorcycles -- Buses - -We can then split these three categories into their propulsion system. For instance: - -- Electric vehicles -- Conventional vehicles - -We can source road transport data from national energy balances such as from the IEA, and divide this between cars, motorcycles and buses based on the split of transport by mode in the country. - -We can then run a MUSE model with the different parameters and see the effect of these different parameters on agent investment decisions. These parameters could be fuel prices, technology costs or performance parameters. We can also run the model with a carbon limit, which places a tax on carbon emissions, allowing us to work out how to pick a desirable policy depending on what we are trying to achieve. +Similar to the previous mini-lecture, this disaggregation is largely dependent upon your requirements and the data available to you. There is no one solution for all areas, or even for the same area and different case studies. For example, one case study may only require the modelling of a country as a single region. Another case study, however, may require the modelling of that same country by many regions. It all depends on the question you are trying to answer and the data available to you. It must be noted, that a more disaggregated case study will take longer to run in MUSE. # Summary -In this mini-lecture we have considered the transport sector and how we can model this within MUSE. We discussed the emissions of the transport sector, and how different technologies can be used to reduce these emissions. - +In this mini-lecture we explored reasons for disaggregating a case study. We discovered that disaggregation (and aggregation) of regions depends largely on the data available to you and the questions you want to answer for your case study. However, we found out that the greater the disaggregation, the more detail the model may reveal, but the longer the model will take to run. +  diff --git a/docs/lecture_06/Lecture_6.3.md b/docs/lecture_06/Lecture_6.3.md deleted file mode 100644 index 0102300..0000000 --- a/docs/lecture_06/Lecture_6.3.md +++ /dev/null @@ -1,47 +0,0 @@ ---- -title: Mini-Lecture 6.3 -- The industrial and commercial sectors -keywords: -- Industrial sector -- Commercial sectors -- MUSE modelling -authors: -- Alexander J. M. Kell ---- - -This mini-lecture reflects on - -# Learning objectives - -- The main characteristics of the industrial and commercial sectors -- How these can be modelled within MUSE - -# Overview of the industrial and commercial sectors - -Next, we will explore the industrial and commercial sectors and their respective energy demands. Figure 6.3.1. shows the energy consumption for different sectors, including industrial, by OECD (generally high-income countries) and non-OECD countries (generally low- and middle-income countries). It is evident that the industrial sector is responsible for a large share of energy consumption across the world. The industrial sector is forecast to rise in non-OECD countries significantly. We must also consider this growing expected demand in the modelling process and during policy design. - -![](assets/Figure_6.3.1.png){width=100%} - -**Figure 6.3.1:** Energy consumption by sector, OECD and non-OECD [@world1020007]. - -Energy is used in industry for a number of different purposes. For instance, heating and cooling, running machinery and chemical processes. These processes use a large variety of fuels and depend on the purpose, location and the technoeconomics. - -The commercial sector has a lower energy demand when compared to the industrial sector. This is because commercial processes, typically, are less energy intense and on smaller scales. This demand is often lighting, heating and to run office equipment and appliances. - -## Industrial and commercial technologies - -Commercial activities use many different technologies which require energy inputs. For example, office electronics, lighting and heating systems. Many of these technologies use electricity. However, for some demands natural gas is used, for example for heating commercial buildings. - -The industrial sector uses a wide range of technologies. This includes heavy machinery, boilers, heating and air conditioning. Again, a wide variety of fuels can be used for this. However, there exist a number of processes, such as steel manufacturing which requires very high temperatures. This is usually only done by burning fossil fuels, as it can be difficult to reach these high temperatures with electricity. - -## Modelling industrial and commercial sectors in MUSE - -Similarly to the residential and transport sectors, we can use an energy balance [@iea_world_energy_balance] to estimate industry demands -- for instance, for industry heating demands. There are different technologies available for industrial heating. These can be grouped in a way that makes sense for your case study. However, as an example we can group these into high heat and low heat, which are modelled as separate demands. This is because generating very high temperatures requires different technologies and processes to generating low heat. - -Again, we can group the technologies by their input fuel, such as biomass, coal, oil products or electricity with the `CommIn.csv` file. Through modelling with MUSE we can understand the emissions and economics of different technologies. - -In addition, the commercial sector will have a different demand load profile to the residential sector. This is because, typically, the demand will follow office times for the specific region, whereas the residential sector will follow the inverse of the office schedule. - -# Summary - -In this mini-lecture we explored the industrial and commercial sectors. We learnt the difference between these two sectors in terms of demand and the different types of technologies used in these sectors. We saw that demand for the industrial sector is expected to rise significantly in non-OECD countries. Finally, we learnt how we can model different technologies in MUSE. - diff --git a/docs/lecture_06/Lecture_6.4.md b/docs/lecture_06/Lecture_6.4.md deleted file mode 100644 index 470ba05..0000000 --- a/docs/lecture_06/Lecture_6.4.md +++ /dev/null @@ -1,36 +0,0 @@ ---- -title: Mini-Lecture 6.4 -- Sector coupling -keywords: -- Preset sectors -- Service demand -authors: -- Alexander J. M. Kell ---- - -In this mini-lecture we will investigate the role of electrification in different sectors, as well as find out what sector coupling is. - -# Learning objectives - -- Understand the importance of sector electrification -- Understand the need for sector coupling - -# Sector electrification - -Electrification is becoming increasingly important in all sectors of the economy in order to achieve decarbonisation goals. As we saw earlier, electrification can be used to decarbonise the residential, transport, industrial and commercial sectors. However, some sectors are likely to be easier to electrify than other sectors. We have seen rapid progress with electric vehicles in parts of the transport sector, but sectors such as shipping and steel, which are harder to decarbonise, still have a way to go. - -However, different options exist for the decarbonisation of steel, for example. This can be done by retrofitting blast furnaces and adding carbon capture and storage (CCS) or scaling up hydrogen-based direct reduced iron. However, this will require innovation and further research on the key technologies, such as CCS. - -## Sector coupling - -We have seen that we must decarbonise to meet global climate targets. However, this is not a straightforward process. A large reason for this is the inflexibility of intermittent renewable resources such as solar and wind technologies. One method of mitigating this variability and inflexibility is through sector coupling. Sector coupling is where we connect energy demands and processes across differing sectors and increase the efficiency and flexibility of energy use. This would allows us to use renewable energy for all sectors. - -One way this could be achieved is through power to gas conversion. When there is a high supply of renewable power, excess electricity could be used to produce hydrogen and methane. This would allow us to store this energy for later use across multiple sectors. This would enable sectors that are difficult to electrify to be based on renewable energy. - -It is possible to model this sector coupling process within MUSE and to understand the tipping points which would make sector coupling possible. This could be based on the price and capacity of renewable energy, as well as the price of generating hydrogen or methane compared to the incumbent technologies. - -# Summary - -In this lecture we have covered the importance of electrifying different sectors to reduce carbon emissions and meet some of the Sustainable Development Goals. We have also learnt of the importance of sector coupling to address hard to decarbonise sectors. - - - diff --git a/docs/lecture_06/assets/Figure_6.1.1.png b/docs/lecture_06/assets/Figure_6.1.1.png deleted file mode 100644 index 80234e3..0000000 Binary files a/docs/lecture_06/assets/Figure_6.1.1.png and /dev/null differ diff --git a/docs/lecture_06/bibliography.bib b/docs/lecture_06/bibliography.bib index 0d1365b..178e9fd 100644 --- a/docs/lecture_06/bibliography.bib +++ b/docs/lecture_06/bibliography.bib @@ -1,62 +1,274 @@ +@article{Bloemendaal2019storm, + author = {Nadia Bloemendaal and Ivan Haigh and Hans {de Moel} and S. Muis and Reindert Haarsma and Jeroen Aerts}, + date-added = {2021-08-13 12:34:08 +0200}, + date-modified = {2021-08-13 12:34:14 +0200}, + doi = {10.4121/uuid:82c1dc0d-5485-43d8-901a-ce7f26cda35d}, + month = {11}, + title = {{STORM IBTrACS present climate synthetic tropical cyclone tracks}}, + url = {https://data.4tu.nl/articles/dataset/STORM_IBTrACS_present_climate_synthetic_tropical_cyclone_tracks/12706085}, + year = {2019}, + Bdsk-Url-1 = {https://data.4tu.nl/articles/dataset/STORM_IBTrACS_present_climate_synthetic_tropical_cyclone_tracks/12706085}, + Bdsk-Url-2 = {https://doi.org/10.4121/uuid:82c1dc0d-5485-43d8-901a-ce7f26cda35d}} -@article{en12112055, - author = {Mancini, Francesco and Lo Basso, Gianluigi and De Santoli, Livio}, - title = {Energy Use in Residential Buildings: Characterisation for Identifying Flexible Loads by Means of a Questionnaire Survey}, - journal = {Energies}, - volume = {12}, - year = {2019}, - number = {11}, - article-number = {2055}, - url = {https://www.mdpi.com/1996-1073/12/11/2055}, - issn = {1996-1073}, - abstract = {This work shows the outcomes of a research activity aimed at the energy characterization of residential users. Specifically, by data analysis related to the real energy consumption of sample buildings, the flexible loads amount has been identified so as to investigate on the opportunity to implement a demand/response (DR) program. The most meaningful input data have been collected by an on-line questionnaire created within an Excel spreadsheet allowing one to simulate and compare the calculations with the actual dwellings’ consumption; 412 questionnaires have been used as statistical sample and simulations have been performed based on single-zone dynamic model. Additionally, once the energy consumptions have been sorted by the different services, reference key performance indicators (KPIs) have been also calculated normalising those ones by people and house floor surface. From data analysis, it emerges how the Italian residential users are not very electrified. Furthermore, the flexible loads are low and, implementing minor maintenance interventions, the potential of flexibility can decrease up to 20%. For that reason, the current research can be further developed by investigating on suitable flexibility extensions as well as on the automation system requirements which is needed managing the flexible loads.}, - doi = {10.3390/en12112055} -} - - -@article{en13020432, - author = {Arens, Stefan and Schlüters, Sunke and Hanke, Benedikt and Maydell, Karsten von and Agert, Carsten}, - title = {Sustainable Residential Energy Supply: A Literature Review-Based Morphological Analysis}, - journal = {Energies}, - volume = {13}, - year = {2020}, - number = {2}, - article-number = {432}, - url = {https://www.mdpi.com/1996-1073/13/2/432}, - issn = {1996-1073}, - abstract = {The decarbonization of the energy system will bring substantial changes, from supranational regions to residential sites. This review investigates sustainable energy supply, applying a multi-sectoral approach from a residential site perspective, especially with focus on identifying crucial, plausible factors and their influence on the operation of the system. The traditionally separated mobility, heat, and electricity sectors are examined in more detail with regard to their decarbonization approaches. For every sector, available technologies, demand, and future perspectives are described. Furthermore, the benefits of cross-sectoral integration and technology coupling are examined, besides challenges to the electricity grid due to upcoming technologies, such as electric vehicles and heat pumps. Measures such as transport mode shift and improving building insulation can reduce the demand in their respective sector, although their impact remains uncertain. Moreover, flexibility measures such as Power to X or vehicle to grid couple the electricity sector to other sectors such as the mobility and heat sectors. Based on these findings, a morphological analysis is conducted. A morphological box is presented to summarize the major characteristics of the future residential energy system and investigate mutually incompatible pairs of factors. Lastly, the scenario space is further analyzed in terms of annual energy demand for a district.}, - doi = {10.3390/en13020432} -} - - -@article{world1020007, - author = {Mendoza, Daniel L. and Bianchi, Carlo and Thomas, Jermy and Ghaemi, Zahra}, - title = {Modeling County-Level Energy Demands for Commercial Buildings Due to Climate Variability with Prototype Building Simulations}, - journal = {World}, - volume = {1}, - year = {2020}, - number = {2}, - pages = {67--89}, - url = {https://www.mdpi.com/2673-4060/1/2/7}, - issn = {2673-4060}, - abstract = {The building sector accounts for nearly 40% of total primary energy consumption in the U.S. and E.U. and 20% of worldwide delivered energy consumption. Climate projections predict an increase of average annual temperatures between 1.1–5.4 °C by 2100. As urbanization is expected to continue increasing at a rapid pace, the energy consumption of buildings is likely to play a pivotal role in the overall energy budget. In this study, we used EnergyPlus building energy models to estimate the future energy demands of commercial buildings in Salt Lake County, Utah, USA, using locally-derived climate projections. We found significant variability in the energy demand profiles when simulating the study buildings under different climate scenarios, based on the energy standard the building was designed to meet, with reductions ranging from 10% to 60% in natural gas consumption for heating and increases ranging from 10% to 30% in electricity consumption for cooling. A case study, using projected 2040 building stock, showed a weighted average decrease in heating energy of 25% and an increase of 15% in cooling energy. We also found that building standards between ASHRAE 90.1-2004 and 90.1-2016 play a comparatively smaller role than variation in climate scenarios on the energy demand variability within building types. Our findings underscore the large range of potential future building energy consumption which depends on climatic conditions, as well as building types and standards.}, - doi = {10.3390/world1020007} -} - - - -@article{owidco2andothergreenhousegasemissions, - author = {Hannah Ritchie and Max Roser}, - title = {CO₂ and Greenhouse Gas Emissions}, - journal = {Our World in Data}, - year = {2020}, - note = {https://ourworldindata.org/co2-and-other-greenhouse-gas-emissions} -} - -@article{iea_world_energy_balance, - author = {IEA}, - title = {World Energy Balances: Overview}, - journal = {IEA}, - year = {2021}, - note = {https://www.iea.org/reports/world-energy-balances-overview} -} +@report{Dawson2016, + author = {R.J. Dawson and D. Thompson and D. Johns and S. Gosling and L. Chapman and G. Darch and G. Watson and W. Powrie and S. Bell and K. Paulson and P. Hughes and R. Wood}, + city = {London}, + institution = {Adaptation Sub-Committee for the Committee on Climate Change}, + publisher = {Paul Hughes}, + title = {UK Climate Change Risk Assessment Evidence Report: Chapter 4: Infrastructure}, + year = {2016}} + +@article{Steinbuks2010, + abstract = {This paper attempts to identify the underlying causes and costs of own generation of electric power in Africa. Rigorous empirical analysis of 8483 currently operating firms in 25 African countries shows that the prevalence of own generation would remain high (at around 20%) even if power supplies were perfectly reliable, suggesting that other factors such as firms' size, emergency back-up and export regulations play a critical role in the decision to own a generator. The costs of own-generation are about three times as high as the price of purchasing (subsidized) electricity from the public grid. However, because these generators only operate a small fraction of the time, they do not greatly affect the overall average cost of power to industry. The benefits of generator ownership are also substantial. Firms with their own generators report a value of lost load of less than US$50 per hour, compared with more than US$150 per hour for those without. Nevertheless, when costs and benefits are considered side by side, the balance is not found to be significantly positive. {\copyright} 2009 Elsevier B.V.}, + author = {J. Steinbuks and V. Foster}, + doi = {10.1016/j.eneco.2009.10.012}, + issn = {01409883}, + issue = {3}, + journal = {Energy Economics}, + keywords = {Africa,Electricity,Generators,Ownership,Reliability}, + month = {5}, + pages = {505-514}, + publisher = {North-Holland}, + title = {When do firms generate? Evidence on in-house electricity supply in Africa}, + volume = {32}, + year = {2010}, + Bdsk-Url-1 = {https://doi.org/10.1016/j.eneco.2009.10.012}} + +@generic{Hall2015, + abstract = {The impacts of extreme events are triggering action and reaction --- sometimes in unexpected ways. Confronted by 'adaptation emergencies', the private sector is rapidly innovating climate risk management, but governments must also fulfil their responsibilities.}, + author = {Jim W. Hall and Frans Berkhout and Rowan Douglas}, + doi = {10.1038/nclimate2467}, + issn = {17586798}, + issue = {1}, + journal = {Nature Climate Change}, + keywords = {Business and industry,Governance,Government}, + month = {12}, + pages = {6-7}, + publisher = {Nature Publishing Group}, + title = {Responding to adaptation emergencies}, + url = {http://dx.doi.org/10.5270/OceanObs09-FOO}, + volume = {5}, + year = {2015}, + Bdsk-Url-1 = {http://dx.doi.org/10.5270/OceanObs09-FOO}, + Bdsk-Url-2 = {http://dx.doi.org/10.1038/nclimate2467}} + +@webpage{Loew2019, + author = {Petra Loew}, + title = {The natural disasters of 2018 in figures | Munich Re Topics Online}, + url = {https://www.munichre.com/topics-online/en/climate-change-and-natural-disasters/natural-disasters/the-natural-disasters-of-2018-in-figures.html}, + year = {2019}, + Bdsk-Url-1 = {https://www.munichre.com/topics-online/en/climate-change-and-natural-disasters/natural-disasters/the-natural-disasters-of-2018-in-figures.html}} + +@article{Thacker2019, + abstract = {Infrastructure systems form the backbone of every society, providing essential services that include energy, water, waste management, transport and telecommunications. Infrastructure can also create harmful social and environmental impacts, increase vulnerability to natural disasters and leave an unsustainable burden of debt. Investment in infrastructure is at an all-time high globally, thus an ever-increasing number of decisions are being made now that will lock-in patterns of development for future generations. Although for the most part these investments are motivated by the desire to increase economic productivity and employment, we find that infrastructure either directly or indirectly influences the attainment of all of the Sustainable Development Goals (SDGs), including 72% of the targets. We categorize the positive and negative effects of infrastructure and the interdependencies between infrastructure sectors. To ensure that the right infrastructure is built, policymakers need to establish long-term visions for sustainable national infrastructure systems, informed by the SDGs, and develop adaptable plans that can demonstrably deliver their vision.}, + author = {Scott Thacker and Daniel Adshead and Marianne Fay and St{\'e}phane Hallegatte and Mark Harvey and Hendrik Meller and Nicholas O'Regan and Julie Rozenberg and Graham Watkins and Jim W. Hall}, + doi = {10.1038/s41893-019-0256-8}, + issn = {23989629}, + issue = {4}, + journal = {Nature Sustainability}, + keywords = {Civil engineering,Developing world,Sustainability}, + month = {4}, + pages = {324-331}, + publisher = {Nature Publishing Group}, + title = {Infrastructure for sustainable development}, + url = {https://doi.org/10.1038/s41893-019-0256-8}, + volume = {2}, + year = {2019}, + Bdsk-Url-1 = {https://doi.org/10.1038/s41893-019-0256-8}} + +@article{Koks2019, + abstract = {Transport infrastructure is exposed to natural hazards all around the world. Here we present the first global estimates of multi-hazard exposure and risk to road and rail infrastructure. Results reveal that ~27% of all global road and railway assets are exposed to at least one hazard and ~7.5% of all assets are exposed to a 1/100 year flood event. Global Expected Annual Damages (EAD) due to direct damage to road and railway assets range from 3.1 to 22 billion US dollars, of which ~73% is caused by surface and river flooding. Global EAD are small relative to global GDP (~0.02%). However, in some countries EAD reach 0.5 to 1% of GDP annually, which is the same order of magnitude as national transport infrastructure budgets. A cost-benefit analysis suggests that increasing flood protection would have positive returns on ~60% of roads exposed to a 1/100 year flood event.}, + author = {E. E. Koks and J. Rozenberg and C. Zorn and M. Tariverdi and M. Vousdoukas and S. A. Fraser and J. W. Hall and S. Hallegatte}, + doi = {10.1038/s41467-019-10442-3}, + issn = {20411723}, + issue = {1}, + journal = {Nature Communications}, + keywords = {Environmental impact,Natural hazards}, + month = {12}, + pages = {1-11}, + pmid = {31239442}, + publisher = {Nature Publishing Group}, + title = {A global multi-hazard risk analysis of road and railway infrastructure assets}, + url = {https://doi.org/10.1038/s41467-019-10442-3}, + volume = {10}, + year = {2019}, + Bdsk-Url-1 = {https://doi.org/10.1038/s41467-019-10442-3}} + +@report{Hall2019, + author = {Jim W. Hall and Jeroen C.J.H. Aerts and Bilal M. Ayyub and Stephane Hallegatte and Mark Harvey and Xi Hu and Elco Koks and Caroline Lee and Xiawei Liao and Michael Mullan and Raghav Pant and Amelie Paszkowski and Julie Rozenberg and Fulai Sheng and Vladimir Stenek and Scott Thacker and Elina Vaananen and Lola Vallejo and Ted I.E. Veldkamp and Michelle van Vliet and Yoshihide Wada and Philip Ward and Graham Watkins and Conrad Zorn}, + institution = {Global Commission on Adapatation}, + title = {Adaptation of Infrastructure Systems}, + url = {https://gca.org/reports/adaptation-of-infrastructure-systems/}, + year = {2019}, + Bdsk-Url-1 = {https://gca.org/reports/adaptation-of-infrastructure-systems/}} + +@article{THACKER201730, + abstract = {The complex and interdependent nature of modern critical national infrastructures provides the conditions for which localized failures can propagate within and between network systems, resulting in disruptions that are widespread and often unforeseen. Within this study, we characterize critical national infrastructures as a system-of-systems and develop methodology to perform a multi-scale disruption analysis. To achieve this, we map functional pathways between network source and sink assets across a range of operational scales. Customer demands are attributed to these pathways and are used to build a weighted network. The resultant functional path set and weighted network are used to perform a disruption analysis that encodes information on the long-range functionality within and between infrastructures, providing insights into failure propagation and the functional dependencies that exist between assets from multiple sectors. We supplement the methodological development with a detailed national scale demonstration for England and Wales using a unique representation of the integrated electricity network and the domestic flight network. The results highlight the potentially large disruptions that can result from the failure of individual electricity assets from a range of different sub-systems.}, + author = {Scott Thacker and Raghav Pant and Jim W. Hall}, + doi = {https://doi.org/10.1016/j.ress.2017.04.023}, + issn = {0951-8320}, + journal = {Reliability Engineering & System Safety}, + keywords = {Infrastructure, Network, Interdependence, System-of-system, Multi-scale, Disruption analysis}, + note = {Special Section: Applications of Probabilistic Graphical Models in Dependability, Diagnosis and Prognosis}, + pages = {30-41}, + title = {System-of-systems formulation and disruption analysis for multi-scale critical national infrastructures}, + url = {https://www.sciencedirect.com/science/article/pii/S0951832017304994}, + volume = {167}, + year = {2017}, + Bdsk-Url-1 = {https://www.sciencedirect.com/science/article/pii/S0951832017304994}, + Bdsk-Url-2 = {https://doi.org/10.1016/j.ress.2017.04.023}} + +@article{Rinaldi2001, + author = {S. M. Rinaldi and J. P. Peerenboom and T. K. Kelly}, + doi = {10.1109/37.969131}, + journal = {IEEE Control Systems Magazine}, + number = {6}, + pages = {11-25}, + title = {Identifying, understanding, and analyzing critical infrastructure interdependencies}, + volume = {21}, + year = {2001}, + Bdsk-Url-1 = {https://doi.org/10.1109/37.969131}} + +@article{mann2009strategic, + author = {Mann, B}, + journal = {Cabinet Office, London< http://www.cabinetoffice.gov.uk/media/308367/sfps-consultation.pdf}, + title = {Strategic Framework and Policy Statement on Improving the Resilience of Critical Infrastructure to Disruption from Natural Hazards}, + year = {2009}} + +@book{Hall2016, + address = {Cambridge}, + author = {Hall, J.W. and Tran, M. and Hickford, A. and Nicholls, R.}, + publisher = {Cambridge University Press}, + title = {{The future of national infrastructure: A systems-of-systems approach}}, + year = {2016}} + +@book{field2014climate, + author = {Field, Christopher B and Barros, Vicente R}, + publisher = {Cambridge University Press}, + title = {Climate change 2014--Impacts, adaptation and vulnerability: Regional aspects}, + year = {2014}} + +@incollection{oppenheimer2015emergent, + author = {Oppenheimer, Michael and Campos, Maximiliano and Warren, Rachel and Birkmann, Joern and Luber, George and O'Neill, Brian and Takahashi, Kiyoshi and Brklacich, Mike and Semenov, Sergey and Licker, Rachel and others}, + booktitle = {Climate Change 2014 Impacts, Adaptation and Vulnerability: Part A: Global and Sectoral Aspects}, + pages = {1039--1100}, + publisher = {Cambridge University Press}, + title = {Emergent risks and key vulnerabilities}, + year = {2015}} + +@article{pant2018critical, + author = {Pant, Raghav and Thacker, Scott and Hall, Jim W and Alderson, David and Barr, Stuart}, + journal = {Journal of Flood Risk Management}, + number = {1}, + pages = {22--33}, + publisher = {Wiley Online Library}, + title = {Critical infrastructure impact assessment due to flood exposure}, + volume = {11}, + year = {2018}} + +@article{watts2002simple, + author = {Watts, Duncan J}, + journal = {Proceedings of the National Academy of Sciences}, + number = {9}, + pages = {5766--5771}, + publisher = {National Acad Sciences}, + title = {A simple model of global cascades on random networks}, + volume = {99}, + year = {2002}} + +@article{PANT2014183, + abstract = {While early research efforts were devoted to the protection of systems against disruptive events, be they malevolent attacks, man-made accidents, or natural disasters, recent attention has been given to the resilience, or the ability of systems to ``bounce back,'' of these events. Discussed here is a modeling paradigm for quantifying system resilience, primarily as a function of vulnerability (the adverse initial system impact of the disruption) and recoverability (the speed of system recovery). To account for uncertainty, stochastic measures of resilience are introduced, including Time to Total System Restoration, Time to Full System Service Resilience, and Time to α%-Resilience. These metrics are applied to quantify the resilience of inland waterway ports, important hubs in the flow of commodities, and the port resilience approach is deployed in a data-driven case study for the inland Port of Catoosa in Oklahoma. The contributions herein demonstrate a starting point in the development of a resilience decision making framework.}, + author = {Raghav Pant and Kash Barker and Jose Emmanuel Ramirez-Marquez and Claudio M. Rocco}, + doi = {https://doi.org/10.1016/j.cie.2014.01.017}, + issn = {0360-8352}, + journal = {Computers & Industrial Engineering}, + keywords = {Resilience, Infrastructure systems, Vulnerability, Recoverability}, + pages = {183-194}, + title = {Stochastic measures of resilience and their application to container terminals}, + url = {https://www.sciencedirect.com/science/article/pii/S0360835214000333}, + volume = {70}, + year = {2014}, + Bdsk-Url-1 = {https://www.sciencedirect.com/science/article/pii/S0360835214000333}, + Bdsk-Url-2 = {https://doi.org/10.1016/j.cie.2014.01.017}} + +@article{HOSSEINI201647, + abstract = {Modeling and evaluating the resilience of systems, potentially complex and large-scale in nature, has recently raised significant interest among both practitioners and researchers. This recent interest has resulted in several definitions of the concept of resilience and several approaches to measuring this concept, across several application domains. As such, this paper presents a review of recent research articles related to defining and quantifying resilience in various disciplines, with a focus on engineering systems. We provide a classification scheme to the approaches in the literature, focusing on qualitative and quantitative approaches and their subcategories. Addressed in this review are: an extensive coverage of the literature, an exploration of current gaps and challenges, and several directions for future research.}, + author = {Seyedmohsen Hosseini and Kash Barker and Jose E. Ramirez-Marquez}, + doi = {https://doi.org/10.1016/j.ress.2015.08.006}, + issn = {0951-8320}, + journal = {Reliability Engineering & System Safety}, + keywords = {Resilience, Engineering systems}, + pages = {47-61}, + title = {A review of definitions and measures of system resilience}, + url = {https://www.sciencedirect.com/science/article/pii/S0951832015002483}, + volume = {145}, + year = {2016}, + Bdsk-Url-1 = {https://www.sciencedirect.com/science/article/pii/S0951832015002483}, + Bdsk-Url-2 = {https://doi.org/10.1016/j.ress.2015.08.006}} + +@article{hickford2018resilience, + author = {Hickford, Adrian J and Blainey, Simon P and Hortelano, Alejandro Ortega and Pant, Raghav}, + journal = {Environment Systems and Decisions}, + number = {3}, + pages = {278--291}, + publisher = {Springer}, + title = {Resilience engineering: theory and practice in interdependent infrastructure systems}, + volume = {38}, + year = {2018}} + +@article{trigg2016credibility, + author = {Trigg, MA and Birch, CE and Neal, JC and Bates, PD and Smith, A and Sampson, CC and Yamazaki, D and Hirabayashi, Y and Pappenberger, F and Dutra, E and others}, + journal = {Environmental Research Letters}, + number = {9}, + pages = {094014}, + publisher = {IOP Publishing}, + title = {The credibility challenge for global fluvial flood risk analysis}, + volume = {11}, + year = {2016}} + +@article{wing2020toward, + author = {Wing, Oliver EJ and Quinn, Niall and Bates, Paul D and Neal, Jeffrey C and Smith, Andrew M and Sampson, Christopher C and Coxon, Gemma and Yamazaki, Dai and Sutanudjaja, Edwin H and Alfieri, Lorenzo}, + journal = {Water Resources Research}, + number = {8}, + pages = {e2020WR027692}, + publisher = {Wiley Online Library}, + title = {Toward Global Stochastic River Flood Modeling}, + volume = {56}, + year = {2020}} + +@article{gassert2015aqueduct, + author = {Gassert, Francis and Reig, Paul and Shiao, Tien and Landis, Matt and Luck, Matt and others}, + title = {Aqueduct global maps 2.1}, + year = {2015}} + +@misc{oh2019addressing, + author = {Oh, Jung Eun and Espinet Alegre, Xavier and Pant, Raghav and Koks, Elco E and Russell, Tom and Schoenmakers, Roald and Hall, Jim W}, + publisher = {World Bank}, + title = {Addressing Climate Change in Transport: Volume 2: Pathway to Resilient Transport}, + year = {2019}} + +@article{pant2016vulnerability, + author = {Pant, Raghav and Hall, Jim W and Blainey, Simon P}, + journal = {European Journal of Transport and Infrastructure Research}, + number = {1}, + title = {Vulnerability assessment framework for interdependent critical infrastructures: case-study for Great Britain's rail network}, + volume = {16}, + year = {2016}} + +@article{jafino2020transport, + author = {Jafino, Bramka Arga and Kwakkel, Jan and Verbraeck, Alexander}, + journal = {Transport Reviews}, + number = {2}, + pages = {241--264}, + publisher = {Taylor \& Francis}, + title = {Transport network criticality metrics: a comparative analysis and a guideline for selection}, + volume = {40}, + year = {2020}} + +@manual{QGIS_software, + author = {{QGIS Development Team}}, + organization = {QGIS Association}, + title = {QGIS Geographic Information System}, + url = {https://www.qgis.org}, + year = {2021}, + Bdsk-Url-1 = {https://www.qgis.org}} diff --git a/docs/lecture_07/Lecture_7.1.md b/docs/lecture_07/Lecture_7.1.md index a650a3f..5d6d7e0 100644 --- a/docs/lecture_07/Lecture_7.1.md +++ b/docs/lecture_07/Lecture_7.1.md @@ -1,27 +1,37 @@ --- -title: Mini-Lecture 7.1 -- Agents in energy systems models +title: Mini-Lecture 7.1 -- Timeslicing in energy systems modelling keywords: -- Agent-based model +- Timeslices +- Energy modelling +- Energy demands authors: - Alexander J. M. Kell --- -In this mini-lecture we will describe the importance of agents within MUSE and also within an energy systems modelling context. +This mini-lecture provides an overview of timeslicing in energy systems modelling. # Learning objectives -- Understand why agents are important in an energy modelling context -- Understand how we can characterise these agents within MUSE +- Learn why we use timeslices in energy systems models +- Understand the importance of representative days -# Agents overview +# Introduction -Within real-life energy systems there are many different objectives that investors or consumers have. These objectives may differ by sector, by investor type or by proportions of the population. For instance, a certain percentage of the population may be willing to be spend more money on heating their homes than others. +With energy systems models we must model how demand is met by supply. However, over the course of a year, or even over the course of 30 years we have large variations in demand and supply. For instance, the weather changes between years, seasons, and days. This all has an effect on the amount of energy that can be supplied by renewable energy sources such as solar and wind. -It is straightforward to specify these objectives and characteristics within MUSE. For instance, you may want to split a population based upon their geospatial and economic characteristics. This could be done by, for example, splitting a population into rural and urban categories. That would provide us with two groups. However, it is possible to go further, and we may want to split the rural and urban groups into different socioeconomic demographics, such as disposable income. +It is also true that this variation in demand has a large impact on the demand. In a particularly cold year, or on a particular cold day, energy demand may significantly increase as consumers use more energy for heating. The same may be true during a particularly warm period if people need energy for cooling systems. We therefore need to model this variability. -Say for example, we only split the population into rural and urban. We can specify these groups as two agents within MUSE. Once we have specified the two agents, we would have to give them characteristics which differentiate them from the each other and define the proportion of the population that they make up. It must be noted, at this stage, that we do not need to have a separate agent for each individual or entity. It is perfectly fine to group and aggregate similar individuals or agents. +## Representative days + +As you can probably imagine, matching supply and demand for every 30 minutes in a year is very costly in terms of computation time. If we must match supply and demand for every 30 minutes for 30 years (or more), we may end up with a very slow model in return for some gains in accuracy. + +However, it may be the case that we do not need to model a year in such high detail. In most cases, for long-term energy systems models, we can reduce the amount of detail to significantly increase the speed of the model, without losing significant accuracy [@Kell2020]. + +A common approach is to model 4 days for each year. Each day corresponds to a season of the year and is split into 24 timeslices (which equates to a timeslice representing one hour). Therefore, we maintain the variability within a day, but also within seasons. We will lose some of the extremely hot or cold days, but that matters less when we're considering the long-term planning horizon. + +We do not always have to take into account entire days, to reduce the complexity further. For instance, we could have 8 days, but with only 2 timeslices (day and night). This will make the model run quickly, but may lose some detail. It is up to you, as the modeller, to find a sweet spot between accuracy and speed of computation. Various papers have been published to find this sweet spot, which you can look into in your own time [@Poncelet2017]. # Summary -In this mini-lecture we understood the concept of agents and how they relate to an energy modelling context. We briefly understood how we can translate these concepts into MUSE. Urban populations might have greater energy needs or rural populations may not have access to the same energy sources. Giving the model a bit more detail will allow you to make sure that the model is both more accurate, and that its projections take into account different parts of society. In the hands-on we will learn how to add a new agent. +In this mini-lecture we discovered why long-term energy models consider timeslices and representative days. Through this approach we are able to maintain high accuracy whilst also reducing computation time. diff --git a/docs/lecture_07/Lecture_7.2.md b/docs/lecture_07/Lecture_7.2.md index 4633f42..177d032 100644 --- a/docs/lecture_07/Lecture_7.2.md +++ b/docs/lecture_07/Lecture_7.2.md @@ -1,38 +1,37 @@ --- -title: Mini-Lecture 7.2 -- How to relate agent representations to the real world +title: Mini-Lecture 7.2 - Technologies by timeslice keywords: -- Agent-based modelling -- Characterisation +- Energy technologies +- Energy modelling +- Timeslices authors: - Alexander J. M. Kell --- -In this mini-lecture we will introduce some methods to translate socioeconomic data into MUSE with a quantitative approach. +In this mini-lecture we describe how different technologies can have different characteristics by timeslices. # Learning objectives -- Discuss surveys and socioeconomic data and how these can relate to MUSE -- Discover ways that surveys can be used in quantitative modelling +- Understand the different characteristics of technologies by timeslice +- Understand how to characterise technologies by timeslice -# Qualitative representation in agent-based models +# Introduction -Through the use of qualitative data, such as using qualitative surveys, it is possible to gain greater insight into the different characteristics of consumers or investors. One example of how this can be done was by Moya et al. (2020). In this paper the authors explore fuel-switching investment in the long-term energy transitions of India's industry sector. They inform the modelled agents through a questionnaire that was carried out to inform MUSE. +In the previous lecture we discovered the importance of timeslices. In this mini-lecture we will learn about how different technologies have different characteristics when it comes to timeslices, and how this can be modelled within MUSE. -Some of the types of questions asked in the questionnaire to industrial companies are listed below: -- Geographical location -- Financial details -- Investment plans -- Type of fuels used -- Willingness to switch fuels +# Technologies by timeslices -Once these data have been collected, they can be used to find similar groups of investors and to start characterising the agents. For instance, if from the data it is clear that geographical location is an important consideration, the decision could be made to group companies by geographical region and form an agent on this basis. If the more important consideration is the investment plans, then a group can be made there. +Different technologies and supply sectors have different characteristics when it comes to timeslices. For instance, solar photovoltaics do not produce any energy when it is dark (for instance, at night) and produce less in the winter. Wind, on the other hand, has a completely different profile and is largely dependent on geography. Therefore, it would make sense to provide a maximum output of the technologies at different times. For instance, it would be useful if the model limited solar output at night time in the form of a maximum utilization factor. Where utilization factor is the ratio of average amount of energy output to total possible output of an energy technology if it were to run 100% of time. -This approach is a more than efficient method of better understanding the characteristics of agents of a system, and it can help to inform a better modelling process. The work by Moya et al. ([@Moya2020]) finds that the results represent the unique heterogeneity of fuel-switching industrial investors with distinct investment goals and limited foresight on costs. In other words, the survey results have an impact on the outcome of the energy system over the long-term. +However, it can be very difficult to turn off some technologies, such as a nuclear power plant. Nuclear power plants are expensive to turn on and can be unsafe if constantly varying their power. Also, their marginal cost, or the cost to produce 1MWh of electricity excluding capital costs, is usually much lower than other power plants such as gas or coal plants. It, therefore, makes sense that we place a minimum service factor, or minimum output allowed, on nuclear, to ensure their output does not fall below a certain level. +Other technologies, however, such as gas power plants, can be turned on and off readily; therefore we can simply leave an average utilization factor for all the timeslices. + +All of these features exist in MUSE, and during this lecture's hands-on, we will show you how to do this within MUSE. # Summary -In this mini-lecture we explored how surveys can be used to inform agents within MUSE. We also discovered how these results can affect the modelling outcomes of energy systems. +In this mini-lecture we have explored the importance of characterising technologies not just by their economic data, but also by their physical characteristics. We discovered that different technologies have different outputs at different times, such as solar and wind. We also found out that nuclear power, for instance, must output a certain level to remain within a safety range. + -  diff --git a/docs/lecture_07/Lecture_7.3.md b/docs/lecture_07/Lecture_7.3.md index 68351e5..51b6fd1 100644 --- a/docs/lecture_07/Lecture_7.3.md +++ b/docs/lecture_07/Lecture_7.3.md @@ -1,38 +1,39 @@ --- -title: Mini-Lecture 7.3 -- Agents by sector +title: Mini-Lecture 7.3 - Different energy demands by timeslice keywords: -- Sectors -- Agent differentiation -- Key agent parameters +- Energy demands +- Timeslice +- Energy modelling authors: - Alexander J. M. Kell --- -In this mini-lecture we will cover how agents and their characteristics can differ between sectors. We will also investigate the similarities between agents and sectors and consider the key parameters that make up agents. +This mini-lecture will continue exploring the importance of timeslices in energy modelling; however, it will have a particular focus on energy demands, and how these can change by timeslice and over the years. -# Learning objectives - -- Understand the differences between agents of different sectors -- Understand the key parameters that differentiate agents +In the previous lecture we explored energy demands and timeslices. In this lecture we will have a brief recap of this, and explore how energy demand can be represented within MUSE. +# Learning objectives -# Agent parameters +- Understand how energy demand can change by timeslice +- Learn how energy demand is represented in MUSE -Different sectors may mean having agents with different characteristics. For instance, within the residential sector socioeconomic data can be used to characterise the agents. We could use wealth to characterise our agents in different geographic locations. For example we could place a constraint on the `Budget` parameter for residential users, and split these agents into different proportions. For example, we could prohibit 70% of residential users from spending more than a certain amount on heating which could affect their technology choice. The other 30% of users would form an agent that was not constricted in this way, and thus their choices may end up being differet in the model. +# Energy demand -Another way we could classify residential agents is through the `Maturity` parameter. This would limit investments in novel technologies until the specified technology had a certain market share. This could be informed by the innovation adoption lifecycle, as shown by Figure 7.3.1. Where, for example, innovators make up 2.5% of the population but have no `Maturity` constraints. As we work our way up the curve from innovators to laggards, this `Maturity` constraint increases. +Energy demand can come in various forms. For instance, the demand we model can be for heating or cooling in the residential sector. It is the case that these demands have different characteristics. For instance, they may have different magnitudes and different technologies which serve these demands as well as they may be able to run at different times. -![](assets/Figure_7.3.1.png){width=100%} +Within MUSE, similarly to the supply sectors, we can model this time varying capability with timeslices. For instance, if we have 4 representative days which refer to the different seasons, we can model the high heating demand in winter and cooling demand in summer. On top of this we can vary these demands by time of day. -**Figure 7.3.1:** Innovation adoption lifecycle +To do this, we must edit the demand in the `preset/Residential2050Consumption.csv` sector. An example of which is shown in Figure 7.3.1. -# Sectors +![](assets/Figure_7.1.1.png){width=100%} -In this mini-lecture we have focused on the residential sector and seen the way we can characterise agents. Although these characteristics may not directly translate to the power sector, in some cases investors in the power sector can have similar characteristics. For instance, some companies are larger, and are more willing to invest their capital, reflecting a larger `Budget` parameter. Others may be less willing to invest in new technologies. The differing objectives of agents will often be the reason behind differences with other agents. For instance, some agents may only want to minimise their costs, whereas others may want to reduce their capital expenditure. It is easy to change these characteristics within MUSE to create diverse energy scenarios. +**Figure 7.3.1:** Example input for the preset sector. +In this small example we see that there is only a demand for `heat` in the residential sector. However, this demand changes per timeslice (which are listed in the leftmost column). For instance, there is low demand for heat in timeslice 0 and a high demand for heat in timeslice 4. These timeslices refer to a single representative day, and therefore timeslice 4 has the highest demand for heat as it is in the late-evening, when people generally come home from work and turn on their radiators. +In your models you can use datasets to disaggregate the demand into different types, or you can aggregate demand to include all gas or electricity utilised in the residential sector. This is largely dependent on the data available and the complexity of the model you would like. # Summary -In this mini-lecture we covered the differences between agents and the different parameters that can be used to inform these differences. We saw how the `Maturity` constraint maps to the innovation adoption lifecycle and how the `Budget` parameter can be informed by socioeconomic characteristics. These parameters lead to a large amount of possible scenarios that can be tested and run. -  +In this mini-lecture, we explored the importance of timeslicing for modelling demand in energy models. We also covered how this can be done within MUSE using the preset sector. + diff --git a/docs/lecture_07/Lecture_7.4.md b/docs/lecture_07/Lecture_7.4.md index 153f323..6bba663 100644 --- a/docs/lecture_07/Lecture_7.4.md +++ b/docs/lecture_07/Lecture_7.4.md @@ -1,33 +1,34 @@ --- -title: Mini-Lecture 7.4 -- Agent parameters +title: Mini-Lecture 7.4 -- Timeslicing and climate policy keywords: -- Agent parameters -- MUSE +- Climate policy +- Timeslicing authors: - Alexander J. M. Kell --- -This mini-lecture explores all the major parameters that can define agents within MUSE. +This mini-lecture explores the relevance of timeslicing to climate policy. We will explore how different timeslicing can affect modelling results, why it is important to consider realistic timeslicing and how these can affect policy decisions. # Learning objectives -- Understand the different agent parameters and their role within MUSE +- Understand the impact of timeslicing on modelling outputs +- Learn how timeslicing can affect policy decisions -# Overview agent parameters +# Timeslicing and policy -Within MUSE each agent can have their own objectives. MUSE is flexible enough to allow for up to 3 objectives, which can be summed together at various weightings. To input these objectives into MUSE one would use the `Objective1`, `Objective2` and/or `Objective3` parameters and select an objective such as `comfort`, `lifetime_levelized_cost_of_energy` or `fixed_costs`. +Timeslicing is a core component of an energy systems model as we have previously discussed. If one were to use an inappropriate number of timeslices in an energy systems model, it is likely that this would have major implications on the model outputs. -Then we would select the weight of each of the objectives using the `ObjData1`, `ObjData2`, `ObjData3` inputs. For example, if we had 3 objectives, we could make the objective of `Objective1` dominant by setting `ObjData1` to 0.5. This would mean it would make up 50% of the final objective. +Let's look at an example: if we were to model solar panels with an average capacity factor for the entire time horizon of the model this would assume that the solar panels can be used at night and could displace other technologies, such as gas turbines. However, in reality, solar panels contribute to the grid during the day and produce nothing at night. Therefore, we need some sort of flexibility in the system to ramp up after the sun sets. This needs to be modelled explicitly within MUSE, so to allow gas (or other technologies) to fill this gap in supply. -We can edit the `SearchRule` to reduce the space of technologies that those agents are likely to consider. For example, we could fill this with `same_fuels`, or `same_enduse`. +If we take this conclusion further, it is possible to see scenarios where the intermittency of solar and wind are not modelled, and therefore we observe scenarios with a majority in solar or wind. With current technologies this is not possible, and this therefore underscores the importance of timeslicing. -The rest of the parameters include the parameters discussed in the previous lecture: - -- `MaturityThreshold` -- `Budget` +If we do not use accurate timeslicing then the model outputs can skew resulting policy, and so due care must be taken for sourcing data from different geographies. # Summary -In this mini-lecture we discovered the main parameters that are used by agents within MUSE. For a full breakdown of the parameters please refer to the MUSE documentation that can be found online. +In this lecture we have looked into the implications of different timeslicing decisions made when creating an energy systems model. We learnt that if we do not get this right, the investments made could be skewed and unrealistic. + + + diff --git a/docs/lecture_04/assets/Figure_4.1.1.png b/docs/lecture_07/assets/Figure_7.1.1.png similarity index 100% rename from docs/lecture_04/assets/Figure_4.1.1.png rename to docs/lecture_07/assets/Figure_7.1.1.png diff --git a/docs/lecture_07/assets/Figure_7.3.1.png b/docs/lecture_07/assets/Figure_7.3.1.png deleted file mode 100644 index 04d9d66..0000000 Binary files a/docs/lecture_07/assets/Figure_7.3.1.png and /dev/null differ diff --git a/docs/lecture_07/bibliography.bib b/docs/lecture_07/bibliography.bib index 4fd8773..a0fb1e6 100644 --- a/docs/lecture_07/bibliography.bib +++ b/docs/lecture_07/bibliography.bib @@ -1,15 +1,22 @@ -@article{Moya2020, - abstract = {This paper presents the formulation and application of a novel agent-based integrated assessment approach to model the attributes, objectives and decision-making process of investors in a long-term energy transition in India's iron and steel sector. It takes empirical data from an on-site survey of 108 operating plants in Maharashtra to formulate objectives and decision-making metrics for the agent-based model and simulates possible future portfolio mixes. The studied decision drivers were capital costs, operating costs (including fuel consumption), a combination of capital and operating costs, and net present value. Where investors used a weighted combination of capital cost and operating costs, a natural gas uptake of $\sim$12PJ was obtained and the highest cumulative emissions reduction was obtained, 2 Mt CO2 in the period from 2020 to 2050. Conversely if net present value alone is used, cumulative emissions reduction in the same period was lower, 1.6 Mt CO2, and the cumulative uptake of natural gas was equal to 15PJ. Results show how the differing upfront investment cost of the technology options could cause prevalence of high-carbon fuels, particularly heavy fuel oil, in the final mix. Results also represent the unique heterogeneity of fuel-switching industrial investors with distinct investment goals and limited foresight on costs. The perception of high capital expenditures for decarbonisation represents a significant barrier to the energy transition in industry and should be addressed via effective policy making (e.g. carbon policy/price).}, - author = {Moya, Diego and Budinis, Sara and Giarola, Sara and Hawkes, Adam}, - doi = {10.1016/j.apenergy.2020.115295}, - file = {:Users/alexanderkell/Downloads/1-s2.0-S0306261920308072-main-2.pdf:pdf}, - issn = {03062619}, - journal = {Applied Energy}, - keywords = {Agent-based,Decarbonisation,Energy survey,Energy systems modelling,Investment metrics,Iron and steel}, - pages = {115295}, - publisher = {Elsevier}, - title = {{Agent-based scenarios comparison for assessing fuel-switching investment in long-term energy transitions of the India's industry sector}}, - url = {https://doi.org/10.1016/j.apenergy.2020.115295}, - volume = {274}, - year = {2020} +@article{Kell2020, + author = {Kell, Alexander J. M. and Forshaw, Matthew and McGough, A. Stephen}, + isbn = {9781450366717}, + journal = {The Eleventh ACM International Conference on Future Energy Systems (e-Energy'20)}, + keywords = {agent-based modelling,en-,energy market simulation,ergy models,genetic algorithm,long-term,op-,policy,simulation,validation}, + mendeley-groups = {Energy Modelling}, + title = {{Long-Term Electricity Market Agent Based Model Validation using Genetic Algorithm based Optimization}}, + year = {2020} +} +@article{Poncelet2017, + abstract = {Due to computational restrictions, energy-system optimization models (ESOMs) and generation expansion planning models (GEPMs) frequently represent intra-annual variations in demand and supply by using the data of a limited number of representative historical days. The vast majority of the current approaches to select a representative set of days relies on either simple heuristics or clustering algorithms and comparison of different approaches is restricted to different clustering algorithms. This paper contributes by: (i) proposing criteria and metrics for evaluating representativeness, (ii) providing a novel optimization-based approach to select a representative set of days and (iii) evaluating and comparing the developed approach to multiple approaches available from the literature. The developed optimization-based approach is shown to achieve more accurate results than the approaches available from the literature. As a consequence, by applying this approach to select a representative set of days, the accuracy of ESOMs/GEPMs can be improved without increasing the computational cost. The main disadvantage is that the approach is computationally costly and requires an implementation effort.}, + author = {Poncelet, Kris and Hoschle, Hanspeter and Delarue, Erik and Virag, Ana and Drhaeseleer, William}, + issn = {08858950}, + journal = {IEEE Transactions on Power Systems}, + keywords = {Energy-system planning,generation expansion planning,power system economics,power system modeling,wind energy integration}, + mendeley-groups = {Electricity Market Simulations/Selecting Representative RES Days}, + number = {3}, + pages = {1936--1948}, + title = {{Selecting representative days for capturing the implications of integrating intermittent renewables in generation expansion planning problems}}, + volume = {32}, + year = {2017} } diff --git a/docs/lecture_08/Lecture_8.1.md b/docs/lecture_08/Lecture_8.1.md index 5304ae4..cdf4879 100644 --- a/docs/lecture_08/Lecture_8.1.md +++ b/docs/lecture_08/Lecture_8.1.md @@ -1,29 +1,95 @@ --- -title: Mini-Lecture 8.1 -- Introduction to regions and aggregation +title: Mini-Lecture 8.1 -- Energy demands in energy systems modelling keywords: -- Regions -- MUSE +- Energy demand +- Energy systems models authors: - Alexander J. M. Kell --- -This mini-lecture provides an overview of different regions within energy systems models and how these can be represented within MUSE. +To begin lecture 3, this mini-lecture provides an overview of energy demands within an energy system. We will cover differences in energy demands by sector, time and population classes. We will also begin to explore why these differences are important within energy models. Lecture 3 will take you through the basics for modelling energy demand in MUSE, the different options available to do so, and some specific examples # Learning objectives -- When to aggregate data into different regions +- Learn what energy demands are in an energy modelling context +- Understand how demands can change based on different variables -# Aggregation +# Introduction -Regions within energy models play an important role. We often want to aggregate technoeconomic data from multiple regions into one. For example, the UK is made up of many different counties with different energy demands and supply. However, it could be the case that we do not have comprehensive data for each of these counties. We may, however, have plentiful data for the UK as a whole, or even for England, Scotland, Northern Ireland and Wales. We can therefore aggregate these data and make assumptions about the geographical locations of supply and demand. +Everyone needs energy for many different purposes. The form in which this energy should be delivered is dependent on the specific application. These demands for energy come from all sectors of society such as: -This is an example of aggregation and can make the modelling process more straightforward, whilst losing a small amount of accuracy. This is because we do not need to model each individual power plant, demand centre or end-use sector. This means we can use aggregated data which are often easier to access. +- The residential sector (rural and urban) + - Cooking + - Heating + - Cooling + - Lighting + - Appliances +- Industry + - Chemical processes + - Steam production + - Heating +- Commerce + - Lighting + - Heating + - Cooling buildings + - Keeping products at low temperatures +- Transport + - Cars + - Trucks + - Buses + - Aviation + - Shipping + - Trains +- Agriculture + - Tractors + - Machinery + - Pumping water -We can also aggregate multiple countries into regions. For example, we can merge the European continent together. This would be especially useful if we are considering a global model. However, it must be noted that we would lose significant detail by aggregating up to a supranational level. It is up to you, the model user, to consider the trade-offs between aggregation and disaggregation. For example, if you only wanted to model a single country, it would be possible to have a single region. However, if you had good access to data at the local level, you could disaggregate the data further. It does not matter whether the region is a single country, a number of counties or at a supranational level. The regions depend on your case study and the data you have access to. +## Variations in daily energy demand +These energy demands can vary on hourly, daily, weekly and monthly timescales. This mainly reflects the schedule of consumers' activities. For example, on a monthly timescale more cooling will be used in summer and more heating in winter. However, these energy demands can also vary by sector, as shown by Figure 8.1.1. +![](assets/Figure_8.1.1.png){width=100%} -# Summary +**Figure 8.1.1:** Variations of energy demand by sector in a hypothetical example [@Taliotis2018]. + +Figure 8.1.1 shows us that the magnitude of demand varies by sector, with agricultural demand significantly lower than residential and commercial demand, in this example. The reason that the commercial and residential sectors consume more is because their activities are more energy intensive or they are simply larger. + +We can also see that the daily profile of demand varies by sector. For example, in Figure 8.1.1 we can see that there is a clear evening peak in residential demand, whereas agricultural and industrial demand remains flat throughout the day. This is because agricultural and industrial demands are consistent throughout the day. This is likely because the industrial and agricultural sector operate constantly, whereas energy use in homes peaks in the evening when consumers use more electricity for cooking, lighting and appliances when they return from work or other business. + +## Sector specific demands + +The differences between sectors means that it can sometimes be important to model demands separately by each sector. This feature allows the models to consider the specific characteristics of each demand. + +Within each of these sectors, the energy demand varies over time and across different types of consumers. For example, within the residential sector, demands can differ between rural and urban households, as shown in Figure 8.1.2. This can also be true between grid-connected and off-grid areas. Energy planners must ensure that energy demand is always met for all types of consumers. Therefore, it is important that the key characteristics of different demands are represented in energy models. + +![](assets/Figure_8.1.2.png){width=100%} + +**Figure 8.1.2:** Variations of energy demand for the residential sector by population types [@Olaniyan2018] + + +## Long-term variations in energy demands + +A major challenge in energy planning is that energy demands can change over time. This could be due to population growth or the creation of new industries. Figure 8.1.3 displays historical variations in energy demands. It is likely that these demands are correlated to changes in society. For example, increases in energy demand likely reflect increased industrial activity. For energy planning, we must also think about how energy demands are likely to change in the future. -In this mini-lecture we learnt about the trade-offs between aggregation and disaggregation when defining regions. We learnt that the more aggregated the model, the less granular data are required. This can be helpful in cases where the data are not available at a local level, but available at a national level. +We can often forecast energy demand, such as with future projections as shown in Figure 8.1.3. These forecasts can be created using estimates of the key influencers of energy demand, such as population growth and economic activity. Future projections are often based on how energy demands have changed historically. + +![](assets/Figure_8.1.3.png){width=100%} + +**Figure 8.1.3:** Long-term energy consumption by source + + +## Capacity expansion planning + +One of the key purposes of MUSE is for capacity expansion. Figure 8.1.4 displays this key issue which MUSE can address. Essentially, if total demand increases (green line) and existing system capacities are retired (blue line), how can we invest to meet the energy capacity needed to supply demand (red line)? + +![](assets/Figure_8.1.4.png){width=100%} + +**Figure 8.1.4:** Capacity expansion [@Taliotis2018] + +You may notice that the red line is higher than the green line at all points. This is due to losses due to lower generating efficiencies. The gap between the red and blue lines demonstrates the required capacity expansion over time. MUSE enables us to plan such a capacity expansion whilst considering technical, economic and environmental constraints. + + +# Summary +In this mini-lecture we covered the differences between energy demands in different population types, sectors and timescales. We learnt why it is important to model these differences in demand in energy systems models. We also explored how energy systems models can be used to meet a changing demand profile in the future. diff --git a/docs/lecture_08/Lecture_8.2.md b/docs/lecture_08/Lecture_8.2.md index 636cbf4..a2622c2 100644 --- a/docs/lecture_08/Lecture_8.2.md +++ b/docs/lecture_08/Lecture_8.2.md @@ -1,32 +1,57 @@ --- -title: Mini-Lecture 8.2 -- Disaggregation of regional data +title: Mini-Lecture 8.2 -- Energy demands in modelling keywords: -- Disaggregation -- Regions +- Energy demands +- Scenario analysis authors: - Alexander J. M. Kell --- -This mini-lecture introduces the concept of disaggregation of regions in further detail. +Mini-lecture 8.2 outlines the general requirements for defining energy demands and how modelling different scenarios can help assess potential future energy demand. # Learning objectives -- When to disaggregate regional data in MUSE and energy systems models +- Understand how to define energy demands +- Understand why we need scenario analysis +# Introduction -# Disaggregation +Within modelling we can break up the previously defined energy demands by sector. Electricity comes from the power sector and can be used to fulfil demand from each of the final service sectors. For example, the residential, commercial or industrial sector. -Disaggregation of regions can often be a good way of gaining a deeper understanding of the interactions between regions. For example, if you have a lot of technoeconomic data on the locations of supply and demand, then it may make sense to disaggregate regions. This will also allow the modeller to understand where there may be issues within a specific region or country. +These sectors can have different electricity demands and needs and which can evolve over time as was seen in the last mini-lecture. We will now explore how these energy demands can be defined. -An interesting example of this would be for the Southeast Asia region. Laos has a good amount of hydropower availability, whereas Thailand has more solar and wind resources. If we modelled the Southeast Asia region as a single region in MUSE, we would lose information on the potential for trade between these two countries. +## Defining energy demands -It is also interesting to see energy flows between regions within a country, similar to the Southeast Asian example. For example, if a country has a large demand centre in the south of the country, but large energy resources in the north, it could be interesting to disaggregate this country into those two nodes. +When defining an energy demand for energy systems models, it is important to identify the following: -Similar to the previous mini-lecture, this disaggregation is largely dependent upon your requirements and the data available to you. There is no one solution for all areas, or even for the same area and different case studies. For example, one case study may only require the modelling of a country as a single region. Another case study, however, may require the modelling of that same country by many regions. It all depends on the question you are trying to answer and the data available to you. It must be noted, that a more disaggregated case study will take longer to run in MUSE. +- The energy carrier which the demand arises for. For example, electricity, gasoline for transportation or biomass for cooking. +- The sector the demand arises from. For example, residential (urban and/or rural, off- or on-grid), industrial or commercial. +- The average variability of the demand within a year. This is usually expressed using average demand profiles, which are explained in more detail later in this lecture. +- The current and expected future annual average demand. + +However, it is very difficult to predict future demand, and there will always be uncertainty in our predictions. Due to this it is important to model different scenarios. + +## Defining our own energy demand + +As has just been seen, when we want to define our own energy demand, we need to identify a number of different features. Let's say, for example, that we want to define the demand for electricity in urban homes. To do this, we need to define: + +- The energy carrier for which the demand arises for. In this case it is electricity. +- The sector the demand arises from. In this example it is the residential sector, or the urban residential sector if you would like to be more specific. +- The average variability of the demand over the year. In this example we can look at daily and yearly electricity demand profiles for a residential urban area. This will tell us how the demand varies on a daily and seasonal scale. +- Current and predicted future demand. For this, we can look at an energy balance (covered in more detail later) to get data for the current and historical residential electricity demand. We can use these data as a baseline, and we could combine it with an estimate of population growth to create a future projection for the demand. + +## Scenario analysis + +Within energy systems modelling, we must explore different possibilities of what could happen in the future. This is known as scenario analysis. We do this as the future is uncertain, particularly over the long-term horizon. We therefore might want to consider multiple scenarios to assess how demand could vary in the future. + +For example, for different scenarios, key predictors of energy demand, such as population growth, economic development and energy policy can be varied across the scenarios. This would mean that each scenario has a different energy demand projection. + +Since we can not be certain of the scenario which will be the best predictor of the future, it is useful to model several scenarios and consider the implications of each of them to give useful insights for policymaking. This allows policy makers to assess which of the different policies and mixes suit their needs based upon likelihoods and risk tolerances. # Summary -In this mini-lecture we explored reasons for disaggregating a case study. We discovered that disaggregation (and aggregation) of regions depends largely on the data available to you and the questions you want to answer for your case study. However, we found out that the greater the disaggregation, the more detail the model may reveal, but the longer the model will take to run. +Mini-lecture 8.2 provided an overview of energy demands, how we can define them and the details which make them up. We also explored how we can perform scenario analysis with energy demands, to understand what could happen in the future. + + -  diff --git a/docs/lecture_08/Lecture_8.3.md b/docs/lecture_08/Lecture_8.3.md index 4939b37..87002e9 100644 --- a/docs/lecture_08/Lecture_8.3.md +++ b/docs/lecture_08/Lecture_8.3.md @@ -1,40 +1,62 @@ --- -title: Mini-Lecture 8.3 -- Communicating research +title: Mini-Lecture 8.3 -- Energy demand in MUSE keywords: -- Science communication -- Visualisation +- Energy demand +- MUSE authors: -- Alexander J. M. Kell +- Alexander J. M. Kell --- -In this mini-lecture, we will explore the different ways that research can be communicated effectively. +## Short description -# Learning objectives +Following mini-lecture 8.2, this mini-lecture provides an insight into how to model service demand within MUSE. There are two possible methods to model service demand in MUSE, from user input and by correlation. In this mini-lecture we will learn what the difference is between these. -- Understand how to communicate the results of your research to influence policy development +## Learning objectives -# Effective communication of research +- Understand how to input exogenous service demand +- Understand what service demand by correlation is -Throughout this course, we have explored the useful insights and analysis that can be provided by energy systems models. We have also explored the types of results that can lead to changes in the planning of energy systems, for example by taking a more holistic approach to investment planning. +# Lecture content -However, it is important that these results are communicated effectively to ensure that decision makers can fully understand the implications of these results. Effective communication also allows the methodology of the study to be better understood, which allows for the positives and limitations of the model to be explored. +## Service Demand -## Presenting figures +A service demand is a term used to describe the consumption of energy by human activity. This could be, for instance, energy for lighting or cooking in the residential sector, personal vehicles in the transportation sector or machine usage in the industrial sector. The service demand drives the entire energy system, and it influences the total amount of energy used, the location of use and the types of fuels used in the energy supply system. It also includes the characteristics of the end-use technologies that consume energy. -It is crucial to present figures in an understandable way. Figures are often the first thing that the audience will look at and try to understand. Figures can be used to convey the key results from your study in an impactful way. There are therefore some things that should be considered. +## Exogenous service demand -The first of these is to design the figure with the target audience in mind. For example, if the audience is made of non-specialists, it may be sensible to ensure figures focus on the message without lots of technical jargon. For any audience, it is important that they understand the content of the figure, and so it is important to always include a figure caption, a legend (explaining colour coding and any symbols) and axis titles where appropriate. Finally, the colours chosen can have a large impact and so the colours should be chosen carefully with sufficient distinction between the colours. +Within MUSE we must set the energy demand exogenously. That means that the model does not calculate how much the service demand is. Effectively, this means that the user must make an assumption on how much electricity is consumed in, for example, the residential sector for a particular region in the model. -## Common mistakes +We can change this per scenario, but these values will not change during a simulation run, even if the price for all fuels increases significantly, for instance. We are able to define the exogenous service demand by year, sector, region and timeslice. -This section focuses on the commonly made mistakes when presenting figures in research. It can often be the case that figures are too confusing and contain too much data. This can often result in the message of the figure being unclear. It may be the case that by confusing your audience you reduce the impact of your research findings. Therefore, it is advisable to make figures as simple as possible to ensure that they are understandable. +## Service demand by correlation. -Other common mistakes include: -- The use of inappropriate axis for graphs which can distort results -- Lack of figure captions, axis titles, labels or legends +In the previous section we learnt about the exogenous service demand. That is, we can explicitly specify what the demand would be per year, sector, region and timeslice. However, it may be the case that we do not know what the electricity demand is per year, especially in the future. We may instead conclude that our electricity demand is a function of the GDP and population of a particular region, as previously discussed. + +To accommodate such a scenario, MUSE enables us to choose a regression function that estimates service demands from GDP and population projections, which may be more predictable or have more accessible data in your case. A regression function is simply a mathematical model which fits a linear model to your data to predict what may happen in the future. + +## Sources for energy demand data + +We can get publicly available energy balance data and/or demand projections from the following sources: + +- International Energy Agency +- International Renewable Energy Agency +- United Nations Statistics +- Asia-Pacific Economic Cooperation + +Energy balances tell us the amount that each energy commodity is used in a country or region in a given year. This is usually broken down by sector. + + +## Summary + +In this mini-lecture we introduced service demands, and the way we can input these into MUSE. The two ways we can input service demands are: +- Exogenous service demand +- Service demand by correlation + +We also learned where we can get energy data from for various countries. + +In the hands-on we will see how we can actually do this within MUSE. -# Summary -In this mini-lecture we explored the different ways that we can communicate our research for maximum impact and ways to make figures understandable to our target audience.   + diff --git a/docs/lecture_08/Lecture_8.4.md b/docs/lecture_08/Lecture_8.4.md index 3de3018..3af0cf8 100644 --- a/docs/lecture_08/Lecture_8.4.md +++ b/docs/lecture_08/Lecture_8.4.md @@ -1,50 +1,69 @@ --- -title: Mini-Lecture 8.4 -- Oral presentations +title: "Mini-Lecture 8.4 -- Demand examples and units" keywords: -- +- Infrastructure performance authors: -- Alexander J. M. Kell +- Alexander J. M. Kell --- -In this mini-lecture we will focus on effective oral communication of research. +# Short description + +Mini-lecture 8.4 explains how we can use timeslices to approximate the real-world demand profile. We will look into the difference between power and energy. Finally, we will learn how to convert units to ensure we are consistent within MUSE. # Learning objectives -- Implement tips for improved oral presentations to influence policy development +- Understand how timeslices can be used in the context of demand +- Understand the difference between power and energy +- Know the units to use within MUSE and how to convert these + +# Demand profile + +Figure 8.1.5 shown an example demand profile for electricity that could be used in MUSE. In this demand profile there are 96 bars: one for each of the timeslices used in MUSE. These timeslices are split into 16 different sections – seasonal and into day and night. This is because there are four different seasons, which are split into day and night (twice). The demand profile is used to represent the proportion of demand occurring in each timeslice. + +![](assets/Figure_8.1.5.png){width=100%} + +**Figure 8.4.1:** Example demand profile for MUSE + +The chart shows us that electricity demand, in this example, is highest during the day in winter, while it is lowest during the night in spring. However, it is important to note that this is a simplification: in reality demand varies in the season and with each hour of the day. This simplification means that we model one representative day for each season, and we assume equal demand within days and nights of those seasons. -# Key features of presentations +Whilst this is a simplification, it allows us to consider the variation in demand across seasons and days without having an incredibly complex model structure. This reduces the amount of time required to run a full model relative to having timeslices for each hour and day of the year, as well as reducing the data input requirements. -The key features of presentations are: +## Units -- Entry point: capture the audience's attention -- Aim: focus on what you want to achieve with the presentation -- Structure: ensure consistency across the slides and tell a coherent story from beginning to end -- Audience: plan for your audience and their background -- Impact: identify key take-home points that the audience should remember +We must ensure that during our data input process we are consistent with our units. Usually we will use the petajoules unit as this is the unit for energy for different sectors. If you were just modelling the power sector, you could use megawatt hours. -Firstly, there is the entry point of the presentation. It is important to focus the audience's attention. This ensures that they are interested in the presentation and understand what will be presented. This could take the form of presenting a question that you know will interest your audience, and telling them that by the end of the presentation they will know the answer. +## Power vs. Energy -Throughout the presentation it is important to have the aim in mind. For example, you could be trying to increase engagement with a new department. For example, if you wish to demonstrate the advantages and disadvantages of building a new coal-power plant in a particular country, the figures and data you present should be focused on this particular situation, rather than providing information about scenarios that are not affected by a new coal power plant. +When using energy modelling tools it is important to remember the difference between power and energy. Sometimes these terms are used interchangeably. However, there is an important difference between the two: -The structure of the presentation can be tailored to your aim. It is important to have a clear beginning, middle and end. There should be consistency across the presentation to maximise the audience's understanding. +- Energy is the total amount of work done or the total capacity for doing work +- Power is the rate at which this energy is supplied or used. -To further ensure that the audience understand and engage with the presentation, it should be designed with the audience's backgrounds and motivations in mind (see more below). +Therefore, energy and power have different units. For example, energy is often measured in Joules, while power is often measured in Joules per Second (or Watts). -Finally, it is important to consider the impact of the presentation and identify key points or policy recommendations that you would like the audience to remember. +For example, providing the weight stays the same, lifting a weight requires the exact same amount of energy no matter how quickly we lift it. However, if we lift the weight more quickly, the power has increased. We used the same amount of energy, but over a shorter amount of time. -## Audiences +## Units for demand -It is important to understand the types of audience that you will be presenting to. For instance, they may be generalists or non-specialists. Or they could be scientists from different disciplines, or even scientists from the same discipline, but focusing on different topics. +It is important that we convert our data from different sources to petajoules (PJ) when we include it in MUSE. -The presentation should be adapted depending on your audience in order to increase the audience's understanding and engagement. Technical content, for example, can be explained in a simple and understandable manner if the audience contains non-specialists. If you think that your audience, on the other hand, will have technical expertise, you can spend less time on explaining technical content. The amount of technical detail you provide may also change: if you are speaking to a policymaker they may be more interested in the results and recommendations than the modelling process. +Here are some example conversion factors: -The purpose of the presentation should be optimised throughout. For example, if you are aiming to create a partnership with a new department, the presentation should have a focus on the implications of your research for that department and the benefits of the proposed partnership for the audience. +- 1 Petajoule (PJ) = 1000 Terajoules (TJ) +- 1 Petajoule (PJ) = 1,000,000 Gigajoules (GJ) +- 3.6 Petajoules (PJ) = 1 Terawatt hour (TWh) +- 0.0036 Petajoules (PJ) = 1 Gigawatt hour (GWh) +We must ensure that we are consistent with the units we use within MUSE. # Summary -In this mini-lecture we introduced some key tips for oral presentations. We explored why understanding your audience of importance, especially when introducing technical content. We also learnt that we can be strategic in our presentation planning and should optimise for the aims for which we want to achieve. +In this lecture we have learnt the difference between power and energy. We have also learnt how to use timeslicing to speed up our model and reduce complexity. Finally, we learnt that we must use consistent units. + + + + + + -This is the final lecture of the Agent-based energy systems modelling: MUSE course. After this lecture you should be in a good position to develop your own models through MUSE, which can then be used to assess the impact of different policy options. -Thank you for engaging with this course, and we hope you have enjoyed the lectures, found them valuable, and find practical uses for MUSE in your research. diff --git a/docs/lecture_08/assets/Figure_8.1.1.jpeg b/docs/lecture_08/assets/Figure_8.1.1.jpeg deleted file mode 100644 index 0de3d2c..0000000 Binary files a/docs/lecture_08/assets/Figure_8.1.1.jpeg and /dev/null differ diff --git a/docs/lecture_08/assets/Figure_8.1.1.png b/docs/lecture_08/assets/Figure_8.1.1.png new file mode 100644 index 0000000..486c983 Binary files /dev/null and b/docs/lecture_08/assets/Figure_8.1.1.png differ diff --git a/docs/lecture_03/assets/Figure_3.1.2.png b/docs/lecture_08/assets/Figure_8.1.2.png similarity index 100% rename from docs/lecture_03/assets/Figure_3.1.2.png rename to docs/lecture_08/assets/Figure_8.1.2.png diff --git a/docs/lecture_03/assets/Figure_3.1.3.png b/docs/lecture_08/assets/Figure_8.1.3.png similarity index 100% rename from docs/lecture_03/assets/Figure_3.1.3.png rename to docs/lecture_08/assets/Figure_8.1.3.png diff --git a/docs/lecture_03/assets/Figure_3.1.4.png b/docs/lecture_08/assets/Figure_8.1.4.png similarity index 100% rename from docs/lecture_03/assets/Figure_3.1.4.png rename to docs/lecture_08/assets/Figure_8.1.4.png diff --git a/docs/lecture_03/assets/Figure_3.1.5.png b/docs/lecture_08/assets/Figure_8.1.5.png similarity index 100% rename from docs/lecture_03/assets/Figure_3.1.5.png rename to docs/lecture_08/assets/Figure_8.1.5.png diff --git a/docs/lecture_08/assets/Figure_8.2.1.png b/docs/lecture_08/assets/Figure_8.2.1.png deleted file mode 100644 index ba5ab8f..0000000 Binary files a/docs/lecture_08/assets/Figure_8.2.1.png and /dev/null differ diff --git a/docs/lecture_08/assets/Figure_8.2.2.png b/docs/lecture_08/assets/Figure_8.2.2.png deleted file mode 100644 index 2053c70..0000000 Binary files a/docs/lecture_08/assets/Figure_8.2.2.png and /dev/null differ diff --git a/docs/lecture_08/assets/Figure_8.2.3.png b/docs/lecture_08/assets/Figure_8.2.3.png deleted file mode 100644 index 92b2694..0000000 Binary files a/docs/lecture_08/assets/Figure_8.2.3.png and /dev/null differ diff --git a/docs/lecture_08/assets/Figure_8.2.4.png b/docs/lecture_08/assets/Figure_8.2.4.png deleted file mode 100644 index d41a1e8..0000000 Binary files a/docs/lecture_08/assets/Figure_8.2.4.png and /dev/null differ diff --git a/docs/lecture_08/assets/Figure_8.3.1.png b/docs/lecture_08/assets/Figure_8.3.1.png deleted file mode 100644 index f9518f9..0000000 Binary files a/docs/lecture_08/assets/Figure_8.3.1.png and /dev/null differ diff --git a/docs/lecture_08/assets/Figure_8.3.2.png b/docs/lecture_08/assets/Figure_8.3.2.png deleted file mode 100644 index b83fab0..0000000 Binary files a/docs/lecture_08/assets/Figure_8.3.2.png and /dev/null differ diff --git a/docs/lecture_08/assets/Figure_8.4.1.png b/docs/lecture_08/assets/Figure_8.4.1.png deleted file mode 100644 index 14eae78..0000000 Binary files a/docs/lecture_08/assets/Figure_8.4.1.png and /dev/null differ diff --git a/docs/lecture_08/assets/Figure_8.4.2.png b/docs/lecture_08/assets/Figure_8.4.2.png deleted file mode 100644 index 9643edf..0000000 Binary files a/docs/lecture_08/assets/Figure_8.4.2.png and /dev/null differ diff --git a/docs/lecture_08/assets/Figure_8.4.3.png b/docs/lecture_08/assets/Figure_8.4.3.png deleted file mode 100644 index 306293d..0000000 Binary files a/docs/lecture_08/assets/Figure_8.4.3.png and /dev/null differ diff --git a/docs/lecture_08/assets/Figure_8.4.4.png b/docs/lecture_08/assets/Figure_8.4.4.png deleted file mode 100644 index 0ddf2a0..0000000 Binary files a/docs/lecture_08/assets/Figure_8.4.4.png and /dev/null differ diff --git a/docs/lecture_08/bibliography.bib b/docs/lecture_08/bibliography.bib index 178e9fd..16a3e17 100644 --- a/docs/lecture_08/bibliography.bib +++ b/docs/lecture_08/bibliography.bib @@ -1,274 +1,24 @@ -@article{Bloemendaal2019storm, - author = {Nadia Bloemendaal and Ivan Haigh and Hans {de Moel} and S. Muis and Reindert Haarsma and Jeroen Aerts}, - date-added = {2021-08-13 12:34:08 +0200}, - date-modified = {2021-08-13 12:34:14 +0200}, - doi = {10.4121/uuid:82c1dc0d-5485-43d8-901a-ce7f26cda35d}, - month = {11}, - title = {{STORM IBTrACS present climate synthetic tropical cyclone tracks}}, - url = {https://data.4tu.nl/articles/dataset/STORM_IBTrACS_present_climate_synthetic_tropical_cyclone_tracks/12706085}, - year = {2019}, - Bdsk-Url-1 = {https://data.4tu.nl/articles/dataset/STORM_IBTrACS_present_climate_synthetic_tropical_cyclone_tracks/12706085}, - Bdsk-Url-2 = {https://doi.org/10.4121/uuid:82c1dc0d-5485-43d8-901a-ce7f26cda35d}} - -@report{Dawson2016, - author = {R.J. Dawson and D. Thompson and D. Johns and S. Gosling and L. Chapman and G. Darch and G. Watson and W. Powrie and S. Bell and K. Paulson and P. Hughes and R. Wood}, - city = {London}, - institution = {Adaptation Sub-Committee for the Committee on Climate Change}, - publisher = {Paul Hughes}, - title = {UK Climate Change Risk Assessment Evidence Report: Chapter 4: Infrastructure}, - year = {2016}} - -@article{Steinbuks2010, - abstract = {This paper attempts to identify the underlying causes and costs of own generation of electric power in Africa. Rigorous empirical analysis of 8483 currently operating firms in 25 African countries shows that the prevalence of own generation would remain high (at around 20%) even if power supplies were perfectly reliable, suggesting that other factors such as firms' size, emergency back-up and export regulations play a critical role in the decision to own a generator. The costs of own-generation are about three times as high as the price of purchasing (subsidized) electricity from the public grid. However, because these generators only operate a small fraction of the time, they do not greatly affect the overall average cost of power to industry. The benefits of generator ownership are also substantial. Firms with their own generators report a value of lost load of less than US$50 per hour, compared with more than US$150 per hour for those without. Nevertheless, when costs and benefits are considered side by side, the balance is not found to be significantly positive. {\copyright} 2009 Elsevier B.V.}, - author = {J. Steinbuks and V. Foster}, - doi = {10.1016/j.eneco.2009.10.012}, - issn = {01409883}, - issue = {3}, - journal = {Energy Economics}, - keywords = {Africa,Electricity,Generators,Ownership,Reliability}, - month = {5}, - pages = {505-514}, - publisher = {North-Holland}, - title = {When do firms generate? Evidence on in-house electricity supply in Africa}, - volume = {32}, - year = {2010}, - Bdsk-Url-1 = {https://doi.org/10.1016/j.eneco.2009.10.012}} - -@generic{Hall2015, - abstract = {The impacts of extreme events are triggering action and reaction --- sometimes in unexpected ways. Confronted by 'adaptation emergencies', the private sector is rapidly innovating climate risk management, but governments must also fulfil their responsibilities.}, - author = {Jim W. Hall and Frans Berkhout and Rowan Douglas}, - doi = {10.1038/nclimate2467}, - issn = {17586798}, - issue = {1}, - journal = {Nature Climate Change}, - keywords = {Business and industry,Governance,Government}, - month = {12}, - pages = {6-7}, - publisher = {Nature Publishing Group}, - title = {Responding to adaptation emergencies}, - url = {http://dx.doi.org/10.5270/OceanObs09-FOO}, - volume = {5}, - year = {2015}, - Bdsk-Url-1 = {http://dx.doi.org/10.5270/OceanObs09-FOO}, - Bdsk-Url-2 = {http://dx.doi.org/10.1038/nclimate2467}} - -@webpage{Loew2019, - author = {Petra Loew}, - title = {The natural disasters of 2018 in figures | Munich Re Topics Online}, - url = {https://www.munichre.com/topics-online/en/climate-change-and-natural-disasters/natural-disasters/the-natural-disasters-of-2018-in-figures.html}, - year = {2019}, - Bdsk-Url-1 = {https://www.munichre.com/topics-online/en/climate-change-and-natural-disasters/natural-disasters/the-natural-disasters-of-2018-in-figures.html}} - -@article{Thacker2019, - abstract = {Infrastructure systems form the backbone of every society, providing essential services that include energy, water, waste management, transport and telecommunications. Infrastructure can also create harmful social and environmental impacts, increase vulnerability to natural disasters and leave an unsustainable burden of debt. Investment in infrastructure is at an all-time high globally, thus an ever-increasing number of decisions are being made now that will lock-in patterns of development for future generations. Although for the most part these investments are motivated by the desire to increase economic productivity and employment, we find that infrastructure either directly or indirectly influences the attainment of all of the Sustainable Development Goals (SDGs), including 72% of the targets. We categorize the positive and negative effects of infrastructure and the interdependencies between infrastructure sectors. To ensure that the right infrastructure is built, policymakers need to establish long-term visions for sustainable national infrastructure systems, informed by the SDGs, and develop adaptable plans that can demonstrably deliver their vision.}, - author = {Scott Thacker and Daniel Adshead and Marianne Fay and St{\'e}phane Hallegatte and Mark Harvey and Hendrik Meller and Nicholas O'Regan and Julie Rozenberg and Graham Watkins and Jim W. Hall}, - doi = {10.1038/s41893-019-0256-8}, - issn = {23989629}, - issue = {4}, - journal = {Nature Sustainability}, - keywords = {Civil engineering,Developing world,Sustainability}, - month = {4}, - pages = {324-331}, - publisher = {Nature Publishing Group}, - title = {Infrastructure for sustainable development}, - url = {https://doi.org/10.1038/s41893-019-0256-8}, - volume = {2}, - year = {2019}, - Bdsk-Url-1 = {https://doi.org/10.1038/s41893-019-0256-8}} - -@article{Koks2019, - abstract = {Transport infrastructure is exposed to natural hazards all around the world. Here we present the first global estimates of multi-hazard exposure and risk to road and rail infrastructure. Results reveal that ~27% of all global road and railway assets are exposed to at least one hazard and ~7.5% of all assets are exposed to a 1/100 year flood event. Global Expected Annual Damages (EAD) due to direct damage to road and railway assets range from 3.1 to 22 billion US dollars, of which ~73% is caused by surface and river flooding. Global EAD are small relative to global GDP (~0.02%). However, in some countries EAD reach 0.5 to 1% of GDP annually, which is the same order of magnitude as national transport infrastructure budgets. A cost-benefit analysis suggests that increasing flood protection would have positive returns on ~60% of roads exposed to a 1/100 year flood event.}, - author = {E. E. Koks and J. Rozenberg and C. Zorn and M. Tariverdi and M. Vousdoukas and S. A. Fraser and J. W. Hall and S. Hallegatte}, - doi = {10.1038/s41467-019-10442-3}, - issn = {20411723}, - issue = {1}, - journal = {Nature Communications}, - keywords = {Environmental impact,Natural hazards}, - month = {12}, - pages = {1-11}, - pmid = {31239442}, - publisher = {Nature Publishing Group}, - title = {A global multi-hazard risk analysis of road and railway infrastructure assets}, - url = {https://doi.org/10.1038/s41467-019-10442-3}, - volume = {10}, - year = {2019}, - Bdsk-Url-1 = {https://doi.org/10.1038/s41467-019-10442-3}} - -@report{Hall2019, - author = {Jim W. Hall and Jeroen C.J.H. Aerts and Bilal M. Ayyub and Stephane Hallegatte and Mark Harvey and Xi Hu and Elco Koks and Caroline Lee and Xiawei Liao and Michael Mullan and Raghav Pant and Amelie Paszkowski and Julie Rozenberg and Fulai Sheng and Vladimir Stenek and Scott Thacker and Elina Vaananen and Lola Vallejo and Ted I.E. Veldkamp and Michelle van Vliet and Yoshihide Wada and Philip Ward and Graham Watkins and Conrad Zorn}, - institution = {Global Commission on Adapatation}, - title = {Adaptation of Infrastructure Systems}, - url = {https://gca.org/reports/adaptation-of-infrastructure-systems/}, - year = {2019}, - Bdsk-Url-1 = {https://gca.org/reports/adaptation-of-infrastructure-systems/}} - -@article{THACKER201730, - abstract = {The complex and interdependent nature of modern critical national infrastructures provides the conditions for which localized failures can propagate within and between network systems, resulting in disruptions that are widespread and often unforeseen. Within this study, we characterize critical national infrastructures as a system-of-systems and develop methodology to perform a multi-scale disruption analysis. To achieve this, we map functional pathways between network source and sink assets across a range of operational scales. Customer demands are attributed to these pathways and are used to build a weighted network. The resultant functional path set and weighted network are used to perform a disruption analysis that encodes information on the long-range functionality within and between infrastructures, providing insights into failure propagation and the functional dependencies that exist between assets from multiple sectors. We supplement the methodological development with a detailed national scale demonstration for England and Wales using a unique representation of the integrated electricity network and the domestic flight network. The results highlight the potentially large disruptions that can result from the failure of individual electricity assets from a range of different sub-systems.}, - author = {Scott Thacker and Raghav Pant and Jim W. Hall}, - doi = {https://doi.org/10.1016/j.ress.2017.04.023}, - issn = {0951-8320}, - journal = {Reliability Engineering & System Safety}, - keywords = {Infrastructure, Network, Interdependence, System-of-system, Multi-scale, Disruption analysis}, - note = {Special Section: Applications of Probabilistic Graphical Models in Dependability, Diagnosis and Prognosis}, - pages = {30-41}, - title = {System-of-systems formulation and disruption analysis for multi-scale critical national infrastructures}, - url = {https://www.sciencedirect.com/science/article/pii/S0951832017304994}, - volume = {167}, - year = {2017}, - Bdsk-Url-1 = {https://www.sciencedirect.com/science/article/pii/S0951832017304994}, - Bdsk-Url-2 = {https://doi.org/10.1016/j.ress.2017.04.023}} - -@article{Rinaldi2001, - author = {S. M. Rinaldi and J. P. Peerenboom and T. K. Kelly}, - doi = {10.1109/37.969131}, - journal = {IEEE Control Systems Magazine}, - number = {6}, - pages = {11-25}, - title = {Identifying, understanding, and analyzing critical infrastructure interdependencies}, - volume = {21}, - year = {2001}, - Bdsk-Url-1 = {https://doi.org/10.1109/37.969131}} - -@article{mann2009strategic, - author = {Mann, B}, - journal = {Cabinet Office, London< http://www.cabinetoffice.gov.uk/media/308367/sfps-consultation.pdf}, - title = {Strategic Framework and Policy Statement on Improving the Resilience of Critical Infrastructure to Disruption from Natural Hazards}, - year = {2009}} - -@book{Hall2016, - address = {Cambridge}, - author = {Hall, J.W. and Tran, M. and Hickford, A. and Nicholls, R.}, - publisher = {Cambridge University Press}, - title = {{The future of national infrastructure: A systems-of-systems approach}}, - year = {2016}} - -@book{field2014climate, - author = {Field, Christopher B and Barros, Vicente R}, - publisher = {Cambridge University Press}, - title = {Climate change 2014--Impacts, adaptation and vulnerability: Regional aspects}, - year = {2014}} - -@incollection{oppenheimer2015emergent, - author = {Oppenheimer, Michael and Campos, Maximiliano and Warren, Rachel and Birkmann, Joern and Luber, George and O'Neill, Brian and Takahashi, Kiyoshi and Brklacich, Mike and Semenov, Sergey and Licker, Rachel and others}, - booktitle = {Climate Change 2014 Impacts, Adaptation and Vulnerability: Part A: Global and Sectoral Aspects}, - pages = {1039--1100}, - publisher = {Cambridge University Press}, - title = {Emergent risks and key vulnerabilities}, - year = {2015}} - -@article{pant2018critical, - author = {Pant, Raghav and Thacker, Scott and Hall, Jim W and Alderson, David and Barr, Stuart}, - journal = {Journal of Flood Risk Management}, - number = {1}, - pages = {22--33}, - publisher = {Wiley Online Library}, - title = {Critical infrastructure impact assessment due to flood exposure}, - volume = {11}, - year = {2018}} - -@article{watts2002simple, - author = {Watts, Duncan J}, - journal = {Proceedings of the National Academy of Sciences}, - number = {9}, - pages = {5766--5771}, - publisher = {National Acad Sciences}, - title = {A simple model of global cascades on random networks}, - volume = {99}, - year = {2002}} - -@article{PANT2014183, - abstract = {While early research efforts were devoted to the protection of systems against disruptive events, be they malevolent attacks, man-made accidents, or natural disasters, recent attention has been given to the resilience, or the ability of systems to ``bounce back,'' of these events. Discussed here is a modeling paradigm for quantifying system resilience, primarily as a function of vulnerability (the adverse initial system impact of the disruption) and recoverability (the speed of system recovery). To account for uncertainty, stochastic measures of resilience are introduced, including Time to Total System Restoration, Time to Full System Service Resilience, and Time to α%-Resilience. These metrics are applied to quantify the resilience of inland waterway ports, important hubs in the flow of commodities, and the port resilience approach is deployed in a data-driven case study for the inland Port of Catoosa in Oklahoma. The contributions herein demonstrate a starting point in the development of a resilience decision making framework.}, - author = {Raghav Pant and Kash Barker and Jose Emmanuel Ramirez-Marquez and Claudio M. Rocco}, - doi = {https://doi.org/10.1016/j.cie.2014.01.017}, - issn = {0360-8352}, - journal = {Computers & Industrial Engineering}, - keywords = {Resilience, Infrastructure systems, Vulnerability, Recoverability}, - pages = {183-194}, - title = {Stochastic measures of resilience and their application to container terminals}, - url = {https://www.sciencedirect.com/science/article/pii/S0360835214000333}, - volume = {70}, - year = {2014}, - Bdsk-Url-1 = {https://www.sciencedirect.com/science/article/pii/S0360835214000333}, - Bdsk-Url-2 = {https://doi.org/10.1016/j.cie.2014.01.017}} - -@article{HOSSEINI201647, - abstract = {Modeling and evaluating the resilience of systems, potentially complex and large-scale in nature, has recently raised significant interest among both practitioners and researchers. This recent interest has resulted in several definitions of the concept of resilience and several approaches to measuring this concept, across several application domains. As such, this paper presents a review of recent research articles related to defining and quantifying resilience in various disciplines, with a focus on engineering systems. We provide a classification scheme to the approaches in the literature, focusing on qualitative and quantitative approaches and their subcategories. Addressed in this review are: an extensive coverage of the literature, an exploration of current gaps and challenges, and several directions for future research.}, - author = {Seyedmohsen Hosseini and Kash Barker and Jose E. Ramirez-Marquez}, - doi = {https://doi.org/10.1016/j.ress.2015.08.006}, - issn = {0951-8320}, - journal = {Reliability Engineering & System Safety}, - keywords = {Resilience, Engineering systems}, - pages = {47-61}, - title = {A review of definitions and measures of system resilience}, - url = {https://www.sciencedirect.com/science/article/pii/S0951832015002483}, - volume = {145}, - year = {2016}, - Bdsk-Url-1 = {https://www.sciencedirect.com/science/article/pii/S0951832015002483}, - Bdsk-Url-2 = {https://doi.org/10.1016/j.ress.2015.08.006}} - -@article{hickford2018resilience, - author = {Hickford, Adrian J and Blainey, Simon P and Hortelano, Alejandro Ortega and Pant, Raghav}, - journal = {Environment Systems and Decisions}, - number = {3}, - pages = {278--291}, - publisher = {Springer}, - title = {Resilience engineering: theory and practice in interdependent infrastructure systems}, - volume = {38}, - year = {2018}} - -@article{trigg2016credibility, - author = {Trigg, MA and Birch, CE and Neal, JC and Bates, PD and Smith, A and Sampson, CC and Yamazaki, D and Hirabayashi, Y and Pappenberger, F and Dutra, E and others}, - journal = {Environmental Research Letters}, - number = {9}, - pages = {094014}, - publisher = {IOP Publishing}, - title = {The credibility challenge for global fluvial flood risk analysis}, - volume = {11}, - year = {2016}} - -@article{wing2020toward, - author = {Wing, Oliver EJ and Quinn, Niall and Bates, Paul D and Neal, Jeffrey C and Smith, Andrew M and Sampson, Christopher C and Coxon, Gemma and Yamazaki, Dai and Sutanudjaja, Edwin H and Alfieri, Lorenzo}, - journal = {Water Resources Research}, - number = {8}, - pages = {e2020WR027692}, - publisher = {Wiley Online Library}, - title = {Toward Global Stochastic River Flood Modeling}, - volume = {56}, - year = {2020}} - -@article{gassert2015aqueduct, - author = {Gassert, Francis and Reig, Paul and Shiao, Tien and Landis, Matt and Luck, Matt and others}, - title = {Aqueduct global maps 2.1}, - year = {2015}} - -@misc{oh2019addressing, - author = {Oh, Jung Eun and Espinet Alegre, Xavier and Pant, Raghav and Koks, Elco E and Russell, Tom and Schoenmakers, Roald and Hall, Jim W}, - publisher = {World Bank}, - title = {Addressing Climate Change in Transport: Volume 2: Pathway to Resilient Transport}, - year = {2019}} - -@article{pant2016vulnerability, - author = {Pant, Raghav and Hall, Jim W and Blainey, Simon P}, - journal = {European Journal of Transport and Infrastructure Research}, - number = {1}, - title = {Vulnerability assessment framework for interdependent critical infrastructures: case-study for Great Britain's rail network}, - volume = {16}, - year = {2016}} - -@article{jafino2020transport, - author = {Jafino, Bramka Arga and Kwakkel, Jan and Verbraeck, Alexander}, - journal = {Transport Reviews}, - number = {2}, - pages = {241--264}, - publisher = {Taylor \& Francis}, - title = {Transport network criticality metrics: a comparative analysis and a guideline for selection}, - volume = {40}, - year = {2020}} - -@manual{QGIS_software, - author = {{QGIS Development Team}}, - organization = {QGIS Association}, - title = {QGIS Geographic Information System}, - url = {https://www.qgis.org}, - year = {2021}, - Bdsk-Url-1 = {https://www.qgis.org}} +@article{Taliotis2018, + abstract = {Defining final energy demands }, + author = {Taliotis, Constantinos and Gardumi, Francesco and Shivakumar, Abhishek and Sridharan, Vignesh and Ramos, Eunice and Beltramo, Agnese and Rogner, Holger and Howells, Mark}, + file = {:Users/alexanderkell/Downloads/Defining final energy demands in OSeMOSYS.pdf:pdf}, + keywords = {Demand,Energy,Energy System,Modelling,Osemosys}, + number = {January}, + title = {{Defining final energy demands in OSeMOSYS}}, + year = {2018} +} + +@article{Olaniyan2018, + abstract = {Considering the challenge of accessing reliable household metering data in Nigeria, how can electricity consumption levels be determined? And how do disparities in electricity consumption patterns across the country affect the pursuit of sustainability, universal access and energy transition objectives? This study combined household-reported data on ownership of electrical appliances and energy expenditure with online sales records of household appliances to estimate current and future residential electricity demand in Nigeria, as well as the required generation capacity to achieve 100% electricity access, under various scenarios. Median residential electricity consumption was estimated at 18-27 kWh per capita but these estimates vary between the geographical zones with the North East and SouthWest representing extremes. Under a universal access scenario, the future electricity supply system would be expected to have installed generation capacity sufficient to meet the estimated residential demand of 85 TWh. To further understand the required infrastructure investment as a whole and the approaches that might be preferred in rural versus urban areas, the disaggregated, zone-by-zone and urban/rural data may offer more insight than a whole-of-country approach. The data obtained is useful for identifying specific transitions at the sub-national level that can minimize the required investment while maximizing households' energy access.}, + author = {Olaniyan, Kayode and McLellan, Benjamin C. and Ogata, Seiichi and Tezuka, Tetsuo}, + doi = {10.3390/su10051440}, + file = {:Users/alexanderkell/Downloads/sustainability-10-01440.pdf:pdf}, + isbn = {8180378314}, + issn = {20711050}, + journal = {Sustainability (Switzerland)}, + keywords = {Electricity access,Energy transition,Household survey,Nigeria,Sustainability}, + number = {5}, + title = {{Estimating residential electricity consumption in Nigeria to support energy transitions}}, + volume = {10}, + year = {2018} +} diff --git a/docs/lecture_09/Lecture_9.1.md b/docs/lecture_09/Lecture_9.1.md new file mode 100644 index 0000000..f238c04 --- /dev/null +++ b/docs/lecture_09/Lecture_9.1.md @@ -0,0 +1,40 @@ +--- +title: Mini-Lecture 9.3 -- Communicating research +keywords: +- Science communication +- Visualisation +authors: +- Alexander J. M. Kell +--- + +In this mini-lecture, we will explore the different ways that research can be communicated effectively. + +# Learning objectives + +- Understand how to communicate the results of your research to influence policy development + +# Effective communication of research + +Throughout this course, we have explored the useful insights and analysis that can be provided by energy systems models. We have also explored the types of results that can lead to changes in the planning of energy systems, for example by taking a more holistic approach to investment planning. + +However, it is important that these results are communicated effectively to ensure that decision makers can fully understand the implications of these results. Effective communication also allows the methodology of the study to be better understood, which allows for the positives and limitations of the model to be explored. + +## Presenting figures + +It is crucial to present figures in an understandable way. Figures are often the first thing that the audience will look at and try to understand. Figures can be used to convey the key results from your study in an impactful way. There are therefore some things that should be considered. + +The first of these is to design the figure with the target audience in mind. For example, if the audience is made of non-specialists, it may be sensible to ensure figures focus on the message without lots of technical jargon. For any audience, it is important that they understand the content of the figure, and so it is important to always include a figure caption, a legend (explaining colour coding and any symbols) and axis titles where appropriate. Finally, the colours chosen can have a large impact and so the colours should be chosen carefully with sufficient distinction between the colours. + +## Common mistakes + +This section focuses on the commonly made mistakes when presenting figures in research. It can often be the case that figures are too confusing and contain too much data. This can often result in the message of the figure being unclear. It may be the case that by confusing your audience you reduce the impact of your research findings. Therefore, it is advisable to make figures as simple as possible to ensure that they are understandable. + +Other common mistakes include: +- The use of inappropriate axis for graphs which can distort results +- Lack of figure captions, axis titles, labels or legends + + +# Summary + +In this mini-lecture we explored the different ways that we can communicate our research for maximum impact and ways to make figures understandable to our target audience. +  diff --git a/docs/lecture_09/Lecture_9.2.md b/docs/lecture_09/Lecture_9.2.md new file mode 100644 index 0000000..302b22f --- /dev/null +++ b/docs/lecture_09/Lecture_9.2.md @@ -0,0 +1,50 @@ +--- +title: Mini-Lecture 9.4 -- Oral presentations +keywords: +- +authors: +- Alexander J. M. Kell +--- + +In this mini-lecture we will focus on effective oral communication of research. + +# Learning objectives + +- Implement tips for improved oral presentations to influence policy development + +# Key features of presentations + +The key features of presentations are: + +- Entry point: capture the audience's attention +- Aim: focus on what you want to achieve with the presentation +- Structure: ensure consistency across the slides and tell a coherent story from beginning to end +- Audience: plan for your audience and their background +- Impact: identify key take-home points that the audience should remember + +Firstly, there is the entry point of the presentation. It is important to focus the audience's attention. This ensures that they are interested in the presentation and understand what will be presented. This could take the form of presenting a question that you know will interest your audience, and telling them that by the end of the presentation they will know the answer. + +Throughout the presentation it is important to have the aim in mind. For example, you could be trying to increase engagement with a new department. For example, if you wish to demonstrate the advantages and disadvantages of building a new coal-power plant in a particular country, the figures and data you present should be focused on this particular situation, rather than providing information about scenarios that are not affected by a new coal power plant. + +The structure of the presentation can be tailored to your aim. It is important to have a clear beginning, middle and end. There should be consistency across the presentation to maximise the audience's understanding. + +To further ensure that the audience understand and engage with the presentation, it should be designed with the audience's backgrounds and motivations in mind (see more below). + +Finally, it is important to consider the impact of the presentation and identify key points or policy recommendations that you would like the audience to remember. + +## Audiences + +It is important to understand the types of audience that you will be presenting to. For instance, they may be generalists or non-specialists. Or they could be scientists from different disciplines, or even scientists from the same discipline, but focusing on different topics. + +The presentation should be adapted depending on your audience in order to increase the audience's understanding and engagement. Technical content, for example, can be explained in a simple and understandable manner if the audience contains non-specialists. If you think that your audience, on the other hand, will have technical expertise, you can spend less time on explaining technical content. The amount of technical detail you provide may also change: if you are speaking to a policymaker they may be more interested in the results and recommendations than the modelling process. + +The purpose of the presentation should be optimised throughout. For example, if you are aiming to create a partnership with a new department, the presentation should have a focus on the implications of your research for that department and the benefits of the proposed partnership for the audience. + + +# Summary + +In this mini-lecture we introduced some key tips for oral presentations. We explored why understanding your audience of importance, especially when introducing technical content. We also learnt that we can be strategic in our presentation planning and should optimise for the aims for which we want to achieve. + +This is the final lecture of the Agent-based energy systems modelling: MUSE course. After this lecture you should be in a good position to develop your own models through MUSE, which can then be used to assess the impact of different policy options. + +Thank you for engaging with this course, and we hope you have enjoyed the lectures, found them valuable, and find practical uses for MUSE in your research.