Possible incorrect price calculation if flow coeff != 1.0

[alexdewar]

See comment here: #660 (comment)

[alexdewar]

No rush on this @ahawkes, but if you can prove that the calculation is/isn't correct with another example, that would be great. It would be good to fix this (if it is indeed a bug) before the first release.

[ahawkes]

I think this should be easy to test. Just change one of the flow out from 1 to like 0.95, ideally of some downstream process so that the change doesn't propagate too far. Will need to re-do the calculation to see if result is still correct.

[ahawkes]

@alexdewar I have just checked this, by changing the GASCGT asset in the simple_mc model such that input fuel = 3 and output electricity = 2 coeffs. Also doubled the var_opex (as we now get 2 units of electricity per unit of activity), and doubled the CO2 out as the gas in has been doubled.

Prices remained the same after these changes. Therefore, the coeff!=1.0 is not a problem.

Just FYI, I also did one other unrelated test. I added the bioenergy supply chain with capacities constrained to very close to zero (as zero is not allowed in input data - fair enough). I can confirm that HiGHs correctly generated prices for each commodity in that supply chain. So adding all candidate assets with zero capacity in the last dispatch should work, at least as far as generating prices goes.

While doing all this I did have one further concern. I wonder if the non-scarcity pricing will work ok when there is more than one output SED/SVD commodity? Don't have time to check that right now, so could we raise it as a possible issue - to be checked soon?

[alexdewar]

@alexdewar I have just checked this, by changing the GASCGT asset in the simple_mc model such that input fuel = 3 and output electricity = 2 coeffs. Also doubled the var_opex (as we now get 2 units of electricity per unit of activity), and doubled the CO2 out as the gas in has been doubled.

Prices remained the same after these changes. Therefore, the coeff!=1.0 is not a problem.

Great, thank you for confirming! I am still a little confused as to why it's ok for us to add a MoneyPerFlow value to a MoneyPerActivity one here, but if the maths works out ok, we can just circumvent the units check in this one place and stick an explanatory comment in.

Just FYI, I also did one other unrelated test. I added the bioenergy supply chain with capacities constrained to very close to zero (as zero is not allowed in input data - fair enough). I can confirm that HiGHs correctly generated prices for each commodity in that supply chain. So adding all candidate assets with zero capacity in the last dispatch should work, at least as far as generating prices goes.

That's good to know, although I think there may actually be a problem with this that I didn't notice before. I've got a PR open (#671) to include candidate assets in the base year but it's unfortunately giving us a price of zero for BIOPRD and BIOPEL in the simple_mc model.

I'll stick more info on that PR, because I think we may need to dig into this a bit more.

While doing all this I did have one further concern. I wonder if the non-scarcity pricing will work ok when there is more than one output SED/SVD commodity? Don't have time to check that right now, so could we raise it as a possible issue - to be checked soon?

The code will work, in that it will generate some numbers, but I assume you mean that the numbers might be wrong?

[alexdewar]

This seems not to be a problem, so I'll close the issue