GenAI Exchange Dashboard

A multi-service Streamlit application for cement industry AI solutions.

Services

1. Cement GPT

Document-based Q&A system using RAG (Retrieval Augmented Generation) for cement plant operations.

Features:

• Upload multiple documents (PDF, TXT, DOCX)
• Intelligent document indexing
• Chat-based Q&A interface
• Session-based conversation management

API Endpoint: https://jkcementgpt.onrender.com

2. Fuel AI - Cement Kiln Optimizer

AI-powered optimization for cement kiln operations with focus on alternate fuel mix and clinkerization process.

Features:

• JSON-based sensor data upload OR use sample data for testing
• Real-time kiln parameter analysis
• Predictive modeling (T+30m forecast)
• Optimization recommendations for coal feed rate, tire chip feed, and fan speed
• Emission and quality control optimization

API Endpoint: https://kiln-fueloptimization.onrender.com

Project Structure

genaiexchange-dashboard/
├── Home.py                    # Landing page with service selection
├── pages/
│   ├── 1_Cement_GPT.py       # Cement GPT service page
│   └── 2_Fuel_AI.py          # Fuel AI service page
├── sample_kiln_data.json      # Sample data for Fuel AI testing
├── pyproject.toml
└── README.md

Installation

Using UV (Recommended)

# Install dependencies
uv sync

Using pip

pip install streamlit requests

Running the Application

uv run streamlit run Home.py

The application will start on http://localhost:8502

Usage

Landing Page

1. Open the application
2. Select either "Cement GPT" or "Fuel AI" service

Cement GPT Workflow

1. Click "Open Cement GPT"
2. Upload documents using the sidebar
3. Click "Upload & Index Documents"
4. Ask questions in the chat interface
5. Get AI-powered answers based on your documents

Fuel AI Workflow

1. Click "Open Fuel AI"
2. Choose data source:
   • Upload JSON File: Upload your own kiln sensor data
   • Use Sample Data: Test with pre-loaded sample data (5 data points)
3. (Optional) View the data before analyzing
4. Click "Analyze & Optimize"
5. Review optimization results:
   • Current state metrics
   • Predicted state (T+30m)
   • Optimal parameter adjustments
   • Actionable recommendations

JSON Format for Fuel AI

{
  "data_points": [
    {
      "timestamp": "2025-10-22T16:15:00Z",
      "kiln_inlet_temp": 1150.5,
      "burning_zone_temp": 1452.1,
      "o2_level": 2.1,
      "co_level": 0.05,
      "nox_ppm": 410,
      "coal_feed_rate": 5.0,
      "tire_chip_feed": 1.0,
      "fan_speed_rpm": 935,
      "clinker_free_lime": 1.4
    }
  ]
}

Development

Requirements

• Python 3.8+
• streamlit
• requests

Configuration

API endpoints can be modified in the respective page files:

• Cement GPT: pages/1_Cement_GPT.py → API_BASE_URL
• Fuel AI: pages/2_Fuel_AI.py → API_BASE_URL