ChemPilot: AI Process Design Assistant

AI-powered platform for chemical engineers that generates process flow diagrams, performs material balance calculations, predicts product yields, and provides an intelligent engineering assistant.

Comment

🚀 ChemPilot – AI-Powered Process Design Suite for Chemical Engineers

Inspiration

Chemical engineering students often struggle to visualize industrial process flow diagrams, perform material balance calculations, and understand refinery operations without expensive commercial software such as Aspen HYSYS. During our learning and internship experiences, we realized there was no simple, free, browser-based platform that combined these essential tools in one place.

ChemPilot was built to bridge that gap by providing an intelligent learning and process design platform specifically for students, junior engineers, and refinery trainees.

What it does

ChemPilot combines four core engineering modules into a single web application:

  • 🔷 Interactive Process Flow Diagram (PFD) Builder
  • ⚖️ Automated Material Balance Calculator
  • 📈 Machine Learning-based Yield Prediction
  • 🤖 AI Chemical Engineering Knowledge Assistant

Users can design simplified process flowsheets, generate engineering diagrams, calculate steady-state mass balances with recycle loops, predict refinery product yields using machine learning, and ask technical questions about chemical engineering concepts.

How we built it

ChemPilot was developed using Python and Streamlit with a modular architecture.

Key technologies include:

  • Streamlit for the interactive web interface
  • Plotly for engineering visualizations and Sankey diagrams
  • Scikit-learn Random Forest models for yield prediction
  • Pandas & NumPy for engineering calculations
  • Custom engineering databases and knowledge base for the AI assistant
  • Git & GitHub for version control

Challenges we ran into

Some of the biggest challenges included:

  • Creating engineering-style PFD symbols inside a browser
  • Building accurate material balance calculations with recycle loops
  • Integrating machine learning predictions into a real engineering workflow
  • Designing a futuristic yet intuitive UI
  • Optimizing Streamlit performance while keeping the application modular

Accomplishments that we're proud of

  • Built an end-to-end engineering platform with four integrated modules
  • Developed an automated material balance engine with closure validation
  • Created an interactive PFD generator with engineering equipment symbols
  • Trained machine learning models for refinery yield prediction
  • Designed a modern cyber-themed interface suitable for technical users

What we learned

This project strengthened our understanding of:

  • Chemical process engineering
  • Mass balance modelling
  • Machine learning workflows
  • Streamlit application development
  • Data visualization
  • Software architecture for engineering applications

What's next for ChemPilot

Future improvements include:

  • Thermodynamic property calculations
  • Energy balance module
  • Distillation column design
  • Heat exchanger sizing
  • Equipment cost estimation
  • Process simulation integration
  • LLM-powered engineering assistant
  • Cloud deployment with collaborative projects

Built With

Share this project:

Updates

Submission history