🍊 Orange Juice Evaporation Predictor

This interactive tool demonstrates how process simulation can be combined with machine learning surrogate models to enable fast and accurate predictions of complex industrial operations.

The system represents a juice concentration process using an evaporator, where the key operating conditions are:

• Evaporation Temperature (°C) 
• Evaporator Pressure (bar)
   

From these two inputs, the surrogate model instantly predicts:

• Steam Consumption (kg/h) 
• Orange Juice Mass Fraction at Evaporator Outlet(%)
   

The original process was simulated in DWSIM, a powerful open-source process simulator. A Design of Experiments (DOE) was created using Latin Hypercube Sampling (LHS) within the AI4T Suite, followed by training a surrogate model optimized with Optuna. This model is then deployed as a web interface, allowing users to interact with the predictions in real time.

✅ Speed – Results are available instantly, without running a full simulation.
✅ Accessibility – A simple web interface makes advanced process modeling available to anyone.
✅ Cost reduction – Engineers can explore scenarios and optimize conditions without consuming computation time or simulation licenses.
✅ Decision support – Supports quick evaluation of process performance and energy consumption.
✅ Scalability – The same methodology can be applied to any industrial process (chemical, food, energy, etc.).