Artificial Neural Network Control of a Multiple Effect Evaporators via Simulation

Abstract

This research studies the dynamic model and control of multiple effect evaporators of tomato solutions by implementing three control strategies: PID, neural model reference, and neural model predictive controllers. The evaporator’s control is crucial to maintain the product specifications at different operation conditions at minimum operating cost. The model reference control and model predictive control has been designed and evaluated. The simulation results showed that the neural predictive controller is more suitable, has lower overshoot, less offset value, and less integral absolute error.