The Dual Heuristic Dynamic Programming Learning Control of Soft Measurement Model in Wiped Film Evaporation Process

摘要

The wiped film molecular distillation system has the characteristics of multivariable, strong coupling, nonlinearity and large hysteresis. Current modeling and control methods are difficult to achieve the desired results. In order to stabilize production, the online sequential extreme learning machine soft measurement model driven by data is constructed by analyzing the auxiliary parameter variables. The dual heuristic dynamic programming learning control was used to optimize for the molecular distillation system. Comparing with traditional back propagation neural network modeling. The simulation experiments are performed to verify the effectiveness and advantage of dual heuristic dynamic programming learning control with online sequential extreme learning machine soft measurement model.