Design of Neural Network Predictive Controller for a Quadruple Tank System

Saranya Balachandran, T Anitha
The processes in most of the industries are highly non-linear and dynamic. The quadruple tank system is a benchmark system used to analyse the nonlinear effects in a multivariable process. The quadruple tank process is thus used to demonstrate coupling effects and interactions occurring in multivariable control systems. This project presents a neural network predictive controller for a quadruple tank system. The process data will be obtained from the mathematical model of the laboratory scale
more » ... laboratory scale experimental setup. The model obtained from training the system via neural network will be used in controlling the quadruple tank by neural network predictive controller. The simulation results will be compared with the closed loop response and constrained and unconstrained model predictive control algorithm results.