Research On Neural Control And Applications In Chemical Engineering Processes

The neuro-control is one of the forward positions of intelligent control. The design theory and applications in the chemical engineering processes of the neuro-control systems are discussed in this thesis. Several neuro-control methods are proposed and the simulation experiments with the examples of chemical engineering processes are made. The main content of this thesis are as follows:1. The neuron variable structural PID controller and a neuro-controller based on the expert fuzzy model are presented for the pH neutralization processes with strong nonlinearity. In the former method, a variable structural PID controller is combined with the neuron controllers. The first neuron is to replace the structure variable part and the second neuron is used to tune the parameters of the PID controller. And, a neural network inverse controller is combined with a neuron PID controller is suggested. These methods could operate efficiently and they are very convenient to use in practice.2. For the CSTR process, a hybrid neural network model based neuron feedback linearizing control method is presented. A neural network is used to model the part with uncertainty parameters, and a feedback linearizing control strategy is combined with the neuron, which can adjust the parameters of the controller according to the performance. Simulation results demonstrate that the proposed method has excellent performance, evenif the model is dismatched, the proposed method also has satisfactory performance.3. The particle swarm optimization (PSO) algorithm is used to optimize the parameters of the fuzzy-neuro controller. The simulation results of controlling CSTR process show that the system reaches good performance.4. The neuron controller is designed for multivariable chemical engineering processes.The learning ability of the neuron can improve the performance of the system with static decoupling. And the parameters of neuron controllers are tuned by the PSO algorithm off-line and on-line. The simulation results of the Wood-Berry's binary distilation column show the efficiency of the propesed methods..