The Application Research Of Intelligent Control Theory In A Kind Of Process Control System

Aiming at such problems as large inertia, uncertainties and nonlinearity in some industrial process control objects, the thesis discusses using intelligent methods to enhance performance index and improve control effect.This thesis is based on a typical process control system-a temperaturecontrol system of medicament. Because the controlled object of the system often shows characteristics such as time delay, large inertia and uncertainties , an intelligent PID method with fuzzy logic is proposed in this thesis to tackle the problems above. Moreover, PLC+Touch Panel are applied to realize intelligent operation of the system. The practical results after system running show a good control effect. They indicate that the controller can reduce large overshoot, increase control accuracy and make system effective in robustness to model uncertainties , meanwhile prove the superiority of this intelligent tuning method of PID controller over the classic one in the application of a kind of process plant.Besides, due to strong nonlinearity and parameter uncertainties in many industrial plants, the thesis also presents a modelling method and a kind of control structure based on Neural Networks (NN). The simulation object is still the temperature system of medicament. The series-parallel identification structure is adopted to reflect characteristics of the controlobject well and truly. Furthermore, the improved BP (Back Propagation) algorithm and floating-point genetic algorithm (GA) are respectively employed to learn the weights of NN positive model. Finally, STC structure is used to form a composite neural networks adaptive control system of temperature and GA is also applied to train NN controller. The simulation results indicate the capability of genetic algorithm in fast and steady learning of neural networks, guaranteeing a global convergence and overcoming some shortcomings of traditional error back propagation algorithms, meanwhile prove that this neural networks adaptive control structure is effective to many control problems and it is easy for us to programme and employ the method in the practical system.