Intelligent Control And Simulation In The Rubber & Plastic Vulcanization Model

Vulcanization is one of the most important crafts in the course of manufacturing rubber & plastic products, and in this process, it is hard to confirm the control parameters. With the development and progress in rubber & plastic products the requirements of accuracy, reliability and costs should be improved. Based on the plat vulcanization machine status and its current progress and the status of products, this thesis presents a new kind of control strategy to be with through the limits in the traditional control analyzing and proving all the relative theories. All of these works offer some consults for the actual vulcanization. This thesis is supported by the important research project of ministry of education named as multi-functional numeric technology and system research of rubber & plastic products (Authorization No.104155). As one part of this project, this thesis makes a brief introduction of the basic vulcanization knowledge and 630 QD ?SY flat vulcanization machine used in the model test and simulation. Based on the theoretical analysis, it is confirmed that the theoretical parameters only have instructive functions in principle. For the problem of accuracy and real time, this thesis presents a intelligent control model based on AI knowledge, in which, all the noise are unified as one kind of noise. During the course of controlling this system accompanied with self-adaptive learning. The model includes two parts: one is NN and the other is self-adaptive system. At the same time, for the demands of intelligent control and cooperation, symbol calculation function is added in the model. In this model, the potential problems can be solved without the helps of the worker during the process of the vulcanization. For the intelligent model, relative theories are proved particularly, and a simulation according to this has been finished in this thesis based on some samples. In the course of simulation, the knowledge of NN, AI and pattern recognizing are employed in the model, and the theories of calculability and calculative complex are used in the works of curling process. On the reason of that it is hard to control the precision of calculation, especially for the self-adaptive behaviors, the calculating works are finished under the helps of the Matlab6.5 used calculating server which is connected with the applications of Delphi6.5.