The Application And Study Of Rough Set Theory In Boiler Steam Pocket Water Level Control

Boiler is a kind of common equipment providing power and heat used either in industry production or ours'daily life. The productivity and security of the factory and the quality of life are affected by boiler control level. The problem of the boiler control has always been the hot issue in this reign. The boiler water level control system is a complex system which is nonlinear and has large time delay process. The mathematical model can not be obtained easily, so the conventional control method can not function properly.We can describe the panorama of a system from different lateral of the system using data fusion method, which is suitable for the water level multi-parameter control system. Rough Set Theory is a new kind of soft computing methods dealing with problems of uncertainty. In contrast to Fuzzy Set Theory and Dempster Shafer evidence theory, Rough Set Theory dose not any the prior knowledge, which makes it popular among scholars. Rough Set Theory plays an important role in decision support and knowledge acquisition. So it is viable to apply rough set in the control rule acquisition of boiler water level control system.In this paper, we research the basic theories such as Neural Network Theory and Rough Set Theory. The main points and control process of control system are illuminated in details in this paper. A new method is provided that designed the control rule of boiler water level control system using rough set theory. Distributed fusion is adopted in the system. At first, ART-2 Neural Network Method is applied respectively in the time sequence fusion of feedwater flux, steam flux and water level. Then, the second fusion of the three parameters is performed using Rough Set Theory to get the current water level working state. Then, we reduct the control decision table and get control rules using Rough Set Theory. At last, we simulate the control process in the MATLAB software. The result shows the control system based on Neural Network Theory and Rough Set Theory outshines the conventional three element PID control system with regard to the object in this paper. ARM-2410S is adopted in the control system hardware designing.