The Rough-Fuzzy Controller Based On FCMA Algorithm

As the most general control method in industrial control, PID controller has many advantages, such as its simple structure, strong robustness, stable state without steady-state error and easy to operate etc. Thus its characters make PID control has been used in industry for a long time. But the limitation for the application of complex controlled object is one of important factors of restricting its development. With the development of IC (Intelligent Control), fuzzy control theory has been mature and fuzzy control technology has been come into being. This brings new vitality for the development of PID control.On the base of studying the development course of the fuzzy control system and its characters, in this thesis aimed at the defect of the pure Fuzzy control we use RS (Rough Set) theory to extract the fuzzy rules reasonably and make FCMA (Fuzzy C-Means adaptive) algorithm as the clustering algorithm for the data from PID controller . By this method the entire control system performance is further raised.RS can analyze, inference and excavate the implicit knowledge and rules from numbers of data. It is resolving this difficult problem that intelligent control handles the knowledge, especially the uncertain knowledge. RS is promoting IC forward. The original data can be handled by use of knowledge reduction of RS in the thesis. The fuzzy rules can be extracted from data directly, which reduces the difficulty of obtaining rules in fuzzy controller.RS can deal with the discrete data; however, the data from Industry process is continuous. The clustering numbers are set as you like in the traditional discretization methods, which leads to the subjective and optional clustering results, lacking of scientific proofs and easily trapping into partial optimal result. FCMA used to cluster in this paper can make sure the clustering number adaptively, which avoids the subjective choice on clustering number and solves the optimal problem. The improving makes the clustering algorithm have well anti-disturbing performance.The fuzzy controller is PD controller in nature, and it has well dynamic performance. The fuzzy controller has no integrating term, so the control system is difficult to eliminate the static error. Therefore, an integrating factor and NLJ are used to get the best control effect.Several optimization methods introduced in the upper article are blend in tradition fuzzy control to realize the fuzzy control system. The simulation results compared with PID controller optimized by ant colony optimization. Results prove this method superiority on improving the controller performance and show the super control performance of this intelligent control system.