Fuzzy Modeling And Control Of Complex Industrial Processes

With the development of technology and productivity, the systems of the industrial process c ontrol become more c omplex due to the lack of precise, formal knowledge about system, strongly nonlinear behavior, the high degree of uncertainty, time varying characteristics, close coupled and high dimensional system, etc. Owing to the lack of precise mathematical model, it is difficult, even impossible, to control such complex systems. From the point view of modeling and control, we discuss the application of fuzzy theory in complex industrial process control in this paper. The main contents are concluded as follows:(1) Based on improved fuzzy clustering algorithm, we research on fuzzy modeling method. The notion of considering both the precision and the interpretation of fuzzy models were presented at first. Then, we define an input variable selection criterion to select and sequent candidate variables. An objective function was introduced to determine the structure of fuzzy models. At last, we present a specific fuzzy modeling method based on modified fuzzy clustering algorithm combined with least-square estimator and L-M optimize algorithm.(2) The stability analysis and controller design of fuzzy systems based on T-S state space model were proposed. We present the principle of parallel distributed compensation, and detail the controller design and stability analysis of fuzzy system using linear system theory (pole placement). Linear matrix inequality was introduced to cope with the puzzle of determining common positive matrix. Theory of stability analysis and controller design was established based on LMI method.(3) Nonlinear fuzzy p redictive c ontrol t heory b ased on T-S s tate s pace m odel was explored. Linear predictive control techniques can be applied to nonlinear system for it can be regard as linear time-variant system. Multi-model was used in predictive horizon to decrease the predictive error and reflect true state of system. Objective function was converted to linearquadratic programming and thus avoid huge computational burden. (4) Hardware and software of glass melting furnace were detailed. The method of fuzzy modeling based on fuzzy clustering and the algorithm of nonlinear fuzzy predictive control m entioned above were used to model and control the temperature of glass melting furnace. The result shows their validity.