Intelligent design of predictive control for typical industrial processes

An intelligent design scheme of predictive control for typical industrial processes is presented. First, the GDMC (generalized dynamic matrix control) algorithm is applied to a typical industrial process system, and three main performance indices of the controlled system are analyzed rigorously. The quantitative relationship between performance indices and design parameters is then aggregated. Based on this result, a comprehensive fuzzy design scheme is presented in accordance with the comprehension of system requirements and the redundancy of performance indices. A software package is developed to search for design parameters to achieve the most satisfactory performance indices. This intelligent design scheme can avoid human intervention or reversal trials and is especially convenient for industrial applications. >