Reserch Of Identication And Internal Model Contorl For Non-Square Multivariable System With Time Delay

In the actual industrial processes, the multivariable systems with complex features are widespread. The control effects of these complex industrial pro-cesses directly affect many important targets such as product quality, yield and energy consumption because that these complex industrial processes usually are the core part of the production process. So it has important theoretical sig-nificance and application value to research control method of complex indus-trial processes. With the advantages of simple to design, excellent robustness, and good prediction ability, the internal model control (IMC) becomes one of the most effective control methods of the complex processes. The model iden-tification methods and the internal model control methods for non-square mul-tivariable systems with time-delay are researched in this paper. The main work includes:1. As a parallel optimization algorithm, fireworks explosion optimization algorithm (FWA) is introduced in this paper. Modified fireworks explosion optimization algorithm (MFA) is proposed based on analysis of the character- istics and problems of FWA. Comparative tests show that MFA has faster convergence and higher accuracy than standard FWA and classical optimiza-tion algorithms. MFA is used in the model identification methods and parame-ter optimization in the internal model control methods.2. Multivariable system identification methods of any input signal based on intelligent search are researched. A new identification method suitable for any input signal named MFA-FRE is proposed in this paper. Modified fire-works explosion optimization algorithm (MFA) and system frequency re-sponse identification method (FRE) are used in this method. MFA-FRE identification method applied successfully in the multi-variable systems by using sequential step test method. A multi-variable system is decomposed into multiple single-variable system by using sequential step test method firstly. And then every single-variable system is identified by using MFA-FRE identi-fication method. Two identification mathods for non-square system by using MFA-FRE are provided in this paper also. Experiment results verify the valid-ity of the identification methods.3. The iterative least squares algorithm and the iterative stochastic gradi-ent algorithm are researched to sovle parameter identification problem of a kind of complex linear output error model. Experiment results verify the va-lidity of the identification methods.4. Design method of IMC controller which is used to decoupling and control for multivariable system with time delay based on V-norm is re- searched in this paper. P-norm is used to describe the controlled process and V-norm is used to describe the decoupling IMC controller in the proposed de-sign method. The controller designed by this method can satisfy the require-ments of decoupling and control. The IMC controller can satisfy the require-ments of overshoot and response time because a new performance index and MFA are used to optimize the filter parameter.5. Internal model control method based on MEOTF-MFA is proposed in this paper to control non-square system with time-delay. A complex multivar-iable non-square system can be decomposed into multiple single-variable sys-tem by using the modified open-loop transfer function (MEOTF). And then every single-variable system is controlled by using IMC method. For the cal-culation of higher order items, Taylor-MFA model reduction method is used to reduce the order of model. The effective control of non-square system with time-delay can realize by using MEOTF-MFA controller design method, and can also perform well in the.Simulation results show that the proposed method have good robustness and control performance when model mismatch and in the interference environment.