Adaptive decoupling control systems based on SVM for large supercritical CFB boilers combustion system

An α th-order inversed decoupling control method based on least square support vector machines (LS-SVM) is presented to resolve the difficulties of inverse modeling with the traditional inverse control methods in this paper. The nonlinear offline inverse model of plant is built by LS-SVM, which is cascaded before the original system to decouple a complex multivariable input and output system into several independent single input single output pseudo-linear sub-systems. In order to make it have better robustness, with single neuron adaptive PID as subsystem additional linear controller to constitute a closed-loop system. The characteristics of combustion system of circulating fluidized bed boilers also are analyzed and control system based on presented methods is presented. The simulation results demonstrated that the presented method could achieve accurate decoupling control with robustness.