Shell gasifier inverse system decoupling control based on LS-SVM

On basis of the characteristics of gasifier process principle in Shell IGCC plants, inverse system decoupling control method based on least squares support vector machine (LS-SVM) is proposed. LS-SVM is used for establishing the inverse model of nonlinear controlled object, which will be connected in series with the original object. Then the original system will decoupled into multiple independent single variable pseudo linear composite subsystem. In order to overcome the modeling error of composite system, a additional single neuron adaptive PID controller is proposed that the improved supervised Hebb learning rule is used for controller parameters self regulating. The simulation results show that this method has decoupling ability, high control precision and advantages of fast responses, and strong adaptability and robustness.