Application of FEL based adaptive inverse control on wood drying kiln temperature

A adaptive inverse control approach based on Feedback-Error-Learning (FEL) is presented in this paper; it's employed for temperature control of the wood drying kiln. It combines the traditional feedback controller PID with a Nonlinear Auto-Regressive Exogenous Input (NARX) neural network to achieve the control. PID is used to stabilize the system, while NARX is acted as dynamic inverse feed-forward controller, and employs the output of the PID to realize adaptive online, it improves the performance of the traditional controller. The adjustment of the control parameters are adaptive adjust by the input and output online information, the information is applied to learn the parameter transformation and unmodeled dynamics of the system, pre-training of the neural network is not needed by using this method. Experiments have been done on JBGZ-1.8 homemade experimental drying kiln and control precision improved to 0.5degreesC, results show that besides the high control precision feature of the traditional PID, this control method has the advantages of high feed-forward control respond, short rise time and low overshoot as well.