Coke Oven Temperature Control Based on Fuzzy Model Reference Learning Control

Coke oven is a kind of nonlinear,time-varying and parameter-distributed system with great inertia and pure time-delay,whose model is difficult to be constructed.The coke oven's temperature control based on linear system theory is lack of precision and couldn't adapt to the change of operating conditions.The fuzzy controller may generate limit cycle on steady state due to the roughness of the fuzzy rules and the membership functions.A fuzzy model reference learning control(FMRLC) was used to control the coke oven system,which avoided identifying the order of the coke oven system.The FMRLC incorporated the advantages of both fuzzy control and adaptive control,and it resolved those problems mentioned above through on-line learning algorithm.Simulation results are presented to illustrate that FMRLC has the good dynamic performances and steady state performances;furthermore,the temperature fluctuation is reduced.The successful simulation verifies the feasibility and validity of FMRLC in the coke oven system,and it is worthy of extension.