Comparison of two different model free fuzzy control architectures based on inverse plant modeling

Robust control performance in heating/cooling systems is a challenging issue. The control scheme should be able to reject all sorts of disturbances while maintaining the set point at the desired level, irrespective of the nonlinearities associated with the plant. Fuzzy control is best suited for uncertain and nonlinear plants. In this paper two different model free fuzzy control architectures based on inverse plant modeling are discussed. The first architecture is based on model reference adaptive control (MRAC) in which feedback error learning is used to develop an inverse plant model by incorporating the control error i.e. the difference between the controlled output and the set point. The second control architecture is based on controller output error method (COEM) in which the inverse model is developed by using the controller error i.e. the error between the actual control signal and the predicted control signal. Both the schemes utilizes the same fuzzy adaptive algorithm. The controllers have been tested on simulation as well as on experimental test rig. It has been successfully demonstrated that for the particular type of application MRAC offers tighter control than its counterpart.