Resistance Furnace Control System Design Based On Fuzzy Neural Network Compensation And Realization

Resistance furnace is commonly used in industry, temperature control results will directly affect quality of the work pieces and the heating efficiency, thus a higher demand is presented in temperature control system. With the improvement of control theory and power electronics technology, the intelligent controller in industry has been accomplishing. The digital controller based on single-chip computer has been applied widely, because of its cabinet cubage, low-cost, abundant function, simple and convenient. But it's very difficult to establish accurate mathematical model of this plant, because it has the characteristics of non-linear, time-delay, time-varying. Nowadays, PID is widely used in this area, but the control result is not precise because of its fixed parameters.Fuzzy logic translates human's knowledge to control rules, and it can efficiently use in the systems which structures are uncertain. But it's difficult to get the exact control rules. In allusion of this, many authors presented Fuzzy Neural Network. It can use the advantage of Neural Network to obtain the accurate control rules of Fuzzy Logic System. Because the control and the learning are working at the same time, this on-line control method has the defect of long setting time.Aimed to this defect, this thesis analysis of the causes of this phenomenon and improve the traditional Fuzzy Neural Network by compensation. And the method for determining the amount of compensation is given. The printed circuit board of resistance furnace temperature control system which ATmega16 microcontroller is the core control unit was designed. Combined with the host computer software, the resistance furnace can be controlled by the improved Fuzzy Neural Network control method. The detail of circuit module, software module and control algorithm is described by this paper.The resistance furnace temperature controller has been utilized in the laboratory resistance furnace, and the results of experiment show that this method is effective and feasible, and the desired goal has been achieved.