The Resistance Furnace Intelligent Temperature Control System Based On Single Neuron PID

Resistance furnace is applied extensively in industry production. The control result of temperature control system will directly affect production efficiency and product quality, so the high requirement of control system is needed. But for a control system with the characteristic of non-linear, pure time delay, large inertia, time variation, and unidirectional rising, such as resistance furnace, it is difficult to use mathematical methods to establish a precise model. Hence, the traditional control theory and algorithms are usually failed to obtain a better control effect.This subject comes from the heat treatment processing control problem of Box-type resistance furnace. The action of heat treatment is to enhance the mechanical properties of material, relieve residual stress, and improve machinability of metal. The subject is requested to design a computer monitoring system, which includes sensors, signal processing, power supply, hardware designing, and monitoring software development, to complete the heat treatment for workpiece.This paper focuses on how to apply algorithm to control the temperature. The key of temperature control lies on two aspects, including measuring and controlling. Temperature measuring is the base of temperature controlling, which is fully mature. However, for the control object is complex, time-varying and the request to the control operation is more and more harsh, how to improve the accuracy of temperature controlling for easy manual control has been an important problems of industry temperature control.This paper designed a resistance furnace temperature control system based on Single Neuron PID control algorithm. Conventional PID is first introduced briefly. Then the Single Neuron PID is derived and described in detail. Finally a conclusion is proposed to analyze the stability of Single Neuron PID and discussed the selection rule of its adjustable parameter. The configuration software has been secondary developed, which fulfills many functions such as configuration, supervision, alarm, table, etc. The temperature control system is operated simply with the characteristic of active pictures, high adjustability, high reliability, and high stability. Basing on the temperature rising curve, which is obtained via experiments, the mathematical model of resistance furnace is established, and an improved Hebb learning algorithm of the Single Neuron PID Controller is designed. The simulation results proved the feasibility of this algorithm and the superiority of Single Neuron PID. Based on the simulation, the temperature control of resistance furnace has been debugged. The result shows that the control effect is better than that of Conventional PID, it has small overshoot, high control precision, strong resistance to interference and good application prospect.