摘要：温度在工业生产中是非常重要的一个因素，它关系到产品的许多项重要品质，也关系到产品的生产效率等等，所以温度的控制不仅仅必要并且很重要。本文选取的电阻炉模型，在工业应用中非常广泛，而与之相关的工作，最重要的一点就是温度的掌控。电阻炉模型具有非线性，大惯性，大滞后性等特点，要达到理想的控制效果需要一个好的控制方案。

经典的PID控制由于结构简单，操作容易，直到现在仍然被广泛应用。但这种传统的PID控制是基于被控对象可以被抽象成一个准确的数学模型，电阻炉温度系统是不确定的，高度非线性的，所以传统方法就不适用了，存在很大的局限性。用这种方法将很难达到令人满意的效果。

近年来智能控制得到了飞速的发展，而作为智能领域的两大基础模糊控制和神经网络更是突飞猛进，它们在温度的控制方面取得了良好的效果。本设计以电阻炉为被控对象，分别采用常规PID控制算法，模糊自适应PID控制算法，单神经元PID控制算法，BP神经网络PID控制算法对电阻炉进行温度控制，运用Matlab软件仿真部分控制过程，通过在控制过程中不断改变相关参数来达到温度控制的目的。

我们通过仿真结果可以很清楚的看到，无论是抑制系统的超调量还是响应的快速性上，BP神经网络比其他控制算法都更为优越。尽管现在的主流控制依然是经典控制理论和模糊控制，但相信不久的将来神经网络控制方法将逐渐替代它们占据主导地位。本论文以实际对象进行控制，起到了良好的控制效果，对现实也具有一定的借鉴意义。

关键词：电阻炉，模糊PID控制，常规PID控制，MATLAB，BP神经网络

Abstract：Temperature is a very important factor in industrial production. It is related to many important qualities of products, as well as the efficiency of products and so on. Therefore, temperature control is not only necessary but also important. The resistance furnace model selected in this paper is widely used in industrial applications, and the related work is the most important point is the control of temperature. The model of resistance furnace has the characteristics of nonlinearity, large inertia and large lag, so a good control scheme is needed to achieve the ideal control effect.

The classical PID control is still widely used until now because of its simple structure and easy operation. But the traditional PID control is the object can be abstracted into a

precise mathematical model ， the system of resistance furnace temperature is   uncertain,

highly nonlinear, so the traditional method is helpless, there is great limitation. In this way it will be difficult to achieve satisfactory results.

In recent years, intelligent control has been developed rapidly, and as the two major foundation of intelligent field, fuzzy control and neural network are advancing by leaps and bounds. They have achieved good results in temperature control. The design of the resistance furnace as the object, using the conventional PID control algorithm, fuzzy adaptive PID control algorithm, single neuron PID control algorithm, BP neural network PID control algorithm to control the temperature of the resistance furnace, using Matlab software simulation control process, by changing the relevant parameters in the control process to achieve the purpose of temperature control.

Through the simulation results we can clearly see that, whether it is fast overshoot suppression system or response, BP neural network is superior to other control algorithms.Although the mainstream control is still classical control theory and fuzzy control, but believe in the near future, neural network control method  will  gradually replace them, occupy the leading position.In this paper, the actual object is controlled, which has a good control effect, and also has certain reference significance to reality..