工业生产过程锅炉温度控制仿真

Simulation of Boiler Temperature Control in Industrial Production Process

温度控制系统是工业电阻炉的重要组成部分。由于温度测量元件和电阻炉温度传导产生的延时,使得电阻炉温度控制存在超调量大,温度控制精度不高,调节时间过长等问题。传统的PID控制以及基于模糊控制理论整定的PID控制方法在温度控制中适应性较差、效果不理想,而人工神经网络作为智能控制的一个重要领域,具有较强自学习和自适应的能力,因此在此基础上提出模糊径向基(RBF)神经网络的PID算法,将模糊PID算法和神经网络相结合,可根据温度的偏差及变化对PID三个参数进行优化,达到最佳组合的PID控制,并利用MATLAB的Simulink工具箱做了常规PID、模糊PID以及模糊RBF神经网络PID的仿真对比实验。仿真表明,在添加了RBF神经网络模块之后,系统响应的超调量和稳态误差大大减小,具有较高的稳定性。

The temperature control system is an important part of the industrial resistance furnace. Due to the delay of temperature measurement and the temperature conduction of the resistance furnace, the temperature control of the resistance furnace has many problems, such as large overshoot, low temperature control precision and too long adjustment time. The traditional PID control and PID control method based on fuzzy control theory are not suitable for temperature control, as an important field of intelligent control, artificial neural network has the ability of self-learn- ing and self-adaptation, therefore, a PID algorithm with fuzzy RBF neural network is proposed which combines fuzzy algorithm with neural network, it can optimize the three parameters of PID according to the deviation and variation of temperature, so as to achieve the best combination of PID control. Also, the paper makes a contrast between the conventional PID, fuzzy-PID and fuzzy RBF neural network PID by using the Simulink simulation toolbox of Matlab. Simulation results show that, after adding the RBF neural network module, the overshoot and steady-state error are greatly reduced.