摘要
为了提高热电偶检定炉温度的控制性能,研究了检定炉模型的在线辨识方法和控制器参数的自整定方法,设计了检定炉温度智能控制系统.由于检定炉是具有非线性和时变性的复杂对象,首先利用RBF神经网络对其输入、输出关系进行在线辨识,然后依据偏差最小准则,采用梯度下降法对控制器的PID参数进行整定,从而实现检定炉温度的智能控制.试验结果表明基于RBF神经网络的控制器在200～1 200℃之间对检定炉温度控制的性能指标优越于传统PID控制器,达到了国家标准中对控温误差和温度波动度的要求.
In order to improve the control property of the calibration furnace （for thermocouple verification） temperature,studies on method of model on-line identification of calibration furnace based on RBF neural network and self-tuning parameters of the PID controller,designs an intelligent control system about calibration furnace temperature.Since the calibration furnace is a nonlinear and non-stationary complex object.Firstly,using RBF neural network to identify the relationship between the calibration furnace input and output,then,based on the minimum standard deviation,regulates the PID parameters by the method of gradient descent,finally,realize intelligent control of the calibration furnace temperature.The test results show that the controller that based on RBF neural network has better performance in regulating the calibration furnace temperature in the 200 ～ 1 200 ℃ interval than the traditional PID controller,and up to the requirements that about temperature error and temperature fluctuation degree which described in the national standard.
出处
《仪表技术与传感器》
CSCD
北大核心
2014年第1期61-63,共3页
Instrument Technique and Sensor
基金
教育部高等学校博士学科点专项科研基金资助项目(20114101110005)
关键词
径向基函数神经网络
在线自整定
检定炉
温度控制
RBF neural network
online self-tuning
calibration furnace
temperature control
