摘要
对热电偶的非线性误差补偿方法进行了研究,提出了一种基于Profibus-DP的新型热电偶温度监测系统,系统对热电偶采集从站的非线性误差在上位机进行在线补偿。Matlab编写的应用程序采用DDE协议实现与MCGS软件的数据交换,它从MCGS实时数据库获取训练温度样本集,并采用遗传算法训练建立的BP神经网络模型,利用训练好的BP网络对待补偿的温度数据进行校正,并把校正结果传递给MCGS在线显示。实验表明,补偿模型把补偿数据的最大相对误差从9.96%降低到1.40%。新系统减轻了热电偶从站的负担,具有模型可靠、实时性强、计算精度高等特点。
The methods of compensation for the nonlinear error of thermocouple are studied and a monitoring system based on Profibus-DP bus is proposed, in which the thermocouple nonlinearity of slave station is on-line rectified in master station. The Matlab application software, which exchanges data with the MCGS software through DDE protocol, gets the temperature samples from the real-time database of MCGS and trains the BP neural network using the genetic algorithm. The measured temperatures are rectified by the trained network and then transferred to the MCGS for on-line display. The load of the temperature acquisition slave station is thus reduced. The new model is more reliable,more accurate and better in real-time performance.
出处
《电力自动化设备》
EI
CSCD
北大核心
2009年第1期124-126,共3页
Electric Power Automation Equipment
关键词
热电偶
非线性补偿
遗传算法
神经网络
thermocouple
nonlinearity compensation
genetic algorithm
neural network
