神经网络在热电偶特性校准中的应用

热电偶传感器,在工业测温系统中得到了广泛的应用。但热电偶的特性,即输出的热电势E与温度t之间的关系是非线性的。在热电偶检定系统中,传统的做法是采用查表法和建立热电偶E-t数学模型来实现热电势与温度的转换,但传统做法计算量较大,适应性不好。为了解决以上缺点,论文引入了神经网络,分析了RBF神经网络的原理,并将虚拟仪器技术和神经网络技术相结合,建立了热电偶E-t特性RBF神经网络模型,用于热电偶的在线测量,结果证明这种方法有效。

Application in calibration of thermocouple character based on neural network

Sensors of thermocouple are used in measure temperature system widely.However,the relation between thermocouple voltage E and temperature T,which is also called character of thermocouple,is non-linear.In automatic calibration system for thermocouple,means to realize transfer between thermocouple and temperature are method of tables and establish mathematical model of thermocouple character E-t,but these methods have shortcoming of larger calculation and worse adaptability.In order to solve these problems,this paper introduces neural network and analyzes principle of neural network.Then combine technology of virtual instrument with neural network to found model on thermocouple character of E-t.Then use it to measure thermocouple on-line.The result shows that system has a good effect.