基于神经网络的铂电阻温度传感器非线性校正方法

Approaches to Non-linearity Compensation of Platinum Resistor Based on Neu-ral Network in Temperature Measurement

简单介绍了当前铂热电阻应用存在的问题,提出了应用前向多层神经网络建立热电阻的逆模型进行非线性补偿,使得铂电阻的静态特性线性化。文中采用MATLAB为工具,对神经网络进行训练,获得权值、阈值,并以SIEMENS的S7-200为控制器,对加热炉进行控制。设计及实际应用表明,这种方法简单、实用、有效,大大方便了铂电阻温度传感器在测控系统中的应用。

The problem of using platinum resistor is introduced. Here bring forward one way, to compensate non-linearity using adverse model of platinum resistor founded by forward multilayer neural network. Which makes static characteristic of platinum resistor linearity. The paper trains the neural network using MATLAB, gets the weights and biases, controls the heater by S7-200 PLC of SIEMENS. It has been shown in practical applications that the methods proposed in the paper not only are correct, practical and precise, but also make it convenient for platinum resistor to be applied on measurement and control systems.