NTC热敏电阻温度标定方法研究

Study on temperature calibration method of NTC thermistor

NTC热敏电阻作为温度传感器,被广泛用于医疗、家电、汽车、军事、航空航天等领域,如何提高NTC热敏电阻测温精度是研究的热点。选择5支同型号NTC热敏电阻,在0~120℃温度范围内,对热敏电阻阻值进行了测试,并采用指数经验公式、Steinhart-Hart方程、多项式、神经网络算法对热敏电阻温度进行了标定。结果表明:利用测温区间上、下限温度点电阻值计算指数经验公式中的B值来代替厂家给出的B值可以改善NTC热敏电阻测温精度;相对于B值标定法,采用Steinhart-Hart方程作为标定方程可以显著改善标定精度;采用5次多项式作为标定方程的标定残差为-0.06~0.03℃,标定精度进一步提高;采用BP神经网络进行非线性补偿后的温度标定残差为-0.02~0.02℃,平均残差为0.01℃,可以满足大部分应用场合精度要求。

NTC thermistor is widely used in medical,home appliances,automotive,military,aerospace and other fields as a temperature sensor.How to improve its temperature measurement accuracy is an active area of research.5 NTC thermistors of the same model were selected to test the resistance in the temperature range of 0-120℃.Temperature was calibrated for NTC thermistors by using exponential empirical formula,Steinhart Hart equation,polynomial and neural network algorithm.The results show that temperature calibration accuracy was improved by using the value of B in the exponential empirical formula calculated by the resistance values at the upper and lower limits of temperature measurement range to replace the value of B given by the manufacturer.The calibration accuracy is further improved by using the 5th order polynomial as the calibration equation,calibration residual of which is-0.06-0.03℃.The range of calibration residual and the average residual are-0.02-0.02℃and 0.01℃respectively by using BP neural network,which can meet the requirements of most applications.