基于神经网络的NTC热敏电阻的校正模型

Correction model of the NTC thermistors based on neural network

为了提高NTC热敏电阻的测温精度以及简化校正过程的步骤。选取型号为10K的NTC热敏电阻,对NTC热敏电阻的R-T关系进行标定,利用BP神经网络对其R-T关系进行非线性校正,将OFFICE数据库与MATLAB工具箱进行有机结合使用,利用误差评判原理求得绝对误差、平均误差以及均方差,对NTC热敏电阻温度计的BP神经网络非线性校正模型的精度进行评价,得到BP神经网络非线性校正模型精度比传统的校正方程的校正精度高。并且实际验证了此校正方程的测温精度可以满足更高精度要求的需要。

In order to correct the NTC（nagative temperature coefficient）thermistors calibration equation to improve the precision of temperature measurement and simplify the correction process of the NTC thermistors. A 10K NTC thermistor is selected to calibrate which the relationship of R-T. Organic combination of database and toolbox used the BP neural network to correct the non-linear characteristic of NTC thermistor, the use of error, judging non-linear characteristic correction model accuracy of NTC thermistor. It can be derived that the accuracy of BP neural network is higher than the traditional correction model. The results of experiments confirm that the precision of measuring temperature by this calibration can meet the needs of even higher precision.