摘要
为了提高热电阻温度测量精度,用径向基(RBF)神经网络建立了Pt100热电阻温度测量模型并与BP网络模型进行了比较.用分度表数据训练神经网络,研究了模型的分布密度范围,讨论了热电阻检定方法.结果表明:提出的模型精度高、稳定性好;470~560℃范围内,样本数据分度间隔为1℃时,适宜的分布密度范围在0.3~1.0.
In order to improve accuracy of temperature measurement, a model of Pt100 thermal resistor is built by using radial basis function (RBF) neural network; and comparison with BP method is given. Network training is done based on data scale division. The proper scope of spread parameter for RBF model and some calibration methods are discussed. The results indicate that the model presented is accurate and steady. At 1℃ interval of specimen in 470-560℃, the proper scope of spread parameter for the model is at 0.3-1.0.
出处
《三峡大学学报(自然科学版)》
CAS
2008年第4期55-58,共4页
Journal of China Three Gorges University:Natural Sciences
基金
湖北省教育厅自然科学研究计划项目(D200513001)
关键词
传感器
热电阻
温度测量
RBF
神经网络
sensor
thermal resistor
temperature measurement
radial basis function
neural network