磁力仪温度误差的径向基神经网络补偿模型  

Temperature compensation model of fluxgate magnetometers based on RBF neural network

作　　者：庞鸿锋[1] 罗飞路[1] 陈棣湘[1] 潘孟春[1] 罗诗途[1] 

机构地区：[1]国防科学技术大学机电工程与自动化学院,长沙410073

出　　处：《仪器仪表学报》2012年第3期695-700,共6页Chinese Journal of Scientific Instrument

基　　金：部委级基金(914A170501090)资助项目

摘　　要：磁通门磁力仪参数受温度影响明显,直接影响传感器测量精度,需要研究补偿方法,提高测量精度。采用无磁高低温试验箱测量磁通门传感器温度特性;提出基于径向基神经网络的温度误差补偿方法,分别建立磁通门磁力仪零漂误差补偿模型和刻度因子误差补偿模型。结果表明,径向基神经网络能良好逼近磁通门传感器参数的温度特性;与BP神经网络相比,径向基神经网络在零漂补偿中训练时间更短,精度更高,重复性更好,零漂误差的抑制能力更强。补偿后,磁通门磁力仪零漂误差从7.105 5 nT减少到0.766 1 nT;刻度因子误差从6.3E-3减少到7.2E-5;测量值温度误差由213.6 nT补偿到9.1 nT。提出建立通用的温度补偿模型,在不同磁场环境下经过反复测试,采用训练过的模型补偿后,温度误差均降低一个数量级,提高了磁通门磁力仪温度性能和精度。Accuracy of fluxgate magnetometers is badly influenced by temperature change.It is necessary to research a compensation method to improve the measurement accuracy.Temperature characteristic of magnetometer output values is studied in nonmagnetic temperature experiment box.RBF neural network is used to compensate the temperature error.Then,the offset drift error compensation model and scale factor error compensation model are established.Result indicates that RBF neural network has better performance in temperature drift compensation compared with BP neural network.After compensation,the offset drift temperature error is decreased from 7.105 5 nT to 0.766 1 nT,and the scale factor temperature error is decreased from 6.3E-3 to 7.2E-5.In addition,the temperature error of measurement value is decreased from 213.6 nT to 9.1 nT.In different magnetic circumstance,it is proved that the temperature error is reduced by one order of magnitude.A generalized compensation model is proposed,and the temperature characteristic and measurement accuracy of fluxgate magnetometers are improved using RBF neural network.

关 键 词：磁通门磁力仪 径向基神经网络 刻度因子 零偏 温度误差 补偿模型 

分 类 号：TH762.3[机械工程—仪器科学与技术]