三轴磁通门传感器受温度影响明显,严重影响其测量准确度,需要研究补偿方法,提高测量准确性.在不同磁场环境下,利用无磁高低温试验箱对传感器输出值温度特性进行了测试,并采用BP神经网络的方法进行温度补偿.分别阐述了设备操作过程及数据处理方式.采集传感器在不同温度下的测量数据样本;将BP神经网络应用于温度误差模型的非线性辨识,训练出了有效的温度补偿网络;在不同外加激励磁场下分别进行补偿;对BP神经网络、径向基神经网络和曲线拟合的逼近效果进行了对比.结果表明,传感器温度误差从195.6 nT,203.2 nT,213.6 nT分别补偿到17.18 nT,18.89 nT,18.04 nT.温度误差明显减少,证明了BP神经网络在磁通门传感器校正中的良好性能;经过对比,证明了BP神经网络具有更高的逼近精度.%Accuracy of fluxgate magnetometers is badly influenced by temperature change. Temperature characteristic of magnetometer output was researched in different magnetic circumstances by nonmagnetic temperature experiment box, and then, BP Neural Network was used to compensate error caused by temperature. Experimental process and data processing method were introduced in detail. Firstly, sample data were obtained at different temperatures. Then, BP Neural Network was used in temperature error modeling and compensation network training, and temperature error was compensated in different magnetic circumstances.Finally, testing data temperature error was compensated by the trained network. It demonstrates that temperature error is reduced from 195.6 nT, 203.2 nT, 213.6 nT to 17.18 nT, 18.89 nT, 18.04 nT, respectively. Obviously, temperature error is suppressed greatly, which proves good performance of BP neural network in fluxgate magnetometers calibration.
展开▼
