摘要
在温度变化的环境中,石英振梁加速度计(QVBA)的输出会发生漂移。为了改善QVBA的温度稳定性,提出了基于改进鲸鱼算法(IWOA)优化BP神经网络的温度补偿模型。IWOA通过优化BP神经网络的初始权值和阈值,克服了BP神经网络易陷入局部最优的缺点,增强了BP神经网络在训练中的准确性和鲁棒性。全温实验表明,该方法能够明显抑制QVBA因温度而产生的漂移。经过补偿,全温零偏稳定性从4.161 mg下降至0.196 mg,全温标度因数稳定性从59.676 ppm下降至35.751 ppm,验证了模型的有效性。
The output of a quartz vibrating beam accelerometer(QVBA)drifts in an environment with varying temperatures.In order to improve the temperature stability of QVBA,a temperature compensation model based on the improved whale optimization algorithm(IWOA)to optimize the BP neural network was proposed.By optimizing the initial weights and thresholds of the BP neural network,the IWOA overcame the shortcoming that the BP neural network was easy to fall into local optimum,and enhanced the accuracy and robustness of the BP neural network in training.Full-temperature experiments show that this method can significantly suppress the temperature-induced drift of QVBA.After compensation,the bias stability decreased from 4.161 mg to 0.196 mg,and the scale factor stability decreased from 59.676 ppm to 35.751 ppm,which verified the validity of the model.
作者
张晗
林盛受
毛志成
梁金星
ZHANG Han;LIN Shengshou;MAO Zhicheng;LIANG Jinxing(School of Instrument Science and Engineering,Southeast University)
出处
《仪表技术与传感器》
CSCD
北大核心
2024年第3期110-115,共6页
Instrument Technique and Sensor
关键词
石英振梁加速度计
温度补偿
WOA算法
BP神经网络
quartz vibrating beam accelerometer
temperature compensation
whale optimization algorithm
back propagation neural network