摘要
由于压阻式压力传感器存在温度漂移,而现有的软件温度补偿方法依赖额外的温度传感器获取温度信号。为了简化这一流程,提出了一种基于BP神经网络的压力传感器原位温度补偿方法。利用多参数测量方法,仅依赖压力传感器自身的电学信号,无需引入新的传感器,便能实现对传感器原位温度及压力的测量;进一步通过BP神经网络实现压力传感器温-压解耦及温度补偿。结果显示,补偿后传感器输出误差降低至±0.5%FS以内,零位温度漂移从0.021%FS/℃降低到0.002 5%FS/℃,灵敏度温度漂移从0.15%FS/℃降低到0.005 5%FS/℃,显著降低了零位温度漂移和灵敏度温度漂移。
Due to the temperature drift of piezoresistive pressure sensors,existing software temperature compensation methods rely on additional temperature sensors to obtain temperature signals.To simplify this process,a pressure sensor in-situ temperature compensation method based on BP neural networks is proposed.Utilizing a multi-parameter measurement method,it can achieve in-situ temperature and pres-sure measurements of the sensor based solely on its electrical signals,without the need to introduce new sensors.Furthermore,it achieves pressure sensor temperature-pressure decoupling and temperature com-pensation through BP neural networks.The results show that the sensor output error is reduced to within±0.5%FS after compensation,with the zero-position temperature drift reduced from 0.021%FS/℃to 0.0025%FS/℃,and the sensitivity temperature drift reduced from 0.15%FS/℃to 0.0055%FS/℃,significantly reducing the zero-position temperature drift and sensitivity temperature drift.
作者
刘雨桥
张姝
雷程
余建刚
唐梦璇
王涛龙
梁庭
LIU Yuqiao;ZHANG Shu;LEI Cheng;YU Jiangang;TANG Mengxuan;WANG Taolong;LIANG Ting(State Key Laboratory of Dynamic Measurement Technology,North University of China,Taiyuan 030051,China;Tianjin Jinhang Institute of Technical Physics,Tianjin 300308,China;Chengdu Spaceon Electronics Co.,Ltd.,Chengdu 610037,China)
出处
《测试技术学报》
2025年第1期13-19,共7页
Journal of Test and Measurement Technology
基金
国家重点研发计划资助项目(2023YFB3208500)
山西省重点研发计划资助项目(2023020302010)
中央引导地方科技发展资金资助项目(YDZJSX20231B006)。
关键词
压力传感器
温度补偿
多参数测量
BP神经网络
pressure transducer
temperature compensation
multi-parameter measurement
BP neural network
