光纤压力传感器测量精度的研究

发布时间：2018-02-21 12:00

  本文关键词： 光纤压力传感器 数据融合 二元回归 粒子群优化BP神经网络 无线传输　出处：《东华大学》2017年硕士论文　论文类型：学位论文

【摘要】：传感器技术,是信息科学的基础,也是现代物联网信息技术重要支柱之一。现代工业持续发展,对传感器的测量精度提出的要求越来越高。随着光纤工艺的发展,利用聚合物封装的反射式光纤压力传感器,因其重量轻、尺寸小、传输距离远、耐腐蚀性能好、抗电磁干扰性能强的特点,获得广泛应用。由于温度对反射式光纤压力传感器测量精度的影响比较大,所以对反射式压力传感器的温度补偿进行研究。本文采用数据融合技术解决温度对反射式光纤压力传感器测量精度影响的问题,其主要研究内容如下:(1)分析了反射式光纤压力传感器的工作原理,并对几种常用数据融合算法如多元回归、传统BP神经网络以及粒子群优化BP神经网络方法进行了论述。(2)针对反射式光纤压力传感器测量精度容易受温度影响较大的问题,采用多传感器信息融合技术,对反射式光纤压力传感器进行温度补偿。将反射式光纤压力传感器设置为主传感器,DS18B20温度传感器设为辅助传感器,并进行了二维温度标定实验。根据标定实验数据,求出未补偿前的灵敏度温度系数和温度附加误差的值分别为8.4808×10~(-3）/℃和32.3120%。(3)建立了多元回归、传统的BP神经网络以及粒子群优化BP神经网络算法的温度补偿模型,对反射式光纤压力传感器进行温度补偿,并且比较几种融合算法的补偿效果。(4)本文采用MSP430F149单片机设计出一个具有无线传输功能的实时压力测量及补偿系统。本系统主要包含三个部分:第一部分,将反射式光纤压力传感器和温度传感器采集数据送到上位机;第二部分,nRF24L01短距离无线通信传输模块;第三部分,下位机对无线模块传来的数据进行处理,并且包括了报警、显示等外围硬件电路。在IAR软件开发平台上运用C语言对各个功能模块进行程序的设计,通过主程序调用各个子模块,使系统能够正常实现工作。研究结果是,通过利用二元回归和传统BP神经网络算法对光纤压力传感器进行温度补偿后,其温度灵敏度系数从原来的8.4808×10~(-3）/℃分别改善为1.9412×10~(-3）/℃和1.0991×10~(-3）/℃,温度附加误差由原来的32.3120%,改善为7.4615%和4.1877%,两种算法对反射式光纤压力传感器性都有一定改善;而用粒子群优化BP神经网络方法求出补偿后的灵敏度温度系数和温度附加误差,分别为2.2734×10-4/℃和0.8622%。从以上数据可知,粒子群优化BP神经网络把反射式光纤压力传感器的灵敏度温度系数提升了1个数量级,温度附加误差提升了2个数量级。与前两种算法比较可知,粒子群优化BP神经网络算法对温度补偿效果更佳。
[Abstract]:Sensor technology is the foundation of information science and one of the important pillars of information technology in modern Internet of things. With the development of modern industry, the requirement of measuring precision of sensor is higher and higher. With the development of optical fiber technology, The reflective fiber optic pressure sensor encapsulated by polymer has the advantages of light weight, small size, long transmission distance, good corrosion resistance and strong resistance to electromagnetic interference. It has been widely used. Because of the influence of temperature on the measuring accuracy of reflective fiber optic pressure sensor, Therefore, the temperature compensation of reflective pressure sensor is studied. In this paper, data fusion technology is used to solve the problem of the influence of temperature on the measurement accuracy of reflective fiber optic pressure sensor. The main research contents are as follows: (1) the working principle of reflective fiber optic pressure sensor is analyzed, and several commonly used data fusion algorithms such as multivariate regression are discussed. The traditional BP neural network and particle swarm optimization BP neural network are discussed. Aiming at the problem that the measurement accuracy of reflective optical fiber pressure sensor is easily affected by temperature, the multi-sensor information fusion technique is adopted. The reflective fiber optic pressure sensor is used to compensate the temperature. The main sensor, the DS18B20 temperature sensor, is set up as the auxiliary sensor, and the two-dimensional temperature calibration experiment is carried out. The values of sensitivity temperature coefficient and temperature additional error before compensation are 8.4808 脳 10 ~ (-3) / 鈩，

本文编号：1521921