FBG Based Optical Weight Measurement System and Its Performance Enhancement Using Machine Learning
FBG Based Optical Weight Measurement System and Its Performance Enhancement Using Machine Learning
复制标题
DOI：
10.1109/jsen.2022.3144173
复制
发表时间：
2022-03-01
影响因子：
4.3
通讯作者：
Thangaraj, Jaisingh
中科院分区：
文献类型：
--
作者：
Pal, Deep;Kumar, Amitesh;Thangaraj, Jaisingh
工业重量测量对质量和生产成本的监控和优化至关重要。光纤布拉格光栅（fbg）是使用最广泛的光纤传感器之一，因为它们对温度和应变波动具有很高的灵敏度，使其成为各种工业应用的理想选择。本文提出了一种光学称重系统，该系统包括四个基于FBG的称重传感器，并使用不同的机器学习算法对其性能进行了详细分析，以提高精度和测量范围。在这里，由悬臂式测压元件上产生的应变引起的布拉格波长的位移是使用光学询问器获得的，给出了波长编码的FBG数据。数据在PC上获取，其中特定的机器学习模型，如线性回归（LR），支持向量机（svm），神经网络（NN），高斯过程回归（GPR）被优化选择，并用于直接学习称重传感器（布拉格波长）变化与应用权重之间的映射。实验分析表明，GPR算法比其他任何标准方法都能提供更精确的重量测量。该模型保证了该方法的实用性，提高了基于光纤光栅的重量测量系统的测量速度和精度。
Industrial weight measurement is essential to monitoring and optimizing the quality and production cost. Fiber Bragg gratings (FBGs) are among the most widely used fiber optic sensors because they have a high sensitivity to temperature and strain fluctuations, making them ideal for various industrial applications. This paper proposes an optical weight measurement system that includes four FBG based load cells with a detailed analysis of their performance using different machine learning algorithms to improve accuracy and measurement range. Here, the shift in Bragg wavelength caused by strain generated on cantilever type load cell is acquired using an optical interrogator, giving the wavelength encoded FBG data. The data is acquired on PC, where specific Machine Learning models such as Linear Regression (LR), Support vector machines (SVMs), Neural networks (NN), Gaussian Process regression (GPR) are optimally selected and are utilized to directly learn the mapping between changes in load cell (Bragg wavelength) and applied weight. The experimental analysis shows that the GPR algorithm provides more accurate weight measurement than any other standard approach. The model ensures the practicality of the proposed method with improvement in measurement speed and accuracy of the FBG based weight measurement system.