Jianqiang, M., Shan, C., & Cheng, C. (2025). Sensor fault diagnosis using Principal Component analysis and convolutional neural network for offshore structural health monitoring. Measurement: Sensors, 38, 101465. https://doi.org/10.1016/j.measen.2024.101465
Abstract:
In this the paper, a data driven methodology for sensor fault diagnosis using Principal Component Analysis (PCA) and Convolutional Neural Network (CNN) is proposed for in-situ sensing data quality assurance during the offshore structural operations. A sensor network consisting of different type of sensors including strain gauge, accelerometer, force sensor and displacement sensor was built up, for structural health monitoring (SHM) of an offshore structural tubular joint. The sensor faults were tested at laboratory using the tubular joint under a cyclic loading condition emulating the excitations from ocean waves. The proposed methodology using the PCA and CNN was examined with the experimental measurement data and results of the sensor faults. It was demonstrated that the proposed methodology was capable of the classification accuracy up to 99% for sensor fault diagnosis in sensor network for offshore structural health monitoring.
License type:
Publisher Copyright
Funding Info:
This research / project is supported by the Agency for Science, Technology and Research - RIE2020 Industry Alignment Fund - Pre-Positioning -ENSURE Programme
Grant Reference no. : A19F1a0104