Detection and quantification of temperature sensor drift using probabilistic neural networks

作者：

Highlights：

• New machine learning method for detection of long-term temperature sensor drift.

• New anomaly detection method based on the trinomial distribution.

• Data-driven prediction of temperature in concrete structures is greatly improved.

• The method was successfully tested on 7-years data of a real structure.

摘要

•New machine learning method for detection of long-term temperature sensor drift.•New anomaly detection method based on the trinomial distribution.•Data-driven prediction of temperature in concrete structures is greatly improved.•The method was successfully tested on 7-years data of a real structure.

论文关键词：Long-term structural health monitoring,Data validation,Temperature sensor drift,Fiber optics,Machine learning,Probabilistic neural networks,Data-driven prediction,Temperature prediction,Anomaly detection

论文评审过程：Received 23 April 2022, Revised 16 August 2022, Accepted 18 September 2022, Available online 28 September 2022, Version of Record 3 October 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2022.118884