Temperature Uncertainty Reduction Algorithm Based on Temperature Distribution Prior for Optical Sensors in Oil Tank Ground Settlement Monitoring

Abstract

Ground settlement (GS) in an oil tank determines its structural integrity and commercial service. However, GS monitoring faces challenges, particularly due to the significant temperature differences induced by solar radiation around the tank in daytime. To address this problem, this paper digs out a prior and proposes a temperature uncertainty reduction algorithm based on that. This prior has a spatial Gaussian distribution of temperature around the tank, and numerical simulation and practical tests are conducted to demonstrate it. In addition, combining uniformly packaged sensor probes and the spatial prior of temperature, the temperature uncertainty is verified to be Gaussian-distributed too. Then, the overall temperature uncertainty can be captured by Gaussian fitting and then removed. The practical test verified a 91% reduction rate in temperature uncertainty, and this approach enables GS sensors to effectively perform daytime monitoring by mitigating temperature-related uncertainties.