A spatiotemporal prediction approach for a 3D thermal field from sensor networks

Abstract

Thermal fields exist widely in engineering systems and are critical for engineering operation, product quality and system safety in many industries. An accurate prediction of thermal field distribution, that is, acquiring any location of interest in a thermal field at the present and future time, is essential to provide useful information for the surveillance, maintenance, and improvement of a system. However, thermal field prediction using data acquired from sensor networks is challenging due to data sparsity and missing data problems. To address this issue, we propose a field spatiotemporal prediction approach based on transfer learning techniques by studying the dynamics of a 3 D thermal field from multiple homogeneous fields. Our model characterizes the spatiotemporal dynamics of the local thermal field variations by considering the spatiotemporal correlation of the fields and harnessing the information from homogeneous fields to acquire an accurate thermal field distribution in the future. A real case study of thermal fields during grain storage is conducted to validate our proposed approach. Grain thermal field prediction results provide a deep insight of grain quality during storage, which is helpful for the manager of grain storage to make further decisions about grain quality control and maintenance.

Keywords:

• 3D thermal field
• spatiotemporal prediction
• transfer learning

Additional information

Funding

This work was supported in part by the National Science Foundation of China under Grants 71771004, 71932006, 71690232, 71571003, and in part by the National Science Foundation under Grant 1637772.

Notes on contributors

Di Wang

Di Wang received the B.S. degree in industrial engineering from Nankai University, Tianjin, China, in 2015, and the Ph.D. degree in management science and engineering from Peking University, Beijing, China, in 2020. She is currently an Assistant Professor with the Department of Industrial Engineering and Management, School of Mechanical Engineering, Shanghai Jiao Tong University, Shanghai, China. Her research interests include statistical modeling of spatiotemporal data and artificial intelligence of complex engineering systems. Dr. Wang is a member of INFORMS and IISE.

Kaibo Liu

Kaibo Liu received the B.S. degree in industrial engineering and engineering management from Hong Kong University of Science and Technology, Hong Kong, in 2009, and the M.S. degree in statistics and the Ph.D. degree in industrial engineering from the Georgia Institute of Technology, Atlanta, GA, USA, in 2011 and 2013, respectively. He is currently an Associate Professor with the Department of Industrial and Systems Engineering, University of Wisconsin–Madison, Madison, WI, USA, where he is also the Associate Director of the UW–Madison IoT Systems Research Center. His research interests include system informatics, big data analytics, and data fusion for process modeling, monitoring, diagnosis, prognostics, and decision making. Dr. Liu is a member of ASQ, INFORMS, SME, and IISE.

Xi Zhang

Xi Zhang received the B.S. degrees in mechanical engineering and business administration from Shanghai Jiao Tong University, Shanghai, China, in 2006, and the Ph.D. degree in industrial and management systems engineering from the University of South Florida, Tampa, FL, USA, in 2010. He is currently an Associate Professor with the Department of Industrial Engineering and Management, Peking University, Beijing, China. His research interests include physics-based engineering data integration and analytics for process monitoring, diagnosis, control and optimization in complex engineering, and service systems. Dr. Zhang is a member of ASQ, INFORMS, and IISE.