Modeling of a three-dimensional dynamic thermal field under grid-based sensor networks in grain storage

Abstract

Thermal management is a major task in granaries, due to the essential role of temperature in grain storage. The accurate acquisition and updating of thermal field information generates a meaningful index for grain quality surveillance and storage maintenance actions. However, given the unknown mechanisms of local uncertainties, including local grain degradation and fungal infections that may significantly vary the thermal field in granaries, the appropriate modeling of field dynamics remains a challenging task. To address this issue, this article combines a three-dimensional (3D) nonlinear dynamics model with a stochastic spatiotemporal model to capture a 3D dynamic thermal map. To best harness the temperature data from the grid-based sensor network, we integrate the Kriging model into the Gaussian Markov random field model by introducing an anisotropic covariance function. Both simulation and real case studies are conducted to validate our proposed approach, and the results show that our approach outperforms other alternative methods for field estimation.