Evaluation of Reconstruction Strategies of 3D Temperature Fields for Detecting Hotspots in Grain Bulk

Abstract

A hotspot in a grain bulk is a localized high temperature zone and normally spoilage begins in this location. The reconstruction strategies of a 3D space are investigated to detect hotspots in grain bulk by acoustic tomography. Since the size of a grain silo is usually large, the whole space to be reconstructed is divided into three blocks and two reconstruction ways: whole reconstruction and block reconstruction, are considered. Three sound ray path selection rules including intra-block, generalized intra-block and unrestricted paths are investigated. These reconstruction ways and path selection rules are combined into four reconstruction strategies. To fully compare the abilities of these strategies in detecting hotspots in grain bulk, 2304 single-hotspot temperature fields and a three-hotspot temperature field are reconstructed by using noise-free and noisy travel-time data. Based on reconstruction results and difficulty in travel-time measurement, the strategy using whole reconstruction way and generalized intra-block paths is the optimal choice.