Volume 19: Sustainable Energy Solutions for a Post-COVID Recovery towards a Better Future: Part II

Prediction of thermal conductivity of underground tar-rich coal seam based on support vector machine Xing Ning, Chang'an Wang, Lei Deng, Tao Zhu, Xiangyu Xue, Defu Che*

Abstract

The in-situ pyrolysis conversion of coal for extracting the tar is carried out to decrease the solid waste, reduce environmental pollution, and ensure energy safety. However, thermal conductivity, the most key thermal parameter, is quite indistinct for underground tar-rich coal seam under actual conditions. To obtain the thermal conductivity of underground tar-rich coal seam under actual conditions, the non-linear regression algorithm model of support vector machine was constructed. The results show that the training model demonstrates favorable generalization ability for predicting in-situ thermal conductivity of tar-rich coal seam. Moreover, the trained model subsequently predicts thermal conductivity of underground tar-rich coal seam with positive matching and reliability in the testing sets. The predicted study may promote further elucidation of the thermal conductivity evolution during the in-situ pyrolysis of tar-rich coal seam.

Keywords tar-rich coal seam, in-situ pyrolysis, thermal conductivity, support vector machine