文献类型：期刊文献

英文题名：AutoML for calorific value prediction using a large database from the coal gasification practices in China

作者：Guo, Yuchao[1];Liu, Xia[1];Gao, Yunfei[1];Wang, Xiaoyu[1];Ding, Lu[1];Pan, Weitong[1];Hua, Cheng[2];He, Yulian[3,4];Chen, Xueli[1];Dai, Zhenghua[1];Yu, Guangsuo[1];Wang, Fuchen[1]

机构：[1]East China Univ Sci & Technol, Inst Clean Coal Technol, Shanghai 200237, Peoples R China;[2]Shanghai Jiao Tong Univ, Antai Coll Econ & Management, Shanghai 200230, Peoples R China;[3]Univ Michigan Shanghai Jiao Tong Univ Joint Inst, Shanghai 200240, Peoples R China;[4]Shanghai Jiao Tong Univ, Sch Chem & Chem Engn, Shanghai 200240, Peoples R China

年份：2025

卷号：12

期号：1

外文期刊名：INTERNATIONAL JOURNAL OF COAL SCIENCE & TECHNOLOGY

收录：;EI(收录号：20252818745826);WOS:【SCI-EXPANDED(收录号:WOS:001526604400001)】；

基金：This work was supported by National Key Research and Development Program of China (2022YFB4101900), National Natural Science Foundation of China (22208104), Shanghai Yangfan Program (22YF1410300) and Shanghai Chenguang Program (21CGA35).

语种：英文

外文关键词：Coal calorific value; Big data; Automated machine learning; Model interpretability; Model adaptability

摘要：Calorific value is one of the most important properties of coal. Machine learning (ML) can be used in the prediction of calorific value to reduce experimental costs. China is one of the world's largest coal production countries and coal occupies an important position in its national energy structure. However, ML models with a large database for the overall regions of China are still missing. Based on the extensive coal gasification practices in East China University of Science and Technology, we have built ML models with a large database for overall regions of China. An AutoML model was proposed and achieved a minimum MSE of 1.021. SHAP method was used to increase the model interpretability, and model validity was proved with literature data and additional in-house experiments. The model adaptability was discussed based on the databases of China and USA, showing that geography-specific ML models are essential. This study integrated a large coal database and AutoML method for accurate calorific value prediction and could offer key tools for Chinese coal industry.