杨宜坤,吴震,刘洪皓,张早校
YANG Yikun,WU Zhen,LIU Honghao,ZHANG Zaoxiao

• 1:西安交通大学化学工程与技术学院
• 2:绿色氢电全国重点实验室

摘要(Abstract)：

金属氢化物材料具有储氢/热密度高，工作温度区间广，无污染无腐蚀性的优点，被认为是理想的储氢/热材料。金属氢化物储氢/热材料可以通过掺杂不同元素形成多元合金，以开发具有不同目标性能的材料。这种方法依赖实验合成，十分耗费时间和经济成本。数据驱动的机器学习性能预测模型可以解决这一问题，通过测试对比最小二乘回归、最小绝对收缩和选择操作符回归、岭回归、弹性网络回归、支持向量回归和随机森林回归多种回归算法，成功建立了金属氢化物微观材料性质与宏观形成能之间的关系。测试结果显示随机森林回归具有最好的预测性能，在训练集和测试集上相对误差均较小，仅为3.078和8.201 1，且决定系数较高，具有良好的回归能力和泛化能力。SHAP分析中表明组成金属氢化物的基态原子体积的均值和最值具有高达5.56和1.26的SHAP值，这2个因素很大程度上决定了金属氢化物材料的形成能大小。对Mg基，Ca基，AB、AB2及AB5型金属氢化物材料预测结果分析显示预测相对误差均在9%以下，证明了模型准确性及普适性，可用于未知数据集的形成能预测。
Metal hydride thermal/hydrogen energy storage material is considered ideal candidate due to high energy density,wide working temperature range and lack of corrosive pollutants. Multi-component metal hydride alloys can be formed by doping with different elements to obtain various target properties. However, conventional material development method relies on experimental synthesis,having the disadvantages of time-consuming and costly. Data-driven machine learning prediction model is capable of addressing this problem. By comparing varieties of regression algorithms such as least squares regression,least absolute shrinkage and selection operator regression,ridge regression,elastic net regression,supporting vector regression,and random forest regression,the relationship between the microscopic properties of metal hydride materials and their macroscopic formation energy are established. Results show that random forest regression have the best prediction performance,with lowest relative errors on both the training and test sets of 3.078 and 8.201 1,high R-squared values,and great generalization and regression abilities. SHAP analysis reveals extreme and mean value of ground state atom of metal hydride exhibit the greatest SHAP value of 5.56 and 1.26,suggesting their significant influence on the formation energy.Analysis for the prediction value of Mg-base,Ca-base,AB type,AB2 type,and AB5 type metal hydrides shows the highest relative error below 9%,further proving the accuracy and universality of the model for all types of metal hydride. This model can be used to predict the formation enthalpy of unknown datasets.

关键词(KeyWords)： 太阳能热利用;金属氢化物;储氢储热;机器学习;性能对比
thermal application of solar energy;metal hydride;hydrogen and heat storage;machine learning;performance comparison

Abstract:

Keywords:

基金项目(Foundation): 国家自然科学基金面上资助项目（52376208;52176203）

作者(Author): 杨宜坤,吴震,刘洪皓,张早校
YANG Yikun,WU Zhen,LIU Honghao,ZHANG Zaoxiao

DOI: 10.13226/j.issn.1006-6772.HH24103101

参考文献(References)：

• [1]BELA?D F,AL-SARIHI A,AL-MESTNEER R. Balancing climate mitigation and energy security goals amid converging global energy crises:The role of green investments[J]. Renewable Energy,2023,205:534-542.
• [2]姚玉璧，郑绍忠，杨扬，等.中国太阳能资源评估及其利用效率研究进展与展望[J].太阳能学报，2022,43(10):524-535.YAO Yubi,ZHENG Shaozhong,YANG Yang,et al. Progress and prospects on solar energy resource evaluation and utilization efficiency in China[J]. Acta Energiae Solaris Sinica, 2022, 43(10):524-535.
• [3]BEZDUDNY A V,BLINOV D V,DUNIKOV D O. Single-stage metal hydride-based heat storage system[J]. Journal of Energy Storage,2023,68:107590.
• [4]MALLESWARARAO K,DUTTA P,MURTHY S S. Applications of metal hydride based thermal systems:A review[J]. Applied Thermal Engineering,2022,215:118816.
• [5]DANGWAL S,IKEDA Y,GRABOWSKI B,et al. Machine learning to explore high-entropy alloys with desired enthalpy for roomtemperature hydrogen storage:Prediction of density functional theory and experimental data[J]. Chemical Engineering Journal,2024,493:152606.
• [6]AGAFONOV A, PINEDA-ROMERO N, WITMAN M, et al.Destabilizing high-capacity high entropy hydrides via earth abundant substitutions:From predictions to experimental validation[J].Acta Materialia,2024,276:120086.
• [7]YIN Y, LI B, YUAN Z M, et al. Microstructure and improved hydrogen storage properties of Mg85Zn5Ni10 alloy catalyzed by Cr2O3 nanoparticles[J]. Journal of Physics and Chemistry of Solids,2019,134:295-306.
• [8]ZHONG H C,HUANG Y S,DU Z Y,et al. Enhanced Hydrogen Ab/De-sorption of Mg(Zn)solid solution alloy catalyzed by YH2/Y2O3 nanocomposite[J]. International Journal of Hydrogen Energy,2020,45(51):27404-27412.
• [9]RKHIS M,LAASRI S,TOUHTOUH S,et al. New insights into the electrochemical and thermodynamic properties of AB-type ZrNi hydrogen storage alloys by native defects and H-doping:Computational experiments[J]. International Journal of Hydrogen Energy,2023,48(27):10089-10097.
• [10]GU X D, WANG F, CHENG J L, et al. Positive correlation of Nb/Cr doping with dehydrogenation performance of ZrCo-based hydrides[J]. International Journal of Hydrogen Energy, 2023,48(67):26276-26287.
• [11]DONG S Y,WANG Y Y,LI J Y,et al. Exploration and design of Mg alloys for hydrogen storage with supervised machine learning[J]. International Journal of Hydrogen Energy, 2023,48(97):38412-38424.
• [12]MOOSAVI S M,JABLONKA K M,SMIT B. The role of machine learning in the understanding and design of materials[J]. Journal of the American Chemical Society,2020,142(48):20273-20287.
• [13]ZHAI S, XIE H P, CUI P, et al. A combined ionic Lewis acid descriptor and machine-learning approach to prediction of efficient oxygen reduction electrodes for ceramic fuel cells[J]. Nature Energy,2022,7:866-875.
• [14]MOOSAVI S M,NANDY A,JABLONKA K M,et al. Understanding the diversity of the metal-organic framework ecosystem[J].Nature Communications,2020,11(1):4068.
• [15]GHEYTANZADEH M, RAJABHASANI F, BAGHBAN A, et al.Estimating hydrogen absorption energy on different metal hydrides using Gaussian process regression approach[J]. Scientific Reports,2022,12(1):21902.
• [16]SAAL J E, KIRKLIN S, AYKOL M, et al. Materials design and discovery with high-throughput density functional theory:The open quantum materials database(OQMD)[J]. JOM, 2013,65(11):1501-1509.
• [17]KIRKLIN S, SAAL J E, MEREDIG B, et al. The open quantum materials database(OQMD):Assessing the accuracy of DFT formation energies[J]. NPJ Computational Materials, 2015, 1:15010.
• [18]JAIN A,ONG S P,HAUTIER G,et al. Commentary:The Materials Project:A materials genome approach to accelerating materials innovation[J]. APL Materials,2013,1(1):011002.
• [19]WARD L, AGRAWAL A, CHOUDHARY A, et al. A generalpurpose machine learning framework for predicting properties of inorganic materials[J]. NPJ Computational Materials, 2016, 2:16028.
• [20]RODRIGUEZ-GALIANO V,SANCHEZ-CASTILLO M,CHICAOLMO M,et al. Machine learning predictive models for mineral prospectivity:An evaluation of neural networks, random forest,regression trees and support vector machines[J]. Ore Geology Reviews,2015,71:804-818.
• [21]AKIBA T,SANO S,YANASE T,et al. Optuna:a next-generation hyperparameter optimization framework[C]//Proceedings of the25th ACM SIGKDD International Conference on Knowledge Discovery&Data Mining. Anchorage AK USA. ACM, 2019:2623–31.10. 1145/3292500.3330701.