中南大学学报（自然科学版）2025，Vol.56Issue(8)：3186-3199,14.DOI:10.11817/j.issn.1672-7207.2025.08.009
基于SVR和NSGA-Ⅱ算法的烧结矿余热回收运行参数优化
Optimization of operating parameters for sinter waste heat recovery based on SVR and NSGA-Ⅱ algorithms
摘要
Abstract
Achieving near-zero emissions in sintering annular coolers is the key to realize the ultimate energy efficiency in the steel industry.The synergistic effect of temperature exergy and pressure exergy constrains the waste heat recovery capacity.An optimization method integrating CFD and intelligent algorithms were proposed to determine the operating parameters of the annular cooler when temperature exergy and pressure exergy were at the maximun value.Firstly,a CFD model was used to obtain the benchmark dataset.The sparrow search algorithm was employed to optimize the hyperparameters of the support vector regression,constructing a high-precision exergy prediction model.Furtherly,the nonlinear influence of parameters was analyzed through feature importance and partial dependence plots.Finally,the NSGA-Ⅱ algorithm was applied to obtain the Pareto-optimal solution set.The results reveal that the cooling gas inlet flow rate is the most significant factor affecting exergy.Increasing the cooling gas inlet flow rate and the material layer height can significantly increase the predicted temperature exergy and decrease the predicted pressure exergy.Maximizing flow rate and minimizing particle size enable exergy to reach the extreme values.The best parameter combination of multi-objective optimization yields is cooling gas inlet flow rate of 37.44×104 m3·h-1,particle size of 0.028 m,material layer height of 1.42 m,and power recovery section length of 81.20 m.The temperature exergy of the optimal point of multi-objective optimisation is improved by 53.3％compared with the optimisation results of orthogonal experiments,and the pressure exergy is reduced by 74.1％.关键词
烧结矿/余热回收/机器学习/多目标优化Key words
sinter/waste heat recovery/machine learning/multi-objective optimization分类
能源科技引用本文复制引用
郝学智,赵亮,吴问昌,张晓虎,冯军胜,董辉..基于SVR和NSGA-Ⅱ算法的烧结矿余热回收运行参数优化[J].中南大学学报（自然科学版）,2025,56(8):3186-3199,14.基金项目
中国宝武低碳冶金创新基金资助项目(BWLCF202307)(Project(BWLCF202307)supported by the Baowu Low Carbon Metallurgy Innovation Foundation of China) （BWLCF202307）
