基于遗传算法优化的LightGBM浮选尾煤灰分预测模型研究  

Research on a Prediction Model of Ash in Flotation Tailings Based on LightGBM Optimized by Genetic Algorithm

作　　者：韩杰 王然风[1] 付翔[1] 王珺 魏凯 张茜[1] HAN Jie;WANG Ranfeng;FU Xiang;WANG Jun;WEI Kai;ZHANG Qian(School of Mining Engineering,Taiyuan University of Technology,Taiyuan,Shanxi 030024,China)

机构地区：[1]太原理工大学矿业工程学院,山西太原030024

出　　处：《矿业研究与开发》2025年第1期269-276,共8页Mining Research and Development

基　　金：国家自然科学基金项目(52274157);内蒙古自治区重点专项项目(2022EEDSKJXM010);山西省重点研发计划项目(202102100401015)。

摘　　要：为了提高煤泥浮选过程灰分在线检测的水平,研究了煤泥浮选过程中尾煤灰分的预测,并建立了基于遗传算法优化的LightGBM浮选尾煤灰分预测模型。通过采集的生产过程数据,包括矿浆流量、浓度、起泡剂量、捕收剂量和干煤泥量,构建了模型训练数据集,并对模型进行了测试验证。试验结果显示,该模型预测结果的平均绝对误差为0.72,比未优化的LightGBM模型提升了11.1%的预测精度,相较于其他模型,GA-LightGBM模型预测结果的平均绝对误差降低了15.8%,进一步证明了所建模型在尾煤灰分预测精度上的有效性,为实现智能化浮选提供了新的技术支持。In order to improve the level of on-line detection of ash in coal slime flotation process,the prediction of ash content of tailings in coal slime flotation process was studied,and the prediction model of ash content of flotation tailings based on LightGBM optimized by genetic algorithm was established.Through the collected production data,including slurry flow,concentration,foaming dose,collection dose and dry slime amount,the model training data set was constructed and the model was tested and verified.The experimental results show that the average absolute error of the prediction results of the model is 0.72,which is 11.1%higher than that of the unoptimized LightGBM model.Compared with other models,the average absolute error of the prediction results of the GA-LightGBM model is reduced by 15.8%,which further proves the validity of the model in the prediction accuracy of ash in coal tailings and provides a new technical support for the realization of intelligent flotation.

关 键 词：煤泥浮选 尾煤灰分 遗传算法 LightGBM 预测建模 

分 类 号：TD94[矿业工程—选矿]