施龙青, 吴洪斌, 李永雷,等.导水裂隙带发育高度预测的PCA-GA-Elman优化模型[J].河南理工大学学报（自然科学版）,2021,40(4):10-18.

SHI L Q, WU H B, LI Y L, et al.Optimization model of PCA-GA-Elman for development height prediction ofwater-conducting fissure zone[J].Journal of Henan Polytechnic University(Natural Science) ,2021,40(4):10-18.

导水裂隙带发育高度预测的PCA-GA-Elman优化模型

施龙青1, 吴洪斌1, 李永雷2, 吕伟魁3

1.山东科技大学 地球科学与工程学院，山东 青岛 266590;2.济宁能源发展集团有限公司，山东济宁 272000;3.山东新巨龙能源有限责任公司，山东 荷泽 274918

摘要:针对矿井水害防治中导水裂隙带发育高度难以准确预测的问题，结合煤矿生产实际和工程地质理论，选取采深S、硬岩岩性比例系数b、采高（煤层厚度）M、工作面斜长I、顶板单轴抗压强度为主要影响因素，运用灰色关联分析法（GRA）分析各主要影响因素与导水裂隙带发育高度的相关性，并将主成分分析（PCA）、遗传算法（GA）和优化的Elman神经网络相结合，建立导水裂隙带发育高度预测的PCA-GA-Elman优化模型。结果表明：PCA-GA-Elman优化模型能有效消除因素间的相互影响，并能优化初始权值和阈值，使导水裂隙带发育高度预测更加准确。与PCA-Elman和PCA-BP预测模型相比，PCA-GA-Elman优化模型预测的导水裂隙带发育高度相对误差仅为-6.34% ~0.18%。

关键词:导水裂隙带;发育高度;主成分分析;遗传算法;Elman神经网络;PCA-GA-Elman  优化模型

doi:10.16186/j.cnki.1673-9787.2020050031

基金项目:国家自然科学基金资助项目（51804184 ，41807283 ）；山东省自然科学基金资助项目（ZR2020KE023 ）

收稿日期:2020/05/11

修回日期:2020/09/04

出版日期:2021/07/15

Optimization model of PCA-GA-Elman for development height prediction ofwater-conducting fissure zone

SHI Longqing1, WU Hongbin1, LI Yonglei2, LYU Weikui3

1.College of Earth Sciences and Engineering， Shandong University of Science and Technology， Qingdao  266590 ， Shandong， China;2.Jining Energy Development Group Co. ， Ltd. ，Jining  272000 ， Shandong， China;3.Shandong New Julong Energy Co. ，Ltd. ， Heze  274918 ， Shandong， China

Abstract:For the problem of difficult prediction of the development height of water-conducting fissure zone in mine flood prevention and control， by combining the actual coal production with engineering geology theory， and by combining the principal component analysis （ PCA ） with the genetic algorithm （ GA） and with the optimization of Elman neural network ， the mining depth s ， the hard rock lithology ratio coefficient b ， the mining height （ coal thickness） M， the face inclined length l， and the roof uniaxial resistance pressure strength were selected as the main influence factors to established the optimization model of PCA-GA-Elman for development height prediction of water-conducting fissure zone. Grey relation analysis （ GRA ） was used to analysis the correlation of main influence factors and the development height of water-conducting fissure zone. The results showed that the optimization model of PCA-GA-Elman could effectively eliminate the mutual influence between factors， and could optimize the initial weights and threshold so as to make the development height prediction of water-conducting fissure zone more accurate. Compared with the PCA-Elman and PCA-BP prediction models， the relative error of the water-conducting fissure zone development height predicted by the optimization model of PCA-GA-Elman was only 一6.34% 〜0.18%.

Key words:water-conducting fissure zone;development height;principal component analysis;genetic algorithm;Elman neural network;optimization model of PCA-GA-Elman

导水裂隙带发育高度预测的PCA_GA_Elman优化模型_施龙青.pdf