采矿与安全工程学报2025，Vol.42Issue(2)：273-281,9.DOI:10.13545／j.cnki.jmse.2022.0679
基于PCA-SSA-BPNN模型的矿井突水水源识别方法
Identification method of mine water inrush source based on PCA-SSA-BPNN
摘要
Abstract
The traditional similarity identification method of hydrochemical characteristics is of low ac-curacy under some complex hydrogeological conditions.In view of this fact,principal component analy-sis(PCA)was adopted to reduce the dimension of hydrogeochemical sample data,and sparrow search algorithm(SSA)was introduced to optimize the weight parameters of back propagation neural network(BPNN).In this way,an identification model of mine water inrush source based on PCA-SSA-BPNN was constructed in the hope of accurately identifying water inrush sources in coal mines.Furthermore,both the constructed model and the similarity identification method of hydrochemical characteristics were applied to the water inrush source identification project of the 111804 working face in Xinshanghai No.1 Coal Mine.The results show that PCA can effectively reduce the information redundancy in the original data and eliminate the correlation between different hydrochemical types.Meanwhile,SSA remarkably enhances the overall optimization ability and prediction accuracy of BPNN.PCA-SSA-BPNN achieves accuracy of 96.7％in identifying the water inrush source,being 13.4％higher than BPNN.As the 111084 working face lies under complex hydrogeological conditions,the similarity identification method of hydrochemical characteristics fails to identify its water inrush source accurately.In this case,the PCA-SSA-BPNN model was employed to identify the water inrush source,and it identifies the Zhiluo Forma-tion aquifer as the source,which is consistent with the actual situation.This study can provide reference for research on accurate identification of mine water inrush sources.关键词
突水水源识别/水化学特征/主成分分析/麻雀算法/BP神经网络Key words
water inrush source identification/hydrogeochemical characteristics/principal component analysis/sparrow search algorithm/BP neural network分类
矿山工程引用本文复制引用
马莲净,王颂,赵宝峰,吕玉广,张阳,卢才武..基于PCA-SSA-BPNN模型的矿井突水水源识别方法[J].采矿与安全工程学报,2025,42(2):273-281,9.基金项目
国家自然科学基金项目(42307073,51974223) （42307073,51974223）
陕西省重点研发计划项目(2023-YBSF-345) （2023-YBSF-345）
国家重点研发计划项目(2016YFC0501104) （2016YFC0501104）
