基于GA-BP神经网络岩石单轴抗压强度预测模型研究  

Research on Prediction Method of Rock Uniaxial Compressive Strength based on GA-BP Neural Network

作　　者：张奥宇 杨科[1,2] 池小楼 张杰[1,2] ZHANG Aoyu;YANG Ke;CHI Xiaolou;ZHANG Jie(Joint National-Local Engineering Research Centre for Safe and Precise Coal Mining,Anhui University of Science and Technology,Huainan 232001,China;State Key Laboratory of Mining Response and Disaster Prevention and Control in Deep Coal Mines,Anhui University of Science and Technology,Huainan 232001,China)

机构地区：[1]安徽理工大学煤炭安全精准开采国家地方联合工程研究中心,安徽淮南232001 [2]安徽理工大学深部煤矿采动响应与灾害防控国家重点实验室,安徽淮南232001

出　　处：《煤》2025年第1期6-10,17,共6页Coal

基　　金：国家重点研发计划资助项目(2023YFC2907502);安徽省煤炭安全精准开采工程实验室开放基金项目(ESCMP202301)。

摘　　要：为探究更为精确的上覆岩层砂岩和泥岩单轴抗压强度与其弹性模量之间的关联性,结合胡家河矿56组砂岩和泥岩单轴抗压强度与弹性模量历史数据,运用遗传算法优化了BP神经网络的结构参数和学习参数,得到了最佳的网络结构和参数设置,利用GA-BP神经网络对煤矿砂岩与泥岩单轴抗压强度进行了预测,并与传统的BP神经网络和非线性回归分析法进行了比较。研究结果表明,GA-BP神经网络预测模型在预测砂岩和泥岩单轴抗压强度与弹性模量间关系上具有较高的精度和泛化能力,能够有效地解决传统BP神经网络的局部最优和过拟合问题,相较于非线性回归分析,拥有更强的非线性关系建模能力,是一种适用于砂岩与泥岩单轴抗压强度预测的有效方法。To investigate the correlation between the uniaxial compressive strength and elastic modulus of overlying strata sandstone and mudstone with greater accuracy,historical data consisting of 56 sets of uniaxial compressive strength and elastic modulus of sandstone and mudstone from Hujiahe Coal Mine were analyzed.Genetic algorithms were utilized to optimize the structure and learning parameters of a BP neural network,resulting in the identification of the optimal network structure and parameter settings.The GA-BP neural network was then applied to predict the uniaxial compressive strength of coal mine sandstone and mudstone.Comparisons were made with traditional BP neural networks and nonlinear regression analysis methods.The research findings indicate that the GA-BP neural network prediction model achieves higher accuracy and generalization capability in predicting the relationship between the uniaxial compressive strength and elastic modulus of sandstone and mudstone.It effectively addresses the local optimum and overfitting issues associated with traditional BP neural networks and exhibits superior nonlinear relationship modeling capabilities compared to nonlinear regression analysis.Therefore,it is considered an effective method for predicting the uniaxial compressive strength of sandstone and mudstone.

关 键 词：岩石力学参数 非线性回归 BP神经网络 遗传算法 预测模型 

分 类 号：TD313[矿业工程—矿井建设]