基于BP神经网络的花岗岩单轴抗压强度预测  

Prediction of Uniaxial Compressive Strength of Granite Based on BP Neural Network

作　　者：杨奇超 王朋姣 Yang Qichao;Wang Pengjiao(Henan Geological Bureau of CCGMB,Zhengzhou,Henan,450000,China;Office of Zhongmou County People's Government,Zhengzhou,Henan,451450,China)

机构地区：[1]中化地质河南局集团有限公司,河南郑州450000 [2]中牟县人民政府办公室,河南郑州451450

出　　处：《化工矿产地质》2025年第1期78-82,共5页Geology of Chemical Minerals

摘　　要：为研究花岗岩的力学行为,通过物理力学试验和薄片鉴定,获取物理力学参数和矿物组成;以矿物组分、密度、纵波波速为基本指标,构建BP神经网络预测模型,研究了花岗岩的矿物组成、物理特征与其力学性质的关系。结果表明:采用长石含量、石英/云母、纵波波速和密度4个指标,借助BP神经网络,能够较准确预测花岗岩的单轴抗压强度;可见在不进行力学试验的情况下,通过矿物组成、物理参数初步估计花岗岩的单轴抗压强度。In order to study the mechanical behavior of granite,through physical and mechanical tests and section identification,physical and mechanical parameters and mineral composition are obtained.Using mineral composition,density,and longitudinal wave velocity as basic indicators,BP neural network prediction model is constructed to study the relationship between mineral composition,physical characteristics,and mechanical properties of granite.The results show that using the four indicators ofeldspar content,quartz/mica,longitudinal wave velocity,and density,the uniaxial compressive strength of granite can be accurately predicted with the help of the BP nerral network.It is concluded that without conducting mechanical tests,the uniaxial compressive strength of granite can be preliminarily estimated through mineral composition and physical parameters.

关 键 词：花岗岩 单轴抗压强度 BP 神经网络 预测 

分 类 号：P642.3[天文地球—工程地质学]