CAJ | 학술논문

通过金沙江某水电站的野外地质调查获取了丰富的基础地质资料后，针对其地质特点，提炼出岩石单轴抗压强度（R）、岩石质量指标（RQD）、岩体风化程度、节理组数（Jn ）、节理粗糙系数（Jr ）、节理蚀变系数（Ja ）和地下水状态7个对岩体质量起控制作用的因素作为神经网络模型的输入变量。基于MATLAB软件平台设计的人工神经网络，通过具有较强代表性的116组样本数据的训练得到了稳定性好、可信度高的岩体质量分级网络模型。在对36组测试数据分析后，发现该模型的仿真结果具有较高的准确度和良好的吻合度，能够满足实际工程需要。
통과금사강모수전참적야외지질조사획취료봉부적기출지질자료후，침대기지질특점，제련출암석단축항압강도（R）、암석질량지표（RQD）、암체풍화정도、절리조수（Jn ）、절리조조계수（Jr ）、절리식변계수（Ja ）화지하수상태7개대암체질량기공제작용적인소작위신경망락모형적수입변량。기우MATLAB연건평태설계적인공신경망락，통과구유교강대표성적116조양본수거적훈련득도료은정성호、가신도고적암체질량분급망락모형。재대36조측시수거분석후，발현해모형적방진결과구유교고적준학도화량호적문합도，능구만족실제공정수요。
After the field geological investigation of a power station on the Jinsha River,we had obtained basic geological data. According to its geological characteristics,we chose seven influential factors as the input variables of the Artificial Neural Network Model,such as one axis compressive strength of rock(R),rock quality designation (RQD),rock weathering degree,set of joints(Jn ), joint roughness coefficient (Jr ),fissure anomaly factor(Ja )and groundwater state. The artificial neural network based on the technology of MATLAB was trained,by 116 samples,to be an artificial neural network model which was high stability and reliability. According to the analysis of 36 test data,the artificial neural network model,of which the simulation results was very accurate,was proved to meet fully requirements of the engineering practice.