【摘要】 本文提出一种根据岩石点荷载强度、密度、岩石类型、孔隙率和粒度预测单轴抗压强度(Rc)的神经网络方法。以反向传播算法和包含96个样本的训练集训练网络模型,以包含71个样本的试验集检验网络。最后,以本模型预测黑沟铁矿钙质千枚岩的30个试件的抗压强度。结果表明,对于预测的单轴抗压强度,神经网络模型能给出比回归模型高得多的精度,比试验方法节省成本和时间。更多还原

【Abstract】 Compressive strength is the most widely used design parameter in the construction industry and in rock engineering. This paper presents the application of a neural network to the prediction of the uniaxial compressive strength based on the point load, density, rock type, and porosity and grain size tests. A training set containing 96 rock sample records was used to train the network with back propagation algorithm. A testing set including 71 samples to test the network. A predicting set involving 30 samples was used to predict the uniaxial compressive strength values of calcareous phyllite at Heigou mine. The results show that the neural network model has higher accuracy than the regression models, and lower cost and time than the experimental test methods.更多还原