冻粘土三轴流变特性的神经网络预测

Neural network prediction of triaxial creep property of frozen clay

针对冻粘土特殊的物理力学性质,通过详细的冻粘土三轴应力条件下长期蠕变实验,讨论了温度、轴压、围压三种因素对冻粘土蠕变特性的影响,并利用人工神经网络(ANN)方法对冻粘土三轴流变特性进行了预测,建立了BP网络模型,预测结果与实验值吻合,因而具有一定的理论和实用价值。

Based on physical and mechanical property of frozen clay, the authors carried out long-term creep test in detail on frozen clay under triaxial stress, the effects are discussed on the creep property of frozen clay under the conditions of different temperature, axial stress and confining stress. A new method for testing creep property by using of artificial neural network is presented and BP network model is established in this paper. Test data and forecasting result is coincided very well, which indicates that the method has certainly theoretical and practical value.