论文部分内容阅读
黄土的湿陷起始压力是评价黄土湿陷性的重要指标之一。影响黄土湿陷起始压力的因素有很多,且各因素间并非独立,通过分析各物性指标间的相关性确定塑性指数、含水率、干密度作为影响黄土湿陷起始压力的因素。本文提出并建立了黄土湿陷起始压力的人工神经网络预测模型,选取新疆伊犁地区黄土的数据作为神经网络模型的学习和预测样本,将神经网络模型的预测结果与实际结果对比可知二者误差小于10%。利用陕西彬县黄土数据验证了网络模型的通用性,说明用人工神经网络方法计算黄土湿陷起始压力准确、可靠,建立了一种计算湿陷起始压力的新方法。
The initial pressure of loess collapsibility is one of the important indexes to evaluate the loess collapsibility. There are many factors that affect the initial pressure of loess collapsing, and the factors are not independent. The plastic index, water cut, and dry density are determined as the factors that influence the initial pressure of loess collapsible collapse by analyzing the correlation between various physical indexes. This paper proposes and establishes an artificial neural network prediction model of the initial pressure of loess collapsibility, selects the loess data from the Yili region of Xinjiang as a learning and prediction sample of the neural network model, and compares the prediction results of the neural network model with the actual results Less than 10%. This paper verifies the universality of the network model by using loess data from Binxian, Shaanxi Province. It shows that the initial pressure of loess collapses calculated by artificial neural network is accurate and reliable, and a new method of calculating the initial pressure of collapse is established.