论文部分内容阅读
土的本构关系模型的研究是土力学研究中的一个关键问题。但由于土的性质受多方面因素的影响 ,如何找到一个既能准确描述土的本构关系又具有工程实用价值的模型一直是学者们探讨的课题。基于此 ,本文利用人工神经网络的方法建立了一个土的非线性本构关系模型 ,该模型考虑了剪应力对体应变和静水压力对剪应变的相互耦合作用 ,并成功的避免了传统本构模型的经验假设和简化 ,同时充分利用了实验的全部数据 ,提高了模型的精确度。模型亦具有很大的容错性。
Soil constitutive model research is a key issue in soil mechanics research. However, as the nature of soil is affected by many factors, how to find a model that can accurately describe the constitutive relationship of soil and have practical value of engineering has always been a subject discussed by scholars. Based on this, a nonlinear soil constitutive model of soil is established by using artificial neural network method. This model considers the mutual coupling between shear stress and body strain and hydrostatic pressure on shear strain, and successfully avoids the traditional constitutive Assumptions and simplifications of the model, while taking full advantage of the experimental data, improve the accuracy of the model. The model also has a great deal of fault tolerance.