摘要
非饱和膨胀土的本构关系非常复杂,涉及的因素很多,且多为因素间的耦合作用.本文探讨了用进化神经网络建立GA-NN本构模型,用8个物理力学指标拟和主应力差与轴向应变的关系,取得了较好的效果.
The constitution of unsaturated expansive soil is extremely complex and involves quite a few factors,and,in many cases,the factors are coupled with each other.This paper puts forward a discussion on the building of a GA-NN constitutive model by use of evolution neural network.Eight physicomechanical indicators are applied to the fitting of relations between the principal stress difference and the axial strain.Fairly good results are obtained.
出处
《内蒙古工业大学学报（自然科学版）》
2010年第1期68-72,共5页
Journal of Inner Mongolia University of Technology：Natural Science Edition
