基于模糊神经网络的高寒路基纵向裂缝危险度综合评价  

Comprehensive Risk Evaluation on the Longitudinal Cracks of Permafrost Subgrade Based on Fuzzy Neural Network

作　　者：叶敏[1] 王铁权[1] 

机构地区：[1]长安大学公路养护装备国家工程实验室,陕西西安710064

出　　处：《重庆交通大学学报（自然科学版）》2015年第4期68-72,106,共6页Journal of Chongqing Jiaotong University(Natural Science)

基　　金：交通运输部科技项目(201231879210)

摘　　要：以高寒地区自然环境因素、设计因素、冻土因素三类致灾因子指标作为输入变量,根据模糊推理规则构建路基纵向裂缝评价的非线性映射关系,通过对输入知识的预处理和输出知识的后处理,将模糊逻辑推理融入神经网络的非线性计算中,建立了综合评价模型。以青藏公路典型路段K3+020段为例,基于11类致灾因子评价该路段纵向裂缝的危险度为2级。结果表明:高寒路基纵向裂缝危险度模糊神经网络综合评价模型可用于评价纵向裂缝的发育程度,经济性好、实用性强。The natural environment factor,the design factor and the frost soil factor for the permafrost subgrade crack damage evaluation were selected as the input vector. According to the fuzzy rule,the nonlinear mapping relation of the longitudinal crack evaluation of subgrade was established. Through the combination of input pretreatment and output post treatment,the fuzzy rule was embedded into the nonlinear calculation of the neural network. It evaluated the longitudinal cracks risk degree in the section of K3 ＋ 020 of Qinghai-Tibet Highway and the disk degree was defined as second grade based on 11 kinds of damage factors. The results show that the model can quickly,accurately and objectively evaluate the subgrade longitudinal cracks risk degree in permafrost regions.

关 键 词：道路工程 路基工程 纵向裂缝 致灾因子 模糊神经网络 

分 类 号：TU433[建筑科学—岩土工程]