基于BP神经网络的道路预防性养护决策研究  

Research on highway preventive maintenance decision based on BP neural network

作　　者：翟登攀 张熙颖[1] ZHAI Dengpan;ZHANG Xiying(Jilin Jianzhu University,Changchun 130118,China)

机构地区：[1]吉林建筑大学交通运输学院,吉林长春130118

出　　处：《交通科技与经济》2020年第5期63-66,共4页Technology & Economy in Areas of Communications

摘　　要：吉林的夏季炎热,冬季严寒的气候条件对于具有温敏性的沥青路面极为不利。此区域沥青路面病害类型丰富、发生很早且发育很快,因此,投入到沥青路面养护维修中的经费数额巨大。采取预防性养护措施防患于未然是解决上述问题的“钥匙”。传统的人工养护方法决策过程效率较低且受工作人员主观因素影响较大。在对季冻区沥青路面病害综合调查基础上,对其进行量化并建立“病害库”,同时,结合季冻区养护规范建立沥青路面“预防性养护措施库”,采用MATLAB工具箱中的BP神经网络实现对病害的分类分级,并以此将“病害库”与“养护措施”进行样本训练学习,建立对应各种类各程度道路病害的预防性养护模型。The extreme climatic of Jilin Province is disadvantageous for asphalt pavement.The types of asphalt pavement diseases in this area are rich,early occurring and rapidly developing.Therefore,the amount of money has been invested in the maintenance of asphalt pavement.To solve these problems,preventive maintenance is the"key".The preventive decision made in the traditional method was affected by the subjective factors and inefficiency.In this paper,diseases are quantified with their geometric and“disease library”is established in seasonal frozen areas.At the same time,a“preventive maintenance measures library”for asphalt pavements is established on account of the maintenance specifications of seasonal frozen areas.With the BP neural network of MATLAB toolbox,the classification of diseases is realized,the samples are trained to learn between the"disease database"and"maintenance measures",and eventually,preventive maintenance model corresponding to various types and degrees of highway are established.

关 键 词：预防性养护 MATLAB BP神经网络 病害分类分级 

分 类 号：U418.6[交通运输工程—道路与铁道工程]