【作者】 侯静；

【导师】 魏连雨；

【作者基本信息】 河北工业大学 ， 结构工程， 2004， 硕士

【摘要】 近几年，我国公路建设飞速发展，传统的管理手段已不能满足迅速增加的公路养护管理工作的需要。为提高公路管理工作的效率，本文将人工神经网络技术应用于路面性能预测中，进行了基于WebGIS的路面管理系统的结构和功能设计，研究内容分为以下两个方面： 首先，以天津市外环线为研究对象，通过对系统的目标、用户的业务需求分析，进行空间数据库和属性数据库的设计，选择适当的WebGIS软件，在此基础之上通过路段编码进行空间数据和属性数据的动态链接，建立路面管理系统专业模块，在实现对外环线路面信息管理的同时，利用这些信息对路面性能做出评价和预测。 其次，在路面性能预测模块中，为解决以往模型中难以考虑环境和荷载等因素影响的问题，本文应用了人工神经网络，利用其在处理复杂的非线性问题时所具有明显的优点建立路面性能预测模型，应用效果较好。人工神经网络不需要研究对象具有精确的数学模型，并能够通过神经网络结构的可变性，逐步适应外边环境的多因素作用。因此该模块的预测结果将随着数据资料的不断增加而愈加接近实际值，并且具有良好的移植性。更多还原

【Abstract】 Traditional management, with the rapid development of highway construction, has not satisfied the increased demand of highway maintenance. For increasing the efficiency of highway management, the author applies artificial neural networks technology to pavement performance prediction and designs the structure and function of pavement management system based on WebGIS. The study divides into the following two parts.First, the author makes the Waihuan Road in Tianjin as the study object. Through analyzing of system target and user’s demands, the author designs the spatial database and attribute database on the base of WebGIS software selected. The author connects the two kinds of data by the road section number and constructs the professional module not only to realize the management of Waihuan Road but also to evaluate and predict the pavement performance.Second, in order to consider the influence of environment and load on pavement performance, the author applies artificial neural networks (ANN) in predicting module and the effect is better for ANN’s remarkable ability of dealing with complex nonlinearity problem. Appling ANN need not to know the exact form of mathematical module and it can gradually adapt to the affect of multi-factors in external environment through it’s variability of neural networks’ structure. So the model will have more and more exact results with the increasing of data and can be easily transplanted to other system.更多还原