Study On The Framework Of Pavement Management System Based On GIS

Recently, with the rapid development of both high-grade highway construction and information technology, pavement management system (PMS) has become one of the most useful tools for planning and investment for road maintenance. Research and application of PMS, including facility management, have been highlighted to apply the functions of PMS on the platform of GIS. Data management is primary element of PMS, which includes classification, organization, encoding, storage, retrieve and maintenance for data. Technology of GIS, with the powerful graphics processing capacity and analysis function for spatial data, makes itself be vital component, of pavement management. Referring to other interrelated literatures, the compositions of PMS and the organization modes of its internal data have been researched, and based on GIS, this thesis has been specified in PMS, and spatial analysis of the system and research and development for functions are emphasized. The main contents are as including:(1) Some widely used methods for pavement performance evaluation have been presented. The principle applying the element-analysis method which is based on extension mathematics theory in the comprehensive evaluation of pavement performance has also been detailed. Advantages and characteristics of the method mentioned above have been proposed by comparing it with that based on fuzzy theory. The result evaluated by element-analysis method is more objective and effective than by fuzzy method which reflects homogeneous result.(2) Based on BP neural network method provided by MATLAB, three influence factors of pavement performance (including temperature, rainfall and axle load) have been introduced, then, through training and simulation for available data, single-index neural network models for each factor are constructed. After optimization of training precision by adjusting weight values and threshold values, single-index simulation forecasts have been achieved. As the "input-output" mode, the "influencing factors-pavement performances" is sselected to train the model comprehensively. Taking the above mentioned factors as input values, another four performance indexes as the output values, a matural integrated network is obtained through training for continuous data within a certain years. With analyses of examples, the method, which has been dissertated in this thesis, is proved to be promising.(3) A new class of object-oriented spatial data is applied for the data organization and establishment of the database. Then, the system databases, which include spatial database, attribute database and some other information databases, are designed and constructed with a unified modeling language. Functional interface of the system is developed with the VBA programming method in ArcMap, and the design and implementation of both model algorithm and the dialogue windows are also completed. Through the development test and attempt for the system, functions of the system such as evaluation and prediction on pavement performances, SQL query, buffer analysis, watching of image data have been accomplished. Then, it can be achieved that the development of PMS on GIS platform could enable the management more visual, reduce users' workloads and time greatly, and improve work efficiency. In conclution, the method is an effective approach for the modern management model.