Computational & Technology Resources
an online resource for computational,
engineering & technology publications
PROCEEDINGS OF THE EIGHTH INTERNATIONAL CONFERENCE ON CIVIL AND STRUCTURAL ENGINEERING COMPUTING
Edited by: B.H.V. Topping
Information Management in a Decision Support System for Pavement Management
A.P. Chassiakos, D.D. Theodorakopoulos and I.D. Manariotis
Department of Civil Engineering, University of Patras, Greece
A.P. Chassiakos, D.D. Theodorakopoulos, I.D. Manariotis, "Information Management in a Decision Support System for Pavement Management", in B.H.V. Topping, (Editor), "Proceedings of the Eighth International Conference on Civil and Structural Engineering Computing", Civil-Comp Press, Stirlingshire, UK, Paper 5, 2001. doi:10.4203/ccp.73.5
Keywords: highway maintenance, decision support system, information management, database, pavement condition, resource allocation, geographic information system.
This paper presents the main features and the information management process of a decision support system for highway pavement management. Pavement management systems have been proved beneficial and cost-effective to those agencies that use them. Specific benefits include money savings, better pavement condition, and enhanced conception of management objectives, procedures, and data requirements. The proposed system consists of a database, a user interface module, and analysis tools for pavement condition prediction, allocation of maintenance funds, and management of maintenance projects.
The database is developed on the SQL Server database system and includes a number of interrelated tables where data are stored. Input data refer to the road network, pavement distresses and condition indicators, maintenance activities, and other information which determines operational scenarios. The road network is divided into large segments according to road functional classification, pavement type, traffic volumes and loads, and environmental conditions. This segmentation provides the means for separate treatment of different segments in terms of desired condition levels, maintenance intensity, priorities in fund allocation, etc. Shorter sections are used for pavement condition assessment and determination of maintenance needs.
The main pavement distress types are cracking (mostly alligator type, longitudinal and transverse), potholes, corrugations and rutting, asphalt bleeding, aggregate ravelling and polish. Typical consequences of such distresses are pavement roughness and low skid resistance. The above defects are considered in the system by four indicators which represent pavement cracking, first crack appearance, roughness and skid resistance For each indicator, a number of levels have been specified to differentiate pavement condition with regard to the specific indicator. Overall pavement condition is then represented by a 36-element state vector which contains probability values of the pavement being in each of the alternative condition states resulting from all indicator combinations. A number of possible pavement treatments have been identified and incorporated in the system. They include routine maintenance, seal coat application, overlay course application, pavement milling and replacement, pavement recycling, and pavement reconstruction. For each maintenance action, information has been stored in the system concerning materials, methods, resources, and cost. Further, an analysis has been performed to reveal the effectiveness of each treatment to restore specific pavement defects.
The pavement performance module is used to assess the pavement condition in each period within the planning horizon. The proposed methodology for condition prediction follows a Markov analysis. In particular, the pavement condition vector at period results from the product of the same vector at period t and the transition matrix which contains probability values of pavement transferring from any condition state to another in one period. Maintenance activities have varying effect on pavement condition improvement and, therefore, a number of transition matrices are accordingly developed.
The resource allocation module determines the optimal maintenance strategy i.e., the appropriate maintenance action and time of application for each road section. Optimisation is performed through a linear programming model in which the objective function minimises the total maintenance cost over all sections and time periods. The linear program employs a set of constraints to model pavement condition deterioration (or improvement, if maintenance is applied) over time. These constraints originate from the transition probability matrices. Other constraints are set to maintain the network condition at a desirable level. Targeted condition levels may vary across network branches according to the importance of the branch, the traffic loads, and the environmental conditions which affect the pavement deterioration rate.
The project management module is used to provide decision support with regard to maintenance application procedures. In particular, individual maintenance needs of similar type and from adjacent sections are combined to form projects of desirable size. Further, time schedules are developed and resources are allocated to concurrent maintenance projects. In addition, a monitoring system has been established to collect and record relevant information and provide feedback to the system for all aspects of the pavement management process.
The user interface module provides the necessary link between the user and the system. Data manipulation is done through window and tabular forms while results can be viewed on a digital map supported by GIS application. System output which can be presented in such a way include geography (cities, type of land terrain), network characteristics (road identification, functional classification, length, number of lanes), pavement characteristics (type, layers, thickness), traffic loads (vehicle counts and composition), pavement condition (type and extent of distresses), proposed maintenance actions (type, resource requirements, cost) and desired condition levels.
purchase the full-text of this paper (price £20)