A SYSTEMATIC KNOWLEDGE-BASED APPROACH TO THE STRUCTURAL ANALYSIS OF PAVEMENTS

Project level structural maintenance of highways relies increasingly on the back analysis of computer models of pavements from which surface deflections and standard loads, such as that of the Falling Weight Deflectometer, are available. The procedures tend to be mechanistic with the best results obtained when used by experienced maintenance engineers. In order to reduce direct reliance on such skills, a prototype knowledge-based system has been developed to "inject" engineering knowledge into the conventional techniques. The system consists of three parts: 1) a finite element computer program, 2) a knowledge base, and 3) a database. The analytical program carries out the analysis of the pavements tested in the field. The knowledge base encodes engineering knowledge, and, using artificial intelligence techniques, it drives the back analysis procedure to select a modulus set for the pavement layers that satisfies two criteria: a) convergence of the model pavement deflection to that of the field and b) compliance with engineering judgment for the known materials and apparent condition of the pavement. The database holds the characteristics of various pavement materials which are used as input data to the analytical model by the user working with the knowledge base of the system.