Integrating Knowledge for the Identification of Cracks in Concrete Using an Expert System Shell and Extensions

摘要

An expert system has been developed for identifying the probable causes of cracks in concrete based on their shape and pattern, density and location. The system uses three forms of knowledge: facts and rules of thumb, databases, and digital imaging capabilities. Knowledge contained in the system is obtained from experts in the field, photographs taken of actual concrete failures and the classification of the failures into a database format. A PC based expert system shell and computer programs written specifically for the integration and use of knowledge were used to implement the expert system. The paper discusses the acquisition of the three forms of knowledge, its representation in the expert system environment and extensions developed to process information and enhance the user's interface. The potential for the use of speech recognition for recording conditions of field sites, and storage of large volumes of information using optical disk technology are discussed.