Building envelope problems are caused by many factors such as improper selection of materials, inadequate design and poor workmanship. Cracking, spalling, staining, fungal growth and several other symptoms may manifest in exterior walls, roof, basement walls and floors. Diagnosing building envelope problems is a complex task requiring the building science knowledge and expertise in interpreting the symptoms, causes and failure mechanisms. Diagnostic problem solving is one of the earliest and most popular applications of knowledge-based expert systems. Many rule-based systems have been developed by representing the heuristics and logical relationships between symptoms and causes. This approach is very simplistic and is suited only to a limited category of diagnostic problems. The nature of building envelope failure diagnosis is much broader in scope and requires many different types of information, in addition to heuristics. A data base of material properties, their functional descriptions, performance requirements, climatic data and analysis procedures are needed to establish the context of building envelope problems. Each building envelope problem and its associated information such as frequency of occurrence, location, materials affected must also be incorporated in the knowledge base. Rule-based representation is inadequate to handle all of the above types of information and hence alternative representation schemes must be considered. The present investigation evaluates the following three techniques namely: rule-based, frame-based and hybrid representations. Three prototype systems are implemented with the aim to identify the best and most efficient representation scheme for long term development of a practically useful diagnostic tool.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £60 +P&P)