Knowledge-based system approach to building envelope design

Abstract

Building envelope design is a decision-making process requiring expertise for the selection of materials and constructional systems. The number of alternatives and performance attributes to be considered in design is so large that the designer cannot efficiently evaluate them without computerized assistance in information handling and decision making. Developments in artificial intelligence and knowledge-representation methodologies have shown that computer programs can be constructed for solving ill-structured design problems. Following this approach, a building envelope design system has been developed to assist the designer in selecting materials and constructional systems. A knowledge base containing information on performance requirements and constraints from building codes is interfaced with a database of material properties to solve the information handling problem. Feasible design alternatives are thus synthesized following a ‘plan—generate—test’ strategy with multiple levels of constraint checking. The designer has the option of modifying or updating the knowledge base, adding more information to the material properties databse or defining preferences or performance attributes during the evaluation process. The designer-specified preferences can be used in ranking the alternatives to identify an overall best design alternative. This paper presents the methodology and implementation details of a knowledge-based design assistant for the selection of materials and constructional systems for the building envelope. A review of the building envelope design process, design alternative generation and evaluation methodologies is also presented.