Intelligent Substitution Models in Structural Design

点击翻译按钮获取中文摘要

The consideration of structural design aspects in early phases of the building planning process leads to a reliable support of planning decisions as well as a significant increase in efficiency. In phase 1 of subproject 5 (TP5), the research approach comprised the elaboration of a knowledge-based system (KBS) that provides structural engineering knowledge depending on the development level. Using the implemented intelligent substitution models, applicable suggestions are determined as options and structural elements of the building design are evaluated. A concept for the identification of compromises and a substitution model for the estimation of modification effects have been developed for the processing of design changes.In the second phase of TP5, basic principles will be developed that utilize the KBS for the completion of the involved knowledge, for the encouragement of interdisciplinary cooperation, for the transparency of planning decisions and finally for the optimization of the design process. A universal applicability of the provided planning support necessitates the integration of knowledge that involves other building materials and construction methods. The integrability into the KBS is verified at the example of wood and wood composite constructions. Material-specific knowledge bases and associated TSK inference systems are implemented and an evaluation variant for TP4 is presented. An early consideration of complex cross-disciplinary design criteria results in a focused planning support that increases the acceptance and efficiency of the design process and promotes interdisciplinary cooperation. The integration of such complex requirements into the intelligent substitution models of the KBS is investigated through acquisition of expert rules for criteria and boundary conditions as well as the verification of their application in the inference systems. With regard to an efficient design process, the early identification and processing of highly complex structural problems is of particular importance. Since the computational efforts of parameter studies are generally very high due to their complexity, natural language rules are used for expert estimations and special engineering experience. For the utilization of these linguistic knowledge sources in the KBS, the adaptivity of the existing knowledge bases as well as the usability of inferences with linguistic variables according to the MA model are investigated. The comprehensibility of planning decisions is ensured through model developments for interdisciplinarity, transparency and adaptivity of the design assessment. For this purpose, an evaluation model for the quality of entire supporting structures, an explanation component for structural justifications and interfaces for the adaptation of the assessments are developed. The resulting knowledge-based system ultimately provides an optimized decision support for the building design process in early phases.