2001-08-012024-05-18https://scholars.lib.ntu.edu.tw/handle/123456789/706557摘要：九二一大地震之後，國人開始關心建築結構物之耐震安全，但是因為以往並無一耐震診斷及維修補強之標準及方式，因此目前並無一較有效之方式解決國人心中之疑慮。而近年來較熱門之案例式推理透過類比式之學習，可有效地保留過去之經驗以做為新問題之解答參考依據。本研究即希望利用類比式學習之優點，將建築結構物耐震設計與評估之相關參數，搭配其受震後之結果，作有效之經驗累積，一方面用以建構有效之耐震評估診斷專家系統，另一方面亦可提供日後之建築耐震設計參考，使之增進更多安全的保障。 案例式推理雖有傳統知識庫法則式推理所沒有之經驗累積的優點，但在工程上當案例不足的情況下，知識庫則能提供案例中所無法得知之知識。本研究希望建立整合案例式推理及知識庫之耐震診斷專家系統，透過二者之溝通、配合，為建築結構物之耐震性提供一快速且有效之診斷系統。近年來，由於WWW之快速發展，web-based之系統可提供方便且快速的使用介面，因此本研究將採用java技術開發一Web-based專家系統，以利將來發展成熟時，供各界之專家們使用並便利後續研究工作之推動。並期能在本研究中建立好系統架構之後，在日後能透過更細部之參數分析及經驗學習，對<br> Abstract: Due to the 921 Chi-Chi earthquake disaster, seismic safety assessment of buildings has become one of the hottest issues recently. However, there are no standard assessment and retrofit approaches at this moment to effectively address people’s concerns about the safety of their buildings. Therefore, there is a great need to conduct more researches on seismic safety assessment of buildings and other related topics. The major objective of this research is to integrate advanced computer and information technologies, especially expert systems and web-based technology, to help assessing the seismic safety of buildings in a more effective, efficient, and consistent way. In the field of Artificial Intelligence (AI) for expert systems, Case-Based Reasoning (CBR) technique has played an increasingly more important role with its analogical learning model to help accumulating useful past experiences for solving new problems. However, the effectiveness of the CBR technique is limited by the cases collected in its database. On the other hand, the Knowledge-Based Reasoning (KBR) technique can process knowledge rules (such as rules in seismic design codes) to help the CBR technique in building safety assessment when there are not sufficient appropriate past examples in the case database. In addition, the Internet and WWW technologies have greatly facilitated sharing of information and software applications among people or organizations that may be far apart in physical distance. Naturally, web-based computer applications have become the mainstream in the software development because they are easier to be maintained and shared. This work proposes the development of a web-based expert system that integrates both the CBR and KBR techniques to help assessing the seismic safety of buildings. A number of practical building examples will be used to investigate the feasibility and effectiveness of the proposed approach and the expert system developed. It is hoped that the completion of this research can greatly help engineers in answering to people’s concerns about the safety of their buildings.耐震診斷案例式推理耐震規範人工智慧類神經網路知識庫專家系統seismic assessmentcase base reasoningseismic codesartificial intelligenceneural networkexpert systems