一個將工業基礎分類檔案轉變為建築本體論的轉換系統

Abstract

建築產業在各個國家都是重要的傳統產業，但卻缺乏與資訊的結合。建築資訊建模(Building Information Modeling, BIM)的出現成為建築、工程、營建產業資訊化的解決方案。在BIM中，Industry Foundation Classes (IFC)是一種開放性未結構化的資料格式，它可以支援BIM資訊的交換與共享、進行資料的處理，避免資料傳遞時發生錯漏造成時間以及財務上的損失。但實際運作上，目前所有支援IFC的軟體間都有失真的問題，從而造成時間及金錢的浪費。建置建築本體論是一個可行的解決辦法。為了將IFC轉換成建築本體論，本研究利用正規概念分析法(Formal Concept Analysis, FCA)設計出一個IFCFCA演算法來建構建築本體論。此演算法將IFC檔案資料結構化，再利用叢集技術與相關工具來建立建築本體論。本研究實作了一個轉換系統，實驗結果顯示本研究可以完整讀出建築圖資料，解決IFC失真的問題，並且能建立準確率高達95%的建築本體論。

Building construction industry in many countries is an important traditional industry. However, this industry lacks of data integration which leads to higher cost and lower productivity. The building information modeling (BIM) was proposed to solve the problem. BIM changes information exchange between architects, engineers and constructors from a traditional paper-based approach to an electronic one. In BIM, the industry foundation classes (IFC) is designed to provide information sharing. However, IFC supports only unstructured data formats. To date, all supported IFC softwares have the problem of distortion, which causes a waste of time and money. The construction of building ontology is a viable solution to this problem. By using the formal concept analysis (FCA), this thesis proposes an algorithm, called IFCFCA to restructure IFC files and subsequently produce building ontology. We have developd a conversion system based on IFCFCA. The experimental results indicate this system can solve the distortion problem and produce the ontology with an up to 95% accuracy.