請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/91343
標題: | 建築資訊模型執行應用於工程設計績效評估 Building Information Modeling Application in Engineering Performance Evaluation |
作者: | 邱文彬 Wen-Bin Chiu |
指導教授: | 張陸滿 Luh-Maan Chang |
關鍵字: | 建築資訊模型,人工智慧,工程統包執行,工程績效,機器學習, building information modeling (BIM),artificial neural networks (ANNs),Engineering, procurement, and construction (EPC),machine learning multilayer perceptron (MLMP),engineering performance, |
出版年 : | 2023 |
學位: | 博士 |
摘要: | 績效管理為營建工程專案重要的經營措施,目的為評核專案執行過程中的各項績效指標,例如成本、進度、品質、安全與顧客滿意度等,主要為提供經營管理者必要的決策參考、早期警示與預防措施、以及加強持續改善的機會。然而,工程績效管理所涉及的層面非常廣泛,除了對績效一詞的看法與定義不盡相同外,其決策者對於專案執行各階段中績效監督的程度,也可能依其經驗與看法不同,而採取不同的決策與措施。
工程績效評估與控制對於專案的執行具有重要的意義,正確而有效的績效評估方式是營建工程專案成功與否的關鍵。根據營建工程業經驗,在專案規劃與執行初期,工程設計的過程就影響了營建工程的生命週期,且直接影響專案執行的成敗。而營建工程中廣泛被應用的成本影響曲線,說明了工程設計執行為最能夠直接影響總造價的階段,在專案執行過程中需要仔細評估與衡量,所以工程規劃與設計對於專案的成功執行有關鍵性影響。儘管在某些營建工程專案中,工程設計成本已接近專案總造價的20%,但至今對於工程設計績效的了解和研究仍不及施工績效普遍與深入。此外工程設計績效是工程專案成本和進度的關鍵決定因素。由於這些原因,評估工程設計以預測專案績效指標與推動持續改善至關重要。 有效而實用的工程營建績效指標必須建立,並將其應用於現行的營建工作流程,然後進一步建立工程執行績效的可預測性。多年來,許多相關研究已經提出了幾種不同的工程績效測量與預測方法,並已經建立專案執行參數變量和績效測量之間的因果關係。然而現在分析測量工程績效的研究和方式,大都主要集中在工時與時程的績效上,並不是一個足夠廣泛的衡量標準來評估工程績效的有效與適用性。 近年來,建築資訊模型(Building Information Modeling, BIM)已成為建築與營建工程領域中快速發展的創新技術,其應用主要是透過多維數位化模型建構與管理,對工程生命週期中的各階段作業進行各種應用分析,除了對工程的實際執行有較佳的掌握外,更能有效整合工程執行過程中的各項設計、採購與施工作業資訊,降低工程成本與錯誤,提升工程品質、效率與安全。現今許多國際工程專案已普遍使用BIM 技術執行,國內一些重大工程也逐步導入 BIM 技術,並積極開發可以整合時程、成本、風險與績效資訊,以期達成對工程執行的有效掌控。然而在國內外各型工程專案一致朝向應用與發展BIM技術同時,現今學術界針對BIM的執行對於專案績效評估相關研究尚未發展成熟,相關研究也僅限於BIM執行本身的績效,未針對其專案工程設計階段績效進行評估與探討。 BIM的應用已改變營建工程統包執行方式。基於了解應用BIM於專案執行所使用的輸入參數變量與專案設計績效成果之間的關係,提高預測工程設計績效的重要性,進一步了解其相對關聯性。本研究建立一套系統分析模型,透過調查收集實際來自60個應用BIM工程執行樣本的專案數據,將BIM使用輸入變量與工程設計績效輸出進行關聯性分析,並依相關係數來檢核BIM輸入因子之間及BIM輸入因子與輸出績效的相互關係,之後進一步採用統計變量遞減技術建立工程績效預測的多元線性回歸模型,並運用人工智慧機器學習技術,建立評估模型,以期達更好的測量和預測營建專案BIM的應用效益。經過嚴格的驗證流程,並採用統計方法評估模型的差異性,實現並達成了最佳的專案工程績效預測,結果證明BIM應用與工程設計績效指標之間存在顯著相關性,可進一步利用建立的多元線性回歸與人工智慧機器學習模型來預測工程設計績效指標,並有效且正確的應用於工程專案執行。 Performance management is an essential task for construction projects. The primary purpose is to evaluate various indicators which impact performance in project execution, such as cost, schedule, quality, safety, and customer satisfaction. These performance indicators provide management stakeholders with necessary decision-making references, early risk warnings, preventive measures, and continuous improvement opportunities. The effective performance evaluation methodology is the key to the success of a construction project. However, engineering performance management involves a wide range of measurement and evaluation details. In addition to the different views and definitions of performance, the decision-makers may also adopt different decisions and measures based on their experience in the level of supervision of each stage at project execution. The engineering design process has fundamentally impacted the life cycle of construction projects, and notably, engineering performance constitutes a critical factor for a project and shall be measured efficiently. The control, measurement, evaluation, and prediction of engineering performance are significant in delivering construction projects, and reliable engineering performance measurement is critical to project performance and continuous improvement. In the project execution life, the engineering design at the early stage is critical for successful execution and can significantly affect the final total cost as illustrated in cost impact curves. Even though engineering costs have increased to reaching around 20% of total installation cost on several construction projects, engineering performance is less well realized and has received less focus compares to construction performance. The implementation of the early-stage engineering design is an essential key for successful execution and the engineering performance evaluation and prediction have a substantial influence on the execution phases and effectiveness of the project. For above reasons, reliable and precision metrics for evaluating real-time performance to drive improvement are significant. Applicable industry engineering performance must be recognized and applied to current engineering work processes before essential improvement and predictability of performance can be developed. Over the past years, several approaches for engineering performance measurement and evaluation methods have been proposed, and the studies have demonstrated the cause-effect relationships between project variables and performance measures. The historical research for engineering performance measurement was analyzed primarily focused on job-hour performance, represented an incomplete picture, and is not broad enough to assess the effectiveness of engineering performance. Recently, building information modeling (BIM) application has been a rapidly developing innovative technology in architecture and construction engineering. In addition to having better control of the actual implementation of the project, BIM can integrate various design, procurement, and construction operations in the project life cycle, reduce project costs and errors, and improve project quality, efficiency, and safety. Many international engineering projects have deployed BIM technology, and some major domestic projects have also gradually introduced BIM technology and actively developed information that can integrate schedule, cost, risk, and performance to control project execution effectively. The application has reformed how owners execute the industry's engineering, construction, commissioning, and operation. While large-scale projects at home and abroad are consistently oriented towards the application of BIM technology, the current academic research on BIM implementation and project performance evaluation has not yet matured, and the relevant research is limited to the performance of BIM implementation. It has not evaluated its project engineering design stage. The application of BIM has changed how design-build or turnkey project are performed. Based on understanding the relationship between the application of BIM use elements and project results, the importance of predicting engineering design performance, and understanding of its relative relevance, improving project engineering performance based on the knowledge of the relationships between BIM use application and performance outcomes becomes essential. This research proposes a system analysis model to correlate BIM use input factors with engineering design performance output analysis by leveraging data from 60 samples. The statistical variable reduction techniques are deployed to develop multiple linear regression models linear regression (LR) analysis and applying artificial intelligence neural network (ANNs) machine learning multilayer perceptron (MLMP) technology of the engineering performance to establish evaluation models to measure and predict the application benefits of BIM in construction projects. The development of the prediction models is based on practical execution data from projects collected through a comprehensive BIM application survey and the best prediction was generated, validated, and implemented. After rigorous verification, the best prediction is obtained and the results prove a significant correlation between BIM application and engineering design performance outcome measures, which can be applied to predict engineering design performance measures using the established models. The study establishes a comprehensive methodology for the proposing models, and the accuracy and reliability of the models are tested validated. Moreover, engineering performance measures can be predicted by BIM uses. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/91343 |
DOI: | 10.6342/NTU202304540 |
全文授權: | 同意授權(全球公開) |
顯示於系所單位: | 土木工程學系 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-112-1.pdf | 3.76 MB | Adobe PDF | 檢視/開啟 |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。