| 摘要: | 由於土木基礎建設在持續的使用下將損壞及劣化，基礎建設管理者的最基本工作乃是在有限的天然資源及維修管理經費下，確保基礎建設能提供正常的服務，而維修決策最佳化即是考量營運成本(包括使用者成本、損壞風險…等)及維修成本之間的平衡，期將基礎建設的生命週期成本降至最低，並在使用年限內提供更優良的服務，而維修決策的內容為維修工作的施作時間、種類、強度等。正確及精準預測設施在生命週期內變化之能力，乃是所有維修決策的基礎，本計畫旨在提供更先進之績效預測模式，以產生最佳化的維修決策。本研究擬使用狀態空間方法建立多變量動態模式以代表設施的劣化機制，針對土木工程之考量包括正確校估維修效果、可接受遺缺值、以及同時納入多個檢測儀器和大數量設施等，而校估所需的平方根濾波及相關程式已初步建置完成。為了更強化預測模式與維修決策之間的結合，本計畫擬進行完整深入的維修效果及非線性模式文獻回顧。其餘研究工作包括：資料庫蒐集與分析、模式建構及程式撰寫、模式測試與比較、實例分析、以及撰寫報告投稿期刊。本計畫之成果將可強化土木基礎建設管理之理論架構，提昇各項土木建設(道路、橋梁、房屋、電力設施、通信網路、水壩…等)之使用效率及年限，節省天然資源及經費，再加上環保意識的高漲，維護現有的建設將比新建更易達成社會共識，並達到土木建設之永續發展，具有極高的研究價值。 ; Civil infrastructure deteriorates over time due to the daily operations, and the key task for government agencies is to maintain the facilities and provide the necessary services that they are designed for under the constraints of natural resources and maintenance budgets. Maintenance decision-making is the process of finding the tradeoff between operation costs and maintenance expenses in order to minimize the life cycle cost of the facilities and provide better service to the general public. In detail, the decisions involved in the process are the timing, type, and magnitude of maintenance actions. It is obvious that infrastructure performance prediction is one of the most important inputs for maintenance decision-making. The purpose of this project is to provide an advanced performance model in order to generate better maintenance decision-making. This project plans to adopt state-space methods to model multivariate dynamic models and describe the deterioration mechanisms of facilities. The features that are attractive in civil engineering include the unbiased maintenance effectiveness, the consideration of missing values, and the possibility of multiple measurement instruments as well as a large number of facilities. The Square Root filter algorithm required for estimation and related procedures have been implemented for the project. To further strengthen the connection between performance prediction and maintenance decision-making, this project plans to conduct a complete literature review for maintenance effectiveness modeling and non-linear dynamic models. The remaining tasks also include collection and analysis of existing infrastructure condition databases, model formulation, program coding, model testing and comparison, numerical analysis, report write-up, and journal paper publication. The results of this project will enhance the overall framework of civil infrastructure management. As a result, the service life and efficiency of civil infrastructure systems such as roads, bridges, buildings, communication networks, electricity facilities, and water dams. It also will lead to savings in natural resources and government funding. Finally, to achieve sustainable development and due to the environmental regulations, it is more attractive to maintain existing constructions than build new ones. Therefore, the contribution of this project will be significant in both academic and practice. ; 研究期間 9710 ~ 9809 |
