建置及應用資料庫以發展台灣混凝土收縮預測公式

Abstract

混凝土結構物設計、興建與維護管理需要考量其強度、變形及耐久性，以達到長期服務之安全及服務性。台灣與國際對於混凝土變形之研究發展已相當長久，產出許多相關之試驗資料，為利國際比較分析與整合，美國西北大學建立NU資料庫、日本建立JSCE資料庫以彙整之。台灣大學於2001年蒐集成立資料庫後，為跟循世界近期科技之發展，本研究待再更新混凝土相關資料及進行探討。本研究從近年試驗資料彙整開始，同時參考國際上資料庫的參數組成、資料架構，並與台灣既有混凝土資料庫結合，建立「台灣混凝土潛變收縮資料庫」。 從台灣混凝土資料庫與國際資料庫比較差異，發現台灣混凝土有著漿體量高、粒料較脆弱及彈性模數較低，導致收縮變形增加之特性。因此，經由預測公式分析比較，本研究選用Bažant之Model B4為預測公式之基礎，經合理之本土化修正，提出台灣普通混凝土預測式。近年來添加高爐石粉、飛灰的混凝土研究增加，本研究以普通混凝土收縮預測式延伸，參考文獻研究及其物理意義，針對添加高爐石粉、飛灰、或者同時添加之混凝土進行參數修正，提出台灣混凝土收縮預測式。 本研究使用MySQL做為資料庫管理系統與並以Python建立分析方法，而非傳統之Excel或者Access，其原因在於使用Python程式語言可以透過SQL語法可攜性進行資料篩選、資料分析，效率更勝一般傳統資料庫分析方法。另外，Python程式語言具備網頁設計的專業套件，使用此資料庫分析程序，也奠下日後欲將台灣混凝土資料庫併入世界混凝土資料庫，與發展至雲端化之基礎。

In order to achieve concrete structure’s safety and servibility, it is important to consider its deformation, strength, and durability in design, construction and maintenance process. Tests and researches on the creep and shrinkage of concrete have been developed over several decades all around the world. For better analysis and integration, Northwestern University (NU) of the USA has established a concrete deformation database which includes data from around the world while Japan Society of Civil Engineers (JSCE) has also established their own database which incorporates data from Japan. National Taiwan University (NTU) began collecting data from Taiwan and established the first-version of database in 2001. This study continues the early work at NTU and follows the framework and formula of the NU database in establishing Taiwan Creep and Shrinkage Database, the purpose of which is in collecting and integrating up-to-date on local shrinkage and creep data. After a comparison between international and Taiwan-only data, some distinct local characteristics of concrete in Taiwan have been discovered. The local concrete produced is generally made with a high amout of cement, low water-cement ratio, as well as low elastic modulus of sandstone aggregate, which may result in higher shrinkage. This study takes the Model B4 shrinkage prediction model as a base for the development of a local prediction model for Taiwan. Modifications to the model, where necessary, mainly concern the underestimation of local aggregate elastic modulus, large deformation due to high paste amount, and concrete with mineral admixture such as slag and fly ash. Rather than traditional methods, this study uses new tools for database establishment and analysis, and combines them into a new procedure. The new procedure consists of MySQL, a Database Management System, for database establishment, and Python, a programming language, for programing of analysis models and post-processing tasks. The result shows that this new database processing procedure is faster and more efficient than traditional procedures. Because SQL (the database programming language of MySQL) is highly portable, which facilitates integration between MySQL and Python and professional and various packages in Python allow one to proceed with many steps of analysis and post-processing automatically. It is foreseen that this study will help incorporate the Taiwan database into world concrete deformation database and will help establish an effective platform based on MySQL and Python. This, in turn, is expected to promote the development of a new era of cloud calculation and user-friendly web systems for deformation prediction.