應用壓力容器參數化建模提升缺陷適用性評估之效率

於石化及化工業中，設備是最直接與化學品接觸的對象，更是造成災難性事故的主要原因之一。因此確保設備的完整性，使設備於生命週期各階段皆有足夠強度能穩定運轉，方能有效降低事故的發生。設備結構產生缺陷是破壞其強度的主因，因此，評估設備於存有缺陷時的殘餘強度是業者相當關切的項目。API 579是現今最常用於評估含有缺陷的設備的國際標準。該標準共分三種評估等級，其中，Level 1與Level 2僅能評估承受內壓的規則設備，對於特殊幾何元件或特殊邊界條件，則須採用Level 3評估，即有限元素法。於有限元素法中，建立設備模型是最耗費時間的程序，有效率的建立模型將是關鍵議題。參數化模型已在設備設計的過程中被廣泛使用，參數化模型係由各種元件特徵組合而成。該方法能同時減少人員建置模型時間及人為因素的繪製失誤。本文以一案例說明參數化建模程序，並結合API 579 Level 3執行有限元素法，分析缺陷位於設備噴嘴區域的承壓能力。本文案例的結果顯示，採用參數化建模確實能顯著降低繪製時間，且能成功結合有限元素法進行分析。

關鍵字

適用性評估；參數化建模；壓力容器設計；最大容許工作壓力

並列摘要

In the petrochemical and chemical industries, equipment is the most direct contact with chemicals, and it is one of the main causes of catastrophic accidents. Defects in the structure of the equipment are the main cause of the damage. Equipment defects are the main cause of damage to its strength. Therefore, it is an item of considerable concern to assess the residual strength of the equipment in the presence of defects. API 579 is the international standard most commonly used today to evaluate equipment with defects. Analysis methods in the standard are divided into three levels, in which Level 1 and Level 2 are confined to evaluate regular devices subject to internal pressure. For special geometry or boundary conditions, Level 3 (e.g. finite element method) is then required. However, model building in the finite element method and analyzing an FFS problem requires a lot of back-and-forth as well as FEM model adjustment. This study will show how we approach such matters and manage to speed up the analysis process. Parametric Modeling have been widely used in the design phase of product development, through which various component models are joined through positioning and boundary settings. This method can simultaneously reduce the time for one to build the model and the human factors to draw errors. Applications of the parametric modeling approach is illustrated via a case study. In combination with API 579 Level 3, fitness of a pressure vessel with a localized defect at the nozzle area can be assessed accurately. The results also indicate that the parametric modeling approach can significantly reduce the rendering time as well as the later on computing time in finite element analysis.