創新人工智慧學習模式預測震後橋梁耐震能力與通行失敗機率之研究－以臺灣橋梁為例

橋梁為臺灣重要交通設施，為避免地震災害發生時，造成損壞導致交通中斷、居民受困或是人員傷亡，對現有橋梁進行全面檢測勢在必行。然而國內橋梁數量高達數萬座，若對所有橋梁進行破壞性檢測或結構分析，將花費許多時間與經費，實務上並不可行。有鑑於此，本研究結合材料劣化、側推分析、人工智慧與地表震動分析，在有限人力及經費下，計算各橋梁考量材料劣化因素下之耐震能力，並建置案例資料庫，藉由人工智慧學習輸入（簡易評估）與輸出（細評結果）之關係，推論求得地震發生時各橋梁之通行失敗機率，供公路主管單位後續處置之決策參考。

關鍵字

橋梁損壞診斷；材料劣化評估；橋梁側推分析；演化式支持向量機推論模式

並列摘要

Bridges are a vital and significant component of Taiwan's transportation infrastructure. Therefore, regular and comprehensive inspections of existing bridges are necessary to prevent damage and traffic disruption and reduce earthquake-related damage and casualties. However, due to the large number of bridges in Taiwan, the time and budget required to perform traditional structural analyses (preliminary assessment, detailed analysis) on every bridge to calculate yield acceleration (Ay) and collapse acceleration (Ac) values make doing so impractical. This paper integrates material degradation, pushover analysis, and artificial intelligence to create a new inference model as an alternative to traditional structural analysis. Historical cases are used to infer Ay and Ac values by mapping relationships between the preliminary assessment factors (input) of historical cases and detailed assessments of Ay and Ac values (output). Using the proposed inference model to predict Ay and Ac values, bridge maintenance planners can quickly and more cost effectively assess bridge earthquake damage probabilities as a guide to identifying priority bridge maintenance projects.

並列關鍵字

seismic assessment ； deterioration of materials ； seismic capacity ； evolutionary support vector machine inference model

國際替代計量

創新人工智慧學習模式預測震後橋梁耐震能力與通行失敗機率之研究－以臺灣橋梁為例

全文下載

主題瀏覽