近幾年氣候變遷迅速、地震頻繁,使得多處地質更加不穩,一遇豪雨即成災害,再者台灣的河川水流湍急,易對橋墩造成沖刷,甚至危及橋梁之安全穩定性。本論文運用數值模擬之方法,將車輛荷載作為外力,建立不同車速於不同橋墩沖刷深度時之橋梁動力反應資料庫,並將橋梁動力反應資料及車速資料放入倒傳遞網路(back-propagation network)加以訓練,用以識別出橋墩之沖刷深度。在不同自變數的組合下,本文使用建設法作為系統自變數的決定方法,以提升識別結果可信度。依據系統識別(system identification In recent years, rapid climate change and frequent earthquakes have led to increasingly unstable geology at various locations, with disastrous consequences when the rains come. Moreover, fast-flowing rivers in Taiwan are the leading causes of scouring aro