應用無人機傳輸橋梁振動頻率於推估淘刷深度之研究

Abstract

台灣的河流大多為東西走向，但台灣的東西向僅142公里，加上中央山脈地勢高聳，導致溪流大多坡陡流急，造成河床沖刷劇烈，以及河道變化等水文現象，進而引發橋梁基礎沖刷的問題，橋梁檢測需要耗費大量的人力物力，而且為一年一度的規畫，每個年度地方政府都需要撥出大量經費來支持橋梁檢測，另外在橋梁檢測時，水面下的損傷例如基礎淘刷較不易進行目視檢測，近年來交通部運輸研究所致力於開發AI影像辨識結合無人機空拍的技術，但目前無人機僅針對水上的構件進行拍攝缺陷辨識，本研究將使用無人機去檢測橋梁基礎的淘刷深度，期望未來能夠使用無人機去代人力目視橋梁檢測。 由結構動力方程式推導可以發現，淘刷深度與自然頻率有關係，本研究將以自製的加速度量測器放置在橋梁端，並且透過無線傳輸的方式將在橋梁上量測到的資料傳輸到無人機上，無人機再將資料帶回研究室分析出自然頻率，最後與SAP2000有限元素模擬出的自然頻率與淘刷深度關係圖進行內插，得出現地的淘刷深度。除了分析以外，開發軟硬體亦為本研究的內容，目前市面上的自然頻率量測儀器雖然能力強大，但是價錢十分昂貴、儀器巨大而且需要人為操作，本研究將使用ESP32、nRF52832開發板加上加速度感測器來組成量測儀器，使用樹梅派放在無人機上來接收資料，並且達成量測自動化，使採集加速度資料時不須人為操作。 本研究在進行室內驗證時發現本研究的設備與市面上的儀器量測出的結果相近，推論本研究自行開發的設備具有一定的可信度，另外取得一壽橋完整竣工圖進行SAP2000建模，分析比對出一壽橋尚未出現基礎淘刷的現象，整體樁帽尚埋在土壤中，驗證本研究之可行性。

Taiwan's rugged terrain and fast-flowing rivers result in severe riverbed scouring, leading to bridge foundation erosion. Bridge inspections require significant human and material resources. Moreover, visually inspecting foundation scouring during bridge inspections is challenging. In recent years, the Ministry of Transportation's Transportation Research Institute has been dedicated to developing AI image recognition combined with unmanned aerial vehicle (UAV) aerial photography technology. However, current UAVs only focus on capturing defects on water-based components. This study aims to use UAVs to detect the scour depth of bridge foundations, with the expectation of replacing manual visual inspections with UAV inspections. In addition to analysis, software and hardware development are also essential components of this study. Although existing natural frequency measurement instruments on the market are powerful, they are expensive, bulky, and require manual operation. This study utilizes ESP32 and nRF52832 development boards along with acceleration sensors to create a measurement instrument. A Raspberry Pi is placed on the UAV to receive the data. During indoor verification, the study found that the developed equipment yielded measurement results similar to those obtained by instruments available on the market, indicating a certain level of credibility. Furthermore, the study obtained the complete as-built drawing of Yishou Bridge for SAP2000 modeling. Analysis and comparison revealed that the bridge's foundation has not experienced scouring, with the overall pile cap still buried in the soil, thereby validating the feasibility of this research.