應用視覺感測器監測船體變形以及裂縫破壞

APPLIED VISION SENSOR INSPECT SHIP DEFORMANTION AND CRACK DAMGE

船舶在不同的海況條件下，會產生不同的結構變形，包括舯拱及舯垂，且波浪為長期負荷，對於船體結構可能會造成疲勞裂縫。本研究針對船體結構進行非接觸式影像監測，找出航行期間，船體結構可能會發生的結構失效模式及破壞位置。現今傳統船體監測方法大多使用應變規，而應變規的分析需要耗費巨大的佈設成本及龐大的人力進行資料採集。本研究採用影像方法進行非接觸式的方式量測，其具有彈性的使用空間，可隨時針對當下海況可能發生的破壞位置進行監控，進而更有效的利用資源。本研究開發的影像追蹤方法可以對船體結構特徵點進行分析，利用特徵點的像素位移換算比例因子得到結構的凹痕變形，除此之外，透過分析結構表面的全域應變場，可使用應變場找出結構表面應力集中處，便可預測裂縫以及破壞可能發生的位置。本研究藉由數位影像相關方法發展影像結構監測的方法，透過像素間的線性關係，可以達到亞像素的追蹤精度，因此可以打破傳統以往影像最高只能夠達到整像素解精度的限制。除此之外，透過建立表面全域的位移場，可以利用全域位移場進行拉格朗日轉換計算出表面的全域應變場，利用全域應變場則可達到在材料表面發生破壞前有效進行預測破壞的目的。

Under different sea conditions, ships encounter different structural deformations, including arching and sagging deformation. Since wave condition is a type of long-term loads, fatigue cracks in the hull structure may occur. In this study, the non-contact image monitoring of the hull structure was carried out to identify the possible structural failure modes and damage locations of the hull structure during the voyage. Most of the traditional hull monitoring methods nowadays rely on strain gauges, and the analysis system of strain gauges requires huge deployment costs and huge manpower to collect data. To this end, the image method is used for non-contact measurement, which possesses flexibility and effectiveness, and can also monitor the possible damage locations of the current sea conditions at any time. The image tracking method developed in this study can be used to analyze the feature points of the hull structure, and employs the pixel displacement of the feature points to convert the scale factor to obtain the dent deformation of the structure. By identifying the stress concentrations on the surface of the structure, it is possible to predict where failure and cracks are most likely to occur. Moreover, the digital image correlation method is used for developing an image structure monitoring method. Through the linear relationship between pixels, the sub-pixel tracking accuracy can be achieved, so it can break the traditional limitation that the previous image can only achieve the highest accuracy of the whole pixel solution. In addition, by establishing the global displacement field of the surface, the global strain field of the surface can be calculated by using Lagrangian transformation of the global displacement field.