칼만필터에 의한 고정식 해양구조물의 미계측응답 추정

Abstract

Monitoring underwater structures is challenging in Structural Health monitoring field. These structures are subjected to tidal current which can easily cause sensor malfunctioning. Estimating structural responses at unmeasured important locations can be a powerful alternative to the direct measurement particularly when installing sensors is difficult and sensors are malfunctioning. This paper targets on Bottom-fixed offshore structures such as mono-piles, tripods, jackets, and gravity-based structures that supports offshore wind turbines and tidal stream turbines.Response estimation at unmeasured locations from a small number of measurements has been attractive to the researchers who seek to overcome the limited instrumentation. Various efforts for response estimation has been made such as finite element model updating with modal expansion (Iliopoulos et al. 2014), natural input modal analysis, time varying auto-regressive model (Yazid et al. 2012), and the model-based Kalman state estimator. Among these efforts, the Kalman state estimator associated with the FE model has been known as an effective tool to estimate the unmeasured responses. Papadimitriou et al. (2009) used strain measurements in the numerical simulation to obtain strain in the entire body, which is subsequently utilized to estimate fatigue remaining life of the structural model. Smyth and Wu (2007) used the Kalman filter to fuse acceleration and diations can be a powerful alternative to the direct measurement particularly when installing sensors is difficult and sensors are malfunctioning. This paper targets on Bottom-fixed offshore structures such as mono-piles, tripods, jackets, and gravity-based structures that supports offshore wind turbines and tidal stream turbines.Response estimation at unmeasured locations from a small number of measurements has been attractive to the researchers who seek to overcome the limited instrumentation. Various efforts for response estimation has been made such as finite element model updating with modal expansion (Iliopoulos et al. 2014), natural input modal analysis, time varying auto-regressive model (Yazid et al. 2012), and the model-based Kalman state estimator. Among these efforts, the Kalman state estimator associated with the FE model has been known as an effective tool to estimate the unmeasured responses. Papadimitriou et al. (2009) used strain measurements in the numerical simulation to obtain strain in the entire body, which is subsequently utilized to estimate fatigue remaining life of the structural model. Smyth and Wu (2007) used the Kalman filter to fuse acceleration and di