Identification of wind loads on a 600 m high skyscraper by Kalman filter. (1st January 2023)

• Record Type: Journal Article
• Authors: Yang, Bin; Zhu, Haitao; Zhang, Qilin; Wüchner, Roland; Sun, Siyuan; Qiu, Jiahui
• Is Part Of: Journal of building engineering. Volume 63(2023)Part A
• Publication Date: 2023-01-01
• Subjects: Wind load inversion -- Kalman filter -- Modal space -- L-curve method -- Two-stage strategy; Building -- Periodicals; 690.05
• DOI: 10.1016/j.jobe.2022.105440
• Languages: English
• ISSNs: 2352-7102

Abstract

Abstract: This paper presents a wind load inversion algorithm for high-rise buildings based on the Kalman filter. With this method, the wind load on a super high-rise building can be identified with acceleration responses obtained from limited measured storeys. According to the wind load obtained from a numerical wind tunnel test and the responses from a dynamic analysis on the finite element (FE) model of Shanghai Tower, which is a 632 m high skyscraper with 127 storeys, the wind load inversion was conducted using the Kalman filter method. A series of discussions on the key issues of this inversion algorithm was carried out in this paper, including the reduction of the modal space, the effect of the measurement noise, the effect of the inaccurate estimation of the initial value, and the effect of the introduction of the mean wind load. Some effective methods are employed in this study, such as the POD method and the L-curve method. A Two-stage Strategy is proposed in this paper and achieves good performance in many different cases. Highlights: The wind load inversion algorithm can identify the responses and wind load of a super high-rise building with only acceleration data from limited measured floors. The Kalman filter-based wind load inversion algorithm can only estimate the fluctuating component of the wind load, which dominates the vibration of the super high-rise building. The minimum dimension of the modal space is a key parameter to the load inversion algorithm. It can be determined by the POD method. An appropriate noise adjustment coefficient γ can reduce the effect of the measurement noise, initial value and the mean wind load. The optimal γ can be determined by the L-curve method. The Two-stage Strategy proposed in this paper is based on the L-curve method. It has good performance in speeding up the convergence of the wind load inversion.