Load identification of a 2.5 MW wind turbine tower using Kalman filtering techniques and BDS data

• Published in: Engineering structures 2023-04, Vol.281, p.115763, Article 115763
• Main Authors: Wei, Da, Li, Dongsheng, Jiang, Tao, Lyu, Pin, Song, Xiaofei
• Format: Article
• Language: English
• Subjects: BDS; Health monitoring; Kalman filter; Thrust and bending moment estimation; Wind turbine
• ISSN: 0141-0296; 1873-7323
• DOI: 10.1016/j.engstruct.2023.115763

•The bending moment estimation method based on PM-AKF is established.•The force estimation algorithm is firstly applied to an actual wind turbine structure using acceleration sensors and BeiDou Navigation Satellite System.•A new load monitoring strategy for wind turbine thrust and tower bending moment is proposed.•The static load fitting equation of the wind turbine is derived. The rapid increase in the number of wind turbine installations and operations, is increasing the importance of the health monitoring of these wind turbines. To assess the remaining service life of a wind turbine, continuous monitoring of the internal forces acting on the critical locations of fatigue in the structure is required. However, a large number of sensors are usually required for the comprehensive determination of the internal force at a site-specific location; additionally, these sensors cannot be placed in locations with strong stress or strain gradients. Therefore, in this paper, a strategy for estimating the thrust at the tower top and the bending moment at any position of the wind turbine tower is proposed, which only requires a limited number of acceleration sensors and the BeiDou Navigation Satellite System (BDS). The estimated load includes static and dynamic components. The former is calculated by fitting the derived function and the BDS data, and the latter is estimated by the time-domain inverse method using the Kalman filter and with acceleration sensors. The performance of the proposed strategy is validated in two case studies, including simulated data and recorded data from a 2.5 MW onshore wind turbine located in China. The results show that the strategy can produce an estimation of thrust and bending moment, with good accuracy.