Extreme winds severely endanger structural integrity of wind turbines. In order to understand failure mechanisms of wind turbine structures under extreme wind conditions, this paper presented a study on structural failure of wind turbines damaged by super typhoon Usagi in 2013. A particular focus was placed on the effect of strong wind speed and large change of wind direction on tower collapse and blade fracture. Post-mortem investigation was conducted at field, and data of local winds, tubular steel tower and composite rotor blade were collected and analyzed. A systematic procedure was developed by integrating wind load calculation and structural modeling. Quantitative investigation on structural response of turbine towers and rotor blades was conducted to identify the root causes of failure. Failure scenarios were studied considering typical stop positions of the wind turbine. The fuse function of the rotor blade whose fracture is able to protect the tower from collapse was also addressed. Based on the findings obtained from this study, some suggestions were proposed to modify the current IEC design standard and a few potential future directions of study were addressed to reduce the risk of wind turbine failure under extreme wind conditions such as typhoon and hurricane.

Conclusions

This paper studied structural failure of wind turbines catastrophically damaged by the super typhoon Usagi in 2013. A particular focus was placed on the effect of strong wind speed and large change of wind direction on failure response of rotor blade and tubular tower when the wind turbine is at an emergency stop state with all blades feathered, yaw locked and mechanical disk brake applied. Through this study, the following conclusion was drawn:

(1). Blade fracture and tower collapse were found to be major structural failure of wind turbines in this study. Following the proposed procedure, failure locations of the blade and the tower were calculated and they were found to be in good agreement with the field observation. Furthermore, the plausible root-cause of the blade fracture was identified to be the blade strain exceeding the failure strain of unidirectional composites in the spar cap, while the plausible root-cause of the tower collapse was found to be local inelastic buckling due to steel yielding.

(2). The rotor blade was found to fracture at a hub wind speed of 71.5 m/s, while the tubular tower was found to collapse at a hub wind speed of 59.8 m/s which is much smaller than the design survival wind speed of 70 m/s for the wind turbine under concern. The failure location of blades occurred at a normalized blade length approximately from 0.26 to 0.48.The failure location of towers occurred at a normalized tower height approximately from 0.20 to 0.22. The tower collapse is possibly associated with a design defect in shell wall thickness which decreases at the failure location of turbine towers.

(3). The stop position of the wind turbines after power loss has vital importance on structural failure due to revolving wind during typhoon impact. A stop position once structurally favorable to the turbine may lead to significant increase of wind loads when the prevailing wind direction changes. Based on the wind direction and the corresponding turbine positions at the time of power loss, the collapse direction of towers was justified by the study.

(4). The blade not feathered as intended was found to be able to protect the tower from collapse taking advantage of its fracture. Upon the fracture of the non-feathered blade, the contributions of rotor wind load and rotor wind moment to the tower stress were reduced by 39% and 3%, respectively. As a result, the stress sustained by the tower was reduced by more than 42% and is decreased to a safe level. Considering the small proportion of the blade cost in the total turbine cost, it is advisable to design a blade as a fuse component in order for the rest of structural components of wind turbines to survive extreme wind conditions.

(5). In extreme wind conditions, the tower wind load is comparable to the rotor wind load and has significant contribution to the tower stress. The contribution of the nacelle wind load to the tower stress is also considerable when the wind turbine subjects to a crosswind. These load sources are not paid enough attention to when wind turbines are at a normal operational state. However, they have to be carefully examined in order for the tower to survive extreme wind conditions.

(6). Based on the findings from structural failure analysis, this study suggested a potential strategy to stop a wind turbine at a stationary state before the impact of a forthcoming typhoon or hurricane to reduce the risk of structural failure. Meanwhile, it is suggested that the tower strength should be given priority over the blade design, and the partial safety factors for the consequence of failure, which are specified to be the same for the blade and the tower in the current IEC design standard, should be differentiated to address the tolerance of the blade failure and the importance of the tower strength during strong winds.

It should be noted that great efforts have been made in this study to ensure the accuracy of the collected data and the information provided by third parties. Assumptions have been introduced in the procedure of failure analysis as presented in Section 4.1 considering that their effects on the results are not significant. As the failure of wind turbines under extreme wind conditions is a complex issue associated with inter disciplines of meteorology, aerodynamics, aero elasticity, structures, materials, controlling, etc., more study is needed to provide more understanding on and accurate answers to the remaining unknowns. Some future directions of study may include: prediction of extreme wind conditions of a given site before typhoon or hurricane attacks, optimal stop position of wind turbines to minimize extreme wind loads, the desired/controlled failure scenario of wind turbine systems, and innovative rotor design with retractable and/or deployable blades to reduce extreme wind loads applied to wind turbines.

The results have been published on Engineering Failure Analysis 60 (2016) 391–404.