Switching strategy of the low wind speed wind turbine based on real-time wind process prediction for the integration of wind power and EVs.

Utilizing the secondary wind resources in cities and countryside is a significant way to promote wind power consumption and sustainable transportation. However, the probability of wind speed near the cut-in wind speed increases in such area and resulting in frequent on/off switches as well as large fatigue load of the wind turbine. To address these problems, this paper proposes a switching strategy of the low wind speed wind turbine based on real-time wind process prediction. First, the wavelet decomposition and neural network are employed to predict the time series of wind speed in a real-time manner. Second, based on the historical and predicted wind speed, the typical wind processes are extracted by using the x-means algorithm. Third, seven indexes are defined to quantify the characteristics of the wind process set before developing the corresponding switching strategy for each type of it. Data from NREL are used to validate the proposed models. The results show that, the proposed strategy increases the energy yield, reduces the number of wind turbine switches as well as the power fluctuation. Therefore, the proposed strategy is beneficial to both wind power projects development and electric vehicles charging. • Low wind speed turbine switching by real-time wind process prediction is proposed. • Typical wind process is established by clustering the historical and future winds. • Strategy adjusts dynamically based on seven presented wind characteristic indexes. • The proposed strategy increases the energy yield and reduces the switches. • The proposed strategy improves quantity and quality of wind power for EVs charging. [ABSTRACT FROM AUTHOR]