Objective To fully explore the data value of lidar in wind turbine wake observation, conduct an in-depth study on the three-dimensional (3D) wake structure and wind speed attenuation characteristics in wake zone, and further provide references for improving engineering models of wake and enhancing the prediction accuracy of mesoscale meteorological models.

Method This paper established a new method for directly locating and analyzing wind turbine wake on radial velocity images of lidar. It extracted characteristic parameters such as wake length, width and wind speed attenuation under different wake superposition states, and verified three engineering models of wake based on these parameters.

Result The results show that the wake zone of a single wind turbine appears as a strip-shaped area with a length of 11.0～13.0 times the rotor diameter behind the rotor swept area on the horizontal azimuth scanning radial velocity image. On the vertical cross-section scanning radial velocity image, it presents an elliptical area with a width of approximately 1.8～2.0 times the rotor diameter, covering the entire rotor height range. When the wake zone of the upstream turbine affects the downstream turbine, the wind speed distribution in the wake zone of the downstream turbine adjusts: the corresponding position of the minimum wind speed in the near-wake region moves forward, the wind speed recovery rate in the far-wake region slows down, and the wake length extends slightly. Under partial superposition of the upstream turbine wake, the wind speed and output power of the downstream turbine decrease by nearly 5% and 15% respectively. Under full superposition, the decreases expand to nearly 27% and 43%.

Conclusion The study indicates that the 3D wake model has good simulation capability for the downstream wake structure under different wake superposition states. However, the single Gaussian distribution wake model cannot reproduce the cross-wind direction wind speed distribution characteristics in the near-wake region. In practical applications, the simulation error should be carefully evaluated when the turbine spacing is less than 3.0 times the rotor diameter.