Range-Doppler domain signal processing to mitigate wind turbine clutter

Abstract

Wind turbines produce clutter signals that can bias estimates of the spectral moments and polarimetric variables of weather signals. These biases can propagate to and negatively influence the output of automatic algorithms, such as severe weather detection and quantitative precipitation estimates. Moreover, existing ground clutter filters are ineffective at removing wind turbine clutter (WTC) contamination because the moving components of the wind turbine produce clutter signals with nonzero Doppler frequency shifts. As the first step in any mitigation scheme, an automatic WTC detection algorithm is necessary and was recently developed by University of Oklahoma and National Severe Storm Laboratory scientists. After successfully detecting the presence of WTC, the goal is to devise signal processing algorithms that mitigate this contamination so that the weather signal can be recovered and used to estimate the spectral moments and polarimetric variables. However, WTC is inherently non-stationary due to the moving wind turbine blades, which makes frequency-domain-filtering based clutter mitigation methods ineffective. In this work, we propose a new signal processing technique to separate the WTC from the weather signal in the range-Doppler domain. This technique exploits the different spatial and spectral characteristics of WTC and weather signals. Real weather signals and WTC data are used to test the effectiveness of the mitigation scheme.

Impacts of wind turbine siting on radar clutter mitigation

“…Spacing of 7D (D as rotor diameter) is conventionally used by many current wind farms, but studies have shown that the optimal average spacing is considerably larger at 15D [17]. The increase of spacing will benefit WTC mitigation algorithms based on signal processing [10,11] because there will be more uncontaminated volumes in between contaminated ones, which will provide more accurate prior information about the weather estimates in the proximity of the wind farm area.…”

“…The difficulty to separate wind turbine and weather return signal in both the spatial and Doppler domains has prompted ideas of combining information in both domains. Range-Doppler processing [10] exploits the difference between wind turbine and weather signals in the range-Doppler domain, alleviates Doppler spectral bins identified as WTC and reconstructs weather signal based on the spatial continuity assumption. The Adaptive Spectrum Processing (ASP) technique [11] makes use of the clustering nature of weather signals and adaptively constructs band pass filters for mitigation of WTC.…”

Wind turbine signal modelling approach for pulse Doppler radars and applications

“…In (13), σ h is the RCS of the WT tower and can be calculated through (10) considering the segments found on the tower of the WT. In addition, mod(·,·) is the modulo operator which controls the unambiguous velocity range while the step function controls the signal level of the WTC considering the scattering properties of the blade, stationary part (i.e., tower) of the WT and DR values.…”

“…The transponder‐based detection method, which exploits a transponder to find the location of WTs, is found in [12]. Furthermore, additional WT detection efforts such as range‐Doppler imaging by using the radar raw data corresponding to the azimuthal sector that WTs are located in [13], adaptive clutter mapping via averaging multiple radar‐scan data [14] and disposition of the radial velocity characteristics of WT and meteorological data [15] have been introduced. The literature also introduces a few design proposals that focus on the mitigation of WTC effects on radar signals.…”

On the mitigation of wind turbine clutter for weather radars using range‐Doppler spectral processing

“…The authors of [20] proposed an algorithm which detects the WTC contamination in spotlight mode and uses interpolation to reduce the bias in reflectivity and radial velocity estimates; however, this suffers from the same problems as the previously mentioned algorithms. Recently, an algorithm based on edge detection was proposed in [21]. This algorithm exploits the spatial discontinuity of the WTC signal in the range-Doppler spectrum (RDS).…”