Weather radar signal processing techniques

Automatic Detection and Mitigation of Wind Turbine Clutter
using Time-Series Data

I collaborate with Dr. Robert Palmer to study the capability of phased array radars to identify and mitigate the weather radar interference generated by wind farms. This project is funded by the US Department of Homeland Security and is focused on three major thrusts: characterization, automatic detection, and mitigation of wind turbine clutter (WTC).

Characterization of WTC is critical to the design of effective detection and mitigation algorithms. Only through the understanding of temporal, spectral, spatial, and polarimetric characteristics of WTC signals, is it possible to attack the problem in a systematic way and eventually create robust detection and mitigation techniques. Over the years, we have obtained a variety of WTC data from polarimetric and non-polarimetric, operational and research, scanning and non-scanning X-, C-, and S-band weather radars. For example, we recently took the NO-XP, a mobile radar built by NSSL and OU, to Weatherford, OK where we got close to an operational wind farm. The polarimetric capabilities of this unique radar allow for identification of the type of scatterers in the radar volume. Through the use Doppler spectra and polarimectric spectral densities we continue to improve our understanding of these signals.

Automatic detection of WTC is a key component of any mitigation scheme. Because the turbines are always at the same location, it would seem easy to identify where WTC contaminates the weather data. However, under certain atmospheric conditions, anomalous propagation of the radar beam can occur such that WTC corrupts weather data without the radar operator knowing of this contamination. In addtion, if the weather returns overpower the WTC contamination is not an issue. Therefore, as a first step in any mitigation scheme, an effective detection algorithm is needed to perform automatic flagging of contaminated data. The flagged data can then be censored or filtered out, thus reducing harmful effects that propagate to other algorithms, such as quantitative precipitation estimates. In a recent AMS Radar Conference paper, both actual and simulated WTC data were used to study the characteristics of WTC to design a WTC detection algorithm. It was shown that unique spectral features of the Doppler spectrum related to WTC signatures can be used to classify the radar return as contaminated by WTC or not. This work was also accepted for publication in an upcoming issue of the AMS Journal of Oceanic and Atmorpheric Technology. The early release of this paper can be found here.

Mitigation of WTC remains a challenge, in particular on operational radars, where the observation times are typically short. Phased-array radars have the unique advantage of allowing adaptive scanning stragies and the ability to combine signals from different channels to change the antenna beam pattern. This process, referred to as spatial filtering, can be exploited to migitate WTC based on spatial characteristics. A recent conference paper on spatial filtering using phased array radar is here. In addition to this, we are currently investigating non-stationary signal processing techniques that use polarimetric signatures to detect and mitigate wind turbine clutter contamination embedded in weather signals. The most recent work is documented in this paper presented at the 2011 IEEE Radar Conference.

Back to WTC