ADVANCED SIGNAL PROCESSING METHODS APPLIED TO ACOUSTIC WIND PROFILING FOR USE IN WIND FARM ASSESSMENT

909894ca-15b8-4e30-a30a-c7d270be3d56

Closed

EPSRC
EPSRC
EPSRC
EPSRC
EPSRC

£1,631,300

Jan. 5, 2009

July 4, 2012

The UK wind energy potential is quoted to be as high as 40% of Europe's entire wind resource. Its exploitation has been high on the political agenda ever since the publication of the government whitepaper on renewable energy in 2003, where a target of 10% of energy from wind farms was defined. With the resulting increasing demand for new wind farm developments, the search for suitable sites becomes both more problematic and more important. New sites are now being populated with maximum capacity wind turbines as tall as 120m. Given the cost of such installations, assessing economic viability requires high-precision wind measurements to forecast the power yield. The current standard uses a measurement technology called cup anemometers. This standard has been questioned for some time. As these small instruments require mast structures which cannot be easily moved it is doubtful whether the data are representative for the proposed turbine blade areas. A promising alternative measurement method uses sound pulses to measure an entire wind profile up to and above the heights of modern wind turbines. These instruments, so called SODARs, consist mainly of an array of loudspeakers and are easy to move around a prospective wind farm site to measure profiles at all proposed turbine locations. One major limitation of conventional SODAR measurements is the loss of data under common atmospheric conditions. This project sets out to overcome this limitation by adapting signal processing techniques which are common in RADAR and SONAR technologies to improve data quality and therefore availability substantially. This approach also promises to enhance the number of data points in a profile. The enhanced spatial data resolution can be particularly important for operators of wind turbines in situations with large wind shear when the load on the blades is unevenly spread across the blade diameter. Even when signal strength is good, another common problem with SODARs is to identify data contamination by sources such as rain and reflections from fixed objects. As the atmospheric signal travels in a different direction than the fixed echoes, we will use the loudspeaker array to locate the directionality of the sound to extract the signal from the noise. In a first step we will simulate the SODAR signals in a computer model to evaluate a number of possible signal processing techniques. At stage two of the project we will implement the most promising ones on a real SODAR instrument. The project will conclude with a comparison between the wind profiles of the new technology with those of an ordinary SODAR to evaluate the extent of the improvements. If successful, the technology can be integrated into commercial SODAR instruments. The enhanced data quality can then also benefit other applications such as air quality studies, the detection of aircraft wake vortices and hazard prevention.