Data Analysis Method

To complete our second goal — developing a failure probability for key components — we had to first investigate the raw data to find what the main component failures are. Then, to determine the reasons for the main component failures, so we broke down the factors that could possibly affect failure. We looked at the effect of:

Additionally, we looked at more specific factors that could affect failure including:


The results of this analysis can be found in the Data Relationships section of the website. To quantify the results and to normalize the data for all the variables, we chose to display the results by Failure per GigaWatt hour per year.
The power generated was calculated by using re-analysis data, which is wind speed recorded every 6 hours, for nodes across the UK. Then by referencing the appropriate power curve, we were able to predict the production of each wind farm.
Due to availability of publicly available data, this calculation of power generated was the most accurate data we could obtain.

As a consequence of using this wind data, the calculated capacity factors are slightly higher than normal. However, we were able to get information for some sites, and our values differed by only three or four percent. By using this Failure per GigaWatt-hour per Year value for the Y axis in our graphs, we could compare the number of failures and the time the windfarm works, and we could compare the results by year.



Sources Used

British Wind Energy Association
Department for Business, Enterprise and Regulatory Reform
ESRU Wind Resource Prediction Project
Blade Failure Images
Gearbox Failure Images courtesy of Design Unit: University of Newcastle upon Tyne
Turbine Diagram and Generator Failure Images courtesy of British Gear Association Member of the Schaeffler Group