Posters
Come meet the poster presenters to ask them questions and discuss their work
We would like to invite you to come and see the posters at our upcoming conference. The posters will showcase a diverse range of research topics, and will give delegates an opportunity to engage with the authors and learn more about their work. Whether you are a seasoned researcher or simply curious about the latest developments in your field, we believe that the posters will offer something of interest to everyone. So please join us at the conference and take advantage of this opportunity to learn and engage with your peers in the academic community. We look forward to seeing you there!
PO206: Gap Filling of Measured Wind Speed Data using Machine Learning Approach
Zia ul Rehman Tahir, Professor, Associate, Department of Mechanical Engineering, University of Engineering and Technology Lahore, Pakistan, Pakistan
Abstract
The onshore or offshore wind power projects require reliable wind speed and direction for power prediction and economic feasibility of the plant. The wind data measured by wind mast and buoys is most reliable, sometimes satellite or reanalysis wind data can be used for preliminary assessment of power potential. The measured data may contain some gaps due to accidents and severe weather conditions, the gaps can be filled by several techniques. This study is a novel technique to fill gaps in mast measured wind speed data using reanalysis data by employing machine learning approach. The mast measured data contains four parameters: Wind Speed (WSm), Temperature (Tm), Surface Pressure (Pm), Relative Humidity (RH). The European reanalysis (ERA-5) wind speed data was used which contains three parameters: Wind Speed (WSE5), Temperature (TE5), Surface Pressure (PE5). The measured wind speed (WSm) was used as output to the network and rest of the parameters (Tm, Pm, RH, WSE5, TE5, PE5) as inputs to the network. The Bayesian Regularization (trainbr) algorithm was selected. The performance of the train network was checked by network training performance (as mean square error, MSE), error histogram and training regression plots. The networks were trained several times, to select net with best performance, the overall training performance MSE was 0.4117 and the overall regression coefficient was 0.9677. The best trained network was used to predict wind speed using a validation set. The rMBE, rMAE, rRMSE and R for predicted time series were 0.18 %, 7.39 %, 10.74 % and 0.976 respectively. The trained network was used to fill gaps in the measured data with less than 1 % error. The technique developed in this study can be used to fill gaps in the measured time series data with very low uncertainty.
No recording available for this poster.
