Wind power generation which serves as a source of renewable energy faces certain challenges due to short-term fluctuations in power output. This led to the addition of a battery system in order to reduce these pitfalls and as a result, the effect of the short-term fluctuations in relation to the battery system needs to be evaluated. One means of evaluating this effect is the use of a power flow simulation.
A group of researchers from , Waseda University in Japan, proposed an innovative method whereby synthetic wind power profiles with high temporal resolutions for power flow simulation can be generated by reproducing plausible statistical behavior of a realistic short-term fluctuation. The work is now published in journal, Energy Procedia.
In order to achieve a realistic short-term fluctuation which occurs in wind power generation, the power flow simulation observes the time-series statistical behaviors. The methods used in achieving the realistic short-term fluctuation include; the previously used autoregressive mean average approach and the block bootstrap approach.
The authors further compared the statistical property of the short-term fluctuations generated from three different approaches; naive bootstrap, autoregressive mean average bootstrap approach and the block bootstrap coupled with evaluations made by finding the autocorrelation functions of the detrended sequence which stands for the typical short-term fluctuation in wind power generation.
Following the generation of ten plausible short-term fluctuations for each approach from a dataset of a case study in Japan, the lowest root mean square error of the autocorrelation functions was observed in the block bootstrap approach. This shows that the block bootstrap approach gave the highest accuracy amongst the three. It improved 26.5% from the autoregressive mean average approach.
The lowest root mean square error for variance sequences was also observed with the block bootstrap approach, which indicates that the generated short-term fluctuations possess realistic volatility.
The block bootstrap approach which exhibited plausible volatility and accuracy of the detrended sequence indicated an imaginative time-series statistical property of the real-world fluctuation in wind power generation which would be of relevance in determining the effects of the short-term fluctuations on battery systems of future wind energy technologies.
Furuya, S., Fujimoto, Y., Murata, N., Hayashi, Y. Reproducing Statistical Property of Short-term Fluctuation in Wind Power Profiles, Energy Procedia 99 ( 2016 ) 130 – 136.
Waseda University, 3-4-1 Okubo, Shinjuku, Tokyo 169-8555, Japan.