마르코프 사슬을 이용한 인공 바람 자료의 생성 및 통계적 검증
Use of Markov Chains for Synthetic Wind Data Generation and its Statistical Verification
0
조회
0
다운로드
3
인용
0
참고문헌
키워드
초록
Long-term time series of wind data are essential to estimate the control and output performance of wind turbine systems. In this work, we propose and verify the use of Markov chains to generate long-term synthetic wind data using measured data at Wido Island. We first identify the Weibull distribution, the daily and monthly wind speed variations of the Wido region. Statistical parameters of the synthetic wind data such as maximum wind speed, minimum wind speed, average wind speed, and standard deviation are compared with those of observed data via met mast. In addition, both plots of autocorrelation and histogram are examined to verify the coincidence between synthetic and observed wind data. The transition matrix reflects the characteristics of wind speed variations over time at the measured site. A cumulative state transition matrix is obtained from the state transition matrix to determine the next state with a random number generator. The size of the synthetic wind data was iteratively determined to minimize statistical parameter errors. Through this study, long-term synthetic wind data can be adopted to simulate both the control performance and capacity factor of wind turbine systems scheduled to be installed in the site of interest.
Abstract
Long-term time series of wind data are essential to estimate the control and output performance of wind turbine systems. In this work, we propose and verify the use of Markov chains to generate long-term synthetic wind data using measured data at Wido Island. We first identify the Weibull distribution, the daily and monthly wind speed variations of the Wido region. Statistical parameters of the synthetic wind data such as maximum wind speed, minimum wind speed, average wind speed, and standard deviation are compared with those of observed data via met mast. In addition, both plots of autocorrelation and histogram are examined to verify the coincidence between synthetic and observed wind data. The transition matrix reflects the characteristics of wind speed variations over time at the measured site. A cumulative state transition matrix is obtained from the state transition matrix to determine the next state with a random number generator. The size of the synthetic wind data was iteratively determined to minimize statistical parameter errors. Through this study, long-term synthetic wind data can be adopted to simulate both the control performance and capacity factor of wind turbine systems scheduled to be installed in the site of interest.
참고문헌 (8)
- Balzter, H., "Markov chain models for vegetation dynamics", Ecological Modelling, 126, pp.139~154 (2000)
- Srikanthan, R., "Stochastic generation of annual, monthly and daily climate data: a review", Hydrology and Earth System Sciences, 5(4), pp.653~670 (2001)
- Balouktsis, A., "Stochastic simulation of hourly and daily average wind speed sequences", Wind Engineering, 10(1), pp.1~11 (1986)
- Poggi, P., "Stochastic study of hourly total solar radiation in Corsica using a Markov model", Int. J. of Climatology, 20, pp.1843~1860 (2000)
- Lei, M., "A review on the forecasting of wind speed and generated power", Renewabe and Sustainable Energy, 13(4), pp.915~920 (2009)
- Research report on development of wind resources map, "Lee, D. I.", Data Management Service Team, Korea Meteorological Administration, ~, pp.Technical Note 2007-9 (2007)
- Dukes, M. D. G., "Estimation of extreme wind speeds with very long return periods", J. of Applied Meteorology, 34, pp.1950~1961 (1995)
- Pesch, T., "A new Markov-chain-related statistical approach for modelling synthetic wind power time series", New Journal of Physics, 17, pp.1~15 (2015)