- Title
- High resolution wind speed modelling of turbulent flow using Markov Chain Monte Carlo simulation
- Creator
- Evans, Samuel; Clausen, Philip
- Relation
- Solar 2014: The 52nd Annual Conference of the Australian Solar Council. Proceedings of the 52nd Annual Conference (Melbourne 8-9 May, 2014) p. 111-120
- Relation
- http://solarexhibition.com.au/past-events/solar-2014/
- Publisher
- Australian Solar Council
- Resource Type
- conference paper
- Date
- 2014
- Description
- Small wind turbines are often located in rural sites close to their load and experience highly turbulent winds. Forecasting wind speeds in these regimes is important to determine the economic viability of the wind resource, and to account for wind turbine maintenance and fatigue life issues. Previous studies have successfully used the Markov chain Monte Carlo method to simulate time series wind speed data, with results showing good statistical agreement to the recorded site data. Wind speeds previously simulated by this method are typically coarse with large time steps and broad wind velocity resolutions which yield poor results in highly turbulent flows. In this study we compare the effect of increasing model resolution from 1 ms-1 to 0.5 ms-1, while using time steps at 1 Hz in conjunction with higher order Markov chains at first, second, and third order. We conclude that increasing the resolution produces wind speed data that is more statistically accurate. The Markov chain method does not account for long-term effects in the wind speed dataset, as shown by comparing the autocorrelation functions. While increasing the chain order slightly improves autocorrelation, further development of this approach is required to fully account for long-term phenomenon.
- Subject
- wind turbines; wind forecasting; Markov Chain
- Identifier
- http://hdl.handle.net/1959.13/1297649
- Identifier
- uon:19500
- Identifier
- ISBN:9480646922195
- Language
- eng
- Reviewed
- Hits: 857
- Visitors: 794
- Downloads: 0
Thumbnail | File | Description | Size | Format |
---|