Advanced Markovian wind energy models for smart grid applications

Markov Chains are widely used for developing wind energy resource models for power system analysis applications. However, the Markovian wind models currently available in the literature cannot capture the temporal variations in wind speed/power output over time periods shorter than 1 hour. This means that they are unsuitable for smart grid applications, which typically require simulations with short time steps, e.g. to the order of minutes or seconds. This paper introduces a novel approach to modelling wind energy resources using “Nested Markov Chains”. It is shown in the paper that this method can accurately capture higher-frequency variations in the wind energy resource. The methodology is demonstrated using recorded onshore and offshore wind data sets. The resulting model can be readily applied for smart grid analysis, allowing the user to replace large historical wind data sets with a simple and efficient analytical model.

[1]  Z. Şen,et al.  First-order Markov chain approach to wind speed modelling , 2001 .

[2]  Kiti Suomalainen,et al.  Synthetic wind speed scenarios including diurnal effects: Implications for wind power dimensioning , 2012 .

[3]  Ron Gallagher,et al.  Monte Carlo Simulations of Wind Speed Data , 2009 .

[4]  Birgitte Bak-Jensen,et al.  Model of a synthetic wind speed time series generator , 2008 .

[5]  James Kirtley,et al.  Pitfalls of modeling wind power using Markov chains , 2009, 2009 IEEE/PES Power Systems Conference and Exposition.

[6]  O.N. Gerek,et al.  The Effect of Markov Chain State Size for Synthetic Wind Speed Generation , 2008, Proceedings of the 10th International Conference on Probablistic Methods Applied to Power Systems.

[7]  F. Y. Ettoumi,et al.  Statistical bivariate modelling of wind using first-order Markov chain and Weibull distribution , 2003 .

[8]  Sasa Z. Djokic,et al.  Modelling of wind generation at all scales for transmission system analysis , 2013 .

[9]  S. Dutta,et al.  Optimal storage sizing for integrating wind and load forecast uncertainties , 2012, 2012 IEEE PES Innovative Smart Grid Technologies (ISGT).

[10]  Robert. S. Weissbach,et al.  Markov based estimation of energy storage requirements accounting for seasonal variations , 2010, IEEE PES General Meeting.

[11]  R. Billinton,et al.  Multistate Wind Energy Conversion System Models for Adequacy Assessment of Generating Systems Incorporating Wind Energy , 2008, IEEE Transactions on Energy Conversion.

[12]  Andrew Kusiak,et al.  Mining Markov chain transition matrix from wind speed time series data , 2011, Expert Syst. Appl..

[13]  Yoram Singer,et al.  The Hierarchical Hidden Markov Model: Analysis and Applications , 1998, Machine Learning.

[14]  F. C. Kaminsky,et al.  A COMPARISON OF ALTERNATIVE APPROACHES FOR THE SYNTHETIC GENERATION OF A WIND-SPEED TIME-SERIES , 1991 .

[15]  G. Papaefthymiou,et al.  MCMC for Wind Power Simulation , 2008, IEEE Transactions on Energy Conversion.