A Two-Tier Wind Power Time Series Model Considering Day-to-Day Weather Transition and Intraday Wind Power Fluctuations

Markov chain‐based wind power time series modelling method considering the influence of the state duration on the state transition probability

“…Forecast [2][3][4] and simulation [5][6][7][8][9][10][11][12][13][14][15][16][17][18][19] of wind power are two main ways to characterise wind power. The former is to predict wind power values in the future, while the latter is to generate synthetic wind power time series (SWPTS) that are highly consistent with the historical ones in characteristics.…”

“…There are two main approaches in wind power simulation modelling: time series (TS) analysis methods and artificial intelligence (AI) techniques. The former includes the Markov chain (MC)-based method [5][6][7][8][9][10][11][12][13][14][15][16][17] and the auto-regressive and moving average (ARMA) method [18,19], while the latter includes neural networks [23], artificial neural fuzzy inference systems [24] and so on. AI techniques usually need massive data to build their intelligence, thus they are mainly used in big data scenarios.…”

“…Compared with AI techniques, the MC-based method requires only limited data to achieve satisfactory results [25], not restricted to relatively small data scenarios. In addition, compared with the ARMA method, the MC-based method can ensure a good fit to the probability distribution (PD) of wind power, leading to a more accurate estimate of wind energy yield [6,7,15,16]. Hence, we focus on the MC-based method in this paper.…”

A meteorological information mining-based wind speed model for adequacy assessment of power systems with wind power

“…There are two main types of wind speed models: probabilistic models [2][3][4] and time series models [5][6][7][8][9][10][11][12][13][14]. Weibull distribution [2][3] and Rayleigh distribution [4] are most widely used in probabilistic models which can reflect the statistical characteristics of wind speed.…”

“…Improved models such as the semi-Markov model [10] and Bayesian Markov model [11] show better accuracy in capturing time evolution characteristics. A two-tier reliability model is proposed in [12], which models the weather types and wind power fluctuations by Markov chains, respectively. Besides, models such as the two-dimensional wind speed statistical model [13] and timedependent clustering model [14] are developed for reliability assessment.…”