【摘要】 为实现风电出力时间序列的高性能模拟，文中提出了一种基于SAGA-KM(simulated annealing and genetic algorithms-K-means)算法实现典型风电场景分类和基于Copula函数进行风电日过程马尔可夫过程建模的风电模拟方法。SAGA-KM算法将传统KM算法与遗传算法和退火算法相结合，能显著提高风电场景分类效果；基于Copula函数建立的马尔可夫链精细概率模型，以去粗粒化方式实现马尔可夫过程蒙特卡洛模拟，克服了粗粒化引起的概率分布偏差。针对甘肃省某风电场数据进行实际模拟，结果表明文中方法生成模拟序列的统计分布特性、自相关函数特性和日均功率的分布特性与实测数据都非常接近，该方法能很好地保留风电序列的概率分布特性和随时间变化的波动特性，具有重要的工程实用价值。更多还原

【Abstract】 In order to achieve high-performance simulation of wind power output time series, this paper proposes a wind power simulation method based on SAGA-KM algorithm to achieve typical wind power scene classification and Copula function for wind power daily process. Markov process modeling. The SAGA-KM algorithm combines the traditional KM algorithm with genetic algorithm and annealing algorithm, which can significantly improve the effect of wind power scene classification; based on the Copula function, the Markov chain fine probability model is used to realize the Markov process Monte Carlo simulation, overcoming the probability distribution deviation caused by coarse-grained. The actual simulation of the data of a wind farm in Gansu Province shows that the statistical distribution characteristics, autocorrelation function characteristics and daily average power distribution characteristics of the simulation sequence generated by the method proposed in this paper are very close to the measured data. This method can well retain the probability distribution characteristics and time-varying fluctuation characteristics of wind power sequence, which has important engineering practical value.更多还原