The random fluctuation characteristics of renewable energy in the autonomous microgrid put forward new requirements for the operating and planning of the spinning reserve capacity. Based on the universal generating function(UGF)method and the Markov technique

a multi-state probability prediction model for each component in the autonomous microgrid is developed. And the probabilistic prediction based evaluation method for the spinning reserve adequacy of autonomous microgrid and its flow are proposed. The inherent characteristics of components and renewable energy are incorporated when modeling

including the uncertainty of the wind speed and the light intensity

the random failure characteristics and the topology of the wind turbine and photovoltaic system. In view of the UGF method for dealing with multi-state may generate combinatorial explosion

merging like terms

an apportion method and a K-means clustering technique are utilized to reduce the amount of states and improve the computing speed. The proposed models are applied to a microgrid test system and the simulation calculation is performed by programming in MATLAB. The results show that increasing the ramp rate of conventional units

setting reasonable wind and solar energy capacities and the initial energy storage condition can help improve the degree of spinning reserve adequacy. This work is supported by National Natural Science Foundation of China(No. 51377162)and Fundamental Research Funds for the Central Universities(No. 1081-15056187).