Owing to the randomness and intermittence of photovoltaic (PV) generation

the risk of power system operation is rapidly increased after integration with large-scale photovoltaic power plant. Traditional spinning reserve estimating methods are no longer able to meet the requirements of power system operation integrated with PV power plant. The probability distribution model of solar irradiance and PV generation output are developed

and the Latin hypercube sampling method is used to simulate PV generation output scenarios. The improved K-means clustering algorithm based on Huffman tree is used to effectively cluster the PV generation output scenarios

which will greatly reduce the scenario number while ensuring the probability distribution characteristics of PV generation output. On this basis

a model for spinning reserve demand estimation in the power system integrated with large-scale PV power plant is proposed to minimize the system comprehensive operation cost

thus balancing the economy and reliability of power system operation. The improved IEEE-RTS 96 system is simulated via the proposed model to prove its rationality and effectiveness. This work is supported by National Basic Research Program of China (973 Program) (No. 2012CB215102)

National Natural Science Foundation of China (No. 51207062

No. 51207092)

and Self-independent Innovation Fund of Huazhong University of Science and Technology (No. 2013TS057).