中文摘要：

随着光伏发电并网容量的不断提高,其对电力系统规划运行,以及可靠性、经济性等方面的影响日益突出。利用自回归滑动平均（auto regression moving average,ARMA）模型对光照辐射强度进行预测,结合光伏出力特性和聚类理论建立了光伏出力多状态随机预测模型,并计入光伏系统自身故障的影响,基于等效电量函数法分时段对含有光伏的电力系统进行随机生产模拟。采用EPRI-36测试系统验证了分时段随机生产模拟方法的正确性,并基于此方法分别从经济性、可靠性等层面分析了光伏并网对电力系统的影响。

英文摘要：

With increasingly increased photovoltaic （PV） generation capacity, the PV generation plays an increasingly prominent role in planning and operation, reliability and economy of power system. Using auto regression moving average （ARMA） model the illumination radiation intensity of photoillumination is predicted, and combining with output characteristics of PV generation and clustering theory a multi-state stochastic prediction model of PV output is established, in which the influence of PV generation system itself is taken into account. Using equivalent energy function method the time-interval based probabilistic production simulation of power system containing PV generation is performed. The correctness of probabilistic production simulation in accordance with time-interval is verified by EPRI 36-machine system. Based on the proposed method, the influences of grid-connected PV generation on power system are analyzed in the aspects of economy and reliability.