采用分布式算法的多区域电网风-光-储容量优化配置

Aiming at a series of problems such as power quality degradation and unreliable operation of the power system caused by the continuous increase of the penetration rate of renewable energy power generation in the entire power grid, this paper proposes a multi-area wind-photovoltaic-storage capacity optimization method for maximizing the average renewable energy accommodation level under the typical renewable energy generation time series based on the stochastic programming theory by comprehensively considering the wind power, photovoltaics and other power sources, as well as the operation constraints of tie lines and energy storage equipment. This method mainly considers that the amount of the generated wind power and photovoltaic power that cannot be accommodated by the power grid after meeting the load demand will be stored without exceeding the capacity of the energy storage equipment and will then be provided by the energy storage equipment during the off-peak period of renewable energy generation, giving full play to the space-time complementarity of the wind power and photovoltaic power and the peak load regulation advantage of energy storage. First, an uncertain set of typical renewable energy power time series characteristic sequences is constructed with the time-space correlation considered; then, the model is decomposed according to the typical sequences, and the distributed optimization solution of the objective function is realized by the distributed penalty primal-dual subgradient algorithm; finally, the case is analyzed based on the IEEE-24, 31-node system to obtain the wind-photovoltaic-storage capacity allocation results, and it is suggested that the energy storage equipment in the multi-area power grid should be considered to be built in the area with installed wind and photovoltaic power capacity. The simulation results show that compared with the traditional centralized optimization method and the Lagrangian relaxation method, the proposed distributed optimization method has more and more obvious advantages in computing efficiency as the number of power grid areas increases. The proposed distributed method, compared with the Lagrangian relaxation method, improves the convergence speed by 57.1% in 10 typical sequences. The proposed method can provide a systematic solution for the optimal allocation of the capacity of the wind-photovoltaic-storage multi-area power grid considering the uncertainty of renewable energy sources.