Dynamic division method of virtual power plant based on the improvement of renewable energy consumption capacity

Abstract

Large-scale wind and photovoltaic integration can reduce fossil-fuel consumption, but renewable output uncertainty and anti-peak-regulation characteristics increase curtailment, dispatch complexity and ancillary-service costs. To improve renewable energy accommodation, this paper proposes a dynamic virtual power plant (VPP) partitioning method based on generalized spinning reserve (GSR) response capability. First, GSR capacity and regulation-rate models are established by incorporating conventional units, pumped-storage energy storage and flexible loads. Then, VPP partitioning models for low-load, normal-load and peak-load periods are formulated to minimize peak-shaving and frequency-regulation service costs under GSR response-capacity and ramp-rate constraints. The method is verified using operating data from a high-renewable regional grid in Northwest China, including three 20-MW photovoltaic stations, five 49.5-MW wind farms, five 100-MW wind farms, a 240-MWh pumped-storage station and 8 MW of flexible load, with grid load ranging from 656.9 to 836.9 MW. Results show that the proposed method forms eight VPPs in the load-valley period and five VPPs in peak-load period I. In the valley period, it reduces the ancillary-service cost by 1,700 CNY compared with individual plant integration and avoids the 225,090 CNY cost of undifferentiated direct integration; in peak-load period I, it reduces the cost by 355,900 CNY and obtains a cluster cost of -248,170 CNY. These results indicate that dynamic VPP partitioning can improve renewable-energy consumption, reduce ancillary-service expenditure and support secure operation of high-renewable power grids.