Abstract:

In order to control the transient over-voltage accurately and deal with the sharp change of voltage in time and effectively, this paper proposes a data-driven power system security and stability analysis and control strategy under the condition of high proportion of photovoltaic power generation connected to the grid. This strategy involves continuous monitoring of key electrical indicators such as voltage fluctuations in the high-penetration PV grid-connected system, upon which a neural network model based on Radial Basis Function-Particle Swarm Optimization (RBF-PSO) is constructed. Using real-time data, the model can predict the transient behavior of the power grid. Once the prediction indicates a risk of transient over-voltage, the system immediately sets the nodes to be controlled according to preset priorities and, in combination with sensitivity analysis at PV connection points, automatically activates the preset transient over-voltage control mechanism by adjusting the reactive power compensation and active power reduction functions of the PV inverters. This aims to swiftly restore system voltage stability and ensure equipment safety. Experimental verification shows that this method not only responds rapidly but also significantly reduces the reactive power required by PV inverters.

Key words: photovoltaic power station, peak voltage, transient overvoltage control