Photovoltaic grid-connected distribution line risk analysis method based on voltage fluctuation characteristics

Abstract

The invention relates to a line voltage abnormity risk assessment technology under a high-permeability photovoltaic grid-connected scene of a power distribution network, and provides an assessment method combining power grid topology, a power flow equation and a deep learning model in order to solve the problems that the physical law of a power grid is ignored and risk attribution is difficult to explain in an existing method. According to the method, multi-source voltage fluctuation data is collected and standardized, structured time sequence data is established in combination with a space structure and an electrical distance, multi-time scale fluctuation and transient characteristics are further extracted, a high-risk area is identified based on topological sensitivity, and a double-branch network composed of an improved Transform and a physical consistency auxiliary network is jointly input. By adopting a gating fusion mechanism and physical perception attention distribution, deep fusion of data driving and physical constraint is realized, a space-time sequence decision interpretable risk score and attribution thermodynamic diagram can be output, anti-fact deduction and operation and maintenance collaboration are supported, and the accuracy, interpretability and decision support capability of risk assessment are improved.

Images