Risk analysis of power surging in novel power systems: a hybrid framework driven by accident chain and WPMixer

To address the complex dynamic power risks in novel power systems induced by the superimposition of "three-high" characteristics (high renewable energy penetration, high direct current infeed ratio, and high load density) and extreme weather, this paper proposes the concept of "power surging" alongside a comprehensive risk analysis and planning methodology. Power surging is defined as the random and drastic variations in power flows across voltage levels or regions, and it is classified into longitudinal and transverse surging based on propagation directions. To overcome the limitations of traditional theories in characterizing cross-level and cross-regional transient power processes, the accident chain theory is introduced and coupled with the proposed concept. Focusing on weak disturbance scenarios, the accident chain is decomposed into four stages: initial trigger, power surging, system instability, and final consequence. Explicit quantitative correlations are established between each stage and key planning parameters, forming an analytical framework that encompasses total risk calculation and key risk point identification. Furthermore, the WPMixer model is proposed to synergistically capture the complex temporal mappings between planning parameters and power surging, enabling efficient surge scale prediction. Simulation results based on an extended IEEE 39-bus system validate that the longitudinal and transverse power surging characteristics align with the theoretical expectations. The quantitative analysis reveals that energy storage capacity and tie-line reserves exert significant regulatory effects on longitudinal upward surges and transverse inflow surges, with sensitivity coefficients of -0.3416 and - 0.0810, respectively. The final consequence and initial trigger stages are identified as the most critical risk points. This study provides an effective quantitative tool for the dynamic risk assessment and resilient planning of novel power systems.