Cascading social risks: A cross-spatial vulnerability analysis of atypical power system failures in China

Atypical large-scale power failure (ALPF) has become more frequent within a risk society, highlighting the need to mitigate cross-spatial cascading risks in order to reduce losses and ensure public safety. This study develops a Cross-Spatial Vulnerability Model (CSVM) to analyse ALPF risks. It combines cascading risk and vulnerability theory with the bow-tie model, forming the basis of a Bayesian network for three ALPF cases in the China scenario, including summer heat and drought, winter cold waves, and coal price fluctuations. The analysis models complete Bayesian networks from official news, policy texts, and expert opinion sources, with three key findings. First, ALPFs can disproportionately impact economic stability, social stability, public health, and safety, with the probability of severe consequences being non-linearly related to the likelihood of ALPF events. Second, effective risk response to ALPFs requires focused attention on several key scenarios, such as delays in emergency response, outdated grid technologies, and restrictions on residential electricity consumption. Third, engineering and non-engineering measures should be integrated, with priority given to ensuring the proper functioning of social and livelihood systems. Unlike previous studies, we focus on the specificity of ALPF events, with a greater emphasis on a case-driven approach, shifting from the engineering and technical tendencies of previous studies to a comprehensive dimension. More consideration is given to the dynamic nature of risk.