Assessing structural resilience of renewable-dominated power networks : a theoretical framework based on stochastic processes and probability theory

Abstract

Global climate change is driving the energy infrastructure transition towards a clean and low-carbon paradigm. Conventional power systems are progressing towards renewable-dominated power systems. Renewable energy generation is predicted to dominate human energy consumption in the future. However, climate change-induced extreme events inevitably pose challenges to power systems. In the foreseeable future, the increasingly extreme events and the expanded system scale render renewable-dominated power systems more liable to suffer from extreme events, highlighting the significance of infrastructure and network structure security. Particularly, the strong correlation between renewable energy and climate conditions exacerbates the inherent volatility and uncertainty, further challenging the operational security of power systems. Moreover, mobile power sources play an increasingly vital role as external, flexible resources in power systems, and their deployment before extreme events emerges as a key challenge in enhancing grid security. Effectively assessing the power system’s ability to withstand extreme events and establishing a resilience assessment framework with wide adaptability become pivotal approaches to address these challenges. This thesis investigates these problems through the following research.

First, to quantify the impact of climate-related extreme events on power network structure, this thesis proposes an effective quantification method to address the structural uncertainty and power transmission uncertainty in power networks induced by extreme events using the point process and stochastic geometry. The proposed distance metric and power connection metric quantify the electrical connections and power connections between network nodes, whose probability distributions are derived to reflect the impact of structural uncertainty induced by extreme events. The point process is utilized to depict potential spatial patterns of power networks under extreme events, and the structural uncertainty of power networks is addressed by investigating its statistical characteristics using powerful tools from stochastic geometry.

Second, to reveal the effect of renewable power uncertainty on power network resilience, this thesis proposes a probability-based analytical approach to characterize wind power uncertainty and load survivability in renewable-dominated power networks under extreme events. The probability distributions of wind turbine output power are obtained in analytical forms. The operating states and potential failures of wind turbines during extreme events are considered. The wind power uncertainty and power network structural uncertainty are utilized to address the uncertainty in power received by loads under extreme events. Numerical characteristics and tractable approximations of metrics related to resilience are presented analytically to reveal the critical factors that influence the structural resilience of power networks.

Last but not least, to assess the performance of mobile emergency generator (MEG) deployment strategies in distribution networks, this thesis proposes a stochastic geometry-based method considering the structural uncertainty of distribution networks under extreme events. Based on the proposed distance metric, the point process and Voronoi tessellation are adopted to depict the spatial patterns of distribution networks and service coverage of MEGs under extreme events. A set of assessment metrics are developed to evaluate the survivability of loads and distribution network resilience, whose analytical expressions are obtained to explore the relationship between MEG deployments, structural features, and distribution network resilience.