Resilient operation of power systems under extreme weather events

Abstract

Extreme weather events may lead to contingencies in power grids and result in significant loss for electricity customers. Although power grids are becoming more reliable, damage of electric components and interruption of electricity supply is still inevitable. Therefore, resilient operation of power grids, which can reduce the influence of weather-related contingencies, is essential for power systems under extreme weather events. In this thesis, a series of approaches are developed to address issues related to the resilient operation of power grids against extreme weather events.

Firstly, since extreme weather mainly affects power systems by increasing the outage rate of components exposed to extreme weather, a weather-related outage rate estimation method is proposed. The proposed method focusses on the outages of overhead transmission lines under wind storms, but the key idea is applicable for other components under different weather conditions. The proposed method contributes to calculating the probability of possible weather-related contingency states, which is the foundation of the following work.

Pre-event risk assessment is an important basis for resilient operation, but conventional risk assessment methods usually require long computational time and may delay operation decision making. Hence, an operational risk assessment method is proposed to quickly obtain the risk level of power systems at different stages of extreme weather events. The proposed method is tested using a large-scale power transmission system to show its feasibility, accuracy and efficiency.

With the rapid development of internet services, internet data centers (IDCs) are becoming the new critical loads in some power distribution systems. Resilient operation, especially preventive operation such as pre-event restoration resource allocation, is important to reduce the utility loss of IDCs under extreme weather events. However, since the utility of IDCs is different from the conventional critical loads and is determined by the emergency operation of IDCs, existing resilient operation techniques may not be directly applicable. Therefore, a pre-event restoration resource allocation method is proposed for distribution systems with critical IDCs to minimize the expected IDC utility loss in possible contingency states. For each contingency state, the IDC utility loss is obtained by co-optimizing the post-event restoration of distribution systems and emergency operation of IDCs. An ADMM based strategy is developed to solve the proposed method.

In reality, the co-optimization of post-event restoration of distribution systems and emergency operation of IDCs may be unachievable since the co-optimization requires exchange of confidential information from IDCs and distribution systems. Therefore, an advanced post-event restoration method of distribution systems with critical IDCs is proposed. In this advanced method, the original co-optimization model is replaced by several sequential subproblems operated by IDCs or distribution systems independently, so that not only the need for confidential information exchange is obviated but also the computational time is reduced.

In summary, this thesis focuses on the resilient operation of power grids under extreme weather events. Deficiencies of existing techniques are addressed, and related approaches are proposed to address these gaps in the literature.