In the last two decades, the number of strategies for planning the maintenance of power systems have increased considerably. As societal dependence on power system infrastructure continues to grow, there is an increased need to identify the best practices in the field of power system maintenance planning to ensure the continued reliable operation of the grid. This paper presents a comprehensive and systematic state-of-the-art review of advances in power system maintenance and the significance for the field of power system reliability. As the main contributions, this paper systematically organizes the published literature, and analyses the most relevant milestones in the context of power systems adequacy and security enhancement, producing a taxonomy of the different maintenance strategies. This includes detailing existing approaches for power system maintenance planning, and providing clear definitions, models, methods, and characteristics of maintenance policy. The review also includes the most relevant standards employed in the power system maintenance field. Finally, areas requiring further research are identified alongside emerging trends in power system maintenance, to inform industry practice and support further research.

Wildfire smoke and other particulate matter can substantially inhibit solar photovoltaic (PV) generation production. While solar PV facilities may not be located in areas with a high fire risk, smoke from wildfires can travel hundreds of kilometers impacting a large number of facilities. This paper proposes a geospatial wildfire PV capacity model to quantify the anticipated temporal reduction in PV capacity due to wildfire smoke. A case study using data for two time periods from the 2020 California wildfires and real utility scale solar generation data evidences the model's high accuracy. Results argue that wildfire smoke can cause significant temporal solar generation capacity reductions over wide geographic regions. Application of the proposed model to inform power system resiliency planning is demonstrated for two use cases: generation scheduling and siting. With meteorological service providers beginning to release smoke forecasts, our geospatial wildfire PV capacity model enables balancing authorities to make use of this information to proactively schedule generation to compensate for reductions in PV capacity. The trained model also produces geospatial derate maps that can enable generation developers to consider historical capacity derates due to smoke when making siting or planning decisions.

Electric vehicle growth is creating a larger share of mobile load in power systems. This can affect system stress in potentially constrained areas of the grid, during a period that may already be an emergency. This paper proposes a novel model to reflect the behavior and spatiotemporal charging demand of electric vehicles during a natural disaster evacuation. The paper integrates the model in an algorithmic approach enabling operators to anticipate grid impacts from the charging of electric vehicles during a wildfire and subsequent evacuation. Geospatial visualization of evacuation charging demand provides a means to enhance operator awareness of spatiotemporal impacts and indicate critical areas of the grid for proactive mitigation. Results for a case study using the RTS-GMLC test system suggest that electric vehicle evacuation can result in dynamic changes to load patterns, adding stress to network components with the most significant load changes occurring during the initial evacuation period.

Extreme weather events can cause significant damage to power distribution network infrastructure, often resulting in power outages. Distribution Network Operators (DNOs) are faced with the challenging task of responding to these outages in real time while maintaining a resilient grid. Our paper presents an innovative approach to alert operators about the potential risk associated with upcoming extreme weather through a normalized fragility curve. The uniqueness of the curve is the ability to capture regional differences across a DNO's territory while presenting operators with a means of setting unified risk thresholds. This can support a proactive response and allow the staging of necessary resources to minimize the threat posed by such events. Our approach captures the changes in failure probability associated with differing wind regimes and demonstrates the benefit of sub‐regional meteorological information. The proposed approach is demonstrated for wind events using 20 years of historical fault records from a DNO in the United Kingdom (UK). While its efficacy is demonstrated for windstorms in the UK, the approach could be applied globally to develop normalized fragility curves for other types of seasonal extreme weather events such as snowstorms, hurricanes, or linked hazards such as wildfires. The approach can also facilitate an understanding of how infrastructure may operate under future climate conditions, supporting proactive adaptation.

The use of Remedial Action Scheme(s) (RAS) is an economic means to facilitate the interconnection of renewable generation while supporting the reliable operation of the transmission system. With the increasing prevalence of RAS on the system, automatic modeling is imperative in order to accurately capture the potential impacts and interaction between these schemes. Within the Western United States, the Western Electricity Coordinating Council (WECC) has been facilitating implementation of RAS modeling in power flow simulation via incorporation of these models into software packages. Beyond WECC's effort, which largely centers on modeling for simulations in the steady state timeframe, is the need for automatic modeling of RAS in the dynamic timeframe. Through implementation of both an event-based and parameter- based RAS, analysis was performed to demonstrate the benefits such modeling can have on simulation accuracy and regional coordination.