We observe suitably located energy storage systems are able to collect significant revenue through spatiotemporal arbitrage in congested transmission networks. However, transmission capacity expansion can significantly reduce or eliminate this source of revenue. Investment decisions by merchant storage operators must, therefore, account for the consequences of potential investments in transmission capacity by central planners. This paper presents a tri-level model to co-optimize merchant electrochemical storage siting and sizing with centralized transmission expansion planning. The upper level takes the merchant storage owner's perspective and aims to maximize the lifetime profits of the storage, while ensuring a given rate of return on investments. The middle level optimizes centralized decisions about transmission expansion. The lower level simulates market clearing. The proposed model is recast as a bi-level equivalent, which is solved using the column-and-constraint generation technique. A case study based on a 240-bus, 448-line testbed of the Western Electricity Coordinating Council interconnection demonstrates the usefulness of the proposed tri-level model.
Grid resilience is a concept related to a power system's ability to continue operating and delivering power even in the event that low probability, high-consequence disruptions such as hurricanes, earthquakes, and cyber-attacks occur. Grid resilience objectives focus on managing and, ideally, minimizing potential consequences that occur as a result of these disruptions. Currently, no formal grid resilience definitions, metrics, or analysis methods have been universally accepted. This document describes an effort to develop and describe grid resilience metrics and analysis methods. The metrics and methods described herein extend upon the Resilience Analysis Process (RAP) developed by Watson et al. for the 2015 Quadrennial Energy Review. The extension allows for both outputs from system models and for historical data to serve as the basis for creating grid resilience metrics and informing grid resilience planning and response decision-making. This document describes the grid resilience metrics and analysis methods. Demonstration of the metrics and methods is shown through a set of illustrative use cases.
FERC order 755 and FERC order 784 provide pay-for-performance requirements and direct utilities and independent system operators to consider speed and accuracy when purchasing frequency regulation. Independent System Operators (ISOs) have differing implementations of pay-for-performance. This paper focuses on the PJM implementation. PJM is a regional transmission organization in the northeastern United States that serves 13 states and the District of Columbia. PJM's implementation employs a two part payment based on the Regulation Market Capability Clearing price (RMCCP) and the Regulation Market Performance Clearing Price (RMPCP). The performance credit includes a mileage ratio. Both the RMCCP and RMPCP employ an actual performance score. Using the PJM remuneration model, this paper outlines the calculations required to estimate the maximum potential revenue from participation in arbitrage and regulation in day-ahead markets using linear programming. Historical PJM data from 2014 and 2015 was then used to evaluate the maximum potential revenue from a 5 MWh, 20 MW system based on the Beacon Power Hazle Township flywheel plant. Finally, a heuristic trading algorithm that does not require perfect foresight was evaluated against the results of the optimization algorithm.
Energy storage (ES) is a pivotal technology for dealing with the challenges caused by the integration of renewable energy sources. It is expected that a decrease in the capital cost of storage will eventually spur the deployment of large amounts of ES. These devices will provide transmission services, such as spatiotemporal energy arbitrage, i.e., storing surplus energy from intermittent renewable sources for later use by loads while reducing the congestion in the transmission network. This paper proposes a bilevel program that determines the optimal location and size of storage devices to perform this spatiotemporal energy arbitrage. This method aims to simultaneously reduce the system-wide operating cost and the cost of investments in ES while ensuring that merchant storage devices collect sufficient profits to fully recover their investment cost. Finally, the usefulness of the proposed method is illustrated using a representative case study of the ISO New England system with a prospective wind generation portfolio.
The SunShot Initiative is focused on reducing cost to improve competitiveness with respect to other electricity generation options. The goal of the Sandia Transmission Grid Integration (TGI) program is to reduce grid access barriers for solar generation. Sandia’s three-year TGI work was divided into five objectives.
This paper outlines the calculations required to estimate the maximum potential revenue from participation in arbitrage and regulation in day-ahead markets using linear programming. Then, we use historical Electricity Reliability Council of Texas (ERCOT) data from 2011-2013 to evaluate the maximum potential revenue from a hypothetical 32 MWh, 8 MW system. We investigate the maximum potential revenue from two different scenarios: arbitrage only and arbitrage combined with regulation. This analysis was performed for each load zone over the same period to show the impact of location and to identify trends in the opportunities for energy storage. Our analysis shows that, with perfect foresight, participation in the regulation market would have produced more than twice the revenue compared to arbitrage in the ERCOT market in 2011-2013. Over the last three years, there has been a significant decrease in the potential revenue for an energy storage system. We also quantify the impact of location on potential revenue.
We propose a mathematical programming-based approach to optimize the security-constrained unit commitment problem with a full AC transmission network representation. Our approach is based on our previously introduced successive linear programming (SLP) approach to solving the non-linear, nonconvex AC optimal power flow (ACOPF) problem. By linearizing the ACOPF, we are able to leverage powerful commercial mixed-integer solvers to iteratively optimize the combined unit commitment plus ACOPF model. We demonstrate our approach on six-bus, IEEE RTS-96, and IEEE 118-bus test systems. We perform a comparative analysis of the relative impacts of singlebus, DC, and AC transmission network models on the unit commitment and dispatch solutions and their associated costs.
This report summarizes findings and results of the Quantifiably Secure Power Grid Operation, Management, and Evolution LDRD. The focus of the LDRD was to develop decisionsupport technologies to enable rational and quantifiable risk management for two key grid operational timescales: scheduling (day-ahead) and planning (month-to-year-ahead). Risk or resiliency metrics are foundational in this effort. The 2003 Northeast Blackout investigative report stressed the criticality of enforceable metrics for system resiliency the grids ability to satisfy demands subject to perturbation. However, we neither have well-defined risk metrics for addressing the pervasive uncertainties in a renewable energy era, nor decision-support tools for their enforcement, which severely impacts efforts to rationally improve grid security. For day-ahead unit commitment, decision-support tools must account for topological security constraints, loss-of-load (economic) costs, and supply and demand variability especially given high renewables penetration. For long-term planning, transmission and generation expansion must ensure realized demand is satisfied for various projected technological, climate, and growth scenarios. The decision-support tools investigated in this project paid particular attention to tailoriented risk metrics for explicitly addressing high-consequence events. Historically, decisionsupport tools for the grid consider expected cost minimization, largely ignoring risk and instead penalizing loss-of-load through artificial parameters. The technical focus of this work was the development of scalable solvers for enforcing risk metrics. Advanced stochastic programming solvers were developed to address generation and transmission expansion and unit commitment, minimizing cost subject to pre-specified risk thresholds. Particular attention was paid to renewables where security critically depends on production and demand prediction accuracy. To address this concern, powerful filtering techniques for spatio-temporal measurement assimilation were used to develop short-term predictive stochastic models. To achieve uncertaintytolerant solutions, very large numbers of scenarios must be simultaneously considered. One focus of this work was investigating ways of reasonably reducing this number.