We develop an energy management system (EMS) for artificial intelligence (AI) data centers with colocated renewable generation. Under a profit-maximizing framework, the EMS of renewable-colocated data center (RCDC) co-optimizes AI workload scheduling, on-site renewable utilization, and electricity market participation. Within both wholesale and retail market participation models, the economic benefit of the RCDC operation is maximized. Empirical evaluations using real-world traces of electricity prices, data center power consumption, and renewable generation demonstrate significant profit gains from renewable and AI data center colocations.

We study the optimal green hydrogen production and energy market participation of a renewable-colocated hydrogen producer (RCHP) that utilizes onsite renewable generation for both hydrogen production and grid services. Under deterministic and stochastic profit-maximization frameworks, we analyze RCHP's multiple market participation models and derive closed-form optimal scheduling policies that dynamically allocate renewable energy to hydrogen production and electricity export to the wholesale market. Analytical characterizations of the RCHP's operating profit and the optimal sizing of renewable and electrolyzer capacities are obtained. We use real-time renewable generation and electricity price data from three independent system operators to evaluate the impacts of market prices and environmental policies on RCHP's profitability.

Distributed energy resource aggregators (DERAs) must share the distribution network together with the distribution utility in order to participate in the wholesale electricity markets that are operated by independent system operators (ISOs). We propose a forward auction that a distribution system operator (DSO) can utilize to allocate distribution network access limits to DERAs. As long as the DERAs operate within their acquired limits, these limits define operating envelopes that guarantee distribution network security, thus defining a mechanism that requires no real-time intervention from the DSOs for DERAs to participate in the wholesale markets. Our auctions take the form of robust and risk-sensitive markets with bids/offers from DERAs and utility's operational costs. Properties of the proposed auction, e.g., resulting surpluses of DSO and the DERAs, and the auction prices, along with empirical performance studies, are presented.

We address the issue of equitable energy access within an energy community consisting of members with diverse socioeconomic backgrounds, including varying income levels and differing capacities to access distributed energy resources such as solar power and storage systems. While optimal energy consumption scheduling is well-studied, integrating equity into decentralized real-time energy access remains under-explored. This paper formulates Equity-regarding Welfare Maximization (EqWM)-a welfare optimization energy scheduling subject to equity constraints. We further develop a decentralized implementation (D-EqWM) as a bi-level optimization, where a non-profit operator designs a community pricing policy aimed at maximizing overall welfare, subject to constraints that ensure equitable access. Community members, in turn, optimize their individual consumption based on these prices. We present the optimal pricing policy along with its key properties.

The problem of the large-scale aggregation of the behind-the-meter demand and generation resources by a distributed-energy-resource aggregator (DERA) is considered. As a profit-seeking wholesale market participant, a DERA maximizes its profit while providing competitive services to its customers with higher consumer/prosumer surpluses than those offered by the distribution utilities or community choice aggregators. A constrained profit maximization program for aggregating behind-the-meter generation and consumption resources is formulated, from which payment functions for the behind-the-meter consumptions and generations are derived. Also obtained are DERA's bid and offer curves for its participation in the wholesale energy market and the optimal schedule of behind-the-meter resources. It is shown that the proposed DERA's aggregation model can achieve market efficiency equivalent to that when its customers participate individually directly in the wholesale market.

Current plans to decarbonize the electric supply system imply that the generation from wind and solar sources will grow substantially. This growth will increase the uncertainty of system operations due to the inherent variability of these renewable sources, and as a result, more reserve capacity will be required to provide the ramping (flexibility) needed for reliable operations. This paper assumes that all of the increased uncertainty comes from wind farms on the grid, and it shows how distributed storage managed locally by aggregators can provide the ramping needed without introducing a separate market for flexibility. This can be accomplished when the aggregators minimize the expected daily cost of the energy purchased from the grid for their customers by submitting optimal bids into the wholesale market with high and low price thresholds for discharging and charging the storage. This model is illustrated using a stochastic multi-period security constrained optimal power flow together with realistic data for a reduction of the network in the Northeast Power Coordinating Council region of the United States. The results show that the bidding strategy for distributed storage provides ramping to the grid just as effectively as storage managed by a system operator.

The objective of this article is to analyze the system benefits of distributed storage at different locations on a grid that has a high penetration of renewable generation. The chosen type of distributed storage modeled is deferrable demand (e.g., thermal storage) because it is relatively inexpensive to install compared to batteries and could potentially form a large component of the peak system load. The advantage of owning deferrable demand is that the purchase of energy from the grid can be decoupled from the delivery of an energy service to customers. Consequently, these customers can reduce costs by shifting their purchases from expensive peak periods to off-peak periods when electricity prices are low. In addition, deferrable demand can provide ramping services to the grid to mitigate the uncertainty of renewable generation. The primary economic issue addressed in this paper is to determine how the storage capacity is allocated between shifting load and providing ramping services. The basic economic tradeoff is between the benefit from shifting more load from peak periods to less expensive periods, and reserving some storage capacity for ramping to reduce the amount of conventional reserve capacity purchased. Our approach uses a new form of stochastic, multi-period Security Constrained Optimal Power Flow (SCOPF) that minimizes the expected system costs for energy and ancillary services over a 24-hour horizon. For each hour, five different levels of wind generation may be realized and these are treated as different system states with known probabilities of occurring. This model is applied to a reduction of the grid in New York State and New England and simulates the hourly load on a hot summer day, treating potential wind generation at different sites as stochastic inputs. The results determine the expected amount and location of conventional generating capacity dispatched, the reserve capacity committed to maintain operating reliability, the charging/discharging of storage capacity, and the amount of potential wind generation spilled. The results show there are major differences in how the deferrable demand at two large load centers, Boston and New York City, is managed, and we provide an explanation for these differences.

This paper builds on the results from our earlier research on the design of electricity markets that have to accommodate the uncertainty associated with high penetrations of renewable sources of energy. The key results show that 1) distributed storage (deferrable demand) is an effective way to reduce total system costs, 2) a simple market structure for energy allows aggregators to meet their customers' energy needs and provide ramping services to the system operator, and 3) using a receding-horizon optimization to dispatch units for the next market time-step benefits from the availability of more accurate forecasts of renewable generation and allows market participants to adjust their bids and offers in response to this new information. In our two-sided market, distributed storage in the form of deferrable demand is controlled locally by independent aggregators to minimize their expected payments for energy in the wholesale market, subject to meeting the energy needs of their customers. In addition, these aggregators are responsible for maintaining a stable power factor by installing local capabilities that automatically deal with local power imbalances. Failure to do this triggers penalties paid to the system operator.Our earlier results have shown that it is optimal for an aggregator to submit demand bids into a day-ahead market that include threshold prices for charging and discharging storage and also ensure that the expected amount of stored energy is consistent with the capacity limits of their storage. Because departures from the expected daily pattern of renewable generation are generally persistent (highly positive serial correlated), it is likely that the system operator determines an optimum pattern of demand for the aggregator that violates the capacity limits of storage by the end of the 24-hour period. If the market uses a receding horizon, the results in this paper show that aggregators can modify their bids to ensure that the capacity limits of storage are never violated in the next market time-step.In an empirical application, a stochastic form of multi-period security constrained unit commitment with optimal power flow (the MATPOWER Optimal Scheduling Tool, MOST) using a receding-horizon optimization determines the optimum dispatch and reserves for the next hour and forecasts of the nodal prices for the next 24 hours. The results show that locally controlled deferrable demand is almost as effective as centrally controlled deferrable demand as a way to reduce system costs and mitigate the variability of renewable generation. The additional advantage from using a receding horizon is that the system operator always charges/discharges the storage managed locally by aggregators within the capacity constraints of the storage.

Many connections between economic efficiency, regulation, the environment and energy markets are evident in the planning for transmission upgrades in an electricity network. Transmission owners have to make decisions about investing in new assets while facing uncertainty in the generation plans, regulatory and environmental constraints, and current system endowments. In this paper, we demonstrate an analytical method for determining the economic value of individual transmission lines in a meshed network by calculating the total welfare effects for the system. While many regulators believe that traditional congestion rents provide the correct incentives for investing in transmission upgrades, we show that the uncertainty in system conditions breaks down this paradigm. The analysis uses an existing Security Constrained Optimal Power Flow (SCOPF) model and a test network to demonstrate how the method can be used to determine the welfare effects of changing the capacity of selected transmission lines. The results show that a substantial portion of the economic benefits for an individual line may come from maintaining system reliability when equipment failures occur. Furthermore, these benefits can change dramatically when inherently intermittent sources of renewable generation are added to a network, and the changes in benefits are not captured effectively by changes in the expected congestion rents.

Our previous research has shown that distributed storage capacity at load centers (e.g. deferrable demand) can lower total system costs by smoothing out and flattening the daily dispatch profile of conventional generating units. The main savings in cost come from the price arbitrage caused by shifting load and from the reduction in the amount of installed conventional generating capacity needed to maintain operating reliability and generation adequacy at the peak system load. However, the full capacity of deferrable demand will not be used to reduce the peak system load whenever the price arbitrage between the peak and off-peak periods is too small to cover the round-trip inefficiency of the storage. If this situation occurs on the peak load day, the outcome is inefficient. The reason is that the presence of deferrable demand makes the peak load endogenous. Since system operators determine the optimal dispatch by minimizing the expected operating costs, they implicitly ignore the potential savings in capital costs associated with reducing the peak system load. This paper presents a mechanism for augmenting the nodal prices during peak load periods to reflect the capital cost of a peaking unit that we call "Peak-System-Load" (PSL) pricing. The first objective is to show that PSL pricing can reduce the total system costs and increase efficiency. However, the relative unpredictability of wind generation makes it harder to identify the timing of the peak net-generation accurately. Since PSL pricing also implies that customers pay higher wholesale prices during peak-load periods, the second objective is to show that paying this extra revenue to generators reduces the amount of missing money caused by the lower wholesale prices associated with generating more from renewable sources. In this sense, PSL pricing may lead to a viable energy-only market. An empirical application illustrates our proposed mechanism using a stochastic form of multi-period Security Constrained Optimal Power Flow (the mops model) and a reduction of the Northeast Power Coordinating Council (NPCC) network to simulate operations on representative days.

With high penetrations of variable generation from wind turbines in remote locations, transmission capacity may be inadequate to transfer this relatively inexpensive source of generation to demand centers. The major reason is that transmission corridors into load centers are often congested when the system load is high, and additional wind generation is effectively shut out. In contrast, when the system load is low and the wind is blowing, wind generation may be able to meet most of the load throughout the network subject to the specific limitations of the network's topology. This paper compares the system costs of two very different ways of reducing congestion on the network to increase the annual amount of potential wind capacity dispatched. The first way uses the standard supply side solution of upgrading transmission capacity on the network. The second way uses a demand-side approach in which deferrable demand shifts the system load from on-peak periods to off-peak periods. In addition, the deferrable demand can be used to offset the inherent variability of wind generation and reduce the amount of reserve generating capacity needed to maintain operating reliability. In fact, reducing the total amount of conventional generating capacity needed to maintain system adequacy for a given amount of installed wind capacity is a major source of cost savings with deferrable demand. The simulation is based on a multiperiod (24h), stochastic, security constrained optimal power flow (SCOPF) and a reduction of the Northeastern Power Coordinating Council (NPCC) network. This framework includes stochastic forecasts of potential wind generation at multiple sites as inputs as well as deferrable demand (e.g.,thermal storage) at different load centers. It determines the optimum patterns of dispatch, reserves, and ramping to maintain reliability over a set of credible contingencies. The results demonstrate that deferrable demand can effectively (1)lower the average wholesale prices for energy;, (2)reduce the installed generating capacity needed to maintain system adequacy; and (3)mitigate the ramping costs associated with wind variability. With a sufficient amount of deferrable demand, the typical daily pattern of load can be flattened, and all the variability of wind generation can be mitigated. The overall conclusion is that deferrable demand reduces the total annual cost of the conventional system substantially more than upgrading transmission capacity, and it is an effective alternative to the standard supply side solution.

The primary purpose of this paper is to evaluate the benefits of distributed storage capacity in the form of deferrable demand managed centrally by a system operator, and in particular, to determine the savings in the total annual cost of supplying electricity for a system that has a substantial amount of variable generation from wind turbines. Since the objective of a centrally controlled system is to minimize the expected daily operating costs subject to the availability of generating units and storage capacity, the basic economic question is whether the savings in the annual system cost of supply, including the capital cost of installed generating capacity, can offset the capital cost of installing deferrable demand capacity. The analysis uses a new multi-period model of a power grid that treats stochastic generation explicitly and determines the optimum hourly commitment of conventional generators and the charging/discharging of deferrable demand needed to maintain the reliability of supply. A simulation example shows that deferrable demand can reduce system costs by (1) shifting demand from expensive peak periods to less expensive off-peak periods, (2) providing ramping services to mitigate the variability of wind generation, and (3) reducing the amount of installed peaking capacity needed for System Adequacy and the associated capital costs. If customers pay rates for electricity that reflect the true system costs of supplying their patterns of purchases from the grid, customers with deferrable demand will pay lower bills for electricity and their savings will be substantially more than the cost of installing deferrable demand devices. The results also show that if customers pay typical flat rates for electric energy, the economic incentives for installing deferrable demand are perverse.

The authors employed a novel optimization framework coupled with an econometric model of wind to study market performance. One conclusion: in circumstances with high uncertainty in the market – as with high penetration of renewables – relying more on a real-time market, similar to the National Electricity Market in Australia, may be a better way to deal with this uncertainty because it uses updated and more accurate information about the wind variability.

Our previous research has shown that distributed storage capacity at load centers (e.g. Deferrable demand) controlled by a system operator can lower total system costs by smoothing out and flattening the daily dispatch profile of conventional generating units and providing ramping services. Since it is in reality impractical for system operators to control large numbers of customers with deferrable demand directly, aggregators will in all likelihood be responsible for managing the individual sources of deferrable demand using instructions provided by a system operator. The objective of this paper is to compare the performance of deferrable demand when 1) the aggregators act as clients to the system operator and receive physical charge/discharge instructions for managing deferrable demand (i.e. Centralized control), with 2) the aggregators follow their own interests and submit bids for purchasing energy into the wholesale auction using projected prices provided by the system operator (i.e. Hierarchical control). The analysis uses a stochastic form of multi-period Security Constrained Optimal Power Flow (SCOPF) in a simulation using a reduction of the Northeast Power Coordinating Council (NPCC) network for representative days. This model treats potential wind generation and load as stochastic inputs and determines the optimum daily profiles of dispatch and demand for different realizations of hourly wind generation and load. Ramping capacity is acquired to ensure that transitions from the realizations in one hour to the next hour, as well as contingencies, can be supported. The results show that if aggregators receive stochastic forecasts of energy prices for the next 24 hours, their optimum strategy for minimizing the expected cost of their purchases from the grid is to determine a high threshold price for discharging and a low threshold price for charging, and as a result, they provide ramping services as well as benefitting from day/night price arbitrage. However, the results are sensitive to the form of the price forecasts.

With more electricity generated from renewable sources, the importance of effective storage capacity is increasing due to its capability to mitigate the inherent variability of these sources, such as wind and solar power. However, the cost of dedicated storage is high and all customers eventually have to pay. Deferrable demand offers an alternative form of storage that is potentially less expensive because the capital cost is shared between providing an energy service and supporting the grid. This paper presents an empirical analysis to illustrate the beneficial effects of Plug-in Hybrid Electric Vehicles (PHEV) and thermal storage on the total system cost using data for a hot summer day in New York City. The analysis shows how customers can reduce total system costs and their bills by 1) shifting load from expensive peak periods to less expensive off-peak periods, 2) reducing the amount of installed conventional generating capacity needed to maintain System Adequacy, and 3) providing ramping services to mitigate the variability of generation from renewable sources. Moreover, this paper demonstrates economic benefits of different types of customers with different deferrable demand capabilities under two bill payment policies, flat price payment and optimum price payment, and it finally shows how long it takes for customers to fully pay back their initial capital costs of PHEV or thermal storage under two different policies.

The goals of this paper are to 1) simulate the ex-ante riskiness of purchasing a TCC, and 2) evaluate the efficiency of the TCC market in New York State to determine if there is evidence of under-pricing. Three VAR models are estimated using only market data available before the auction. This model is then used to simulate the daily payouts of a TCC for the following summer. A Monte Carlo procedure simulates the daily summer temperatures, the levels of quantity demanded and prices over the summer months. The main empirical result is that the market price paid for the most important TCC, in terms of volume, (the Hudson Valley to New York City) is higher than the mean of the simulated payouts even though the actual payout was higher than the market price. The market prices for the other two TCCs are lower than the means of the simulated payouts, and as a result, there is no consistent evidence of under-pricing in this analysis of the market for six-month TCCs in the summer of 2006. (c) 2014 Elsevier B.V. All rights reserved.

Provides an overview of the technical articles and features presented in this minitrack.