Marina González Vayá

A framework for and assessment of demand response and energy storage in power systems

The shift in the electricity industry from regulated monopolies to competitive markets as well as the widespread introduction of fluctuating renewable energy sources bring new challenges to power systems. Some of these challenges can be mitigated by using demand response (DR) and energy storage to provide power system services. The aim of this paper is to provide a unified framework that allows us to assess different types of DR and energy storage resources and determine which resources are best suited to which services. We focus on four resources: batteries, plug-in electric vehicles, commercial buildings, and thermostatically controlled loads. We define generic power system services in order to assess the resources. The contribution of the paper is threefold: (i) the development of a framework for assessing DR and energy storage resources; (ii) a detailed analysis of the four resources in terms of ability for providing power system services, and (iii) a comparison of the resources, including an example case for Switzerland. We find that the ability of resources to provide power system services varies largely and also depends on the implementation scenario. Generally, there is large potential to use DR and energy storage for providing power system services, but there are also challenges to be addressed, for example, adequate compensation, privacy, guaranteeing costumer service, etc.

Uncertainty in the flexibility of aggregations of demand response resources

Aggregations of demand response resources can provide a variety of services to the power grid. To utilize them effectively, for both planning problems and real-time control, we must estimate their flexibility. However, flexibility estimates are uncertain because of issues such as model error and forecasting error. In this paper, we present a model of uncertain flexibility and describe the many causes of uncertainty. We conduct two case studies, one for electric vehicle aggregations and one for air conditioner aggregations, in order to show specific examples that illustrate the causes and magnitude of uncertainty. We find that uncertainty can be very large for small load aggregations, when models do not capture enough of the underlying dynamics, or when forecasts of other quantities, such as ambient conditions, are bad. Although the focus of the paper is on understanding uncertainty, we also briefly discuss how one might use knowledge of uncertainty distributions in planning problems to derive closer-to-optimal results.

Self Scheduling of Plug-In Electric Vehicle Aggregator to Provide Balancing Services for Wind Power

This paper focuses on the self-scheduling problem of an aggregator of plug-in electric vehicles (PEVs) purchasing energy in the day-ahead market, and offering balancing services for a wind power producer, i.e., committing to compensate the forecast errors of wind power plants. The aggregated charging and discharging flexibility of the PEV fleet is represented by a probabilistic virtual battery model, accounting for the uncertainty in the driving patterns of PEVs. Another source of uncertainty is related to the balancing requests, which are a function of the forecasted wind power output. A scenario-based robust approach is used to tackle both sources of uncertainty in a tractable way. The interdependency between the day-ahead market prices and the aggregators bidding decisions is addressed using complementarity models. A case study analyzes the capability of the PEV aggregation to provide balancing services, for different settings of the balancing contract, and both with and without the use of vehicle-to-grid.

Integrating renewable energy forecast uncertainty in smart-charging approaches for plug-in electric vehicles

Both an increasing share of intermittent renewable energies and an introduction of plug-in electric vehicles (PEVs) are challenging for the electric power system. Nevertheless, PEVs could be used as distributed storage resources to help integrate fluctuating energy sources into the power system. In this paper we analyze the case where PEV batteries are used to compensate the forecast error of a wind power plant. We introduce a day-ahead charging scheduling strategy that minimizes system generation costs, enforces network and PEV end-use constraints, and at the same time enables the fleet to compensate deviations of wind power output from its day-ahead forecast. For this purpose, a probabilistic wind power forecast model is integrated into an Optimal Power Flow based smart-charging scheme. The fleet is modeled as a set of virtual storages whose characteristics depend on individual driving patterns. Results show that with the proposed scheme enough charging flexibility is made available to compensate the forecast error of a wind power plant. However, there is a trade-off between charging flexibility and cost-minimization.

On the interdependence of intelligent charging approaches for plug-in electric vehicles in transmission and distribution networks

This paper proposes a novel approach to assess the impact of different plug-in electric vehicle charging strategies on power generation, transmission and distribution. The proposed impact assessment tool integrates simulations of transport behavior as well as power generation, transmission and distribution. The detailed temporal and spatial information of vehicle behavior obtained from the transport simulation is mapped to the transmission and distribution network models to perform a holistic assessment of the impact of different charging strategies. The charging strategies considered are uncontrolled charging and two types of smart charging schemes, one centralized the other decentralized. The advantage of this integrated approach is that both local and system-wide aspects can be analyzed simultaneously, uncovering effects that would otherwise remain unnoticed.

Combined smart-charging and frequency regulation for fleets of plug-in electric vehicles

In this paper, a co-optimization of smart-charging and frequency regulation for a plug-in hybrid electric vehicle aggregator is proposed. To this end, a day-ahead charging scheduling algorithm that minimizes costs and avoids network overloading while leaving enough flexibility to provide regulation is described. Then, a decentralized approach for the real-time dispatch of the regulation signal is introduced, based on the broadcast of a probability signal. The results show that the aggregated vehicle fleet can contribute significantly to regulation reserves with high response accuracy.

On the price of being selfish in large populations of plug-in electric vehicles

We consider the problem of optimally scheduling the flexible electricity demand of a fleet of plug-in electric vehicles (PEVs). More specifically, we analyze the solutions of the following charging optimization problems: the welfare-optimal problem, where the overall system cost is minimized; the fleet-optimal problem, where the charging cost of the fleet as a whole is minimized by a central agent, that is the PEV aggregator; the selfish-optimal problem, where the noncooperative PEVs aim at minimizing their individual charging cost. For a homogenous PEV fleet and a simplified problem setup, we show that the solutions of the three different approaches correspond to different valley-filling results. A main insight is that, as the population of PEVs grows, the selfish-optimal solution converges to the welfare-optimal solution. On the other hand, we show that the centralized fleet-optimal solution of the PEV aggregation can be recovered via decentralized selfish-optimal solutions with respect to an appropriate price signal as the population size grows. Finally, we demonstrate our technical results on a realistic PEV fleet case study.

Optimal bidding of plug-in electric vehicles in a market-based control setup

This paper presents a market-based control approach to minimize the charging costs of plug-in electric vehicles (PEVs) without impacting their end-use. In the proposed framework, vehicles are modeled as agents actively placing bids to purchase electricity in the spot market. Their bids depend on status variables which represent the urgency to charge. To ensure scalability, the bids of large numbers of PEVs are aggregated. Then, they are cleared with the remaining market supply and demand bids. In this paper we focus on determining the optimal bidding strategies of individual PEVs, taking into account the uncertainty related to market bids and to driving behavior. For this purpose, we formulate a learning process based on Q-learning, where each vehicle adapts its bidding strategy over time according to the market outcomes. We perform simulations with historical market bid data, and realistic vehicle driving patterns from an agent-based transport simulation. Results show that the costs of charging can be significantly reduced compared with an uncontrolled charging approach. Moreover, we compare the results with those of a centralized aggregator-based approach, where an aggregator directly manages charging and purchases electricity on behalf of PEVs on the spot market. We show that the results of the decentralized market-based control approach are just slightly higher than those of the centralized approach.

Integration of PEVs into Power Markets: A Bidding Strategy for a Fleet Aggregator

With a large-scale introduction of plug-in electric vehicles (PEVs), a new entity, the PEV fleet aggregator, is expected to be responsible for managing the charging of, and for purchasing electricity for, the vehicles. This book chapter deals with the problem of an aggregator bidding into the day-ahead electricity market with the objective of minimizing charging costs while satisfying the PEVs’ flexible demand. The aggregator is assumed to potentially influence market prices, in contrast to what is commonly found in the literature. Specifically, the bidding strategy of the aggregator is formulated as a bi-level problem, which is implemented as a mixed-integer linear program. The upper-level problem represents the charging cost minimization of the aggregator, whereas the lower-level problem represents the market clearing. An aggregated representation of the PEV end-use requirements as a virtual battery, with time varying power and energy constraints, is proposed. This aggregated representation is derived from individual driving patterns. Since the bids of other market participants are not known to the aggregator ex ante, a stochastic approach is proposed, using scenarios based on historical data to describe such uncertain bids. The output of the proposed approach is a set of bidding curves, one for each hour of the day. Results show that by using PEV demand flexibility, the aggregator significantly reduces the charging cost. Additionally, the aggregator’s bidding strategy has an important impact on market prices.

Smart Charging of Plug-in Electric Vehicles Under Driving Behavior Uncertainty

An upcoming introduction of plug-in hybrid electric vehicles and electric vehicles could put power systems’ infrastructure under strain in the absence of charging control. The charging of electric vehicles could be managed centrally by a so-called aggregator, which would take advantage of the flexibility of these loads. To determine optimal charging profiles day-ahead, the aggregator needs information on vehicles’ driving behavior, such as departure and arrival time, parking location, and energy consumption, none of which can be perfectly forecasted. In this chapter, we introduce an approach to derive day-ahead charging profiles that minimize generation costs while respecting network and drivers’ end-use constraints, as well as taking into account the uncertainty in driving patterns. The charging profiles are derived by aggregating vehicles at each network node into virtual battery resources and dispatching them with a multiperiod optimal power flow (OPF). To take driving pattern uncertainty into consideration, different possible realizations of individual driving patterns are generated with a Monte Carlo simulation, modeling individual driving behavior with non-Markov chains. This information is integrated into the OPF, where constraints concerning the virtual batteries are modeled as chance constraints, i.e., as constraints that may be violated with a certain probability. Compared with a deterministic approach, this framework increases the chances of not violating the constraints subject to uncertainty.