The value of aggregators in local electricity markets: A game theory based comparative analysis

https://doi.org/10.1016/j.segan.2021.100498Get rights and content

Abstract

Demand aggregators are expected to have a key role in future electricity systems. More specifically, aggregators can facilitate the harnessing of consumers’ flexibility. This paper focuses on understanding the value of the aggregator in terms of aggregation of both flexibility and information. We consider the aggregation of flexibility as the ability to exercise a direct control over loads, while the aggregation of information refers to knowledge of the flexibility characteristics of the consumers. Several game theory formulations are used to model the interaction between the energy provider, consumers and the aggregator, each with a different information structure. We develop a potential game to obtain the Nash equilibrium of the non-cooperative game with complete information and we analyze the system dynamics of consumers using the adaptive expectations method in an incomplete information scenario. Several key insights about the value of aggregators are found. In particular, the value of the aggregator is mainly related to the aggregation of information rather than flexibility, and flexibility is valuable only when it can be coordinated. In this sense, prices are not enough to guarantee an effective coordination.

Introduction

In several countries an important political, social and environmental objective is the integration of a large amount of renewable energy sources [1]. In this new paradigm, several challenges emerge in both market design and system operations [2]. A source for these challenges is the volatile nature of several renewable energy sources, particularly the most widely developed (e.g., solar and wind). These sources impose new flexibility requirements for a reliable and economic operation [3]. Flexibility, defined as the ability of the system to cope with variations on generation and load [4], cannot only be provided by flexible generating units, e.g., hydro or gas, but also by the demand through Demand Side Management (DSM) [5], and Demand Response (DR) [6] schemes. According to [7], DR is a kind of DSM strategy that can be used to manage energy costs. In particular, DR is defined as a tariff or program that motivates changes in the electric usage by end-users, while DSM includes energy efficiency and/or load management programs.

Given the distributed and diverse nature of flexible loads, there has been extensive research regarding the impact of a new participant in charge of aggregating these loads, the so-called aggregator. The aggregator might play a key role in taking advantage of the demand flexibility and could be crucial in the integration of volatile energies resources in large-scale systems [8]. Several papers have studied the benefits of the aggregation of loads in several dimensions including: harnessing load flexibility, minimizing operations costs and reducing consumers’ tariffs. For example, in [6] and [9], the market effects of an aggregator which takes advantage of operational flexibility of loads are examined. In particular, the equilibrium of a wholesale market in a Cournot game setup, where the players are the aggregator, the Independent System Operator (ISO) and generators, is analyzed. In [8], the participation of an aggregator as a strategic agent in the wholesale market is studied. The revenue of the aggregator can be transferred to end-users with the objective of encouraging participation in demand response programs. In [10] a market model in which a set of aggregators act as intermediaries between the utility and the consumers is introduced. Aggregators compete to sell demand response services to the utility and provide compensation to consumers in order to modify their consumption patterns. In [11] a smart pricing policy that encourages EVs to participate in frequency regulation through aggregators is presented and shown to benefit both consumers and the grid. In [12] an offering strategy for a risk-averse aggregator that participates in the balancing market is proposed. As flexibility product, the market accepts asymmetric block offers, which are defined by duration and power effect of the response part, duration and power effect of the rebound part, and the recovery time. A fixed-term DR contract market is proposed in [13]. The Distribution System Operator (DSO) buys flexibility services to DR aggregators in a competitive market, introducing scheduled and conditional contracts. The paper determines the optimal service for the system, allowing DSOs and DR aggregators ensure their profitability. In [14] a local coordinating mechanism for strategic aggregators is proposed. The goal of the mechanism corresponds to maximize the aggregators’ utilities from their participation in the market, considering the physical limits of the distribution network.

In [15] a non-cooperative game is presented, in which aggregators compete between them to sell energy stored in consumers storage devices. In [16] a game theory model is proposed in order to find the equilibrium that maximizes the profit of aggregators and the distribution operator system (DSO). In [17] an offering strategy for wind power producers that includes DR aggregators is proposed. The problem is modeled using a bi-level programming approach, where the upper-level represents the producer’s problem, and the lower-level models the aggregator behavior. In [18] a review of the value of the aggregator under different technological and regulatory scenarios is performed. Some benefits of the aggregator are related to the economies of scale and scope, uncertainty management and market complexities. In [19] a confidential information problem between aggregators and the balancing group manager (BGM) is presented. The paper proposes an iterative distributed algorithm based on non-cooperative game theory to address the confidential information problem considering strategic aggregators. In [20] the competition between DR aggregators to provide DR services to the system operator is modeled as a non-cooperative game with incomplete information. The goal of the aggregators is to obtain their optimal bidding strategies in a deregulated energy market. In [21] a two-stage stochastic bi-level programming approach is proposed to solve the interaction problem between aggregators and electricity retailers in a competitive environment. The upper-level problem corresponds to maximize the retailer’s profit considering a two-stage stochastic programming scheme, while the lower-level problem corresponds to maximize the aggregator’s profit. In [22] a multi follower bi-level framework is presented to model the interaction between the Distribution System Operator (DSO), the aggregator, and the Microgrid Owner (MGO). The upper-level problem minimizes the DSO’s costs, while the lower-level problems maximize the MGO and aggregator’s profits.

The existing literature has mainly focused on the case in which the aggregator is already on the market. In this paper, we study the conditions that might justify the existence of an aggregator in two key dimensions: the ability to exercise a direct control over flexible consumption, which we define as flexibility aggregation, and the ability to aggregate information about the load attributes of the flexible consumption, which we define as information aggregation. In order to understand what the value of the aggregator is, we consider three scenarios with different coordination schemes and levels of information. First, we consider a scenario with complete information in which each agent chooses its consumption knowing the preferences of other consumers, and has beliefs that perfectly anticipate other agent’s actions. While this assumption is difficult to justify in the real world, we use it to study the case in which the only competitive advantage of the aggregator is the coordination of consumers’ demand. Second, we consider a scenario where agents form beliefs about other agents’ future actions based only on past price signals, which could be a result of incomplete information about the attributes of the other market participants. In this case, each agent relies only on its own preferences and price signals to choose consumption, based on the expectations about the actions of other consumers, which are formed based on history. Here, an aggregator may bring crucial advantages to the system by coordinating agents consumption and beliefs. Finally, in order to compare the previous decentralized scenarios, we implement as a benchmark the centralized problem which considers an aggregator that supplies energy to consumers using a direct control method. The optimal value of the centralized problem is compared with the two cases of competition. The higher the cost of the competition scenarios compared to the centralized problem, the greater the value of the aggregator. Regarding the main contributions of this paper, they include:

  • Study of the value of the aggregator. In particular, we focus on understanding the value in terms of both aggregating flexibility and information.

  • The development of a potential game as a solution method to find the Nash equilibrium of a non-cooperative game for demand response. The non-cooperative game is used to understand the value of the aggregator.

  • The study and discussion of the effectiveness of prices as control signals to harness demand flexibility. In particular, we focus on understand conditions for which agents’ response to prices might lead into cycles that remain far from an efficient outcome.

This paper is structured as follows: Section 2 presents the problem definition. Section 3 presents the characteristics of the system and explains the optimization models used. In Section 4 the solution methods for the game theoretical models are explained. The computational experiments and results are explained in Section 5. Section 6 is devoted to the discussion about the results and the impact on market design. Finally, Section 7 provides the conclusions.

Section snippets

Problem definition

We consider a market with three participants: consumers, an energy provider, and potentially, an aggregator. The energy provider establishes a tariff scheme based on price signals. We evaluate the effectiveness of demand coordination in centralized and decentralized scenarios. In the decentralized scenario, the energy provider communicates the price signals directly to consumers, while the centralized scenario considers that the energy provider communicates the price signals to the aggregator,

System and optimization models

The nomenclature used in the optimization models is presented in Sections 3.1 Indexes, sets, and parameters, 3.2 Variables . We consider that consumers have duration-differentiated loads, which are defined by a fixed amount of energy in a fixed period of time. We present further information about this kind of loads in Section 3.3. Section 3.4 presents the market modeling considered in the optimization problems. Finally, Sections 3.5 Non-cooperative game: Individual problem, 3.6 Centralized

Complete information: Potential game

We use the concept of potential game [23] to find the Nash equilibrium with asymmetric players in the non-cooperative game with complete information. Potential games have been used in several papers of electricity markets, such as [28], [29], [30]. The potential game used in the literature corresponds to that presented in [23], which models a Cournot game. In the literature, the potential game is used to find the Nash equilibrium for a Cournot game between generators. In this paper, the

Computational experiments

We present our different case studies and their sensitivities in Section 5.1. The Section 5.2 shows and explains the solution of the non-cooperative game with complete information, and compare this solution with the solution given by the centralized problem. Finally, Section 5.3 explains the solution given by the non-cooperative game with incomplete information, and compare it with the solution given by the centralized problem.

Discussion

The analyzed cases allow us to obtain several insights about market design and demand response implementation schemes. Based on the previous results, we have three cases:

  • First Case:Consumers know the attributes of all flexible loads.

  • Second Case:Consumers have price signals as available information.

  • Third Case:Aggregator uses a Direct Load Control with full information about the flexible loads.

In the first and second case, the consumers could reach the Nash equilibrium of the market, while in the

Conclusions

We study the value of the aggregator in terms of aggregation of flexibility and information. The analysis is carried out by defining several interaction schemes between the consumers, aggregator, and the energy provider. The strategic interactions among the agents were characterized by using game theoretical models. We conclude that information plays an important role in the operational evaluation of the aggregator. Based on the comparison between the performance of the agents and the

CRediT authorship contribution statement

Rafael Rodríguez: Investigation, Writing, Formal analysis, Visualization, Software. Matías Negrete-Pincetic: Conceptualization, Supervision, Project administration, Methodology , Writing - Review editing, Funding acquisition. Nicolás Figueroa: Conceptualization, Methodology, Writing, Editing. Álvaro Lorca: Conceptualization, Writing, Editing. Daniel Olivares: Conceptualization, Writing, Editing.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References (44)

  • HaghifamSara et al.

    Optimal operation of smart distribution networks in the presence of demand response aggregators and microgrid owners: A multi follower Bi-level approach

    Sustainable Cities Soc.

    (2020)
  • MondererDov et al.

    Potential games

    Games Econom. Behav.

    (1996)
  • BischiGian Italo et al.

    Equilibrium selection in a nonlinear duopoly game with adaptive expectations

    J. Econ. Behav. Organ.

    (2001)
  • AlbadiMohamed H. et al.

    A summary of demand response in electricity markets

    Electr. Power Syst. Res.

    (2008)
  • LoHelen et al.

    Electricity rates for the zero marginal cost grid

    Electr. J.

    (2019)
  • AcemogluDaron et al.

    Competition in electricity markets with renewable energy sources.

    Energy J.

    (2017)
  • AmpatzisMichail et al.

    Local electricity market design for the coordination of distributed energy resources at district level

  • CochranJaquelin et al.

    Flexibility in 21st Century Power SystemsTechnical Report

    (2014)
  • ChenChen et al.

    How will demand response aggregators affect electricity markets?—A cournot game analysis

  • QdrQJUDE

    Benefits of Demand Response in Electricity Markets and Recommendations for Achieving Them, Vol. 5Tech. Rep

    (2006)
  • HenríquezRodrigo et al.

    Participation of demand response aggregators in electricity markets: Optimal portfolio management

    IEEE Trans. Smart Grid

    (2018)
  • GkatzikisLazaros et al.

    The role of aggregators in smart grid demand response markets

    IEEE J. Sel. Areas Commun.

    (2013)
  • Cited by (16)

    • Temperature Control Strategy to Mitigate Electrical Energy Cost for Air Conditioning

      2024, e-Prime - Advances in Electrical Engineering, Electronics and Energy
    • A cooperative game theory-based approach to formulation of distributed slack buses

      2022, Sustainable Energy, Grids and Networks
      Citation Excerpt :

      In [18], a game theory-based comparative analysis has been presented and the valuation of demand aggregators has been examined in terms of aggregation of both flexibility and information. Also, a potential game has been developed in [18] to analyze the system under the scenarios of complete and incomplete information. The applications of cooperative game theory in power system expansion planning have been reviewed in [19].

    • A charging station planning model considering electric bus aggregators

      2022, Sustainable Energy, Grids and Networks
      Citation Excerpt :

      This agent is decisive in managing geographically dispersed EVs, which have to be grouped since their load is relatively low for the power grid [43]. Various works considered the participation of EV aggregators in power systems [44–50]. The objectives are related to mitigating grid issues while minimizing electricity costs, but EV aggregators can also participate in power markets such as spinning reserves or regulation services [51].

    View all citing articles on Scopus

    This work was supported by CONICYT/FONDECYT/ 11170630, CONICYT/FONDECYT/ 1190460, CONICYT/FONDAP/ 15110019, and the Complex Engineering Systems Institute, ISCI (CONICYT: FB0816).

    View full text