Virtual Power Plant Optimization in Smart Grids: A Narrative Review

Abstract

Virtual power plants (VPPs) are promising solutions to address the decarbonization and energy efficiency goals in the smart energy grid. They assume the coordination of local energy resources such as energy generation, storage, and consumption. They are used to tackle problems brought by the stochastic nature of renewable energy, lack of energy storage devices, or insufficient local energy flexibility on the demand side. VPP modeling, management, and optimization are open to research problems that should consider, on one side, the local constraints in the operation of the energy resources and power flows and the energy grid’s sustainability objectives on the other side. There are multiple goals to create a VPP, such as to deliver energy services on a market or to the grid operator, to operate a microgrid in autonomy decoupled from the main grid, or to sustain local energy communities. In this paper, we present the results of a narrative review carried out on the domain of VPP optimization for the local energy grid integration. We have defined a search strategy that considers highly rated international databases (i.e., Elsevier, IEEE, and MDPI) in a six-year timeframe and applied objective inclusion/exclusion criteria for selecting articles and publications for the review; 95 articles have been analyzed and classified according to their objectives and solutions proposed for optimizing VPP integration in smart grids. The results of the study show that VPP concepts and applications are well addressed in the research literature, however, there is still work to be done on: engaging prosumers and citizens in such a virtual organization, developing heuristics to consider a wider range of local and global constraints and non-energy vectors, and to decentralize and make transparent the services delivery and financial settlement towards community members. This study can help researchers to understand the current directions for VPP integration in smart grids. As a next step we plan to further analyze the open research directions related to this problem and target the development of innovative solutions to allow the integration of multi-energy assets and management of cross energy sector services in energy communities.

1. Introduction

Energy has been and will always be one of the resources without which the community, and society as a whole, could not be sustained and evolve. Nowadays, no industry is not dependent on the energy regime. The adoption of renewable and decentralized energy sources transforms the energy grid. The management processes become highly dependent on the metering and ICT technology to secure and optimize the power distribution. Renewable energy is generated intermittently and highly dependent on the weather, thus making the planning of energy resources usage complex and exposed to uncertainty [1]. The energy storage capabilities installed at the edge of the grid are insufficient and not able to mitigate the variations in energy production. The surplus or deficit in energy generation may threaten the security of consumers’ energy supply, thus distribution system operators (DSO) must incorporate new techniques to avoid system overload and power outage situations. The distribution of the energy has to offer stability to the consumers, minimize the costs for the producer, and avoid CO₂ emissions, and all in a secure way. The economy and energy grid need to remain stable, but these goals are impacted by factors such as energy resources, policies, politics, markets, the number of customers, and the surface to cover [2].

In this context, the concept of virtual power plants (VPPs) has been proposed and used lately as a solution for assuring an affordable, secure, and steady supply of energy in the smart grid [3]. It aims at combining and coordinating energy production with storage and consumption resources featuring controllable loads in an optimal way to meet renewable integration or energy costs minimization objectives. The associated optimization problem is usually addressed in the literature using various heuristics [4]. VPPs consider the predicted energy demand and production values for different resources, the operational, power flow, and cost constraints. They virtually aggregate resources to participate in energy markets, plan their operation in advance to deliver energy services, aggregate energy profiles for a steady supply and provide demand response services.

In the state-of-the-art literature, there are many works that that focus on creating such VPPs and optimally interacting them in various smart energy grid scenarios [5,6]. At the same time there are multiple ways to create a VPP that offer solutions for the energy market such as energy service delivery, energy autonomy, energy community, energy management or energy optimization, each of them with different purposes and steps to achieve a balance in the markets [3]. In order to establish which is the best approach for different situations, a requirements engineering solution has to be used where the starting point is the scenarios that describe the issue, in this case the energy consumption, and how to ensure stability and balance in the energy market, and the solution for it, represented in this case by the energy that can be provided from renewable resources which can sustain the economy and the market if a VPP is created in the architecture.

In this paper, we carry out a narrative review in the domain of VPP optimization for local energy grid sustainability. The objective is to identify and analyze the main grid scenarios that can be successfully addressed using virtual coordination of resources as well as the state-of-the-art solutions used in implementation. The motivation for conducting the present survey is to identify and describe the relevant literature approaches from the past six years that deal with VPP management and applications for smart grid sustainability. For our narrative review, we have defined a methodology that uses a couple of inclusion and exclusion criteria in the search strategy. We have focused on the reference interval 2015–2022 and we have created a list of keywords and phrases used often in the research literature. The initial list of papers was filtered by selecting only journal articles and conference proceedings from highly rated databases (i.e., MDPI, Basel, Switzerland, IEEE, Piscataway, NJ, USA, Elsevier, Amsterdam, The Netherlands) written in English. The articles should be available to access and with good awareness of the number of citations. We have identified 95 articles analyzed and classified them into three main categories: energy services delivery, local energy autonomy, and energy communities’ sustainability.

The rest of the paper is structured following the IMRAD (Introduction, Methods, Results, and Discussion) methodology. This methodology is used for reporting narrative reviews in the social sciences, engineering, and computer sciences domains. It provides a framework for organizing articles agreed and accepted by many of the highest impact factor Web of Science journals. Section 2 presents the methods used in conducting the narrative (non-systematic) literature review. Section 3 and Section 4 illustrate the results of the study by describing the most relevant research works. Section 5 presents a discussion on the survey findings and conclusions, while Section 6 concludes the paper.

2. Materials and Methods

In this study, we will follow the guidelines for narrative reviews as described in the research literature in several highly cited articles [7]. We have adapted them for the article context and the addressed domain related to VPPs and their efficient integration into the smart grid. Even if narrative reviews do not feature a strict reporting guideline and methodology (such as the PRISMA methodology used for systematic reviews) [8], in this section, we will highlight the main methods used for the literature search process. The following points will be discussed:

• Searching strategy: database, search key words/key phrases;
• The main inclusion/exclusion criteria: types of articles, language, time periods, visibility, and availability.

The main objective of the search strategy was to carry out a review and analysis of the status of the literature for VPPs’ concept and technology, as well as on their usage and operation, in the smart grid domain. We are interested in finding the most important VPP and smart grid integration scenarios, use cases, and solutions, and classifying them.

As the main database, we have used the well-known Scopus articles repository. We carried out separate and aggregated keyword manual searches and finally extracted the papers of interest. The search keywords and key phrases to narrow the search results involve the following concepts: energy forecasting, energy modeling and optimization, energy management, renewable integration, local energy autonomy, energy community, and energy services provisioning.

These key concepts were concatenated with VPP in the search process. Additionally, we have considered in the review eight online/magazine articles from the industry to capture the companies’ opinions and advances about the researched topics. We have eliminated the duplicates from the results and defined an initial set of articles.

In the search strategy for filtering the Scopus database search results, as a first inclusion criterion, we have targeted three of the most important publishers, namely MDPI, IEEE, and Elsevier. Additionally, we have considered high-quality articles from other databases such as Wiley, Springer, etc. Conference proceedings and journal articles, and review papers, were the kinds of studies selected. The English language for article redaction was also used as an inclusion criterion. As for the publication time, we have included works newer than 2015, with an emphasis on the past four years (2019–2022).

The exclusion criteria used are mainly on the availability of the articles. If it was not possible to have access to the full text, they were excluded. Additionally, we have excluded papers that feature a low number of citations or the ones that do not cover well the search keywords or phrases.

Table 1. Criteria for including and excluding articles.

Inclusion Criteria	Exclusion Criteria
MDPI, IEEE, Elsevier, or other highly rated databases	Not available as full text
Conference proceedings	Low number of citations/views
Journal articles/review papers	Not connected to the research topic (VPP applications for smart grid)
Usage of English language	Duplicate
Publication date 2015+

presents an overview of the search strategy inclusion and exclusion criteria.

We have filtered the search results and created a list of more than 95 articles that were analyzed in the current review. Figure 1 shows the distribution of the selected papers by using the publisher as discriminant, while Figure 2 shows the research works’ year coverage.

We have grouped them into two categories: (1) VPP concept and technologies presenting the main models and techniques for VPP creation and operation, and (2) VPP application in smart energy grid management. The first category is divided based on the main elements to be addressed in VPP management and operation: (1) modeling of energy assets via digital twins, (2) energy and price prediction, and (3) optimization and coordination in VPP operation. The second is further classified based on the type of application and optimization objective of: (1) VPP for energy services provisioning, (2) local energy autonomy through VPPs, and (3) VPP for energy communities’ sustainability.

Table 2. Narrative review research directions studied.

VPP Research Direction		No. Approaches	References
VPP concepts and technology	Digital twins of energy assets	13	[9,10,11,12,13,14,15,16,17,18,19,20,21]
	Energy forecasting	14	[1,22,23,24,25,26,27,28,29,30,31,32,33,34]
	Optimization and coordination	18	[35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52]
VPP applications in smart grids	Energy services delivery	31	[4,5,22,35,38,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78]
	Local energy autonomy	21	[5,36,44,75,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94]
	Energy communities’ sustainability	18	[51,52,63,64,93,95,96,97,98,99,100,101,102,103,104,105,106,107,108]

presents a summary of the articles included in the study using the three categories.

In Section 3 and Section 4, we will present the results of the study carried out on the research literature following the methods described above.

3. VPP Concept and Technology

Smart grids are complex systems, difficult to manage, control, and operate, especially towards the edge, and when integrating renewable energy from distributed sources [2]. Demand response programs and prosumers’ flexibility and energy trading in energy markets raised an optimization problem for modern smart grids, leading to defining of various virtual aggregation schemes, models, and algorithms [109]. Apart from TSOs and DSOs, new roles in the smart grid ecosystem have been defined, such as that of energy aggregators [101]. This is facilitated by the evolution of energy management systems through new architectures that involve dynamic interactions between the actors and the energy assets and the implementation of new energy services [110].

Virtual power plants (VPPs) deal with the aggregation and management of distributed energy assets to meet specific optimization goals [3]. The creation and operation of VPPs involves the application of complex ICT techniques dealing with the modeling of the energy assets and energy flows, forecasting of local energy demand and generation, operation optimization to meet the defined goals using heuristics, and finally portfolio management and actuation for closing the loop [5,6,73] (see Figure 3).

3.1. Digital Twins’ Models

A fundamental aspect of virtual aggregation of energy assets or resources in VPPs is their modeling to understand operational constraints, energy behavior, and interaction flow [5]. Lately, the concept of the digital twin (DT) has emerged as a virtual representation of energy assets. It can successfully be used to conduct various analyses and simulation processes to study their energy behavior and energy exchanges [9]. In the energy grid, DTs are considered a disruptive technology for enhancing and maintaining smart grids and for building ML models to enhance their performance and efficiency [10].

DTs are used to model all components of the VPP portfolio of assets (energy production, consumption, storage) and simulate behavior through complex analytics with the general objective of facilitating the integration of optimization heuristics to drive the operation closer to the defined goals [11]. The technologies associated with DTs are complex system modeling, big data prediction, ML, optimization, and agent-based techniques [12]. They involve complex data processing and models to digitize and evaluate the grid rules, understand the energy distribution flows, or the impact of decision-making [13]. For example, DT models for net zero energy buildings deal with optimization of the renewable usage considering inhabitants’ comfort and constraints [14]. They are useful for empowering VPPs to manage their building portfolio towards minimizing the energy exchanges with the grid and improving the energy efficiency [15]. They are not limited to electrical energy aspects, but they can model thermal, usage, and cost aspects of various energy assets from the VPP portfolio to increase the amount of committed flexibility [16]. Simulation also can be carried out for the thermal design of a heating and cooling system of buildings and their relationship with energy usage and available energy flexibility to be used by the VPP [16].

DTs can model the energy flexibility profiles of assets and devices such as heat pumps, EVs, PV, hot water, or gas systems that are relevant for VPPs [10]. The main challenge is to couple them with information relating to the user’s wishes in terms of comfort, convenience, and well-being. A user’s DT profile seamlessly incorporates their flexibility profiles, representing the selected flexibility assets, and can increase the “smartness level” of buildings which will lead to increased participation in energy flexibility services to be delivered by the VPP [17]. The modeling techniques’ focus is not just on the average dynamics but also on modeling uncertainty, i.e., a statistical description of stochastic behavior, as it has a severe impact on any decision-making logic on VPP optimization [18]. In holistic terms, the DT can consider the consumer’s profile in terms of behavior and incentive mechanisms of VPPs along with their community aggregation association and flexibility assets reflecting various commodities (electricity, biogas, heat) [19].

Finally, DTs have been used for managing customer profiles and raising awareness about energy services [9]. Data on customer preferences, satisfaction, and behavior can help DTs to generate reports for increasing customer awareness [20]. Indeed, DTs are seen as a key asset for citizens’ engagement which is very relevant for local community management and sustainability [21]. Engagement of individual households and collectively operating households in demand response and energy communities remains a challenge [111]. In addition, since smart meters themselves do not communicate with consumers, their roles in changing consumer behavior will be low. Thus, providing products and services which address concerns and provide enough benefits is pivotal for reaching the potential of VPPs and energy communities [112].

3.2. Energy Forecasting

The forecasting of energy demand, generation, and energy prices are fundamental inputs of the VPP management and optimization problem [22]. The prediction accuracy impacts the quality of the VPP solution; thus the uncertainty should be considered in the optimization problem formulation [55]. The stochastic nature of renewable energy generation and volatility of energy prices are limiting factors for VPP participation in energy markets [1]. Therefore, it is important to improve the quality of the forecasting models and techniques to lower as much as possible the impact of these uncertain parameters on the optimization problem [23].

Nowadays the literature is addressing the forecasting of energy demand, generation, and price using machine learning models [24]. Few approaches can be found in the prediction of the energy flexibility of customers which are relevant for building VPPs and energy communities [25]. Deep learning models are reported to provide good accuracy results for short-term energy use prediction [26]. Most promising in energy forecasting are artificial neural networks using long short-term memory (LTSM), convolutional neural networks (CNN), multi-layer perceptron (MLP), or recurrent neural networks (RNN) to achieve more accurate energy predictions [27]. Lately, ensemble-based approaches are also providing good forecasting results. For example, hybrid energy forecasting models are reported based on CNN and LTSM [28], two-hidden-layer LSTM and two-hidden-layer CNN [29], CNN-LSTM-RNN hybrid networks [30], and LTSM-RNN [31] hybrid models.

Analyzing the current state of the art here is very difficult in comparing existing models and technologies in terms of performance, prediction accuracy, and impact on the virtual aggregation of energy assets [32]. The quality of the prediction outcome is influenced by factors such as seasonality, social factors, the intrinsic physical parameters of the assets, time step, and prediction interval [23]. There is no consensus on which is the most suitable forecasting model, so the exploration of more than one technique on the data at hand, and the comparative evaluation of these, is recommended for VPP optimization. Forecasting models can have different accuracy considering the data at hand [33]. There may also be differences in the quality of the prediction results for different time intervals and this significantly impacts the VPP operation processes.

Finally, the prediction processes can be joined with the characteristics of energy assets and communities determined using physics-informed DT models [34]. They are implemented using big data infrastructure to consider the energy meters’ data as well as a significant number of energy assets in VPP management and operation. Such solutions consider the social and behavioral characteristics of customers, local communities, etc., that have the potential of improving the prediction accuracy and thus reducing the uncertainty related to VPP optimization [23].

3.3. Optimization and Coordination

The optimal virtual aggregation of energy assets is a complex constraint satisfaction problem addressed using different heuristics [62] considering local and energy grid sustainability objectives [35]. Relaxation of time constraints is possible in the case of VPP operation on the day-ahead energy markets allowing the implementation of cloud-based optimization solutions [36]. Various heuristics are defined for minimizing the VPP operational cost and maximizing the energy profit of energy assets [3,37,38]. Stochastic programming models are defined for VPP management and portfolio optimization under various objectives [39,40]. Moreover, multi-criteria optimization heuristics are investigated for VPP portfolio optimization to provide simultaneously cross-sector combined services for increased flexibility provisioning and to provide a transparent, verifiable, and trustworthy management framework [41,42].

Nevertheless, the decentralization of VPP coordination is only partially addressed in the literature even though it is a promising solution for better and timely consideration of local energy constraints [41,43]. Local density, power flows, administrative, and economic or social factors are criteria considered for VPP optimization processes [44]. There is a strong need for decentralized decision support systems on energy assets coordination in community-level VPP energy assets considering data on local sustainability goals, assets size, communication efficiency or latency, local typology, and remuneration schemes [45]. Others are using games theory to achieve a certain degree of decentralization while reducing the uncertainty impact on VPP optimization [46]. The optimal virtual aggregation of energy assets aims to achieve a state of equilibrium among coordinated energy assets while optimizing the service to the DSO [47].

Several authors use edge computing principles for VPP optimal operation while addressing the challenges related to localized decision-making [48]. The authors focus on households’ or citizens’ engagement and aggregation when dealing with limited local renewable energy budgets showing privacy features. The fog computing principles can be applied for VPP optimization heuristics [49] for better using computational resources and improving availability. They bring important features for sustainable energy transition such as reducing operational costs, better integration of renewable energy, local balancing of generation, demands and storage, etc. [50]. Finally, energy as a service platform can be used for microgrid and VPP operation to provide closer to real-time ancillary services [51]. The service-oriented optimization models consider potential congestion, power flows and structures, and privacy of citizens for reliable delivery of VPP services [52].

4. VPP Applications in Smart Grids

The main identified research works are briefly presented and classified according to the VPP usage and applicability in specific smart grid energy management scenarios or strategies:

• VPP coordinates energy resources for collectively providing energy services in different markets or directly to interested stakeholders such as a DSO;
• VPP coordinates energy resources for local energy autonomy to achieve an optimal balance between the demand and supply and to minimize energy exchanges among microgrids and the main grid;
• VPP coordinates energy resources for the optimal implementation of sustainable energy communities considering in addition to energy aspects the local economic and social factors.

4.1. Energy Services Delivery

In modern smart grids, one important problem is system stability [2]. The grid can benefit from energy services as a solution to keep the power supply stable [109]. The services can be used to ensure that the energy demand, flexibility, or storage capabilities are used efficiently. The main role of a VPP in this context is to facilitate the prediction of the energy demand and generation of power from renewable sources such as wind, sun, etc. [22,35]. Using these estimations, the VPP coordinates the remote control of the spread devices that are dealing with these issues to offer energy services [5]. Different services may be provided by the VPPs depending on the market type [73] (see Figure 4). A VPP may interact with the energy market to buy energy when the prices are low and charge energy storage systems and sell the energy surplus when prices are high by adjusting the demand and discharging energy from the batteries [65]. The VPP will act as an intermediary between the energy resources in its portfolio and the energy market while addressing market or policy barriers. In the balancing market, the VPP can provide capacity for power plants that cannot meet their original commitment [65]. In the ancillary services market, the VPP may provide near real-time services such as frequency regulation [65].

At the same time, a VPP can provide energy management towards the DSO, and the constraints of the members need to be satisfied while the stability of the entire system is considered the main objective [74]. The goal of VPPs is to manage the energy assets that are virtually aggregated to provide energy services to the DSO while optimizing the profit [64]. Such a scenario was embraced especially by the markets where the distribution of energy from raw materials is problematic or implies costs that cannot be supported [63]. The distribution of that energy is also important, and in this type of architecture, the storage resources are taking a big part in the distribution system. The focus is on the balance that efficient coordination can offer, and the costs associated that are kept as low as possible using management tools [38].

The delivery of energy service using VPPs was studied in the past decade. The Toshiba VPP offers energy services for multiple storage batteries from 2019 [66] and the focus is to facilitate the demand response based on the storage batteries and the status of the power network. Some of the services offered to consumers are imbalance prevention, storage for excess power, peak-cuts, etc. Multiple elementary and junior high schools will receive energy from the virtual storage battery group and will demonstrate that distributed power supply can be achieved at a low cost [67]. Others allow consumers to create their VPPs featuring infrastructure and services such as energy trading with live data, dispatching software to help the scheduling, load management, and at the same time demand response to control the supply and to ensure balanced services [68]. The customers are helped to aggregate different types of renewables in their VPP and to control the load. Tesla created a VPP that is designed to help the energy consumption in California [69]. Its involvement will facilitate the energy demand during the summer when the temperatures are high, but the consumers must participate voluntarily, without compensation. The beta version is offering services to predict the need for energy and ways to store the remaining energy when it is not used at full capacity. Other providers such as Sonnen invested in VPP energy services delivery for specific areas such as communities from New York [70]. They focus only on solar renewable energy that can supply 200 houses at a lower price. Sonnen provided services to 50 residents with limited income in Los Angeles to use solar energy power in California. Swell created a VPP that offers services in California, Hawaii, and New York focused on solar power energy. Some of the services it provides to customers are efficient system automation, energy backups in case of emergencies, permanent monitoring, and stability [71]. It also offers services where the customer can design a VPP. Battery energy raised a lot of interest among multiple companies such as PG&E, OhmConnect, SPACs, Portland General Electric (PGE), and Vermont. All of them have pilot projects that create renewable energy using batteries. The architectures and maturity of the projects differ, but all these companies provide services for residents that will offer stability and backup energy in case of special events and a decrease in the cost [72].

The impact of the VPP services and the benefits brought by this type of architecture for energy management has been studied in the literature [75]. In [57], a framework for the coordinated management of VPPs is described, considering correlated demand response. The VPP operating costs are minimized while assuring the power system stability and power quality. The optimization problem is addressed using a low-risk two-stage stochastic strategy considering renewable generation and energy price uncertainty. With the aid of the extensive advanced simulations conducted, the benefits of the proposed model in minimizing the VPP costs and improving the power quality have been illustrated. The authors of [60] give attention to a hierarchical model proposed for simultaneous modeling of scheduling and VPP energy management problems. As the stochastic nature of the scheduling inputs is given, power production and load demand waving are modeled method. As a result, the final model is presented as a stochastic mixed-integer linear programming model. By using demand response, it is possible to cover the fluctuations at the minimum cost [76]. Moreover, the possibility of power transactions between the adjacent microgrids is investigated in the scheduling model, preventing unreal power exchanges. From the simulation result, it is possible to reduce the supply costs, thus leading to a significant increase in profit of the VPP. Making a transition towards renewable energy gives small, distributed generators a huge impact, thus leadig to the use of actively stabilized distribution grids [4].

The need for coordinating the distributed generation in VPPs forces their schedules to be managed as a single plant [77]. This problem can be addressed by using agents and flexible control methods [78]. The key for coordinating such systems is the underlying communication infrastructure. In [53], the authors provide a direction where energy assets and batteries are integrated into a VPP that serves as a service layer for the market by analyzing the technical and economic possibilities. They study the economic efficiency of the model, and the influence of energy price and renewables, showing that positive net value can be reached in several years. The authors of [55] highlight the necessity to create alternative electricity management systems and a solution would be the use of VPP services that can provide energy to places where the resources are limited. It may be necessary to combine several distributed renewable energy generation units, and electrical vehicle charging stations. They are coordinated by considering the prediction of consumed power and of the power produced, allowing the management of the volume of energy in a balanced way. The prophet model is implemented for energy prediction and technical analysis purposes. In [62], we have defined a fog computing solution for the creation and management of VPPs to provide ancillary services. Energy meters are used for prosumers’ monitoring while a fog layer manages the local energy systems through cooperative models. The energy services are integrated through a cloud layer considering frequency reserves, local constraints, and profit maximization.

The energy markets allow for efficient and flexible energy trading, in a competitive environment. The sustainability concerns increased the adoption of medium-small renewable VPPs distributed in the grid. However, the lack of central monitoring generated a whole new set of challenges for power system operations. With the implementation of the VPP for energy management concept comes the growth of the power system benefits, as well as the increase in the efficiency of the distributed generation units, which can become more visible and maximize the opportunities regarding incomes when it comes to selling the energy [22]. A VPP is used as a tool to integrate distributed renewable and storage with the needs of consumers in demand response programs [73]. In [61], blockchain and self-enforcing smart contracts are used to allow VPP construction out of prosumers’ pools with the main objective of providing energy services. The prosumer level constraints (energy profiles and requirements) have been modeled and encoded in smart contracts to enable their aggregation in VPPs. The VPP is managed using blockchain offering decentralized energy trading and financial settlement. In [54], a probabilistic model is defined for optimal scheduling of a VPP considering electrical and thermal energy flexibility. It is able to participate as a service provider in different markets, contributing to the demand balancing and efficient distribution of the energy. The uncertainties in VPP optimization generated by energy prices and renewable production [77] are addressed using scenario-based decision-making. The VPP acts as a participant in the market, selling energy, and is can be seen as a controllable demand unit by the main grid. The VPP can not only enable participation in the market of small generation resources but also can provide additional services such as reactive power control.

In [58], the authors depict a VPP model consisting of distributed generation and controllable resources. It will handle the management of the energy market of the VPP, helping to reduce transmission losses and electricity scarcity. The distributed renewables deployment leads to increasing generation variability and uncertainty. A solution to this issue is a VPP that offers services and aggregates a hybrid resources. In this way, [56] proposes a modeling framework that allows a VPP to participate in energy markets. It is divided into three stages (the high, mid, and low level) that allow the optimization of resource scheduling and dispatch. A real case study with electricity and hydrogen is being conducted in South Australia and shows that the VPP can coordinate the resources with promising results. In [59] is depicted a model approach to determine the optimization and coordination of a VPP participation in energy markets for energy services delivery. The authors propose a mixed-integer linear programming optimization model for day-ahead optimal energy and reserve dispatch and optimal real-time energy re-dispatch. A VPP that provides software as services for the community or cross-sector services represents a new direction for the future, and the studies on this topic and possible implementations started in the past decade [78]. There are still a lot of unknowns because renewable energy is influenced by the weather, but this topic can help solve the energy demand problems. The studied approaches focused on possible implementation for these services but also on solutions on how to make them function optimally and securely for all types of renewable energy.

Table 3. Energy services delivery through VPP coordination.

VPP Energy Service	Literature Approach
Imbalance prevention, store for excess power, peak-cut	[58,62,66,67,73,76,77,78]
Load scheduling and balancing,	[54,55,57,58,68,69]
energy trading	[35,38,53,54,56,57,58,59,60,61,65,68,70,74,75]
Energy consumption/production prediction	[5,55,57,60,61,63,64,69,77]
Energy backup and stability	[22,61,71,72,74,75]
Capacity management, frequency regulation	[4,54,58,62,73,74,78]

summarizes the categories of energy services proposed in the research literature for VPP energy management.

4.2. Local Energy Autonomy

Energy autonomy is considered an effective solution for managing local energy systems and represents a relevant development direction for managing decentralized smart grids characterized by sustainability [44]. The VPP may provide support for energy autarky by coordinating the energy generation with the storage and the demand so that the local microgrid can work decoupled from the grid [52]. One problem is how to manage the energy demand and the renewable supply in a balanced way so that the exchanges of energy with the main grid are minimized [78] (see Figure 5). The grid can benefit from the power supply stability at the edge, while the VPP can work fully or partially powered by renewable [36].

In this scenario, the management of the energy reserves has an impact on the VPP, and therefore it is well connected with energy islands [90]. The islands are an example of a local energy system that is facing geographical, technical, environmental, and socioeconomic challenges related to its energy supply. The energy supply of islands often relies on high yet costly dependency on imported fossil fuels, suffers from energy supply constraints due to the lack of interconnections among electricity and gas, and is significantly affected by the high seasonality of demand due to tourism [81]. VPPs can provide a more targeted approach for moving forward towards a sustainable and more decarbonized energy supply, which combines and integrates both technical and socioeconomic components [5]. Such energy systems share a significant wind and solar potential so that renewable energy can provide a good opportunity for decarbonization while reducing the pressure on the local economy.

VPP usage to achieve local autonomy was researched and applied considering renewable such as large-scale solar, rooftop solar, biomass, biogas, mini hydro, etc. The concern, in this case, is related to the amount of flexibility needed to assure the stability of the local power systems [91]. One of the solutions to this problem, proposed in [79], is the battery energy storage system. This technology provides good opportunities in mitigating some of the power system issues. Such energy storage is considered relatively new technology but can provide good levels of local energy autonomy [92]. Achieving renewable energy integration to create local energy autonomy requires the engagement of consumers [111]. In [80] is described the development of a VPP in Western Australia that uses a cloud platform for decision-making and considers a renewable generation redox flow battery [93], heat pump, and management of appliances. The authors estimate that the energy cost is reduced by 24%.

Insular areas took steps in studying the possibility of gathering energy autonomy using renewable energy and VPPs. As mentioned in [87], planning a system to offer autonomy using renewable energy can be challenging due to the change in weather conditions. For this reason, stability may be affected. For example, Ireland and UK do not have optimal interconnectivity with the distributors of energy based on fossil fuels, but, at the same time, the wind represents a resource that could be exploited. The studied and simulated requirements were the static ones that focus on the balance between the demand and the supply that needs to be proactively achieved to avoid offering a frequency that is not sustaining the minimal needs [112]. As renewable energy is unstable, a ramp rate limitation is computed to decrease any issues that may occur in case of a spike of energy, where the gradient is out of the limits [94]. The authors of [81] describe both analytical and experimental infrastructure used on the island of Bornholm. The approach is focused on modeling a low-voltage microgrid for providing grid services and achieving energy autonomy. The potential of a renewable-driven VPP is analyzed considering different energy domains such as electric, heat, gas, transportation. The constraints impacting the energy autonomy, such as the power flow, the dependency on batteries, and the system flexibility are considered. Another island that took steps toward assuring the local energy autonomy using a VPP configuration is Porto Santo Island [82]. The aim is to use all the available renewable resources to generate energy and meet the demand while avoiding the import of energy from the continental energy network, which is expensive. The potential energy that could generate an alternative solution can be taken from wind and sun, the only available resources. An action plan was created, and the characteristic is to increase the system efficiency by using VPP aggregators as load dispatchers that will balance the supply and the demand.

Other studies focused on the energy for households using photovoltaic generation and controllable loads [84]. The objective is to model a system that controls the energy provided, since solar energy is considered cost-effective but at the same time unstable and not secured, but without storing it, since that is an expensive procedure. To accomplish this, the system architecture plays a role. The energy distribution must be managed with controllable loads, and the strategy is computed with quadric interpolation. The model simulation demonstrates the possibility of achieving energy autonomy without the necessity of storage resources. The authors of [83] describe a model for a grid-connected VPP that collects solar and wind and produces renewable energy with low costs targeting the optimization of transportation. To understand the energy demand necessary for a full trip, some factors are considered. For example, the number of passengers is computed using stochastic methods because the number is not distributed uniformly during the entire day. Another factor is the minimum redistribution of vehicles from the origin to the distribution on a specific hour. Multiple cases were simulated with 100 iterations, indicating a reduction of costs by 20% on a dataset from Tokyo city. The authors of [85] discuss VPP energy autonomy and its usage for integrating energy resources, considering the demand on the market. The mathematical methods need to be correlated with the techniques that manage the distribution of the network. The model offers a structure that can handle the diversity of requests, being flexible, but at the same time effective. The authors also accentuate that the geographical position is important even for renewable energy.

Energy autonomy can be reached [86] with a stochastic model that is heavily optimized with multiple levels of scheduling. A day-ahead prediction in a VPP that can provide increases on the income is modeled with a bi-level optimization. It creates the first schedule plan, goes over it again, and analyzes the ratio of costs and improvements. In this type of model, the need of a backup is necessary because the predictions can change based on special events induced by the consumers. The demand curve is analyzed and flattened as much as possible. The system output can be coordinated using the VPP over which charging−discharging power is applied. The model was simulated in the easter China micro-grid, assuring some levels of energy autonomy and a decrease in costs. Another country that created a model for energy autonomy is Portugal. The goal is to fully operate on renewable energy over 30 years [88]. Multiple scenarios are taken into consideration with different perspectives that are evaluating the profit in the VPP in 2020, 2030, and 2050. The model is built on mixed-integer nonlinear programming with discrete variables and equations for nonlinear networks. The model objective is to minimize emissions but at the same time maximize profits. The system is optimized with a multi-objective particle swarm concept [113] where multiple algorithms are combined to obtain profit. Two of these algorithms are the non-dominated sorting genetic algorithm [114] and Pareto archive evolutionary strategy [115]. The results showed that renewable energy consumption increased in 2050 by almost 46%, directly proportional to the profits. India exploits solar energy to provide solutions for environmental threats or raised prices of fuels [89]. A case study was done at a large scale where the government helped in deploying distributed renewable such as solar PV. The VPP helped in correlating and scheduling the PV resources and the storage to offer flexibility and high availability [90]. The model was used with a customer adoption model tool using linear programming. Multiple types of storage were used to create VPP, such as battery energy, ultra-capacitor energy, flywheel energy, and superconductor energy. A VPP for energy autonomy can be realized if the cost minimization, the increase of reliability, and the peak-load reduction are considered as objectives.

Table 4. Energy resources coordination in VPP for energy autonomy.

Resources Integration in VPPs			Approach
Controllable loads	Energy Storage		[78,79,80,81,88,90,92,94]
	Renewable Energy	Hydro	[79,80,85]
		PV	[5,36,44,75,79,80,81,84,85,86,87,88,89,93]
		Wind farms	[5,44,75,81,82,83,85,86,87,88]

summarizes the usage of DERs in different approaches for achieving energy autonomy.

4.3. Energy Communities’ Sustainability

Fundamental for increasing the adoption of VPPs are the customer engagement strategies and underlying measures for voluntary participation [93]. Decentralized renewable energy and digitalization allow new ways for engagement through energy cooperatives and citizen energy communities [111]. The EU energy regulation provides an enabling framework for citizen energy communities as well as renewable energy communities [112]. VPPs are keys to ensuring that the prosumers and local communities take the front seat and co-create innovations that are aligned with their values and expectations (e.g., comfort, well-being, prices, etc.) [52]. These developments further provide opportunities for support of additional values and VPP management and optimization. A concrete example is a VPP of electric vehicle sharing within the local community, powered by their electricity, and used for storage, with multiple users such as citizens, companies, volunteer organizations, municipality personnel, etc. [102] (see Figure 6). The engagement of citizens and communities in such a local energy system increases the trust, identity, and the sense of community [101].

For smaller communities, the costs to use services from other suppliers of VPPs can represent an impediment in obtaining the energy needed [80]. Therefore, the concept of a VPP for the energy community was studied and applied in multiple communities where the needs of the population were the focus, but at the same time the costs must remain affordable, and the system had to offer stability, balance, and high availability [103]. In a VPP that provides energy for communities, the consumers can decouple from the grid and still function as needed using local resources [103]. This scenario was embraced especially by communities where the distribution of energy from raw materials is problematic or implies costs that cannot be supported. Even inside the community, the distribution of that energy is important, and in this type of architecture, the storage resources play a big part in the distribution system [104].

The authors of [96] investigate how VPPs can be used to provide energy for communities and services to the grid. The approach proposes to use an energy storage distributed to energy prosumers so that the energy storage devices are aggregated into a VPP. The daily optimization considers the battery state to improve its lifetime. In [95], the authors propose the use of VPPs that coordinates a set of distributed energy resources to improve the quality and reliability of the power supply. The model described offers a solution for increasing the production of energy and sustainability while, in turn, decreasing the carbon footprint. In [97], a VPP community in Poland is studied. It contains hydropower plants and energy storage systems and addresses the power quality issues. The research considers the frequency of the energy, voltage, and the asymmetry factor. The VPP is formed by two types of voltages of the distribution of energy in the network. The system was tested with real data and is helping the community from Poland to generate energy from renewable resources.

Community VPPs have become one of the most promising solutions to integrate intermittent renewable while considering the local socioeconomic context [51]. In [98], an array of optimization solutions are developed and applied for a community-oriented VPP aggregating renewable energy generation units. The scheduling of the VPP aims to maximize revenues and reduce penalties while addressing market constraints. The authors of [99] explore how modeling and fuzzy logic can be used for VPPs in an open context which is characterized by features such as external components that can be included in the system as needed at run-time without any impact, new interfaces that could facilitate the connection of the system with those components can also be added if needed and all the interfaces or components can be changed at any time if required. Moreover, the context and time dependencies are computed with formal methods and the erroneous information is extracted from the result. The high renewable energy sources produce excess energy which in turn generates a burden that creates a notable residual load [105]. The authors of [63] examine a possible implementation of a VPP to aggregate small-scale renewables to be sold on the energy market. It also investigates and implements the optimal VPP configuration for a RES-based power plant. With the aid of this paper, the capability of VPPs to provide secured power market products derived from EEX/EPEX standards is proved. In [64], solutions to optimize coalitions and communities are proposed. The VPPs are created from the prosumers’ pool using as the main criteria the energy price during an interval. Using load forecasting and blockchain, the approach contributes to the uptake of local communities and will lead to profit for the customers, offering an alternative for participating in energy markets and paving the way for flexible trading. Indeed, blockchain, smart contracts, and peer-to-peer energy trading are considered key technologies for developing VPP for energy communities [107,108], especially due to blockchain features of immutability and enhanced security.

Interest has increased in using VPPs for community management, offering the citizens access to renewable and storage in urban areas [106]. In [100] is presented a VPP solution for managing both the generation and demand sides using game theory to explore the cooperation among prosumers. The profit of the VPP was allocated based on the Shapley value.

Table 5. VPP objective for energy communities.

VPP Objectives		Approach
Coalitions of Prosumers	Energy optimization	[52,95,97,100,101,103,106,108]
	Local sustainability	[51,96,99,102]
	Profit	[63,64,93,98,104,105,107]

summarizes the general objectives of the studied VPP communities’ approaches.

5. Discussion

Many different strategies are reported in the literature for the creation of VPPs. They aim to integrate renewable energy and increase its usage to address current sustainability issues generated by natural fossils. VPPs can be used to offer energy services in different energy markets, to increase the local energy system autonomy, and to support the implementation of energy communities considering the local socio-economic context. Each application case provides advantages for different smart grid management scenarios. Most of them are studied, researched, and applied to pilot infrastructures. All VPP alternatives are characterized by the same model structure and energy assets. The services are offered not only in energy markets but also to different energy operators responsible for the safe and reliable distribution of energy. An example, in this case, is the distribution system operator that manages the energy of a low-voltage grid. The distributed energy resources are the most important components of a VPP. They are the entities that generate, store, and consume energy. Additionally, in the case of energy communities, the citizens or energy prosumers play an important role.

The VPP may coordinate the energy resources to provide different energy services depending on the market type. Among the offered services by VPP, the most important ones covered in the research literature are imbalance prevention, storage for excess power, peak-cut, load management, energy trading, backup and stability, congestion management, and frequency regulation. For charging energy storage systems when energy prices are low, VPPs can buy energy in the wholesale electricity market. On the other hand, VPPs can decide to sell energy that is a surplus and discharge power from the energy storage systems when the prices are growing in order to match the demand of flexible assets. For the balancing market, when a power plant cannot reach its commitment, VPPs can sell replacement capacity on short notice and may be able to increase/decrease production/consumption during a period. In the ancillary services market, VPPs can offer frequency regulation services such as, for example, committing unused capacity. At the same time, a VPP can provide energy management towards the DSO. VPPs can manage all involved resources to deal with the intermittent generation (uncertainty) in smart grids by offering services to the DSO and at the same time optimizing the profit of its participants.

VPPs are used to assure the energy autonomy of local energy systems that may operate decoupled from the main grid. The researched literature deals with VPPs as a solution to provide support for energy autarky. They coordinate the energy generation with the storage and the demand so that the local microgrid energy imports or exports are minimized. In this case, the grid can benefit from the power supply stability at the edge, while the VPP can work fully or partially powered by renewables. Islands are reported as a positive example of a local energy system that is facing technical, environmental, and socioeconomic challenges related to its energy supply that may be addressed using VPPs. The energy supply of islands often relies on high yet costly dependency on imported fossil fuels, suffers from energy supply constraints due to the lack of interconnections among electricity and gas, and is significantly affected by the high seasonality of demand due to tourism. In this case, the VPP simplifies the distribution of energy for that region and uses local resources to the benefit of the community. Batteries, water-based systems, PV, and wind farms are the RES that needs to be available fully or partially for achieving autonomous VPP in islanding conditions.

Finally, the reviewed literature reports on the usage of VPPs for the management of energy communities. In this case, the local socioeconomic context is considered in optimization on top of the energy management aspects. Fundamental for increasing the adoption of VPPs are customer engagement strategies and underlying measures for voluntary participation. VPPs are keys to ensuring that the prosumers and local communities take the front seat and co-create innovations that are aligned with their values and expectations (e.g., comfort, well-being, prices, etc.). These developments further provide opportunities for support of additional values and VPP management and optimization. The engagement of citizens and communities in such a local energy system also increases trust and identity, implementing stronger communities.

One promising research direction in the context of the above is blockchain, which can pave the way to construct decentralized community-oriented VPPs with a view of optimizing flexibility and energy services delivery. Smart contracts can be used as a solution that allows the injection of various prosumers and VPP-level objectives and constraints to meet the provisioning flexibility services. A decentralized VPP optimization solution on top of the blockchain infrastructure can take advantage of the state-of-the-art multi-criteria optimization heuristics to allow for VPP prosumers’ portfolio optimization. Moreover, blockchain technology will provide a transparent, verifiable, and trustworthy management framework for the VPP’s reliable delivery of energy services.

6. Conclusions

In this paper, we have used a narrative review method to study and report the relevant state-of-the-art literature around VPP integration and optimization of smart grids. In our review, we have defined inclusion and exclusion criteria, have set several international databases for pooling articles, and finally selected 30 research papers in the study. All papers have been studied and classified, and the main findings were presented. This paper studies different types of VPPs’ management and integration strategies. It emphasizes the applicability of each solution to address energy grid problems. The outcome of the study shows that VPP integration into smart grid is a hot topic in the research literature. As open research directions we can mention strategies for engaging prosumers and citizens in VPPs, developing new heuristics to consider energy and non-energy data vectors and cross energy sector information, and to employ decentralization techniques such as blockchain for services delivery and financial settlement. This study can offer to researchers a clear overview of the current directions for VPP integration in smart grids. As next steps we plan to take a deeper look at the usage of VPPs for smart grid decentralized management using blockchain as the main innovative technology.