Modeling Electricity Markets and Energy Systems: Challenges and Opportunities

1. Introduction

From the dawn of the Industrial Revolution, energy was predominantly produced by conventional technologies, relying on a heat source (mostly from the combustion of fossil fuels) and the turning of steam turbines. These fossil fuel-based technologies were fairly reliable as long as the fuel supply remained secure [1,2], providing both baseload and load-following power. As the insidious ecological consequences of global warming became evident, governments worldwide decided to reduce greenhouse gas emissions [3,4]. This entailed moving away from using fossil fuels as the main energy source. Today, renewable energy technologies, such as wind turbines and solar photovoltaic systems, are changing the outlook of the climate crisis. Policymakers around the world support these clean technologies through various instruments (e.g., levies and tariffs) to decarbonize the power sector. Energy system models also consistently predict higher deployment of renewable technologies in the forthcoming decades as they become increasingly affordable [5] while the cost of fossil fuels continues to rise [6].

Unfortunately, the seasonal and short-term variability of these renewable sources (when deployed in large quantities) can introduce instability to the power grid [7,8,9], thereby posing tremendous challenges for system operators in sustaining a balance between electricity supply and demand [10]. The imbalance between supply and demand can lead to power outages if operators establish no contingency plan to dampen these fluctuations [11]. Furthermore, the variability of renewables can raise the cost of electricity, which can directly impact future capacity expansion investment decisions [12]. One way to cope with the variability of intermittent renewable sources is to incorporate new technologies and pathways to satisfy energy service demands. There are alternative energy vectors (e.g., hydrogen, ammonia) that can assist us in transferring energy through spatial and temporal boundaries and introduce flexibility to a carbon-neutral power sector [13]. Although these technologies are analyzed extensively in energy system models [14,15], their roles in the future electricity grid are being overlooked. This shortcoming is due to the fact that energy system models have a lower temporal and spatial resolution than power market models [16], enabling them to encompass higher technological explicitness so that researchers can study the interplay between other energy vectors and renewables. Nonetheless, the strategic impacts of the short-term intermittency in power market models should be implemented in energy system models that optimize long-term investments [17].

Combining power market models, which focus on operational and behavioral problems [18,19], with energy system models, which primarily address strategic planning and investments, can make these integrated models intractable [10]. Doing so requires increasing the temporal and spatial resolution of these integrated models from yearly to hourly or even quarter-hourly. Moreover, accessing real measurements at these high resolutions is not always possible, as there are laws to prevent the misuse of information (e.g., the EU General Data Protection Regulation [20]). This is why simulation models can be useful for generating realistic values based on publicly available weather or climate data [21,22]. Most energy system optimization models are linear or mixed-integer models, which allows them to tackle large-scale problems [23]; however, the need to account for the stochasticity of variable renewable energies and social aspects in these integrated energy system and power market models necessitates the creation of new paradigms of models. Modelers also have to consider the mutual impact of anthropogenic climate change and energy system transition [24]. Thus, it has been suggested that we should soft-couple (or hard-couple when possible) various domain-specific models [25,26]. The borderland separating the power market and energy system models is vanishing as computing power proliferates at a considerable rate, which enables the next generation of energy system models to consider operational complexity with high temporal and spatial detail [27,28].

As complex models intertwine to generate an integrated assessment framework, they become increasingly opaque, turning into black boxes. Therefore, to effectively communicate the results of such frameworks to policymakers and laypersons, modelers need to invent new information-dissemination methods that simplify the qualitative results into intelligible stories that facilitate and support decision-making processes [29].

2. A Short Review of the Contributions in This Article Collection

In this Special Issue, Esmaeili Aliabadi et al. [29] discuss the importance of user-friendliness in future energy system models. The authors suggest modelers invent new user interfaces that exploit storytelling techniques for their computational models, facilitating communications among collaborators with a broad spectrum of backgrounds and generating consensus. To analyze renewable energies at high spatiotemporal resolutions, Lehneis and Thrän [21] propose the ReSTEP simulation model, which can realistically simulate power production from wind turbines.

When spatially detailed energy systems are modeled, the operational and strategic behavior of the players should be reflected. For instance, Algarvio et al. [18] investigate the strategic behavior of prosumers in the local energy market, which is difficult to formulate using conventional energy system optimization models, due to their distributed nature. The authors applied the developed model to the Iberian energy market with Portuguese consumers and suggested that consumers can reduce costs by participating in local citizen energy communities. Using stochastic programming, Lauro et al. [30] concentrated on the decision-making process of hydroelectric power plants, taking into account financial risks. They show a positive correlation between risk aversion and participation in forward contracts, thus conveying the risk to consumers. In addition, the authors show that stochastic modeling is more convenient for lower levels of risk aversion.

The implications of these studies are to promote clean, transparent, and just energy systems, which are essential in the era of algorithms and machine learning [19]. Hao et al. [31] reviewed the existing stock and power market indicators to monitor markets and prevent market manipulation. They created a new framework centered on the power, performance, and behavior of the market, which could be crucial to developing a completely new monitoring system in the electricity market.

To conclude, this collection of articles brings together relevant and complementary perspectives on the subject of modeling electricity markets and energy systems, and presents relevant solutions to some of the most important challenges in the domain. However, modeling electricity markets and energy systems in an unified and solid manner is still a considerable challenge that requires continuous developments in a wide range of multidisciplinary fields to cope with the fast-evolving nature of the energy system.