The Profitability of Energy Storage in European Electricity Markets.

• Author: Spodniak, Petr

1. INTRODUCTION

   Existing energy storage is dominated by hydroelectricity, but the rapid growth in generation from variable renewable energy sources (vRES) is pushing the markets to experiment with new storage technologies that could efficiently support the stability of electrical grids. Grid connected electrochemical energy storage (EES) is envisioned to potentially provide high-value energy services (Dunn et al., 2011). (1) At the same time, any commercial investment into a potential energy storage project must be economically feasible, which means covering investment costs and ofering a reasonable rate of return.

   In this study we focus on the value of energy storage by studying temporal energy arbitrage in electricity day-ahead markets. We define arbitrage practiced by energy storage as an operation strategy that maximises profits, i.e. taking advantage of electricity spot price spreads between hours with varying residual demand. We are particularly interested in the fundamental drivers that explain the magnitude and dynamics of energy storage profitability. Among others, we focus on the effects of intermittent generation from wind and solar, which are changing the dynamics of electricity prices and potentially affecting the value of energy storage.

   Heal (2016) highlights two functions that energy storages have to perform, 1. shifting solar power produced in the daytime to the night (assuming insufficient power resources available at night), and 2. smoothing out the variable output of renewable energy. However, Heal (2016) points out the strengths of spatial diversification of renewable energy sources as well as the power of demand-side management, which both reduce the need for energy storage. Similarly, Newbery (2018) stresses other typically cheaper sources of flexibility in contrast to EES, namely peaking generators, demand-side management, and electrical interconnectors.

   Most of the current literature focuses on spatial and sizing issues of diverse energy storages while concentrating on the minimisation of the system-wide costs of operation and investment (Dvorkin et al., 2017; Panzic et al., 2015; Babrowski et al., 2016). Other studies focus on electricity price arbitrage (Connolly et al., 2011; Zafrakis et al., 2016; Bradbury et al., 2014) of specific energy storage technologies while using different profitability measures, such as internal rate of return. Compared to these studies, we consider a generic storage device defined only by storage capacity constraints and efficiency, i.e. besides the investment-related costs, we mainly neglect start-up and ramping costs and constraints, which is an appropriate abstraction, e.g., for battery-electric or hydro storages.

   By abstracting from a specific storage technology with very different fixed and sunk cost assumptions, we can focus on the variable costs and revenues from operating an energy storage device of a particular efficiency, power and energy capacity. In this context, we mainly focus on which drivers have a significant positive or negative impact on the operational profits of energy storages rather than analysing the profitability in absolute terms. Therefore, in our study we measure the profitability by considering the contribution margins, which are defined as the difference between revenues and variable costs in the spot market. The revenues originate from the sale of electricity when the storage is discharging energy and the variable costs arise from the purchase of electricity when the storage is charging. The contribution margins therefore indicate the profits after variable costs, i.e. the amount of revenues available to cover the fixed costs and investments. Negative or low contribution margins will indicate poor economic performance of investment into energy storage technology and vice versa. Our main focus, however, is on understanding what drives the relative development over time rather than the absolute profitability.

   Our motivations, which are at the same time our research questions, are threefold. First, we want to understand how the contribution margins of illustrative 1-13 MWh/MW energy storages evolved during 2006 to 2016 as the shares of vRES increased in many European markets. For this purpose, we consider storages with a maximum power output of 1MW, which can store energy worth of 1-13h of generation at maximum output. Second, we want to understand the fundamental drivers behind the evolution of the contribution margins. Third, we aim to understand what factors affect the number of charging and discharging hours of the sampled energy storage and how the operating hours are related to the development of profits. Our main reason to look at both storage profits and operating hours and analysing the impact of fundamental drivers on each of them is that the operating hours are a good proxy indicating the system's requirement for storage capacity, whereas the profits are a good proxy indicating whether the markets as they exist today reward storage capacity. Studying both together will allow for identifying potential misalignments between system needs and market design.

   To deliver insights into these questions, in the absence of storage operator profits at plant level, we first develop a storage optimisation model which maximises profits earned by arbitraging price differences in hourly electricity spot markets. As a case study we choose three European electricity markets, namely the UK, Germany and the Nordics. (2) Next, we estimate an econometric ARX-type model that explains the relationships between contribution margins and market fundamentals. In particular, we are interested in whether the profits change over time as the shares of vRES increase. Similarly, we build an econometric Poisson regression model to understand the relationships between operating hours of the energy storage and the market fundamentals, such as vRES generation, electricity demand and fuel prices.

   By using an 11 year-long sample we attempt to capture the structural changes the current power markets are going through. Despite our sole focus on Europe, the three power markets chosen for the analysis are quite diverse and ofer examples to other non-European markets. Germany is the largest power system in Europe and represents an interesting case study of a system traditionally dominated by thermal generation (mainly coal and some gas) with limited hydropower (around 6% of installed capacity), while rapidly integrating vRES. The UK is similar to Germany in the sense of being traditionally dominated by thermal generation and rapidly adopting vRES, though relying much more on gas-fired thermal generation. However, the UK has much less cross-border interconnectors and has implemented different policy mechanisms (capacity markets, carbon price foor). In contrast, the Nordic electricity market has an abundance of flexible hydro generation, where Norway alone has 25 000 times more storage in its dams than the entire British pumped hydro storage (Newbery, 2018).

   Our methodological contribution is the combination of an optimisation model and econometric analysis which enables a better understanding of the drivers affecting economic viability and operation decisions of energy storages. The results presented in this study focus on the contribution margins which comprise a part of the overall investment evaluation. However, by abstracting from a technology-specific analysis of profitability, our results can be further used as inputs into capital budgeting accompanied by additional assumptions on fixed costs, capital costs, and operation and maintenance costs of a specific energy storage technology.

   Finally, our study contributes towards the debate on the increasing importance of energy storage (flexibility) in the future electricity systems, which are increasingly dominated by vRES with close to zero marginal costs. Electricity markets based only on energy pricing may not provide sufficient incentives for storage investments. Hence, they may not be sustainable in the long-run. In addition to energy, flexibility, reliability and capacity will play increasingly important roles, which need to be rewarded as such.

   The paper is structured as follows. Section 2 presents the market setting of the three power markets. Section 3 first describes the storage optimisation model, which quantifes the main dependent variables of interest (storage profits and operating hours), which is followed by a data summary of the independent variables. Section 4 specifes two econometric models. Their results are presented and discussed in section 5. The work ends with conclusions in section 6.

2. MARKET SETTING

   This work focuses on three European electricity markets (Nordic, German, and UK) which are set in specific techno-economic environments exerting influence on the types and levels of risks the energy storage operators face. It is therefore essential to first outline and understand the relevant local factors of electricity supply and demand (3) before proposing relevant determinants of power spreads. By the term Nordic we jointly refer to Norway, Sweden, Finland, and Denmark.

   On the supply side, the power systems in Germany and the UK have traditionally relied on thermal generation (coal, gas, nuclear). However, since the introduction of EU targets for reductions in carbon emissions and the promotion of RES, both countries have seen a rapid growth in capacity and power generation from vRES since 2008 (particularly wind and solar). (4) On the contrary, the Nordic electricity market is a hydro-dominated system with a large share of indigenous generation from biomass, making the adoption of vRES less rapid, compared to the two other cases. With respect to the market design, the UK slightly differs from the two other markets in terms of the introduction of a separate carbon price foor and capacity market mechanisms in 2013 and 2014, respectively. The UK and Nordics are generally...