Cost optimal scenarios of a future highly renewable European electricity system: Exploring the influence of weather data, cost parameters and policy constraints
Introduction
In order to meet the ambitious target of reducing carbon dioxide (CO2) emissions in the European Union by 80%–95% in 2050 compared to 1990 values, the electricity system has to undergo a fundamental transformation (see for instance the Energy Roadmap 2050 from the European Commission [1]). Wind and solar power plants are already today both mature and cost-efficient technology options, which can be scaled up to act as the basis of a low-emission future power supply (see Ref. [2] for an analysis of the increasing cost-competitiveness of renewable power generation technologies, and [3] for a discussion of cost-effective renewable energy options for all EU Member States). The challenges presented by the temporal fluctuations in these resources can be met with low-carbon technologies such as existing hydroelectricity power plants, or with storage options like batteries or hydrogen storage, which still have significant potential for further development. The benefit of the flexibility provided by storage has for instance been studied for a simplified model of a highly renewable European electricity system in Ref. [4], or with a focus on pumped hydro storage and wind power generation for the Irish system in Ref. [5]. In Ref. [6] the authors focus on least-cost combinations of renewable generation and storage for a large regional grid, whereas in Ref. [7] also the role of the spatial distribution and dispatch of storage capacities on continental scale in a European electricity system with 80 power production from variable renewable energies has been studied.
With respect to the spatial variability of weather-dependent renewable generation, large-scale power transmission capacities play a decisive role to provide a smoothing effect and to connect generation capacity at favourable distant locations with the load centres. The systemic advantage of aggregating variable renewable power generation over large distances has already been observed in the pioneering study by Czisch in Ref. [8]. This benefit of transmission has been confirmed in studies using more detailed models, which also take into account limitations and costs of the transmission infrastructure (see for instance [9] for a systematic study with regard to the renewable penetration levels and mixes, [10] for additional detailed backup capacity optimisations, or [11] for a techno-economic optimisation study at high renewable shares.
Given the complexity of such a system, in particular with respect to the spatio-temporal patterns and correlations in the renewable generation and load time series, it is a difficult task to deduce a cost-efficient overall system layout from heuristic principles alone. As a consequence, computational models are a central element in the development of policy guidelines for the design of a future low-emission energy system (see Ref. [12] for a recent review). For each model, choices have to be made regarding the methodological scheme and the scope of its system representation (for instance the spatial and temporal scale, or the choice of energy sectors and technology options considered in the model). The resulting diversity of system models represents a challenge for the interpretation and comparison of the corresponding results, in particular if the underlying input data or modelling details are not publicly accessible. In this context, open energy modelling promises to provide a more transparent and comprehensible scientific approach (see the discussion in Ref. [13] or [14]).
However, even for a single system model for which all input data and modelling details are transparent, the deeper understanding of the numerical results that is fundamental for robust policy advice can be hindered by the dependence of the results on the choice of the model parameters. In particular the details of the input data, the cost assumptions, and the constraints employed in the model all affect the properties of the resulting scenarios in a non-trivial way.
In this contribution these issues are addressed by studying in detail the influence of weather data, cost parameters and policy constraints on the properties of cost optimal scenarios of a future highly renewable electricity system taken from Ref. [11]. It is shown that the total system costs are only weakly affected by the choice of the input weather data or by small changes in the capital costs. The optimisation landscape is flat in many directions, which allows system planners to choose between different near optimal system configurations without a significant increase in total costs. With respect to policy constraints, the investigation of a wide range of CO2 emission limits helps to understand the mechanisms in the cost-efficient interplay of different technology options along the pathway towards a future low-emission electricity system.
The approach taken in the present contribution goes beyond the small variations of a few selected input parameters that are commonly used in the literature to test robustness. Here very large changes to input parameters are considered to try to understand fully what role each system component plays in the cost-optimal system.
In the literature, the consideration of different input assumptions is often limited to the focus of the respective study. For instance in Ref. [11], Schlachtberger et al. focus on the role of transmission expansion for the layout and cost structure of a low-emission European electricity system. The parameter variation is then mainly limited to the constraint for the total transmission capacity as the crucial parameter of the study, but not expanded in a similar way to other model dimensions. Similarly, in Ref. [15] the authors analyse the effects of grid extensions in a renewable Europe, with a focus on the influence of the composition of the power generation in the system. The screening of parameter ranges for the renewable penetration and the mix between solar and wind is the central line of the investigation. In other cases, a sensitivity analysis of modelling results is added to the presentation of a numerical study to assess the robustness of the numerical findings. Such an analysis usually addresses only a few key parameters, but is not central to the understanding of the workings of the system. For instance in Ref. [10], the authors model a renewable-based European electricity system including storage options and concentrated solar power. Choosing different investment costs for storage, grid and backup technologies, they find that although the system composition and operation is highly dependent on these parameters, the overall system cost is only slightly affected.
This article starts with a review of the model and the data inputs in Sec. 2. In Sec. 3 results on the sensitivity to different samples of weather and load data and in Sec. 4 to changes in the cost assumptions are presented. The subsequent Sec. 5 explores the role of policy constraints, such as limits on the expansion of onshore wind, or different CO2 limits. Finally in Sec. 6 the limitations of this study are discussed, before conclusions are drawn in Sec. 7.
Section snippets
Methods
The following sections briefly review the underlying model and data as well as properties of some of the resulting scenarios of a future European electricity system discussed in Ref. [11].
Results I: Sensitivity to the input weather and load time series
The base scenarios presented in Ref. [11] build on time series for renewable generation potential and consumption with hourly resolution for the year 2011. Previous investigations in Refs. [29,30] show that the outcomes of energy system models depend to some degree on the sampling and time resolution of the input data as well as the data reduction method. Similar observations have been made in Ref. [31] in the context of the role of sampling and clustering techniques for offshore grid expansion
Results II: Sensitivity to cost assumptions
Estimates for the future development of costs for technologies that are yet to undergo very large-scale deployment are intrinsically uncertain. In particular solar PV and storage costs could potentially drop significantly over the next decades (see Ref. [17] for an overview of prospective electricity generation costs until 2050, and in particular [37] for a discussion of the cost development of solar PV). Since cost assumptions are a crucial input parameter for the optimisation approach in
Results III: Influence of policy constraints
In Ref. [11] the role of different levels of constraints on transmission capacity expansion for cost-efficient layouts of a European electricity system is investigated. The limited geographic potential of different generation and storage technologies as well as a cap on CO2 emissions entered as fixed constraints into the optimisation. However, for onshore wind a further restriction beyond geographical limits due to public acceptance issues is plausible. In a second investigation different CO2
Discussion: Limitations of the study
This contribution studies the sensitivity of cost optimal scenarios to various influences for the model presented in Ref. [11]. The investigations are explicitly not extended to alternative models, which might apply different methodological approaches, incorporate other sectors (like heating and transport) and other technologies (for instance nuclear generation or carbon capture), or consider a finer temporal or spatial scale. Such alternative models will show different sensitivities, depending
Summary and conclusions
Models of the electricity system give important insights into how to cost-efficiently combine different technology options in the framework given by the physical, environmental, or societal constraints of the system. Even if the methodological approach and the scope of a model is fixed, the simulation results will depend on the assumptions concerning the input data, input parameters, and constraints inside the model. Using the techno-economic optimisation model for the European electricity
Acknowledgements
The project underlying this report was supported by the German Federal Ministry of Education and Research, Germany under grant no. 03SF0472C. Mirko Schäfer is partially funded by the Carlsberg Foundation, Denmark Distinguished Postdoctoral Fellowship. David Schlachtberger, Tom Brown and Martin Greiner are partially funded by the RE-INVEST project (Renewable Energy Investment Strategies – A two-dimensional interconnectivity approach), which is supported by Innovation Fund Denmark (6154-00033B).
References (54)
- et al.
Storage and balancing synergies in a fully or highly renewable pan-European power system
Energy Pol
(2012) - et al.
The technical and economic implications of integrating fluctuating renewable energy using energy storage
Renew Energy
(2012) - et al.
Cost-minimized combinations of wind power, solar power and electrochemical storage, powering the grid up to 99.9% of the time
J Power Sources
(2013) - et al.
Electrical energy storage in highly renewable European energy systems: capacity requirements, spatial distribution, and storage dispatch
J Energy Storage
(2017) - et al.
Parametric study of variable renewable energy integration in Europe: advantages and costs of transmission grid extensions
Energy Pol
(2012) - et al.
Integrated modelling of variable renewable energy-based power supply in Europe
Energy
(2017) - et al.
The benefits of cooperation in a highly renewable European electricity network
Energy
(2017) - et al.
Energy systems modeling for twenty-first century energy challenges
Renew Sustain Energy Rev
(2014) - et al.
Opening the black box of energy modelling: Strategies and lessons learned
Energy Strateg. Rev.
(2018) - et al.
The importance of open data and software: is energy research lagging behind?
Energy Pol
(2017)