A Novel Method for Analyzing Highly Renewable and Sector-Coupled Subnational Energy Systems—Case Study of Schleswig-Holstein

Abstract

The energy transition requires an integration of different energy carriers, including electricity, heat, and transport sectors. Energy modeling methods and tools are essential to provide a clear insight into the energy transition. However, the methodologies often overlook the details of small-scale energy systems. The study states an innovative approach to facilitate subnational energy systems with 100% renewable penetration and sectoral integration. An optimization model, the “Open Sector-coupled Energy Model for Subnational Energy Systems” (OSeEM–SN), was developed under the Open Energy Modeling Framework (Oemof). The model is validated using the case study of Schleswig-Holstein. The study assumes three scenarios representing 25%, 50%, and 100% of the total available biomass potentials. OSeEM–SN reaches feasible solutions without additional offshore wind investment, indicating that it can be reserved for supplying other states’ energy demand. The annual investment cost varies between 1.02 and 1.44 bn €/year for the three scenarios. The electricity generation decreases by 17%, indicating that, with high biomass-based combined heat and power plants, the curtailment from other renewable plants can be decreased. Ground source heat pumps dominate the heat mix; however, their installation decreases by 28% as the biomass penetrates fully into the energy mix. The validation confirms OSeEM–SN as a beneficial tool to examine different scenarios for subnational energy systems.

1. Introduction

To help achieve the 1.5 °C targets of the Paris Agreement [1], the European Union (EU) needs a transformation of energy system based on the smart integration of renewable energy across different sectors. In the European Green Deal, the European Commission stated plans to integrate renewables, energy efficiency, and other sustainable solutions across sectors to achieve decarbonization at minimum cost [2]. The integration of energy systems is often referred to as “sector coupling” [3]. It indicates the combination of multiple energy sectors, such as electricity, heat, and transport, so that the integrated energy system can achieve the target of overall climate neutrality [4].

Decarbonization of heat and transport sectors depends on state-of-the-art techniques, such as power-to-heat and power-to-gas. These techniques, used in a sector-coupled network, are expected to increase the energy storage capacity and provide additional flexibility to the energy system [5]. The modeling of multiple energy sectors, especially power, heat, and transport, is becoming popular in the newer energy models [6]. In the past decade, many researchers analyzed the feasibility of integrating other sectors, especially the heat sector, in 100% renewable energy models. These analyses often show that sector coupling decreases the overall system cost; however, the benefits should be further investigated before cross-border transmission in a sector-coupled EU network is implemented [3,4,5,6].

The North Sea (NS) region can become a pioneer in achieving the European energy transition [7]. The area has enormous offshore wind potential and other renewable sources, such as wave energy, ocean thermal energy conversion, carbon capture and storage, etc. [7,8]. It is also possible to strategically transform the region’s existing grids and networks to accommodate future sustainable solutions [9]. Collectively, the NS region is in a frontrunner position in the European energy transition [10]. The area can, accordingly, be seen as a representative region covering the critical challenges for change to a sustainable energy system and the consequences for the incumbent system. Data, models, tools, and possible solutions from the NS region will, therefore, constitute essential knowledge and methodologies that can be transferred to other areas in transition [11]. The combined use of different modeling tools and concepts to understand and describe different actors’ behavior is essential to provide a coherent picture of the necessary transition process of the energy systems over time and for different spatial levels [10,11].

Modeling methods and tools play a crucial role in providing insights into the energy transition [12]. However, the modeling methodologies often ignore the details of small energy systems. Aggregated models provide more holistic pictures but cannot absorb the regional specifics and often fail to deliver meaningful results at lower spatial scales, including demand behavior. In general, models need not be necessarily larger and more complex; instead, using model collaboration, different approaches and tools are used in conjunction. Aggregated modeling should have proper parameterization and level of detail to capture the main system aspects and interactions. Similarly, disaggregated models should have proper interlinkages with potential developments and system changes across scales [13,14]. The linkages between small- and large-scale energy systems need to be addressed better [15]. Developing a methodology for building models with provisions to represent aggregated and disaggregated energy systems will enable users to select the details based on the analysis’s objective. The study develops a novel method to facilitate aggregated and disaggregated subnational energy systems with 100% renewable penetration and sectoral integration. Therefore, the following research question was formulated:

How to develop methodologies for building subnational models of highly renewable-based energy systems within the sector-coupled networks?

The Open Energy Modeling Framework (Oemof) is a useful tool because of its characteristics, such as flexibility, access to collaborative and interdisciplinary modeling, transparency, reliability, open-source, and open data approach to enhance understanding of energy systems and accelerate the energy transition [16,17,18,19]. Based on the findings of [20], this study selects Oemof for developing a novel method to answer the research question. This article describes how an hourly optimization model using the Oemof is developed and how to use the model to analyze a subnational energy system. The model is validated using the case study of Schleswig-Holstein (SH) in Germany. Schleswig-Holstein, the northern-most federal state of Germany, is increasingly becoming an energy hub between Germany and the Scandinavian countries, due to its geographic location and the ongoing expansion of onshore wind energy [21,22].

Section 2 briefly compares 16 open model tools, discusses Oemof’s usability for developing energy models, and presents Germany’s Schleswig-Holstein as a potential region to validate the subnational energy model. Section 3 describes the architecture of the developed model using Oemof. Section 4 describes the application of the model for the case of SH. The input data and the scenarios are presented. Section 5 compares the scenarios and analyzes the results from the SH case study. Section 6 discusses the developed model’s usability for a broader context and compares the tool with other Oemof-based applications. Section 7 presents the current limitations of the model and future development plans and summarizes the study.

2. Background

Modeling methods and tools play a crucial role in providing the insights mentioned above. A broad range of available state-of-the-art energy models portrays a comprehensible picture of the energy transition over different temporal and spatial levels. However, most of these tools are not “open” or “free for educational use,” limiting the models’ quality, transparency, and credibility. In [20], the author identified 16 “open” tools based on the literature of [23,24], which can be used to model 100%-renewable and sector-coupled energy systems in Europe. The proposed list of tools is presented in

Table 1. The 16 tools for modeling 100% renewable and sector-coupled energy systems. Adapted from [20,23,24].

Sl.	Tool	Methodology (LP/ MILP)	Hourly Resolution	Sectoral Coverage	Demand Response	Investment Decision Support	Top–Down & Bottom–Up Approach	All Storage Inclusion	Net Transfer Capacity	Commodities (Electricity & Heat)	Inelastic Demand	Supply–Demand Modeling	CO2 Cost	CO2 Emission	Source/ Reference
1	Calliope	√	√	-	√	√	-	√	√	√	√	√	√	√	[25]
2	DESSTinEE	-	√	-	-	√	-	-	√	-	√	-	√	√	[26]
3	Dispa-SET	√	√	√	√	√	-	√	√	√	√	√	√	√	[27]
4	ELMOD	√	√	√	-	√	-	√	-	√	-	√	√	√	[28]
5	ficus	√	-	√	-	√	-	√	-	√	√	√	√	√	[29]
6	LEAP	-	-	√	-	-	√	√	-	√	-	√	√	√	[30]
7	LUSYM	√	√	-	√	-	-	√	-	-	√	√	√	√	[31]
8	MEDEAS	-	-	√	-	-	-	-	-	√	-	-	√	√	[32]
9	OSeMOSYS	√	√	-	√	√	-	√	-	-	√	√	√	√	[33]
10	PowerGAMA	√	√	-	-	√	-	√	-	-	√	√	-	-	[34]
11	PyPSA	√	√	√	√	√	√	√	√	√	√	√	√	√	[35]
12	RETScreen	-	-	-	-	√	√	-	-	√	√	√	√	√	[36]
13	SIREN	-	√	-	-	-	-	√	√	-	√	√	-	√	[37]
14	SWITCH	√	√	√	√	√	-	√	√	-	√	√	-	√	[38]
15	urbs	√	√	√	√	√	-	√	√	√	√	-	√	√	[39]
16	Oemof	√	√	√	√	√	√	√	√	√	√	√	√	√	[40]

.

2.1. Modeling Energy Systems Using Oemof

The background paper by Hilpert et al. [16] describes how Oemof can facilitate open science in energy system modeling. The article discussed the scientific contribution, concept, architecture, implementation, and usage of Oemof. Figure 1 presents a graphical representation of how to describe an arbitrary energy system using Oemof [16].

There are two types of nodes in Oemof: components, and buses. Every component has to be connected with one or more buses. The connection between a component and a bus is the flow. The main components of Oemof are Sources, Sinks, and Transformers. The Sources have only outflows. For example, solar photovoltaics (PV), wind turbines, and biomass commodities are modeled as Sources. The Sinks only have inflows. Consumer demands such as electricity or heat loads are modeled as Sinks. Transformers have both inflows and outflows. For example, heat pumps can be modeled as Transformers, which receive electricity inflow and convert it to heat outflow. There are also other components, such as ExtractionTurbineCHP, GenericCHP, Link, GenericStorage, GenericCAES, SinkDSM, etc., which are designed in the Oemof Solph package [41].

There are three ways to create an optimization problem based on Oemof:

• The energy system describes a graph with flows on its edges by combining necessary components and buses;
• The basic energy system is adapted by defining additional constraints on top of the aforementioned graph logic; and
• Custom components are added to a model by subclassing from the core or creating from scratch.

The use cases can be separately or combinedly used in an energy model, allowing maximum flexibility. Oemof provides existing functionalities to build energy models for varying scales. In addition, it enables the combination and adaptation of different energy models to create tools with specific research objectives. It is suggested for readers to go through [16] for further details on the usefulness, usability, and applications of Oemof.

2.2. Schleswig-Holstein as a Prospective Subnational Region

For activities and successes in expanding renewable energies, the Schleswig-Holstein achieved first place in the federal state comparison 2019–2020 by the Agentur für Erneuerbare Energien (AEE) [21]. In 2018, electricity generation from renewable energies in SH reached around 150%, which is almost four times Germany’s national average of 38%. SH takes a leading position in the expansion of electricity generation from renewable energies. The share of renewable energies in SH was almost 15.8% in the heating sector, slightly above the Germany-wide share of 14.4%. When it comes to the percentage of renewable energies in gross final energy consumption, SH’s 36.6% is well above the national average of 16.5%. SH aims to generate at least 37 terawatt-hours of electricity from renewable energies by the year 2025. Figure 2 shows the individual energy sources’ shares in the total final energy supply contribution of renewable energies 2018 [22].

Due to its geographical conditions, SH is predestined for the use of wind energy. In SH, wind turbines with a nominal output of around 8.2 GW were installed by the end of 2018, which means that electricity from wind energy makes up the largest proportion of SH’s electricity supply from renewable energies. SH considers the expansion of wind energy to increase to at least 25 GW by 2030. Biomass represents one of the largest shares (32.5%) of renewable energies in SH’s supply contribution. Wood, energy crops, straw, and biogas can sustainably generate a significant proportion of the energy requirement. SH is well-suited for solar systems, as the increased amount of wind between the seas provides natural cooling. The potential of geothermal energy in SH is particularly suitable for space heating and electricity production. Due to the lack of landscape conditions, water traditionally plays a subordinate role as an energy source in Schleswig-Holstein. Since geothermal energy does not depend on the weather or the course of the day, it is ideal for covering the base and medium loads in the heating and electricity markets [21,22].

The geological subsurface is suitable for storing considerable amounts of energy carriers (e.g., hydrogen, synthetic methane), potential energy, or thermal energy. In theory, porous geological layers and cavities in the subsurface can be used for storage. The latest energy models to analyze the SH energy system should consider the compressed air energy storage capacities in geological formations. Due to the problem-free storage and the diverse and flexible application possibilities, hydrogen is a perfect link in the sector coupling. Hydrogen can make a significant contribution to the decarbonization of these areas of application and comply with the climate policy CO₂ reduction targets by 2050 [22].

Two points are clear from the literature review: First, Oemof can be used to develop methodologies for building energy models ranging of varying geographical and temporal scopes for highly renewable energy integrated energy systems with interlinked sectors. Second, Schleswig-Holstein is an ideal subnational region to validate the developed energy model for its prospects in expanding renewable energies across all sectors. Therefore, the study used Oemof to build an open sector-coupled subnational energy model and validated it for SH’s case.

3. Model Architecture

3.1. Elements and Objective Function

The study developed a unique hourly optimization tool using a hybrid approach. The technological capacities are exogenously set, and the investment capacities are endogenously resolved. Technical limits set the boundary of the system so that the solutions are realistic. The subnational model, “Open Sector-coupled Energy Model for Subnational Energy Systems” (OSeEM–SN), was created using Oemof Tabular [42]. Figure 3 illustrates the OSeEM–SN energy model.

The model presents SH as a self-sufficient energy system, where the demands are met using its renewable resources. In the real case, other energy systems are connected with the SH system; for example, the neighboring region’s electricity bus will be connected with SH’s electricity bus using a transmission line (transshipment approach), which uses the Oemof class Link. For visualizing such energy systems, the readers are suggested to look into [43], where the author connected two energy systems using Link. The model also does not consider industrial process heating because of limited data availability and model complexity of high-temperature technologies. The nodes (i.e., components and buses) of the OSeEM–SN model, according to the Oemof Classes, are presented in

Table 2. List of all components and buses of the OSeEM–SN model.

Oemof Class	Nodes	Remarks
Bus	Electric Bus	Represents grid or network without losses.
	Heat Bus
	Fuel Bus
Sink	Electricity	Represents the electricity and building heat demands in the energy system.
	Space Heat
	Domestic Hot Water (DHW)
Source	Offshore Wind	Represents the volatile generators of the energy system.
	Onshore Wind
	Solar PV
	Hydro Run-of-the-river (ROR)
	Biomass	Represents the biomass commodities which are fed into the CHP plants.
ExtractionTurbineCHP	Combined Heat and Power (CHP)	Represents the heat generators of the energy system. The OSeEM–SN model uses extraction turbines and uses only biomass as the fuel.
Transformer	Air Source Heat Pump (ASHP)	Complements CHP for meeting heating demands.
	Ground Source Heat Pump (GSHP)
GenericStorage	Li-ion (Li-ion)	Represents batteries.
	Vanadium Redox Flow (Redox)
	Adiabatic Compressed Air Energy Storage (ACAES)	Simplified model as Generic Storage. Presents electricity storage.
	Hydrogen (H2)
	Pumped Hydro Storage (PHS)	Storage units with constant inflow and possible spillage. The storage capacity is not expandable.
	Thermal Energy Storage (TES)	Simplified model as Generic Storage. Presents heat storage in sensible hot water tanks.

.

The OSeEM–SN model follows the formulation described by Hilpert [44]. The model limits the volatile generators, biomass commodities, and storage capacities by putting maximum limits. However, the limit for using heat pumps depends on the electricity availability. OSeEM–SN uses a perfect foresight approach, indicating the weather and renewable data are provided in advance. Detailed mathematical formulations for the OSeEM–SN model follow the modeling equations presented by Maruf [43]. The model optimizes the operating and investment costs of all the volatile generators, combined heat and power (CHP), heat pumps, and storages. The endogenous variables are presented using $x$, and the exogenous parameters are presented using $c,$ as listed in

Table 3. Variables and parameters for cost optimization of OSeEM–SN.

Variables/Parameters	Description	Technology
$x_v^{flow}$	Flow of volatile generator unit $v$	Offshore Wind Onshore Wind Solar PV Hydro ROR
$x_v^{capacity}$	Capacity of volatile generator unit $v$
$c_v^{marginal\_cost}$	Marginal cost 1 of volatile generator unit $v$ [43]
$c_v^{capacity\_cost}$	Capacity cost 2 of volatile generator unit $v$ [43]
$x_{chp}^{flow}$	Flow of CHP unit $chp$	CHP
$x_{chp}^{capacity}$	Capacity of CHP unit $chp$
$c_{chp}^{marginal\_cost}$	Marginal cost of CHP unit $chp$
$c_{chp}^{capacity\_cost}$	Capacity cost of CHP unit $chp$
$x_h^{flow}$	Flow of heat pump unit $h$	ASHP GSHP
$x_h^{capacity}$	Capacity of heat pump unit $h$
$c_h^{marginal\_cost}$	Marginal cost of heat pump unit $h$
$c_h^{capacity\_cost}$	Capacity cost of heat pump unit $h$
$x_s^{flow}$	Flow of storage unit $s$	Li-ion Redox ACAES H2 PHS (No Investment) TES
$x_s^{capacity}$	Capacity (power) of storage unit $s$
$x_s^{storage\_capacity}$	Storage capacity (energy) of storage unit $s$
$c_s^{marginal\_cost}$	Marginal cost of storage unit $s$
$c_s^{capacity\_cost}$	Capacity cost (power) of storage unit $s$
$c_s^{storage\_capacity\_cost}$	Storage capacity cost (energy) of storage unit $s$

¹ The marginal costs are calculated based on variable operation and maintenance costs, carrier costs, and the efficiency. ² The capacity costs are calculated based on fixed operation and maintenance costs and the annuity.

.

The objective function of OSeEM–SN is created from all instantiated objects, which use all operating costs and investment costs arguments:

$$\begin{array}{l}min:{\displaystyle \sum }_{v,t}\stackrel{operating\_cost Volatile Generator}{\overbrace{x_v^{flow}\left(t\right)·c_v^{marginal\_cost}}}+ {\displaystyle \sum }_v\stackrel{investment\_cost Volatile Generator}{\overbrace{x_v^{capacity}·c_v^{capacity\_cost}}}+\\ {\displaystyle \sum }_{chp,t}\stackrel{operating\_cost CHP }{\overbrace{x_{chp}^{flow}\left(t\right)·c_{chp}^{marginal\_cost}}} + {\displaystyle \sum }_{chp}\stackrel{investment cost\_CHP}{\overbrace{x_{chp}^{capacity}·c_{chp}^{capacity\_cost}}} +\\ {\displaystyle \sum }_{h,t}\stackrel{operating\_cost Heat Pump }{\overbrace{x_h^{flow}\left(t\right)·c_h^{marginal\_cost}}} + {\displaystyle \sum }_h\stackrel{investment\_cost Heat Pump}{\overbrace{x_h^{capacity}·c_h^{capacity\_cost}}} +\\ {\displaystyle \sum }_{s,t}\stackrel{operating\_cost Storage}{\overbrace{x_s^{flow}\left(t\right)·c_s^{marginal\_cost}}} +\\ \hspace{1em}\hspace{1em}\hspace{1em}{\displaystyle \sum }_s\stackrel{investment\_cost Storage}{\overbrace{x_s^{capacity}·c_s^{capacity\_cost}+x_s^{storage\_capacity}·c_s^{storage\_capacity\_cost}}}\end{array}$$

(1)

3.2. Development Methodology

OSeEM–SN considers the existing capacities of the volatile generators and pumped hydro storage (PHS). The model also assumes that the current biomass and biogas capacities are converted to CHPs. The maximum potentials of the resources limit the investment capacities of additional volatile generators and CHPs. The storage capacities are expandable to their utmost limits, except PHS, where no further capacity expansion is possible. Heat pump expansion depends upon the availability of power from renewable electricity resources. The model also uses cost and demand data as inputs of the model. After investment and dispatch optimization, the model outputs such as investment capacities can be obtained. Details of the model input and output are presented in Figure 4.

The OSeEM–SN model was developed using a Python Script, Oemof Solph, Oemof Tabular, and tabular csv data files. The development methodology steps for OSeEM–SN are stated below:

• Importing necessary data packages (Python Script)
• Setting up the input datapath (Python Script)
• Setting up the result directory (Python Script)
• Reading input data (Python Script)
• Creating the energy system (Oemof Solph)
• Creating the buses (Oemof Solph)
• Adding buses to the energy system (Oemof Solph)
• Adding components to the energy system (Oemof Tabular)
• Reading demand data (Oemof Tabular)
• Creating the model (Oemof Solph)
• Solve the optimization problem (Oemof Solph)
• Postprocessing of results (Oemof Tabular)
• Writing results (Oemof Tabular)
• Plotting results (Oemof Tabular)

Data preparation is an essential step of the model development process, where the data are normalized and scaled for use as the inputs of OSeEM–SN. For different scenarios, the input data are varied in the data handling stage.

4. Model Validation: Case of Schleswig-Holstein

4.1. Hourly Renewable Profiles and Demand Data

The OSeEM–SN model is validated using historical data for a full year. According to [45], data from 2011 are used for analyzing 2050 scenarios, except for hydro data, which uses 2016 data.

Table 4. Hourly input data sources for the OSeEM–SN model.

Data	Source	Remarks
Wind profiles	Renewables Ninja project [46]	Based on the MERRA-2 dataset.
Solar PV profiles
Hydro ROR inflow	Dispa-SET project [47]	-
PHS scaled inflow
Electricity demand	OPSD project [48]	Based on the ENTSO-e statistical database [49].
Space heat demand	OPSD project [48]	Based on the When2Heat dataset [50]

shows the hourly input data sources used for validating the OSeEM–SN model.

The onshore wind profile is obtained from the MERRA-2, current fleet dataset for the NUTS-2 region (SH: DEF0). The offshore wind profile represents the offshore profile of Germany based on the MERRA-2 database. The solar PV profile is also obtained for SH (NUTS2, DEF0). The hydro run-of-the-river (ROR) and PHS scaled inflows are obtained from Dispa-SET’s 2016 data. The inflows are defined as the contribution of exogenous sources to the level (or state of charge) or the reservoir. Scaled inflows are normalized values of the inflow concerning the nominal power of the storage unit. The PHS inflows are scaled down to match SH’s inflow profile (in MWh). Germany’s demand data (electricity, space heat, domestic hot water (DHW)) are downscaled based on population to represent SH’s hourly demand profiles. The wind, solar, and hydro normalized profiles do not change in the scenarios and can be visualized as shown in Figure 5.

Figure 6 shows the normalized demand profiles of SH in 2050. The total electricity demand for SH in 2050, based on the representative year, is 18.6 TWh_el. Total space heat demand is 18.6 TWh_th, and the DHW demand is 4 TWh_th. The amount of available biomass is calculated from the Hotmaps project [51]. The study assumes that the existing biomass and biogas power plants are converted to CHP plants by 2050. CHP’s electrical and thermal efficiencies are assumed to be 45%, and the condensing efficiency is assumed to be 50%. The coefficient of performance (COP) of the air source heat pump (ASHP) and ground source heat pump (GSHP) are assumed to be 2.3 and 3.9, respectively [52].

4.2. Capacity and Available Potential

The existing capacities and available potentials for the volatile generators and the storage investments are taken from different sources, namely the Hotmaps project [51], ANGUS II project [52], Deutsche WindGuard [53], AEE [54], and LIMES-EU project [55], as listed in

Table 5. Capacity and available potential for volatile generators and storage in SH in 2050.

Technology	Existing Capacity	Available Potential
Onshore Wind [GWel]	7 [53]	1.9 [54]
Offshore Wind [GWel]	1.7 [53]	25.2 1 [55]
Solar PV [GWel]	1.6 [54]	6.7 [54]
Hydro ROR [MWel]	2 [54]	4 [54]
Biomass & Biogas	1 GWh [54]	21.8 PJ [51]
Li-ion [MWel]	-	782.5 [52]
Redox [MWel]	-	46.5 [52]
H2 [MWel]	-	505 [52]
ACAES [MWel]	-	1715.5 [52]
PHS [MWel]	120 [56]	-
TES [MWth]	-	1000 2

¹ The maximum offshore wind potential, according to the LIMES-EU project, is 83.6 GW_el. The available potential of SH assumes the equal distribution of remaining capacities in the three northern states of Germany. ² Own assumption.

. The available potentials are calculated from the maximum potentials and the existing capacities. The Li-ion, vanadium redox flow (Redox), and hydrogen (H₂) potentials are assumed to be 5% of Germany’s available potentials, as stated in the project databases. The adiabatic compressed air energy storage (ACAES) potential is assumed to be 50% of Germany’s total potential because of its availability in only Northern Germany.

4.3. Cost Data

Table 6. Cost data for OSeEM–SN Model [43].

Technology	Onshore Wind	Offshore Wind	PV	ROR	Biomass	Li-ion	H2	Redox	PHS	ASHP	GSHP	ACAES	TES
Capex (€/kW)	1075	2093	425	3000	1951	35	1000	600	2000	1050	1400	750	0
Lifetime (Years)	25	25	25	50	30	20	22.5	25	50	20	20	30	20
WACC	0.025	0.048	0.021	0.05	0.05	0.05	0.05	0.05	0.05	0.05	0.05	0.05	0.05
VOM Cost (€/MWh)	0	0	0	0	11.3	1	1	1	0	0	0	1	0
FOM Cost (€/kWh)	35	80	25	60	100	10	10	10	20	36.75	49	10	0.38
Storage Capacity Cost (€/kWh)	-	-	-	-	-	187	0.2	70	-	-	-	40	38
Carrier Cost (€/MWh)	-	-	-	-	34.89	-	-	-	-	-	-	-	-

presents the cost data taken from various resources as described in [43].

4.4. Other Input Data

The loss rates are 1% for PHS and 1.4% for thermal energy storage (TES) [52]. Hydro ROR efficiency is 90% [52]. The roundtrip efficiencies are 92%, 80%, 46%, 75%, 73%, and 81% for Li-ion, Redox, H₂, PHS, ACAES, and TES, respectively [52]. The maximum state of charge capacity in terms of hours at full output capacity is 6.5 h, 3.3 h, 168 h, 8 h, 7 h, and 72 h for Li-ion, Redox, H₂, PHS, ACAES, and TES, respectively [52]. Land limitation for onshore wind is 4 MW/km², and offshore wind is 6 MW/km² [55]. The solar PV installations consider the protection of nature reserves and restricted zones.

4.5. Scenarios

The study assumes three scenarios for validating the OSeEM–SN model for Schleswig-Holstein: BM-25, BM-50, and BM-100. The scenarios represent 25%, 50%, and 100% of the total available biomass potentials, respectively. The study aims to investigate how the results change upon varying one parameter of the model. However, the model does not account for all the parametric variations for the input data; rather, it focuses on the model’s usability to create different scenarios and examine different possible pathways.

5. Results

5.1. Supply–Demand Matching

The OSeEM–SN model reached feasible solutions for all three scenario assumptions. Figure 7 shows the supply–demand matching of electricity and heat demands for the three different scenarios over the year 2050.

The energy generation from onshore wind in BM-25 is 26.6 TWh_el, which drops to 20.9 TWh_el in the BM-50 and BM-100 scenarios. Similarly, solar PV generation drops from 8.7 TWh_el in the BM-25 scenario to 6.8 TWh_el in the BM-50 scenario and 4.6 TWh_el in the BM-100 scenario. Offshore wind generation remains the same, 5.5 TWh_el, in all three scenarios. The CHP generation for electricity and heat increases with increasing biomass availability in the scenarios, from 1.5 TWh in the BM-25 to 3 TWh in the BM-50 scenario and 3.8 TWh in the BM-100 scenario. In contrast, heat pump (GSHP and ASHP) generation reduces from 21.5 TWh_th in BM-25 to 18.9 TWh_th in the BM-100 scenario. Therefore, it is obvious from the scenarios that, with increasing biomass penetration in the energy mix, the CHP plant capacities are expanded, increasing electricity and heat generation. This, in turn, reduces the expansion of other power plants and heat pumps to meet the demands.

5.2. Scenario Comparison

5.2.1. Capacity Expansion

According to the optimization from OSeEM–SN, the required investment of different technologies can be obtained.

Table 7. Comparison of capacity expansion for three scenarios.

Technology	Scenario-Wise Investments
	BM-25	BM-50	BM-100
Onshore Wind [GWel]	1.9	0	0
Offshore Wind [GWel]	0	0	0
Solar PV [GWel]	6.7	4.9	2.7
Hydro ROR [MWel]	4	4	4
CHP [GW]	1	1.6	1.9
GSHP [GWth]	5	3.8	3
ASHP [GWth]	1.3	2.1	3.1
Li-ion [MWel]	782.5	782.5	782.5
Redox [MWel]	46.5	46.5	46.5
H2 [MWel]	397	0	0
ACAES [MWel]	357.1	357.1	357.1
TES [MWth]	1000	460.2	0

compares the required investments on top of the existing capacities of Table 5 for the three scenarios.

The results show no need for additional investment in offshore wind plants to meet SH’s energy demand. As a result, the offshore capacities can be reserved for supplying other states’ energy demand, especially those in Southern Germany. CHP investment rises because of the higher availability of biomass over the three scenarios and the high overall efficiency due to the combined production of electricity and heat. This impacts the investment in onshore wind and solar PV capacities and reduces investments in volatile generators. GSHP investment also decreases with increasing CHPs; however, ASHP investment increases to complement the heating demand. For storage, Li-ion, Redox, and ACAES are used to their maximum investment capacities for all three scenarios. Hydrogen storage is used only in the BM-25 scenario, indicating its use only in low biomass availability cases. The need for TES storage decreases over the scenarios with more biomass availability. Therefore, with limited biomass, it is possible to meet the heat demand with a heat storage option.

5.2.2. Investment Cost

The study calculates the annual investment cost (AIV) by multiplying the annuity with the model’s optimized capacity, as shown by Equation (2). The annuity calculation is shown in Equation (3), which considers the capital expenditure (Capex), weighted average cost of capital (WACC), and the lifetime ($n$). The total investment cost (TIV) is obtained by multiplying the Capex and the model’s optimized capacity, as shown by Equation (4). In the case of storage, the investment cost considers both power and energy costs.

$$c^{AIV}=c^{Annuity}·c^{optimized\_capacity}$$

(2)

$$c^{Annuity}=c^{Capex}·\frac{(c^{WACC}·{\left(1+c^{WACC}\right)}^n}{({\left(1+c^{WACC}\right)}^n-1)}$$

(3)

$$c^{TIV}=c^{Capex}·c^{optimized\_capacity}$$

(4)

Figure 8 compares the volatile generators’ investment cost, i.e., wind, solar PV, and hydro ROR plants. We see that the total investment cost for the volatile generators in SH decreases by 76% (4.9 bn € vs. 1.1 bn €) over the scenarios. The annual investment cost decreases from 262.7 mn €/yr to 61.5 mn €/yr with the increasing biomass availability.

Figure 9 compares the investment cost of CHP plants and heat pumps (GSHP and ASHP). Overall, the total investment cost increases by 7% (10.6 bn € vs. 11.3 bn €) over the scenarios. The annual investment cost increases from 819.2 mn €/yr in the BM-25 scenario to 853.7 mn €/yr in the BM-100 scenario.

Figure 10 compares the storages’ investment cost, i.e., Li-ion, Redox, H₂, ACAES, and TES. The total investment cost for the storages in SH decreases by 69% (4.5 bn € vs. 1.3 bn €) over the scenarios. The annual investment cost decreases from 357.6 mn €/yr in the BM-25 scenario to 105.1 mn €/yr in the BM-100 scenario.

The total investment in volatile generators, CHPs, and storages is 20.1 bn € in the BM-25 scenario. The investment decreases by 22% in the BM-50 scenario (15.6 bn €) and 30% in the BM-100 scenario (13.9 bn €). The annual investment cost decreases accordingly, from 1.44 bn €/yr in the BM-25 scenario to 1.02 bn €/yr in the BM-100 scenario. Figure 11 illustrates the total investments for different scenarios from OSeEM–SN optimization results.

5.2.3. Energy Mix

Figure 12 compares the energy mix results of the OSeEM–SN model. Figure 12a compares the electricity generation from the combined (i.e., existing and new) capacities. The onshore wind generation dominates the energy mix because of high wind availability in SH. The onshore wind electricity generation varies between 20.9 TWh_el and 26.6 TWh_el for the three scenarios. The model does not suggest installing new offshore capacities, because of two reasons: (i) the cost of offshore is higher, and (ii) the demand is already met using other resources. However, this is only valid for SH’s subnational case, where plenty of renewable resources are available. The scenario will be different for a larger case with a lack of adequate renewable resources. The offshore electricity generation from the existing capacities is the same for the three scenarios, 5.5 TWh_el. The hydro ROR electricity generation also remains the same, 0.016 TWh_el, for all three scenarios. Solar PV-based electricity varies from 8.7 TWh_el in the BM-25 scenario to 4.6 TWh_el in the BM-100 scenario. Overall, the electricity generation decreases by 17% (42.4 TWh_el vs. 34.9 TWh_el) from the BM-25 scenario to the BM-100 scenario. Therefore, with high biomass-based CHPs in the energy mix, the curtailment from other variable renewable energy plants can be decreased. Figure 12b compares the heat generation from the combined CHP capacities and new heat pump capacities. GSHPs dominate the heat mix; however, the installation of GSHP decreases as the biomass penetrates more into the energy mix. From the BM-25 to the BM-100 scenario, while the CHP-based heat generation increases by 154% (1.51 TWh_th vs. 3.84 TWh_th), the GSHP installation decreases by 28% (19.8 TWh_th vs. 14.1 TWh_th). However, the demand is also complemented by ASHPs, which increase from 1.69 TWh_th in the BM-25 scenario to 4.79 TWh_th in the BM-100 scenario.

The analysis based on OSeEM–SN model results is summarized below:

• SH has adequate renewable resources to meet its electricity and building heat demands.
• The onshore wind dominates electricity generation.
• Electric heat pumps, mainly GSHPs, dominate heat generation.
• The batteries offer short-term storage solutions for electricity storage.
• ACAES, H₂, and TES are promising storage solutions, especially when renewable energy availability is limited.
• Power-to-heat devices, such as GSHP and ASHP, stand out as prominent heating options besides traditional CHPs.
• TES plays an important role in integrating the power and heat sectors.
• Increasing biomass in the system impacts other technologies’ investment costs and can reduce the overall system cost.
• The optimization reached feasible solutions without utilizing the full potential of many resources. Therefore, the high amount of available potential, especially offshore wind resources, emerges as a promising alternative for powering up other parts of the country, especially Germany’s high energy-consuming industrial southern states.

6. Discussion

The increasing penetration of the energy system with predominantly volatile generation units leads to a far-reaching techno–economic and social transformation process. The advancement of the energy transition leads to the fact that power generation plants, instead of being located in the consumer centers’ immediate vicinity, are increasingly located in places with the best site conditions for the respective plants. With the increasing feed-in of supply-dependent renewable energy, consumption and generation are falling apart in time, placing new demands on the system. Due to its special location situation, Schleswig-Holstein already represents a system that can be considered exemplary for future energy systems, with shares of 100% renewable energies. Analyses of future energy systems with a very high share of renewable energies, in which consumption and generation must be coordinated, can already be carried out today using Schleswig-Holstein as an example. Schleswig-Holstein can thus already act as a blueprint for future renewable energy systems.

Model-based analyses play a central role in the analysis of such increasingly complex energy systems. In addition to purely technical modeling, open science approaches for the reproducibility of scientific results have become increasingly important in recent years. Therefore, models must address the diverse issues in the field of energy system analysis and, at the same time, meet high scientific standards in the sense of open science. To accompany the energy transition’s progress and transfer it to the heating sector, an energy model is needed to map the complex interdependencies between the individual sectors. It makes sense to make the model and its databases freely accessible to enable scientists, political decision-makers, entrepreneurs, and citizens, alike, to use and further develop it.

Highly renewable-based energy systems at the subregional level are becoming popular in energy transition studies. Such studies include investigation of 100% renewable energy systems in Hvar (Croatia) [57], Samsø (Denmark) [58], the Åland Islands (Finland) [59], Orkney (Scotland) [58], the Canary Islands (Spain) [60], California (USA) [61], and New York State (USA) [62]. It is more common to analyze electricity systems than integrated energy systems.

The project aims to develop a source-rich and flexible tool for modeling highly renewable and sector-coupled subnational energy systems and validate them in the context of the Schleswig-Holstein system. The OSeEM–SN model, consisting of scripts and tabular database concepts, was developed to achieve the outlined requirements. The basic concept serves the general description of sector-coupled energy systems. It is based on a graph-theoretical approach, which provides a generic basis for different simulation and optimization models. This concept was implemented within the Oemof framework, using the optimization library Oemof Solph. Comprehensive applications using Oemof are renpass [63], openMod.sh [64], and reegis^hp [65]. However, their usability is limited. For instance, renpass is an electricity market model, openMod.sh focuses on one region, and reegis^hp evaluates district heating and CHP [16]. In contrast, the scope and upgradability of OSeEM–SN are broader and simpler. The model’s openness allows other users to actively develop the tool and modification for specific purposes to answer research questions within a subregional framework. The tool is also useful for investigating island energy systems using only renewables. Therefore, the OSeEM–SN model is unique for analyzing power and heat sector-coupled energy systems at the regional level. With the gradual integration of more sectors and upgradation of the model, the tool can be useful for researchers and policymakers to answer energy transition-related questions using a straightforward approach. The development of the OSeEM–SN model confirms the findings by Hilpert et al. [16], i.e., the flexibility in a new tool development using Oemof allows adjustments and changing research objectives, avoiding lock-in effects.

7. Conclusions

7.1. Limitations of the Model

The model is in its early stage of development, and, therefore, the analysis conducted in this study is subject to certain limitations. Since the results highly depend on the inclusion of different technologies and demands, the results may change based on the newer version of the model with more components and demands. This will also broaden the scope of the model to use it for different geographical contexts. The current version of the OSeEM–SN model does not consider

• Geothermal, ocean and wave energy, concentrated solar power plants, etc.;
• Industrial process heating demands;
• Transmission line modeling;
• Latent and thermo–chemical heat storages as TES options;
• Interconnection with neighboring regions;
• Modeling of electric vehicles, coupling of the transport sector, and provision of vehicle-to-grid charging;
• Renewable heating options, such as using solar thermal collectors; and
• Demand response management.

Nevertheless, as a continuous development of the model, future versions will gradually include different technologies, demands, and other components into the model.

7.2. Future Plans and Summary of the Study

The upgradation of OSeEM–SN can be outlined as follows:

• Grid modeling between subnational regions;
• Modeling of process heating components in industries;
• Modeling of district heating components;
• Inclusion of concentrated solar, geothermal, ocean, and wave energy plants;
• Inclusion of latent heat and thermo–chemical heat storage plants;
• Modeling of the transport sector;
• Modeling of demand response activities; and
• Development of a graphical user interface.

The following research questions, among others, can be investigated with the upgraded version of the OSeEM–SN model:

• Which combination of plants is suitable for economic operation?
• What added value does sector-coupling of the energy sources bring?
• Which inconsistencies exist in the subregional energy systems?
• What role may dispatchable loads or CHPs play as flexibility options?
• How can regional energy markets be designed?
• Where do key bottlenecks arise in the power grid of a subnational system?
• How can the industry sector be decarbonized?
• What role will storage play in a highly renewable and sector-coupled subnational energy model?

This study’s main objective is to develop methodologies for building subnational models of 100% renewable and sector-coupled energy systems. The model used Oemof Tabular to develop a state-of-the-art tool to analyze subnational energy systems. The model consists of the basic renewable resources and storage options to meet the electricity and building heat demands. Oemof Solph’s functionalities are used for the core part of the model. Simple Python scripts and tabular data files allow the user to change the model’s input details and analyze energy systems for different scenarios.

The study validates the model for Schleswig-Holstein’s energy system analysis. The OSeEM–SN model optimized three different scenarios for different available biomass potentials. The model could reach feasible solutions for all three scenarios, indicating the feasibility of a 100% renewable and power-building, heat-coupled energy system for Schleswig-Holstein. Analysis of the results shows that, with increasing biomass availability, volatile generator investment decreases. The increasing amount of biomass-based CHP plants impacts both the electricity and heat generation mix.

The study also identifies the current limitations of the model. Since the model is based on Oemof, the model’s upgradation is possible using different functionalities under the Oemof. The model has been validated with SH’s case study, and the results have been analyzed in detail in the study. Therefore, the OSeEM–SN model is presented as a beneficial tool to create different scenarios and examine different possible pathways for subnational energy systems of similar contexts.