ArticlePDF Available

Regionally integrated energy system detailed spatial analysis: Groningen Province case study in the northern Netherlands

February 2023
Energy Conversion and Management 277

February 2023
277

DOI:10.1016/j.enconman.2022.116599

License
CC BY 4.0

Authors:

Somadutta Sahoo

Delft University of Technology

Joost Van Stralen

Christian Zuidema

University of Groningen

Jos Sijm

ECN part of TNO

Show all 5 authorsHide

Regional level energy system analyses and corresponding integrated modeling is necessary to analyze the impact of national energy policies on a regional level, while considering regional constraints related to energy infrastructure, energy supply potentials, sectoral energy demands, and their interactions. Nevertheless, current literature on energy system analysis largely overlooks the regional level. In response, this study provided a systematic approach to refining and improving the spatial resolution of an existing regional energy system modeling framework. The methodology involved creating regions and nodes within the modeling framework under categories corresponding to land use (cities and other regions), energy supply, and energy infrastructure. We established a unidirectional soft linking with geographical information system-based modeling results allocating spatially sensitive elements, such as renewable resources or heat demand. We provided a detailed breakdown of sectoral energy demand, supply options, and energy infrastructure for electricity and heat, including district heating (DH). This framework explicated regional differences in terms of demand–supply mismatch, supply options, and energy infrastructure. Our case study of the Dutch province of Groningen demonstrated clear differences compared to the previous crude regional model, with, e.g., an increased role of biomass (+460 % change) and decreased role of solar (- 59 %), while cities with high heat demand densities and/or compact structures exhibited serious DH penetration, ranging from 11 to 21 %. The systematic steps allow for the replication of the model in other regional analyses. Our framework is complementary for energy system analysis at the national and pan-European levels and can assist regional policymakers in decision-making.

Content uploaded by Somadutta Sahoo

Content may be subject to copyright.

Energy Conversion and Management 277 (2023) 116599

Regionally integrated energy system detailed spatial analysis: Groningen

Province case study in the northern Netherlands

Somadutta Sahoo

, Joost N.P. van Stralen

, Christian Zuidema

, Jos Sijm

, Andr´

e Faaij

Department of Spatial Planning and Environment, Faculty of Spatial Sciences, University of Groningen, the Netherlands

Energy Transition Studies, Netherlands Organization for Applied Scientic Research (TNO), Amsterdam, the Netherlands

Energy Research and Sustainability Institute Groningen, Faculty of Science and Engineering, University of Groningen, the Netherlands

ARTICLE INFO

Keywords:

Regional energy system

Integrated energy system modeling

Energy infrastructure

Geographical information system

Renewable resources

District heating

ABSTRACT

Regional level energy system analyses and corresponding integrated modeling is necessary to analyze the impact

of national energy policies on a regional level, while considering regional constraints related to energy infra-

structure, energy supply potentials, sectoral energy demands, and their interactions. Nevertheless, current

literature on energy system analysis largely overlooks the regional level. In response, this study provided a

systematic approach to rening and improving the spatial resolution of an existing regional energy system

modeling framework. The methodology involved creating regions and nodes within the modeling framework

under categories corresponding to land use (cities and other regions), energy supply, and energy infrastructure.

We established a unidirectional soft linking with geographical information system-based modeling results allo-

cating spatially sensitive elements, such as renewable resources or heat demand. We provided a detailed

breakdown of sectoral energy demand, supply options, and energy infrastructure for electricity and heat,

including district heating (DH). This framework explicated regional differences in terms of demand–supply

mismatch, supply options, and energy infrastructure. Our case study of the Dutch province of Groningen

demonstrated clear differences compared to the previous crude regional model, with, e.g., an increased role of

biomass (+460 % change) and decreased role of solar (−59 %), while cities with high heat demand densities and/

or compact structures exhibited serious DH penetration, ranging from 11 to 21 %. The systematic steps allow for

the replication of the model in other regional analyses. Our framework is complementary for energy system

analysis at the national and pan-European levels and can assist regional policymakers in decision-making.

1. Introduction

National targets related to renewable energy addition and emission

reduction are stringent within the European Union [1]. Efforts to add

infrastructure, particularly related to renewables, need planning at a

regional level, especially in densely populated countries such as the

Netherlands [2]. A region in this context refers to a geographical area

beyond a municipality or city level and to a subnational level. Regional

energy system analysis is necessary to understand regional differences in

supply sources potential, energy demand, and energy infrastructure, i.e.,

integrated energy system. In addition, a regional analysis is necessary in

the decision-making process related to energy systems, for example,

what should be the investment in medium voltage (MV) cables or where

should these cables be expanded, i.e., capacity addition [3]. Answering

these and other regional energy system-related questions requires an

energy system model (ESM) equipped with appropriate spatial detail.

This model can act as a regionally-relevant decision support tool related

to land use and infrastructure planning. In addition, a regional ESM

should have the capability to translate the implications of national

policy choices at the regional level.

Regional energy system modeling can be a crucial addition to energy

system analysis at a pan-European, national, and local level. In addition,

aligning modeling outcomes and inputs of each level allows for a

comprehensive analysis of the overall energy system. Currently, regional

level analyses and their links with the (inter)national level analyses have

received little attention in the energy system-related literature. Regional

models are mostly benecial in identifying regional constraints related

to energy infrastructure, renewable energy supply potentials, and their

interactions. A regional model can support decision-making upon future

policies, both regarding energy systems and the spatial and environ-

mental context in which these are embedded. In response, developing a

* Corresponding author.

E-mail address: somadutta.sahoo@rug.nl (S. Sahoo).

Contents lists available at ScienceDirect

Energy Conversion and Management

journal homepage: www.elsevier.com/locate/enconman

https://doi.org/10.1016/j.enconman.2022.116599

Received 7 September 2022; Received in revised form 2 December 2022; Accepted 14 December 2022

Energy Conversion and Management 277 (2023) 116599

regional ESM can greatly benet if it allows for the inclusion of spatially

sensitive parameters, which may be informed by policy considerations.

Spatially sensitive parameters that have a strong impact on ESM

results, for example, sectoral heat demand or renewable energy poten-

tial, cannot be properly captured with a low spatial resolution [4].

Similarly, having a high spatial resolution leads to computational

complexity and issues with getting high quality data related to the

regional allocation of parameters within an ESM, for example, related to

the built environment (BE) energy demand in every region [4]. The

spatial resolution should be appropriate to represent regional differ-

ences in energy demand–supply mismatches, renewable potentials, and

energy infrastructure, particularly related to district heating (DH),

which can signicantly impact overall regional energy balances, infra-

structure costs, and planning. Additionally, renewable spatial potential

can vary signicantly under different spatial policies and land-use

constraints at a regional level [5]. Our literature review into regional

energy system modeling revealed a lack of information and guidance on

choosing appropriate spatial resolutions for analyzing various important

components of a regionally integrated energy system. Similarly, our

review indicated a gap in providing guidance on systematic steps to

incorporate these regional spatial categorizations into an ESM.

Ideally, regionally integrated energy system models should simul-

taneously analyze heat and electricity, and their corresponding in-

frastructures, with high and low spatial resolutions, respectively. For

example, the national or European level is typically considered suitable

for analyzing electricity infrastructure [6], whereas heat network ana-

lyses, particularly those related to low temperature DH, are highly

geographically detailed and demand regional or local analysis [7]. DH

network feasibility and corresponding investments are dependent on the

spatial proximity of demand and supply due to huge transmission losses

[8]. Again, our literature review showed that regional energy infra-

structure analyses simultaneously considering different spatial domains

and resolutions are currently lacking, which provides us yet another

research gap.

1.1. State-of-the-art review

A geographical information system (GIS)-based tool helps perform a

detailed and spatially sensitive energy system analysis. Some integrated

ESMs and related literature do exist that allow for interaction with GIS.

However, this application is more common in energy infrastructure

analysis. Review of the state-of-the-art shows GIS and the energy system

model MARKAL were linked to study the feasibility of hydrogen (H

)

infrastructure at the pan-country level [9], for example. Similarly, GIS

and the ENERGYPLAN model combination have been used to under-

stand the DH expansion potential within Denmark [10,11]. GIS was

applied to identify locations for installation of heat pumps (HPs) in

Denmark in combination with the TIMES-DK model [12]. Furthermore,

GIS and the MARKAL model have been combined to identify the po-

tential of the carbon dioxide (CO

) infrastructure in the Netherlands,

including offshore [13,14]. These abovementioned studies, however, do

not explain their choice for a certain spatial resolution or geographical

scope. Similarly, they lack interaction of an ESM with other aspects of an

integrated energy system, such as various sectoral demand or supply

options.

Recent literature at varied geographical scopes were reviewed to

further identify GIS and energy system model interaction on other as-

pects of energy system mentioned before. Most of this literature focused

on a city level. For example, a GIS-based platform was conceptualized

[15] and utilized [16] for city where electricity generation and con-

sumption were simulated, along with storage under different scenarios.

Similarly, GIS interacted with thermal and electricity energy system

models to identify energy savings and emission reduction potential of

buildings on a city district level [17]. GIS was coupled with a system

dynamics model to identify wind capacity potential in Latvia [18].

Modeling results from these analyses are focused on a single topic rather

than on holistic approach towards analyzing the entirety of energy

system.

There have been regional analyses at a country level within an ESM

environment. For example, in Italy, analyses have been performed on

power and mobility sector [19], mobility infrastructure [20], renew-

ables deployment [21], and deep decarbonization [22]. Similarly,

related to Great Britain, studies have been performed with emphasis on

analyzing uncertainties related to meeting decarbonization targets [23],

investigating supply sources for power sector [24], and exploring the

role oating offshore wind within electricity system in 2050 [25].

Considering other countries, researches have been carried out with

emphasis on the mobility sector in Germany [26] and focus on renew-

ables penetration in the electricity supply within Greece [27]. Studies

with only regional analysis have been performed with emphasis on

waste and energy crops in central Sweden [28], for example. Models

have also been developed for an analysis of multiple energy demanding

sectors, but the regional categorization is rather crude, for example,

[29]. Literature such as [30] acknowledge the need for regional energy

system analysis to understand the potential of locally available re-

sources, such as geothermal, or identify regional constraints imposed by

spatially-sensitive parameters. Readers are directed to [4] for a review

on further studies related to national models with multi-regions and

Nomenclature

Acronyms

BE Built Environment

CAPEX Capital Expenditure

CBS Central Bureau of Statistics

CHP Combined Heat and Power Plant

DH District Heating

ESM Energy System Model

FBI Food and Beverage Industry

GBPV Ground-based Photovoltaics

GIS Geographical Information System

HP Heat Pump

HV High Voltage

IWH Industrial Waste Heat

KEV National Energy Outlook (in Dutch)

KNMI Royal Netherlands Meteorological Institute (in Dutch)

MV Medium Voltage

MSW Municipal Solid Waste

NG Natural Gas

OPERA Option Portfolio for Emission Reduction Assessment

O&M Operation and Maintenance

OPEX Operational Expenditure

RNL Rest of the Netherlands

STEG Steam and Gas Turbine (in Dutch)

Units

GW gigawatt

km kilometer

square kilometer (km*km)

€

/year million euro per year

PJ petajoule

Formulas

carbon dioxide

hydrogen

S. Sahoo et al.

Energy Conversion and Management 277 (2023) 116599

fully regional models. Still, not all aspects of energy system are covered

simultaneously, including spatial and land-use planning-related issues.

In response, our methodological approach offers novel suggestions upon

how to respond to this gap. Our focus is on a regional integrated system

analysis linked to a national energy system, the advantages of which

have already been documented in [4].

Our research add to GIS-based independent studies that identify

future variable spatial potentials for renewable energy deployment

considering land-use or planning-related constraints, for example for

solar [31], wind [32], and biomass [33]. This addition is not only to

identify capacity potentials, but also to incorporate these spatial po-

tentials into an integrated ESM. Such a linkage allows for studying the

feasibility and constraints of renewable energy deployment and for

realizing their potentials in relation to other parts of the energy system,

such as, regional energy infrastructure or sectoral demand.

1.2. Research objective and questions

Based on the literature, state-of-the-art review, and the research gaps

identied, our objective was to develop a systematic approach that al-

lows for performing a detailed spatial analysis of a regional energy

system within a national integrated modeling context. This was done by

rening and improving the spatial resolution of an existing regional

energy system modeling framework through capturing the key regional

spatial parameters via land-use regions or nodal segregation with

detailed emphasis on energy demand, supply options, and infrastruc-

ture. With this objective, the following research questions were

formulated:

•How can a regional ESM be developed that captures detailed spatial

constraints and boundary conditions regarding sectoral energy de-

mand, supply potentials, energy infrastructure, and their in-

teractions, while targeting both heat and electricity along with their

different preferred spatial resolutions?

•What are the signicant differences between a high spatially detailed

ESM and an existing ESM (with a low spatial detail) on regional

energy balances, energy carrier ows, and cost structures?

A key novelty of our method is improving the spatial representation

of a regional ESM taking into account all the specics of regional de-

mand, energy supply potentials, energy infrastructure, and associated

costs. While doing so, our regional ESM is linked to national and pan-

European level for electricity infrastructure, allowing the analysis of

different future congurations or scenarios on the regional scale. As

such, another key novelty is that our method allows for analyzing the

potential impact of various policies and regional constraints on the

regional energy system and spatial claims, including network capacities,

sectoral demands, and supply potentials. Therefore, this study is a major

improvement in the methodological approach to understanding the

regional energy system. Our regional ESM can function as a key decision

making tool for both energy policies and spatial policies at a national

and regional level.

This paper takes a major step toward improving the state-of the-art of

the integrated energy system modeling and analysis through detailed

regional modeling. Methodologically, this study is novel in offering in-

formation on how to identify an appropriate spatial resolution for

investigating energy infrastructure, land-use, and energy supply regions.

Replicability is also facilitated through introducing a systematic

approach for inclusion of relevant spatial detail on the abovementioned

aspects. Key novelties include the creation of a DH network with a new

unique infrastructure on a pan-provincial level, considering future city

planning regarding placing corresponding centralized heat technology

options. Novelty also lies in soft linking a detailed GIS-based analysis [5]

with an existing crude regionally categorized energy system model [4]

to create a spatially-detailed regional energy system model covering

most of the aspects of a regionally integrated system. Result-wise,

appropriate spatial detail allowed us to observe signicant differences

in overall regional energy balances, energy demand, and supply po-

tentials. Additionally, our regional ESM spatially represent the primary

energy supply mix, interregional electricity ow, and investment in

infrastructure, particularly in the DH network. Thus, our regional ESM

can provide explicit and detailed inputs for both national and provincial

energy system and spatial policies, beyond what was possible through

previous ESMs.

This study used the regional Option Portfolio for Emission Reduction

Assessment (OPERA) model which is a Dutch-based model, already

available, and has a modeling structure for regional and nodal catego-

rization [4]. Our case study was the province of Groningen in the

northern Netherlands. Section 2 presents the proposed method with a

detailed representation of the modeling framework. Section 3 describes

the modeling results related to regional analyses. Section 4 presents a

critical discussion on our methodology and results. Finally, Section 5

Fig. 1. Structure and ow of the method and results section.

S. Sahoo et al.

Energy Conversion and Management 277 (2023) 116599

presents our conclusions and future work. Fig. 1 illustrates the structure

and ow of the method and results section.

2. Method

Our modeling method is based on the use of the optimization model

OPERA [34], which was developed in AIMMS 4.84 software [35],

whereas our GIS model is based on QGIS 3.10 [36] and ArcMap 10.5

Fig. 2. Methodological framework for this study.

Fig. 3. Categorization of regions and nodes created in OPERA, illustrating what these regions represent, what they contain, and what major OPERA activities are

performed in these regions. Numbers within parenthesis represent the number of regions/nodes in each category. Regions explicitly created for this study are

provided inside the dashed square.

S. Sahoo et al.

Energy Conversion and Management 277 (2023) 116599

[37] software. OPERA is a Dutch-based national integrated energy sys-

tem model, where regions within the country are inherent, i.e., hard-

linked, to make it a regional model [4]. The major fundamental opera-

tion of this paper are creating regions and nodes to spatially represent

land-use regions, industries, geothermal doublets, and energy infra-

structure explicitly under corresponding categories.

The second major

operation is soft linking GIS modeling results with the regionalized

OPERA model. These two contributions were essential to allow for a

detailed spatial analysis on regional energy balances, regional primary

energy supply mixes, regional cost structure, and interregional energy

ows for the province of Groningen. Our modeling framework devel-

opment consists of three core activities: (1) creating regions in OPERA,

which is described in Section 2.1, (2) providing regional allocation for

energy-demanding sectors and supplying options, which is described in

Section 2.2, and (3) energy infrastructure modeling along with related

nodal categorization in Section 2.3. Fig. 2 illustrates the methodological

modeling framework used in this study.

2.1. Region creation

The allocation of regions needs to consider both data availability and

computational complexity. Too high a resolution can lead to difculties

in data availability and allocation, while extra effort is needed to con-

nect these regions through appropriate energy infrastructure for linking

demand and supply. Even when data is available and a GIS-based

analysis can provide a high spatial resolution with a large number of

similar regions, these may not necessarily be sensible to use as explicit

regions within a regionally integrated energy system analysis. The

computational complexity increases as more regions are added; for

example, the number of variables increases exponentially in an ESM

with the addition of regions. Hence, combining regions from a GIS

analysis may be preferable. For nding a balance between spatial detail,

data availability, and computational complexity, approximately 100

land-use regions and nodes were included for the province of Groningen,

an area of 3000 square kilometers (km

) and about 600, 000 inhabitants,

in this study (Fig. 3). The regions mostly correspond to energy demand

and supply, whereas the nodes are mainly linked to energy infrastruc-

ture. Nodal structure creation and addition techniques in ESM have been

previously applied [4,6].

Data availability is an important criterion for creating regions. In the

Netherlands, government agencies such as the Central Bureau of Sta-

tistics (CBS) [38] produce data at a minimum resolution of a munici-

pality level. Therefore, the ten municipalities in the province of

Groningen were rst considered as a set of independent land-use regions

(ranging from 50 to 200 km

, which can vary in other regional contexts)

– see Fig. 4 (A). Subsequently, within a municipality, all population

centers with over 10, 000 inhabitants were considered as distinct re-

gions, i.e., municipalities containing such centers would be split into

multiple regions consisting of a city (the population center) and the

remaining part of the municipality (municipality rest) – see Fig. 4 (B).

We did so as these centers had much higher population densities and had

large pockets of land having a heat demand density greater than 1200

gigajoule per hectare [5], which may allow for creating a DH network as

suggested by Persson et al. [39] (Table 1). Furthermore, BE-related de-

mand data are available for these current centers from CBS as most of

these centers were municipalities in the past (see Table 1). Nine inde-

pendent population centers were selected and Groningen city (230, 000

inhabitants) was split into two population centers: Groningen inner city

and outer city, as the inner city has a much higher heat demand density

compared to the outer city, which is relevant for our DH analysis (see

Section 2.3.2).

Each land-use region was balanced for net heat and electricity de-

mand in each time slice (Fig. 3). Our regional map were created in GIS,

in line with OPERA, by intersecting maps of population centers and

municipalities. Groningen inner city was manually delineated by over-

laying the heat density map from [5], the open street map, and the

building map of Groningen city.

In addition to creating these land-use regions using OPERA, other

onshore regions that were created in [4] were included, namely Drenthe,

Fig. 4. (A) Categorization of the province of Groningen municipalities; (B) Shortlisted population centers for analysis.

Regarding time resolution, OPERA (and in this paper) uses a time slice

approach where hours (within a year) having similar characteristics of energy

demand and supply are grouped together – see [34] for further details.

S. Sahoo et al.

Energy Conversion and Management 277 (2023) 116599

Friesland, and the rest of the Netherlands (RNL). Additionally, the model

included the Dutch part of the North Sea, i.e., the northern and western

parts of the region (Fig. 3). Adjoining countries, namely, Germany,

Norway, and Denmark, were created as land-use regions because of their

high voltage (HV) electricity connections with Groningen. These cross-

border connections through electricity infrastructure are described in

detail in Section 2.3.1.

2.2. Regional allocation

All land-use regions created in this study need regional allocation for

OPERA demand and supply-related activities. Accordingly, this section

categorizes sectoral demand (Section 2.2.1) and supply options (Section

2.2.2).

2.2.1. Energy demand

We considered the regional allocation for three energy-demanding

sectors: BE, industries, and agriculture. Mobility is allocated based on

the regional population distribution, similar to [4]. The energy demand

Table 1

Categorizations of regions or nodes created using the OPERA model with in-

formation on the number of regions and nodes created, the underlying reasons

for such a categorization, and their selection criteria.

Category Reasoning and selection criteria (if required)

Municipality This level was selected as most of the regional data for

the Netherlands is available at this level. Further,

adding ten municipalities present within the province of

Groningen in OPERA did not present major issues

considering computational time requirements. In

addition, some decision-making related to investments

in infrastructure related to energy, renewables, and

building stocks takes place at the municipality level.

Municipalities also have a say in the future creation and

demolition of buildings stocks. Also, installing

infrastructure related to wind and solar requires

clearance from the municipality [40].

Cities or population

centers

Larger population centers have been identied in our

study as these have a high concentration of the BE, and a

high BE heat demand density allows us to study the

feasibility of a DH network on a city level. A threshold of

10, 000 inhabitants was set as a criterion for shortlisting

cities for analysis. This led to a selection of 9 centers,

which is still manageable in terms of both data

collection and computational time. Furthermore, these

cities have a future heat demand density greater than

1200 gigajoule per hectare (based on [5]) which is in

line with what Persson et al. [39] considered to be

suitable for DH networks. In addition, CBS provided

historical sectoral demand data on these regions as CBS

provides data on the municipality level as the lowest

level of categorization, and these centers are mostly ex-

municipalities or the only major energy-demanding

regions within a municipality.

The remaining part of a

municipality

These represent regions of a municipality outside of

population centers. If a municipality does not have a

center, then the whole municipality comes under this

category. Therefore, ten regions were created related to

the municipality rests. On the supply side, these regions

have centralized renewable infrastructure related to

mainly electricity production, namely ground-based

photovoltaics (GBPV) and wind, which are expected to

be located in this region in the future. Additionally,

biomass, such as forest, nature, and wet manure are

associated with this region. Infrastructure related to

renewables is suitable in this region. Furthermore, these

regions have the BE, which is not a part of any city, and

are scattered throughout the municipality. No attempt

has been made to consolidate them into new population

centers.

Offshore regions and

abroad

Two offshore regions were created with their explicit

wind proles. These regions’ segregation is based on

planned offshore wind farms in the Dutch part of the

North Sea [4]. Abroad regions were created for

accommodating their electricity import/export prole

with the Netherlands based on a pan-European

electricity market model COMPETES – see [4] for

further detail. A total of four abroad regions were

created.

Industry The industry represents distinct nodes in our research.

important and relevant regional industries were

identied based on [4] and current research. These

industries are scattered throughout the province of

Groningen and exist in clusters or as individual

industries (also see Fig. 5). A total of eleven industrial

nodes were created based on their spatial location.

Creating these explicit nodes allows us to link energy

infrastructure, particularly electricity (Section 2.3.1),

and identify their energy demand structure (Section

2.2.1). Fixing their locations also helped us to link their

industrial waste heat (IWH) potentials to cities with the

help of DH networks (Section 2.3.2). Having industries

as separate nodes allowed us to investigate secondary

energy balances and primary energy use of each of these

nodes.

Geothermal doublets Geothermal heat potential in the province of Groningen,

and the Netherlands, is suitable for spatial heat

Table 1 (continued )

Category Reasoning and selection criteria (if required)

applications of the built environment (BE) [41].

Therefore, we planned to link these regions to major

population centers through the DH network for the

supply of low temperature heat to the BE (Section

2.3.2). For DH linking purposes, we needed specic

geothermal locations, or geothermal doublets, rich in

technical potential with low economic costs. Currently,

there are no doublets in the province and no concrete

plans are there to set up doublets in the near to medium

future [42]. Therefore, technical potential and

economic cost maps available from ThermoGIS [43]

were used to nd suitable doublet locations and identify

their potential. Overlaying these maps showed that only

a small region in Het Hogeland municipality is suitable

from technical and economic viewpoints. A doublet

effective distance is a range of 2–3 kilometers (kms)

[44] also accounting for the distance between

production and reinjection wells. These nodes were

created manually by carefully aggregating the technical

potentials of cells surrounding potential doublets. The

technical and economic potential of various locations

were manually calculated by trial and error method

within the feasible region to pinpoint the locations of

geothermal doublets. Three doublets were created

based on space limitations due to technical and

economic constraints. Having three doublets is also

manageable for linking heat infrastructure in the

OPERA model.

Electricity nodes Electricity nodes, both high voltage (HV) and medium

voltage (MV), were created to adequately represent

electricity infrastructure. HV nodes were used for

connecting cities or population centers, major industrial

clusters, and connections to centralized electricity

supply sources in the remaining part of municipalities.

MV nodes were used for making the nal connections to

the abovementioned regions as the HV network cannot

be directly connected to most of these regions. We

created a total of 18 and 21 HV and MV nodes,

respectively. Section 2.3.1 details the spatial locations

and connections of these nodes.

DH nodes Explicit DH nodes were created to make appropriate DH

connections to the whole of each city or population

center, shortlisted in this research, starting from their

respective city centers, which were ten in number.

Additionally, a few additional nodes were created, a

total of four in number, as explicit centralized DH

supply source locations. Section 2.3.2 explains their

locations. For the remaining cities, either an industrial

node was used as a supply source location, or a heat

connection was established via another major adjoining

city.

S. Sahoo et al.

Energy Conversion and Management 277 (2023) 116599

Table 2

Sectoral distribution of activity along with regional allocation method and

criteria for these activities (data requirements for these activities and assump-

tions associated with methods are provided).

Sector Activity

name/

category

Data

requirement/

availability

Regional

allocation

method and

criteria

Major

assumptions

production

volumes and

energy

production per

unit volume

were updated

- Only for the

newly added

activities in

the OPERA

database,

reference

production

processes were

introduced.

These

processes

competed with

generic

sectoral energy

supply options.

Agriculture Heat

demand

- CBS regional

data on

greenhouse

areas

- 2050 data on

arable land

from [5]

- GIS regional

map

- CBS data on

arable land

- CBS

municipality

level data on

greenhouse

areas is linked

to various

municipality

rests. The heat

and electricity

regional

demand share

is similar to the

regional

greenhouse

share and

corresponds to

that in [4]

- Intersecting

the arable land

map 2050 with

the regional

map to obtain

regional arable

land and

correcting for

mismatch with

the projection

of CBS arable

land historical

data.

- Mobile

machinery

energy

demand is

proportional to

the regional

arable land

- Agriculture

is assumed to

be absent in

cities

Electricity

demand

Mobile

machinery

S. Sahoo et al.

Energy Conversion and Management 277 (2023) 116599

2.2.2. Energy supply

In this section, we provide regional allocations for energy supply

options linked to the following renewables: solar PV, wind, geothermal,

and biomass. We also describe some major energy supply technology

options that produce a variety of secondary carriers not using the

abovementioned renewables. Additionally, we investigate the IWH po-

tential of the industries analyzed in Section 2.2.1. We directly used in-

puts from our previous study on renewable energy potentials for these

Fig. 5. (A) Industrial activity category used in this study and (B) Segregation of already added industries in [4] and newly added industries in this study, showing

industrial clusters, created using OPERA.

Fig. 6. Locations where wind speed proles were considered, overlaid on the wind 2050 progressive scenario map obtained from [5].

S. Sahoo et al.

Energy Conversion and Management 277 (2023) 116599

resources, which was based on considering the existing and future de-

velopments in land use and spatial policies inuencing what may be

possible in the province [5]. We used the most optimistic (least con-

strained) 2050 progressive scenario from this previous study. No addi-

tional region or node was created for energy supply, except for

geothermal. For regional allocation of renewables (except geothermal),

we intersected the progressive scenario and GIS municipality maps, and

the potentials were allocated to the remaining part of each municipality

(Fig. 7) unless stated otherwise. The energy potentials for regions

outside the province of Groningen remained the same as those in [4].

Wind

Apart from the regional distribution of wind future spatial potential

from [5], an additional methodological step was introduced to estimate

the wind speed prole for the remaining part of each municipality

(Fig. 7). For this, one location was identied, which would represent the

central location of the onshore wind feasible region in each municipality

(Fig. 6). These locations should coincide with Royal Netherlands Mete-

orological Institute (KNMI, in Dutch) data points [45], where KNMI

refers to the meteorological agency of the Netherlands.

Explicit data

points were considered instead of the average wind prole of each

feasible region because averaging leads to compromise of actual peaks

and ebbs at various time periods which would have led to a poor esti-

mation of wind farm capacity and wind energy supply potential – refer

to Appendix A for detailed analysis on this method.

Solar

For solar ground-based photovoltaics (GBPV), a common and exist-

ing prole

for the whole of the Netherlands was used as the solar spatial

potential does not vary as much as the wind potential. Regional BE

buildings mapping was used for rooftop PV allocation. The projected

rooftop space of the current buildings was calculated using the area

function in GIS. The method used in [5] was applied for future pro-

jections. Inland oating PVs, mainly on stagnant water bodies, are

gaining attention in the Netherlands. Therefore, regional allocation of

oating PV potential was calculated based on the current inland

standing water proportion, assuming that this space will not change in

the future – see Appendix A for region-specic potential calculation. The

potential of other PV types included in OPERA is considered the same as

in [4], and no attempt was made to regionally allocate these PVs. In

addition to the above-mentioned electricity-producing PV options,

OPERA considers heat production from solar thermal options in

different sectors. However, no regional allocation is made for this

option.

Geothermal

The Upper Rotliegend geothermal stratigraphic layer was considered

for analysis as this is the only shallow layer available for the province of

Groningen with high heat potential [41]. A maximum of three probable

geothermal doublet locations were manually identied as they have

high potential compared to the surrounding locations within the prov-

ince by overlaying the potential recoverable heat and economic poten-

tial maps provided by ThermoGIS [43]. Table 1 further details the

selection criteria of these doublet locations, and Fig. 10 shows these

locations. In terms of modeling, each doublet was made an individual

node in OPERA and linked to the created DH network (see Section

2.3.2). Additionally, these nodes were connected to the MV network

(Fig. 3) for providing electricity to pumps responsible for extracting heat

– see Appendix A for further detail.

Biomass

Before this study, biomass production-related activities were not

well represented in OPERA. All biomass type energy balances occurred

at the national level. A variety of biomass types were introduced in [5]

which could not be tackled by existing biomass-related energy carriers

and the corresponding technology options in OPERA. Four biomass-

related energy carriers were introduced in OPERA: wood chips incor-

porating forest and nature biomass, biomass straw, energy crops, and

grass incorporating grass rening in arable land and grassland, along

Fig. 7. Summary of unidirectional information ows from GIS to OPERA for renewables on the supply side. GIS and OPERA interlinking activities are illustrated in

detail for renewables. The 2050 renewable potential maps were obtained from [5].

This organization is responsible for the Dutch national weather forecasting

service. The primary tasks of KNMI are monitoring of climate changes, weather

forecasting, and monitoring seismic activity.

This prole is the same as the prole used in [4]. The source of this prole is

[75].

S. Sahoo et al.

Energy Conversion and Management 277 (2023) 116599

with verge grasses (Fig. 7). Energy carriers related to manure, both dry

and wet, already existed in OPERA.

All newly added energy carriers were regionally balanced, where a

region is a municipality with a movement distance of 5–25 kms. The

conversion from land-use related to these energy carriers to biomass

energy content is based on [5]. The production costs of each of these

carriers were determined based on a literature review of biomass wood

chips and straw [46], energy crops [47], and grass [48]. Technology

options utilizing these energy carriers were introduced, along with their

technical characteristics, inputs/outputs, and cost structure. The outputs

and the corresponding technology options from energy crops, wood

chips, and straw were heat boilers or combined heat and power plants

(CHPs). Additionally, energy crops can produce ethanol. Grass was

converted into biogas using specic grass-based digesters. These bio-

masses were allocated to municipality rests or industrial nodes because

biogas can be utilized by the BE or industries. The preferred location of

energy crops was industrial nodes and grass and wet manure were

municipality rests. Biomass wood chips and straw were linked to DH

energy supply-related nodes (see Section 2.3.2). As dry manure has a

low production volume at the municipality level, it was not allocated to

any particular region or node, but was considered within the entire

province and was balanced at the national level. GIS-based regional

allocation was performed for most of the carriers (Fig. 7). Household and

municipal solid waste (MSW) produced within the province of Gronin-

gen were allocated to the Delfzijl industrial node because the only MSW

incinerator in the province is located at this node.

Other energy supply technology options

These options can be divided into two categories. (1) These options

are not a part of any energy-demanding sector. For example, electricity

is produced from an H

steam and gas turbine (STEG, in Dutch), which is

a highly efcient electricity-producing turbine with H

as the input.

Similarly, technology options associated with centralized heat supply to

a DH network are also a part of this category. (2) These options are a part

of energy-demanding sectors and are responsible for meeting the

sectoral energy demand. For example, heat is produced from options

such as HPs, hybrid boilers, and electric boilers in the BE or industrial

sectors (see [4] for more details).

Industrial waste heat

The province of Groningen has a strong potential for IWH because of

the presence of a variety of industrial clusters and individual industries

at various spatial locations. IWH can make a signicant contribution to a

provincial DH network – see Appendix A for IWH potential calculation

method. Data for newly added industries in our analysis rely on existing

industrial subsector categorization as it is difcult to nd IWH-related

literature for each activity. Glass ber is a specic case that is not

covered by existing categorization and has a high IWH potential owing

to its high operating temperatures. A separate industrial subcategory

was created in OPERA for glass ber and calculated its IWH potential

based on a case study in Germany [49] and adjusted the production

volumes according to our case.

2.3. Energy infrastructure analysises

Every land-use region and node were connected using energy infra-

structure. Fig. 3 illustrates the detailed OPERA activities related to en-

ergy infrastructure in different nodes and Table 1 lists their selection

criteria. The electricity network is already represented in OPERA [4],

although not in signicant spatial detail, as discussed in Section 2.3.1.

For the heat network, no modeling structure is available, particularly

related to DH, which we modeled, including the key technical and

economic constraints (Section 2.3.2). Fig. 8 illustrates the GIS-related

activities and their interlinkages with OPERA in detail.

2.3.1. Electricity network

The electricity network within Europe is interconnected. Therefore,

electricity ows in the Netherlands are affected by adjoining European

countries with which the Netherlands is connected via the HV network.

As OPERA is an integrated energy system model of the Netherlands, it

Fig. 8. Summary of unidirectional information ows from GIS to OPERA for energy infrastructure. GIS and OPERA interlinking activities are illustrated in detail

related to energy infrastructure.

S. Sahoo et al.

Energy Conversion and Management 277 (2023) 116599

does not perform pan-European power ow analysis. To consider elec-

tricity import–export proles with these countries, OPERA is soft-linked

with a pan-European electricity market model, COMPETES [6]. For

consistency purposes, a similar high renewable scenario (see [4]) is

considered in both models. For the adjoining countries that were added

in this study (see Section 2.1), electricity import–export proles of these

countries were used from the COMPETES model run. OPERA does

analyze electricity ows at the regional and the national level. The

network spatial categorization and representation is presented in more

detail below.

High voltage network

Earlier, OPERA incorporated the HV network cost as a function of

distance, along with capacity constraints between regions and network

losses [4], although these distances were simple straight-line connec-

tions between nodes. This study improved the spatial representation of

the network for the province of Groningen based on the actual physical

electricity network through GIS (Fig. 8). In OPERA, the province of

Groningen has HV network connections with multiple regions within the

Netherlands and with the Dutch part of the North Sea. Therefore, the

deployment of renewables in Groningen will impact the network ca-

pacity of other regions. Henceforth, the capacity ranges for the whole of

the Netherlands were improved, i.e., 1.25 times of the current capacity

(latest values on current capacity was obtained from the HV network

service provider for the Netherlands [50]), to accommodate an increase

in capacity due to the deployment of large scale renewables. The model

determined the actual increase in various connection capacities through

energy system optimization. This increase was compared with the

planned capacity addition of these connections so that network capacity

shortages or constraints can be better identied.

For HV network connections, the priority was to connect all mu-

nicipalities by creating at least one HV node in each of them. The

emphasis was on nodes representing actual major interconnections,

which were responsible for electricity supply to cities via the MV

network, electricity supply from renewable regions located in the

remaining part of municipalities, electricity supply to industrial clusters,

and connection to the offshore electricity network. A total of 18 nodes

was created for the HV network for the province of Groningen (Fig. 9).

Appendix A further details these connections and nodes for the province

of Groningen.

Nodal connections in GIS was relied on maps from Dutch trans-

mission system operator, TenneT, for onshore regions [51] and a future

scenario analysis developed by the Netherlands Environmental Assess-

ment Agency for offshore regions [52] - Fig. 9(A). These maps were

overlaid and manually constructed network connections linking HV

nodes to determine various network lengths in GIS.

Medium voltage network

Before this study, the MV network in OPERA was represented as a

copper plate at the national level, i.e., demand and supply were met

within the country without consideration of network distances. A

framework for MV, similar to HV, was created in this paper, including

network costs and losses as a function of distance. Additionally,

constraint characteristics and region/node-specic connection types

were introduced, all of which are major modeling framework im-

provements on a regional level. The MV network is important on a

regional level because large scale deployment of renewables, growth of

cities, and electricity demand due to electrication will signicantly

Fig. 9. (A) Onshore HV connections of the province of Groningen with other provinces in the northern region of the Netherlands and the RNL with the Dutch offshore

part of the North Sea and the partition related to regional allocation towards the province of Groningen and the RNL and offshore HV connections to this region. (B)

Enlarged view of the province of Groningen, representing onshore HV connections between municipalities, along with the onshore HV connection between Germany

and the province of Groningen, and the offshore HV connections between the province of Groningen and Denmark/Norway and Germany. Municipalities with

shortlisted cities have nodes closer to these cities.

S. Sahoo et al.

Energy Conversion and Management 277 (2023) 116599

affect the future MV capacity and its spatial structure.

The MV network was divided into two major categories in this study.

First, an MV connection was established from an HV network to the

municipality rests for supplying renewable energy to the electricity grid.

The MV network map was obtained from the MV transmission and dis-

tribution system operator for the province of Groningen, ENEXIS [53]. A

part of this network that will be nearer to the space suitable for GBPV

and onshore wind in the 2050 progressive scenario [5] was manually

reconstructed and the network lengths were calculated in GIS (Fig. 10).

The renewable regions within each municipality are directly connected

to an MV network because these sources are large and centralized. As

every municipality has a positive potential for renewables, we created

ten MV nodes to establish MV connections from HV nodes.

As cities cannot be connected directly with an HV network, addi-

tional MV network connections were created to supply electricity to

cities. Establishing this connection through the MV node is helpful in

future situations wherein the electricity demand of the cities increases

owing to, for example, increased electrication to meet heat demand in

the BE. Similarly, the electricity supply can also increase due to the

presence of more BE-related rooftop PVs. This demand and supply may

not occur within the same time interval. By creating an MV network

connection to the cities, the energy-demanding/supplying regions were

segregated from the energy infrastructure, which is our research

contribution. MV connections were established between city MV nodes

and an HV node located in the municipality (Fig. 10). An additional MV

connection was provided to the Delfzijl industrial cluster through a

corresponding node because of the presence of a large number of in-

dustries, and a direct HV connection could not be provided to these

industries. Thus, a total of 21 MV nodes were created in OPERA for

Groningen (Fig. 3).

The MV network does not have a single standardized cost as an MV

network can be responsible for transmission or distribution depending

on the region or connection to which it supplies electricity [54,55].

Therefore, a region connection-specic database was introduced in the

OPERA. Different cost and loss characteristics were set for cities and

other connections [56] as listed in Appendix B. Capacity limits for the

MV network could not be implemented because of a lack of data avail-

ability on the current network capacities in various regions, although

the modeling structure allows us to incorporate these limits. Neverthe-

less, the inclusion of network costs provides a primary, yet simplistic,

representation of the additional investments required to potentially

expand the MV network if needed (see Section 4). The MV network

connection with the HV network allows us to understand how the impact

of electricity transmission at the national level may be translated to

electricity distribution at the regional level.

2.3.2. District heating network

DH network is not common in the Netherlands and has only recently

gained importance as a possible future heat supply option. In the

province of Groningen, the municipality of Groningen is the most active

region, and it is developing a DH network for approximately 10,000

households. The DH structure was almost nonexistent in OPERA. This

study performed the following DH-related activities: adding heat supply

sources, establishing network connections, and modeling DH, all of

which are discussed in the subsequent sections, thereby making a sig-

nicant improvement to the OPERA modeling framework related to the

DH representation. Another major research contribution is the creation

of a pan-provincial DH network spatial structure not found in any DH-

related literature studies or integrated energy system models. In addi-

tion, improving the technical details of this network within the context

of regional energy system modeling is a major step as heat gets less

attention at higher geographical levels.

Heat supply sources

Various centralized heat supply sources and technology options can

potentially support DH networks in the province of Groningen, ranging

from geothermal, and IWH, to HPs. Future geothermal heat and poten-

tial IWH locations were xed, as discussed in Section 2.2.2. The focus

was centralized heat supply options that could provide heat to the whole

city and added those options to the OPERA database, based on [57], as

the existing DH options were insufcient. New nodes were created in

Fig. 10. (A) MV network connections linking various cities (i.e., BE relevant, within the province of Groningen) and (B) MV network connections relevant for

renewables (overlaid on GBPV future (2050) space potential based on [5]).

S. Sahoo et al.

Energy Conversion and Management 277 (2023) 116599

OPERA to spatially allocate these options. From a city planning

perspective, it likely that these nodes would not be positioned within

city centers, notably because of their spatial and environmental impli-

cations. Therefore, all of these centralized options were placed at a

single location on the current outskirts of the cities to ensure proximity

to demand and represented that location as a new node. While doing so,

further considerations were made on whether industries might be

located close to a city, as they are a source of IWH.

If a single industry is located on the city outskirts, then the location of

all heat supply sources is shifted to that location to reduce the additional

piping costs associated with connecting IWH with a DH network. For

multiple industries, a centralized location or node was identied, and a

DH network connection was established. This process identied four

cities without any industries located on the city outskirts, thereby

creating four additional nodes on the city outskirts for Groningen

(Fig. 3). Cities that were too close to one another within a municipality

were awarded one node on the outskirts of the larger city. For example,

Haren and Appingedam cities had no nodes as these are closer to Gro-

ningen and Delfzijl cities, respectively, within the same municipalities.

All these activities assisted in reducing the number of DH-related explicit

nodes.

DH network connections

The route for the DH network is along the road network to avoid the

pipelines laid in undesirable areas, based on [58]. The existing road

network map for the Province of Groningen was used for calculating

network lengths. The DH network is constructed to connect cities to

energy-supplying options, which include the geothermal doublets and

industrial locations for IWH, implemented as nodes in the model. Con-

straints were imposed, for example, related to the capacity or energy

potentials of supply options and related to the capacity and ows of

energy (heat) infrastructure, and in the process, our rst research

question was answered. For DH, two types of network connections were

created and used in this paper: distribution DH and transmission DH.

The distribution DH is useful for short connections within cities and

were added to all included cities for this paper. For all cities, the

distribution network starts from the city center where the heat demand

density is the highest [5] and expands radially outward towards low

density areas, as this structure is identied to be the shortest possible

network length [59]. Hence, starting from the city center also leads to

least heat losses. Ten nodes were created in OPERA to account for the

center of each city (also see Fig. 3). The length of the network was

determined using the intersections of the road infrastructure and pop-

ulation center maps in GIS (Fig. 8). The overall DH network length of

each city was reduced by 30 % to account for the repetitive nature and

bi-directionality of the road network.

The transmission network for DH is useful for long network con-

nections, say >10 kms, but also short unobstructed connections of a few

kms. Compared to a distribution network for DH, a transmission

network has relative lower losses and investment cost per unit of dis-

tance, as well as higher heat ow capacity. These aspects are important

to transport heat over longer distances. The transmission network

created for this paper connects population centers and DH supply

sources such as geothermal doublets or industries for waste heat.

Modeling-wise, the GIS network analysis tool with the shortest route

(road network) criteria [60] was used to obtain the minimum connec-

tion lengths between two DH-related nodes linked to the DH trans-

mission network, for example, geothermal doublets, industries, and city

outskirts. All possible connections between these were created without

repetition by utilizing the same connections wherever feasible and

allowing the OPERA model to choose cost-effective routes through

optimization. The planning related to network connections can be

divided into four categories. (1) Connections were established from

geothermal doublets to city centers or outskirts (whichever was nearer),

either directly or via industries depending on the spatial location of the

cities, including connections between doublets. (2) Connections were

established from industry to city centers or outskirts, with a preference

to outskirts. (3) Connections were established between industries and

between doublets with an emphasis to reduce overall transmission

network length. (4) Connections were established from city outskirts to

city centers, including connections between city centers, for example,

Fig. 11. Transmission DH network for the province of Groningen, including industries (as some connections are exclusively for individual industries), geothermal

doublets locations, and DH nodes used to represent the city center, city outskirts, or industry clusters.

S. Sahoo et al.

Energy Conversion and Management 277 (2023) 116599

between Haren and Groningen city and between Appingedam and

Delfzijl city. Transmission DH connected the source to distribution DH

unidirectionally in the OPERA model. Fig. 11 shows the transmission DH

network created in GIS which is used as the input in OPERA.

This study is a rst-time attempt to spatially connect a whole prov-

ince of 3000 km

through a DH transmission network. Even though the

province of Groningen does not have any concrete plan on the layout of a

DH network, a recent provincial planning document on heat [61] sug-

gested that transmission lines could come up from Delfzijl industrial

cluster to Groningen city, which we have also included. Such policy

developments signify that a regional energy system model for analyzing

regional DH networks is indeed a sensible addition to the toolbox of

energy system analysis.

District heating network modeling

The DH model structure was created similar to the electricity

network to avoid adding too many equations in OPERA. The major

differences are that the cost and losses are dependent in a non-linear

manner on the capacity of the pipeline (or pipe diameter). In addition,

unlike the electricity network, the ows are sequential and unidirec-

tional which is distinguished in the database. The model allows bidi-

rectional ows, but the redundant connections are discarded in the

model preprocessing run. Differences exist between DH activities within

a region/node and between nodes, as shown in Figs. 3 and 8. A generic

cost and loss structure was used for transmission, whereas a connection-

specic structure was used for distribution. Appendix C presents a

detailed description and modeling equations related to the new DH

structure addition in the OPERA model.

3. Results

The scenario denition is similar to that in [4] and emphasizes

extensive renewable deployment and national self-sufciency in 2050

[62]. The renewable potential of the province of Groningen is based on

[5], and those for the other regions are similar to [4]. The sectoral

demands of the province of Groningen and other regions are the same as

in [4]. The total greenhouse gas emissions at the national level were

restricted to zero in the scenario denition.

Some emissions, particu-

larly non-CO

, indirect, and process emissions, were difcult to avoid. In

OPERA, the balance of these emissions occurs at the national level.

Therefore, the energy demand and related costs associated with the

processes or technology options responsible for negative CO

emissions

(and related infrastructure) were excluded from our regional analysis.

These processes and related costs were allocated to the RNL, instead, and

the interregional electricity ows and electricity imports to the province

of Groningen were adjusted.

The results of this study were compared with those in [4] to under-

stand the impact of detailed spatial segregation on energy balances and

costs. For consistency in the comparison of results, similar databases

were used for [4] and this study. The results are further categorized into

energy demand and supply (Section 3.1), interregional energy ow

analyses (Section 3.2), and cost analyses (Section 3.3).

3.1. Energy demand and supply

Section 3.1.1 compares [4] and this study regarding the primary

energy supply mix, followed by overall heat and electricity energy bal-

ances, for the province of Groningen. Heat balances for the BE and in-

dustrial sectors were analyzed in all land-use regions and industrial

nodes in Section 3.1.2 to illustrate the model capabilities regarding the

sector-specic detailed spatial analysis. Section 3.1.3 focused on the

capacity potential of spatially sensitive renewables and their utilization

shares in the regions and nodes mentioned above.

3.1.1. Primary energy mix and secondary energy balances

The changes in the contribution of some primary energy carriers in

this study compared to [4] were signicant (Fig. 12). For example,

biomass supply in this study was 18.4 petajoule (PJ) higher (approxi-

mately 460 %) than that in [4] due to the availability of a variety of

biomass types, such as grass and energy crops, in almost every

Fig. 12. Primary energy mix comparison (data in PJ) between the results of Sahoo et al. [4] and this study for the aggregate of all regions within the province

of Groningen.

Apart from renewables mentioned in the method section, fossil fuel-based

resources, such as coal, oil, natural gas, and uranium, are part of the national

energy system and are nationally balanced. For some of the processes or

technology options, particularly related to industries, these resources are

necessary, and, for some other processes, these might be cost-effective. Hydro-

energy is another renewable option available, which was not discussed in the

method section due to its limited availability in the Netherlands. In addition, a

variety of biofuel options, such as bioethanol, biodiesel, and bio benzene, are

available in this scenario. These options have common utilization in the

mobility and industrial sectors.

This does not mean that the province of Groningen has zero net emissions.

We did not regionally balance CO

emissions.

We did not model the CO

network in this study. Neither did we allocate

spatial locations for CO

storage. The model nds it convenient to allocate

greater than50% of the CO

consuming process and related infrastructure to the

province of Groningen owing to its close connection to the North Sea for cheap

electricity. In reality, this might be an overestimation, as a singular province

might not agree to share such a high proportion of emission reduction share.

S. Sahoo et al.

Energy Conversion and Management 277 (2023) 116599

municipality and their utilization within the corresponding region. In

addition, our analysis demonstrated that the regional biomass potential

is higher than that of [4] and is cheaper to utilize. The contributions

from onshore wind increased slightly (2 % increase), whereas the solar

PV contribution decreased signicantly (59 % decrease). In contrast to

[4], the geothermal heat contribution is negligible because geothermal

energy is suitable for the BE or horticulture applications in the

Netherlands [41]; however, geothermal applications are restricted

owing to the presence of natural gas (NG) elds in the province of

Groningen, the extraction of which can make the land susceptible to

earthquakes. Therefore, geothermal heat had limited application in

horticulture. Additionally, geothermal doublets have less use in the BE

because of high DH network costs owing to large distances from popu-

lation centers (Fig. 11). As expected, the onshore wind potential (38.3

PJ) is signicantly higher compared to the provincial government

assessment for 2030 (the latest year for which the assessment is made)

[63], i.e., 9.5 PJ. Similarly, the large-scale photovoltaic potential in this

report is 1.8 PJ which is much lower compared to 13 PJ of solar potential

in our analysis.

Biofuels, other renewables, such as hydro-energy, and oil have

negligible contributions. Other primary sources, such as synthetic fuels

and coal, have no contributions. The nal energy mix is not represented

separately because this mix is almost similar to the primary energy mix,

as no major conversion process occurs in the province of Groningen,

such as reneries, fuel-based power plants, or electrolysis. NG is utilized

in processes such as methanol production (Bio-MCN plant in Industry

Delfzijl node), although the utilization volume is slightly lower than that

in [4] (Fig. 12). Fig. 13 shows the primary energy supply and nal

sectoral demand balance for this study. The demand was dominated by

the BE (44 PJ), followed by industries (25 PJ).

The BE (23 PJ) and industrial (17 PJ) sectors were responsible for the

majority of the sectoral electricity demand (Fig. 14). Most of the elec-

tricity demand was met by renewable energy supply options, namely,

onshore wind (38 PJ) and solar PV (10 PJ).

The heat demand was

dominated by the BE (38 PJ) for the province of Groningen in this study

(Fig. 15). Among the heat supply options, electric boilers made a sig-

nicant contribution (14 PJ), followed by hybrid boilers (13 PJ) and HPs

(7 PJ).

3.1.2. Sectoral heat balances

Each land-use region has a distinct share of various technology op-

tions for heat supply to meet the heat demand of the BE sector (Fig. 16).

For example, Groningen outer city had the highest heat supply from HPs

(2 PJ), followed by DH (1.2 PJ), whereas Appingedam city had the

Fig. 13. Primary energy supply and nal sectoral demand balance for the province of Groningen (data in PJ).

Fig. 14. Electricity balance for the province of Groningen (data in PJ). H

STEG is a highly efcient electricity-producing turbine with H

as input.

The role of exibility options related to power ow, such as batteries or

thermochemical storage, is minimal at the geographical resolution considered

for our analysis. Therefore, these options are neither mentioned in the method

nor reected in the results related to electricity.

S. Sahoo et al.

Energy Conversion and Management 277 (2023) 116599

highest supply from hybrid boilers (0.3 PJ), followed by electric boilers

(0.25 PJ). Overall, hybrid boilers made the highest contribution (13 PJ),

followed by electric boilers (12 PJ). CHP did not contribute as the in-

vestment cost is high, and the co-produced electricity is not cost-

effective compared to electricity produced from renewable resources.

Groningen outer city had the highest BE heat demand (6 PJ), followed

by Het Hogeland municipality and the rest of Westerkwartier munici-

pality (4 PJ each).

Various industrial nodes have distinct heat demands from different

industrial subsectors and are supplied by diverse technology options

(Fig. 17). The industry Delfzijl node had a variety of heat supply sources

that are not present in other locations, such as MSW incineration,

methanol production, and oil boiler pyrolysis. The overall heat demand

was the highest for the Groningen industry cluster (1.93 PJ), followed by

the Avebe starch FBI in Ter Apelkanaal (1.3 PJ). The sectoral heat de-

mand was the highest in the FBI (4.7 PJ). On the supply side, anaerobic

digestion resulted in the highest overall contributions (5.3 PJ or 60 % of

the total contribution), followed by electric boilers (2 PJ). Some heat

produced in Delfzijl remained unutilized (0.3 PJ) because heat pro-

duction exceeds the demand in this node and the DH heat transmission

to the nearby cities is costly.

Detailed spatial modeling of land-use regions and industrial nodes

provided the sectoral heat balance insight that was not possible to obtain

via crude spatial regional segregation, as performed in [4]. The regional

Fig. 15. Heat balance for the province of Groningen (data in PJ). The hybrid boiler has two components: the major component produces only heat and a minor

component produces both electricity and heat, and the HR-107 boiler is a high-efciency boiler.

Fig. 16. Heat supply technology options for meeting heat demand of the BE in all the land-use regions in the province of Groningen. The area of the pie represents

energy supply volume in PJ.

S. Sahoo et al.

Energy Conversion and Management 277 (2023) 116599

Fig. 17. Heat balance for industries. (A) and (B) represent the heat demanding industrial subsectors and heat supplying technology options, respectively, for the

province of Groningen. The area of the pie represents demand or supply volume in PJ.

Fig. 18. Renewable energy supply mix in each land-use region and industrial node in the province of Groningen. Here, m, h, and e represent multiple, heat, and

electricity energy carriers, respectively. The area of a pie represents energy supply volume in PJ.

S. Sahoo et al.

Energy Conversion and Management 277 (2023) 116599

heat balance analyses of the BE and industrial sectors are illustrative

examples of the regional OPERA model capabilities. Similar balances

can be conducted on other energy carriers, such as electricity, within the

province of Groningen. Therefore, our modeling framework allowed us

to understand the mechanisms of demand and supply of various energy

carriers within an integrated system with spatial details. Even though

some of the overall balances are similar between our spatially-detailed

modeling and crude modeling in [4] on some of the supply options,

these detailed land-use region balances provided insights and deviations

that were not possible to identify before. In addition, similar spatially-

detailed energy balance analyses can be performed for other regions

following our method.

3.1.3. Regional renewable supply mix capacity potential utilization shares

The renewable energy supply mix was diverse in all land-use regions

and industrial nodes within the province of Groningen in this study

(Fig. 18). The supply mix of cities was dominated by rooftop PV

(approximately 70 % contribution), followed by solar thermal (29 %),

with the remainder coming from biomass. The predominant renewable

source in the remaining part of each municipality was onshore wind,

with the highest contribution from Het Hogeland municipality (9.4 PJ).

In addition, Het Hogeland municipality had the highest overall renew-

able contribution (14.2 PJ) among all regions and nodes. Geothermal

heat made a negligible contribution only to Het Hogeland municipality.

Floating PV, agricultural PV, and industrial PV had negligible contri-

butions within the province of Groningen. The detailed biomass-related

regional analysis of sources and energy carriers, including capacity

utilization share, is presented in Appendix D.

In the latest report on the Provincial Strategy for renewables [64],

the cumulative contribution of the municipality Het Hogeland onshore

wind and GBPV is 6.3 PJ for 2030 (this is the latest future year for which

provincial planning has been made), which is 3.6 PJ less compared to

our calculation for 2050. In addition, the provincial strategy report did

not show contributions from any other renewables that were considered

in our analysis. This report [64] also shows the overall provincial energy

supply of wind and GBPV is 23 PJ, which is signicantly lower compared

to our case of 43 PJ. This suggests that signicant efforts are needed to

improve the contribution of wind and GBPV after 2030 towards 2050 if

ambitious renewable targets are to be met. The province should also

seriously consider including other renewables in the primary energy

supply mix, for example, biomass, as this will provide more options and

some of these options have high regional potential as we calculated.

There was a difference between the available capacity and utilization

potentials of renewables in various regions. For example, GBPV

exhibited an average utilization share of 0.21 for all the remaining parts

of municipalities, with the highest share for Veendam (1.0) and the

lowest for Eemsdelta and Westerkwartier (zero utilization) – Fig. 19 (A).

Our GBPV capacity potential utilization of 1.3 GW (GW) is within the

capacity range of 0.7–3.8 GW suggested by [65] and slightly lower than

1.5 GW considered in the national management scenario in [5,62] for

the province of Groningen in 2050. The maximum capacity potential of

onshore wind was fully utilized in all the remaining parts of munici-

palities, except Eemsdelta, where the utilization share is 0.48 – Fig. 19

(B). This suggests that onshore wind provides a cost-effective solution as

a renewable source for meeting future electricity demands as also

identied in [4]. Our onshore wind capacity potential utilization of 2.5

GW for the province of Groningen is comparable to 2.24 GW considered

in the same scenario as GBPV in [5] and slightly higher than that of [65],

where the capacity potential range identied is 0.6–1.7 GW. These

comparisons conrms the reliability of our data and results. The rooftop

PV regional capacity utilization share is detailed in Appendix D.

3.2. Interregional energy ow analyses

This section includes analyses of electricity ow volumes in the

Netherlands, including offshore, with a focus on the province of Gro-

ningen (Section 3.2.1). Furthermore, the HV network capacity potential

utilization share is included in this section. Section 3.2.2 details DH

ows including ow volumes toward various cities from DH supply

points and DH penetration in cities.

Fig. 19. Analysis of capacity utilization share for GBPV and onshore wind in (A) and (B), respectively, for the remaining part of each municipality within the

province of Groningen.

S. Sahoo et al.

Energy Conversion and Management 277 (2023) 116599

3.2.1. Electricity

Our modeling framework of the electricity system considers the

whole of the Netherlands, including the Dutch offshore part of the North

Sea and adjoining countries linked via the HV network. This allowed us

to perform electricity ow analysis on a national scale, along with cross-

border electricity trade. Fig. 20 presents the electricity network ows as

a result of optimization modeling.

A large volume of electricity owed

from the northern Dutch part of the North Sea (245 PJ) – Fig. 20 (A). A

majority of this ow is directed toward Groningen municipality (118

PJ), followed by Eemsdelta municipality (105 PJ), with the rest exported

abroad – Fig. 20 (B). A net annual total of 144 PJ of electricity was

exported from the province of Groningen to Drenthe and 43 PJ to

Friesland. The model also allowed us to investigate the electricity ows

from supply regions via the MV network, which was not possible in the

earlier crude modeling.

Measuring and analyzing the future utilization of the available HV

network capacity is necessary to understand which network needs to be

strengthened if the aim is to achieve a cost-efcient integrated energy

system at the national level. Our model indicated that some network

maximum allocated capacities were fully utilized such as the connection

between municipalities Eemsdelta and Groningen with a capacity of 0.4

GW (Fig. 21). There are no plans for actual capacity improvement of this

network, at least until 2045 based on European electricity network

planning database, which is concerning given the fact that this

connection supplies electricity to the largest city in the province, Gro-

ningen city.

The connection between Eemsdelta and Het Hogeland (offshore

connection) municipalities has a current connection capacity of 7 GW.

Our analysis depicted that the network will fully utilize its maximum

capacity of 8.8 GW; however, the Dutch transmission system operator,

TenneT, does not plan to expand this connection until 2045. Similarly,

the maximum connection capacity between Groningen and Het Hoge-

land (offshore connection) municipalities of 4 GW was also fully uti-

lized. This raises concerns because these connections are responsible for

transporting electricity from the Dutch north offshore and from other

countries to the province of Groningen towards Drenthe and Friesland.

Future offshore wind capacity additions without fast and signicant

upgrades to these network capacities can lead to network congestion and

unexpected price hikes. In addition, as maximum network capacities are

utilized for certain network connections, the model might not nd it

cost-effective to invest in renewables, particularly GBPV, in these re-

gions (Fig. 19) because the corresponding additional electricity pro-

duced cannot be transferred over these networks.

3.2.2. District heating

Even though DH network was considered in all the cities included in

our modeling framework, the regional model (after optimization) found

ve cities to have feasible DH supply, among which Groningen outer city

had the highest penetration (1.2 PJ or 21 %), followed by Appingedam

city (0.14 PJ or 17 % penetration) (Fig. 22). These penetration volumes

are signicantly higher than the current penetration of <1 % [66].

However, these values are lower than that in [7], which obtained an

average penetration of 21 % for various cities in the Netherlands,

including Groningen city. Even though all the cities investigated in our

study demonstrate high heat demand densities [5], cities with DH con-

tributions are either highly compact with less distribution DH lengths or

have higher average heat demand densities compared to other cities. In

general, the distribution DH cost per unit length and losses are consid-

erably higher than the transmission DH cost per unit length and losses.

Another reason for the low DH penetration is heat savings associated

Fig. 20. Interregional net annual electricity ows obtained from the optimization modeling results for the HV and MV networks with the arrow width representing

the net annual ow volumes in PJ, and the arrow directions representing the net ow directions. (A) Flow in the Netherlands including offshore regions. (B) Enlarged

view of ow in the province of Groningen.

The OPERA model optimization details at the national level, including al-

gorithm, design variables, and constraints, are presented in [34]. This is

applicable for the electricity network. Additionally, objective function

including regions and nodes, energy balances at the national and regional

levels, related constraints are documented in [4]. This includes electricity

network connected structure and the corresponding optimization.

S. Sahoo et al.

Energy Conversion and Management 277 (2023) 116599

with changes in energy labels reduce the net heat demand density and

make investing in a citywide DH network less cost-effective. The nd-

ings of our study demonstrate that a DH network might be cost-effective

in small pockets of high heat demand densities within cities, such as

Groningen inner city. However, it was not possible to analyze all of these

pockets within the geographical resolution of our research. The frame-

work offers the possibility to perform a more detailed spatial analysis,

though. This study spatial resolution is still too low for estimating the

true potential of DH networks in cities or population centers.

Even though a pan-provincial DH network having a dense spatial

spread was created (Fig. 11), the network obtained as a result of opti-

mization was sparsely spread throughout the province (Fig. 23), sug-

gesting that most of the potential network routes are not cost-effective

due to long transmission network distances and related high costs and

losses. Considering the interregional DH ow volumes, a signicant net

annual ow occurs through the transmission network from the Industry

Groningen node to the Groningen inner city transmission node (1.7 PJ),

followed by the ow from Groningen inner city to outer city (1.5 PJ).

IWH has the highest contribution to DH (0.6 PJ) because it is available

for “free,” subject to DH costs and losses. The results demonstrate sig-

nicant losses in the distribution network; for example, the Winschoten

city distribution DH network has a loss of approximately 19 % compared

to the net transmission heat supply at the Winschoten city center. These

results show that technically-detailed DH network analysis can be per-

formed at a city level using our regional model along with provincial

analysis.

Fig. 21. Utilization share of the maximum capacity of the HV network in 2050 for the province of Groningen. The thickness of the line represents the share of the

maximum HV network capacity utilized.

Fig. 22. DH supply (data in PJ) and % DH penetration of cities (included in our study) in the province of Groningen.

S. Sahoo et al.

Energy Conversion and Management 277 (2023) 116599

3.3. Cost analysis

Fig. 24 shows the total system cost comparison between that re-

ported in [4] and this study for the province of Groningen. This cost is

further broken down into energy-demanding sectors, namely, agricul-

ture, BE, industry, and mobility, energy supplying options, such as large-

scale onshore wind and GBPV, energy infrastructure, and net import of

secondary energy carriers,

such as electricity, H

, and NG, to the

province of Groningen. The cost associated with these secondary carriers

are taken from Appendix B. The difference in the total system cost be-

tween this study and [4] was the highest for the BE sector, this difference

is 1260 million euro per year (M

€

/year), which is 208 % more than that

obtained in [4]. The main reason for the difference in the BE sector is

that the model in this paper nds it more cost effective to invest in

additional insulation measures, reected in upgrade of energy labels.

The upgrade of energy labels is so signicant that the difference in in-

vestment related to only energy label change is 807 M

€

/year. Detailed

analysis of spatially explicit data of various regions within the province

of Groningen shows that a large proportion of buildings has low energy

Fig. 23. Net annual heat ows in the DH network within the province of Groningen. The thickness of the arrow represents the annual net ow volumes in PJ, and the

direction represents the net ow direction.

Fig. 24. Total system cost comparison between [4] and this study (data in M

€

/year).

We xed the annual costs of these energy carriers’ (see Appendix B for

costs) and multiplied the annual net import of these carriers to determine the

overall annual import cost.

S. Sahoo et al.

Energy Conversion and Management 277 (2023) 116599

labels, such as GFE or DC labels, and improving these labels to the next-

level labels are cheap compared to improving DC or B labels to the

highest level labels such as A or A+[4].

The model as used in [4] could

not pick this regional specicity due to crude regionalization. Therefore,

the model in this paper suggests more investments in the BE compared to

[4] as this is the cost effective solution, even though the overall cost is

higher. Differences in energy infrastructure costs is the second highest, i.

e., 863 M

€

/year or 37 % more than the previous study. The main reason

for this positive difference in energy infrastructure is additional in-

vestment associated with predominantly MV electricity and DH infra-

structure. These networks are not spatially well developed in [4], and

hence could not pick the related investment and operation costs. To

illustrate, in [4], MV network only considers average investment cost as

a function of energy ows and not as a function of distance. Similarly,

for DH network cost, only marginal cost associated with heat ows are

considered and not investment costs per unit distance. Further expla-

nation on energy infrastructure is provided in the subsequent para-

graphs. The aggregated cost associated with the net regional import of

energy carriers, such as, electricity and H

, is also higher for this study,

accounting for 697 M

€

/year, which was 205 M

€

/year in [4]. This

suggests that more import is required to meet regional energy demand,

compared to [4], as regional supply is less and it is less cost effective to

use own supply options.

Fig. 25 shows a further breakdown of total system cost to variable

operation and maintenance (O&M) cost, capital expenditure (CAPEX),

operational expenditure (OPEX), and fuel costs for energy-demanding

sectors, energy infrastructure, and supply options in the province of

Groningen.

CAPEX cost was the highest (6704 M

€

/year), followed by

OPEX (1789 M

€

/year). Energy infrastructure CAPEX and OPEX played a

major role, followed by the energy-demanding sector BE. Appx. Fig. A4

in Appendix D illustrates the abovementioned cost components’ com-

parison between [4] and this study.

The energy infrastructure cost for various energy carriers (heat,

hydrogen, gas, and electricity) were compared between [4] and this

Fig. 25. Cost breakdown into variable O&M, CAPEX, OPEX, and fuel costs for this study (data in M

€

/year).

Fig. 26. Energy infrastructure CAPEX comparison between [4] and this study (data in M

€

/year).

OPERA can calculate the changes in energy labels and the corresponding

changes in investment costs, without considering the costs associated with the

entire building stock. This aspect is consistent in both [4] and this study.

Variable O&M costs refer to the operation and maintenance expenses that

vary with the amount of production of the main product for a process or

technology option. CAPEX represents the expenses associated with adding new

equipment within a process or upgrading a technology option. OPEX or xed

O&M costs represent auxiliary expenses associated with technology options,

such as the expenses in stafng and maintenance or research and development.

Both CAPEX and OPEX are considered over multiple years and annuity is

applied for calculating the cost equivalent for a single year. Fuel costs are the

costs of fuel used as inputs to processes or options.

S. Sahoo et al.

Energy Conversion and Management 277 (2023) 116599

study for the province of Groningen (Fig. 26) because this cost is a major

component of the system cost. The heat and electricity CAPEX costs

increased when compared to [4] owing to a high investment in detailed

spatial structure of these infrastructures with the corresponding cost

increased by 42 M

€

/year (or 330 % increase) and 1243 M

€

/year (96 %),

respectively. DH CAPEX cost for transmission and distribution networks

are 8.3 M

€

/year and 46.2 M

€

/year, respectively, in this study, whereas

the corresponding costs were 12.7 M

€

/year and zero, respectively, for

[4]. The results of this study are likely to be more realistic, as one would

expect DH distribution costs to be much higher than DH transmission

costs. Our detailed DH modeling allowed for better cost allocation in the

spatial structure of heat and electricity networks.

These results show that creating a large number of regions (and

nodes) allowed us to distinguish spatially between energy-demanding

regions, supplying options, and energy infrastructure, which was not

possible in the earlier study. In our previous crude regionalization study,

demand and supply were met within the larger region wherever feasible

and possible without the need of energy infrastructure. In our study, the

mismatch between energy demand and supply within a region is met by

energy infrastructure, thereby increasing the utilization of infrastructure

and optimizing its investment cost. Currently, the regional distribution

system operator and the province of Groningen also consider an

increased need and related expansion of MV [67] and DH [61] networks,

respectively, on a regional level. Additionally, the discrepancies in the

spatial distribution of energy supply sources within the province are

better captured by adding spatial detail, leading to greater differences in

primary energy supply mixes as reected in our model. Our spatially-

detailed model can make a more accurate estimation of future

regional energy supply potentials, particularly related to renewables,

Fig. 27. CAPEX distribution for land-use regions and industrial nodes for electricity network infrastructure within the province of Groningen. The area of the pie

represents the cost in M

€

/year.

Table A1

MV network connection costs applied to this paper based on [56].

MV network type CAPEX

(Euro/m/

MW)

Fixed

O&M

(Euro/

MW/

year)

Loss

(%)

Explanation

1–5 MW

capacity,

suburban

16 2570 3 We multiplied CAPEX

with the length of each

MV network to obtain

cost in Euro/MW. This

cost applies to all cities,

except Groningen inner

city.

5–25 MW

capacity, city

6.76 2747 2.25 This applies only to

Groningen inner city as

Groningen inner city has

a dense MV network

structure compared to

other cities.

Main

distribution,

50/60 kV

electricity

3.9 21.2 0.3 This cost applies to all

other MV network

connections. Fixed O&M

cost is per unit meter. The

losses in these networks

are low as they are meant

for transmission

purposes, for example, a

connection to large-scale

GBPV or onshore wind

farms.

Table A2

Various energy carriers and their net annual price, along with their cost refer-

ence or explanation.

Energy

carrier

Price

[

€

_2015/GJ]

Cost reference or cost explanation

NG 6.667 Berenschot and Kalavasta report [62], considering

the national management scenario

Electricity 12.419 This data is obtained from the COMPETES model

run for the TRANSFORM 2021 scenario [72]. This

scenario is very similar to the national management

scenario from [62].

Hydrogen 18.255 Hydrogen cost is derived from electricity. Since

hydrogen is mainly produced from large-scale

electrolyzers in the Netherlands, we applied

electricity price and the efciency of hydrogen

conversion, i.e., 1/1.47.

S. Sahoo et al.

Energy Conversion and Management 277 (2023) 116599

whereas crude spatial modeling might lead to overestimation or un-

derestimation of these potentials depending upon the scenario.

The CAPEX

costs associated with the HV and MV networks were

further investigated with a detailed spatial resolution (Fig. 27) as the

electricity infrastructure has a high CAPEX cost (Fig. 26). The HV and

MV network combined CAPEX cost was the highest for the Industry

Delfzijl node, 213 M

€

/year, of which the MV network cost was 205 M

€

year, as this node was responsible for connecting the Delfzijl industrial

cluster via an MV network. The HV network exhibited the highest cost

contribution in the Het Hogeland offshore connection, 74 M

€

/year,

owing to the important connections within the province of Groningen

and abroad, including offshore (Fig. 20).

Our regionalized model, with an emphasis on spatial disaggregation,

enabled the investigation of network capacity additions with a detailed

geographical resolution. This showed that national or regional expenses

in energy infrastructure, particularly those related to electricity, will

play a major role in the regional energy system. Regional disaggregation

further demonstrated that investments in efciency improvement in

various building types within the BE are cheaper than investments in the

DH infrastructure. This detailed insight is not possible with low spatial

resolution regional modeling. The diverse set of results, thus, provide a

novel view on the regional energy system, signifying the added value of

our spatially-detailed modeling framework.

4. Discussion

Central to interpreting the modeling results presented in this study is

to discuss uncertainties regarding some of the assumptions made

regarding the future development of energy demand, supply, and

infrastructure. To begin with, the model results demonstrated a signif-

icant potential role of industrial waste heat (IWH) in providing heat to a

district heating (DH) network, as they could be linked to nearby popu-

lation centers. However, there are uncertainties regarding the IWH po-

tentials from various industrial activities as nal product demand

projections and their corresponding energy use may change in the

future. The continued existence of key industries and their exact energy

use cannot be simply assumed in the long term. In addition, energy

carriers, technology options, and processes can change, affecting the

IWH potentials from various activities. The model is robust in accom-

modating such future changes, though. For example, inputs related to

Table A3

Denitions of variables, parameters, and indices related to equations 1 to 11.

Variables Parameters Indices

FlowHeatInterRg The ow of heat between regions or nodes

(interregional/internodal)

AF Availability per time slice [0,1] ts Time slice

CapFlow The capacity of transmission or distribution pipe C2A Capacity to availability conversion factor rhtn,

rhtn1

Transmission-related DH

nodes

FlowIn Amount of heat that enters a node LossInterRg Interregional/internodal heat losses, for

example, between transmission nodes or

between a transmission and a distribution

node

otdh Option for transmission

network related to DH

FlowHeatIntraRg The ow of heat from supply option to

transmission option or from transmission option

to distribution option within a node

(intraregional/intranodal)

LossIntraRg Intraregional/internodal heat losses, for

example, transformer stations

ohsdh Heat supply sources in DH,

including geothermal

doublets and IWH

FlowOut Amount of heat that exits a node Y Fraction of hours assigned to a time slice rhdn Distribution-related DH

node

CostDH DH-related total infrastructure cost, does not

include costs associated with centralized heat

supply options

a Annuity factor. A discount rate of 2.25 %

was considered

oddh Option for distribution

network related to DH

CostTrans Transmission network cost InvestmentTrans Capital expenditure (CAPEX) cost of the

transmission network

obe End-use option in the BE

CostDistri Distribution network cost VarO&MTrans Variable operation and maintenance

(O&M) cost of the transmission network

InvestmentDistri CAPEX cost of the distribution network

VarO&MDistri Variable O&M cost of the distribution

network

C1 Construction cost constant (

€

/m)

C2 Construction cost coefcient (

€

)

Dd Diameter of the distribution pipe (m)

Capad Standard capacity of the distribution

network

Table A4

Lookup table for calculating DH pipe loss, cost, water ow, and capacity ().

Sl.

No.

DN Loss (kWh/

(m*yr.))

Cost

(

€

/m)

Water ow

(m/s)

Capacity

(MW)

1 32 186 195 0.9 0.2

2 40 214 206 1 0.3

3 50 239 220 1.2 0.6

4 65 281 240 1.4 1.2

5 80 289 261 1.6 1.9

6 100 302 288 1.8 3.6

7 125 350 323 2 6.1

8 150 413 357 2.2 9.8

9 200 448 426 2.5 20

10 300 494 564 2.7 45

11 400 509 701 2.8 75

12 500 720 839 2.9 125

Source: [58,73]

Table A5

Fixed parameters values for different region types for the DH distribution

network based on [7].

Region type C1(

€

/m) C2(

€

)

Inner city areas 286 2022

Outer city areas 214 1725

In OPERA, the investment cost of a network between two nodes is equally

allocated to these nodes. It is the case for losses as well. However, in the case of

the connection between the Het Hogeland offshore node and the northern North

Sea, the HV network CAPEX cost is completely allocated to the northern North

Sea offshore region to keep the offshore connection costs separate from the

regional costs.

S. Sahoo et al.

Energy Conversion and Management 277 (2023) 116599

nal products demands can be easily adjusted in the database and the

model can calculate the corresponding changes in processes, capacity of

technology options, and IWH potentials accordingly. Hence, it is feasible

to run the model under different assumptions in scenarios and to

perform detailed sensitivity analyses if needed. Considering changes in

IWH potentials maybe highly policy relevant as these changes are re-

ected in the feasibility of regional DH network linkages to industries

which are currently considered in policies. The integrated nature of our

Fig. A1. Detailed analysis of biomass supply in land-use regions and industrial nodes within the province of Groningen. (A) and (B) Supply sources and secondary

bio-energy carriers resulting from these sources, respectively. The area of the pie represents the energy volume in PJ. The size of smaller pies (<1 GJ) has been

increased for representational purposes.

Fig. A2. Analysis of utilization share for biomass grass and energy crops in (A) and (B), respectively, within the province of Groningen. Cities show biomass grass in

(A) because of the presence of verge grass alongside roads [5]; however, all the allocations were made to the rest of the municipalities.

S. Sahoo et al.

Energy Conversion and Management 277 (2023) 116599

modeling framework would allow for studying this feasibility under

different scenarios.

Secondly, geothermal heat contributed minimally to the DH

network, even though the model allows us to do so. Geothermal-related

technology options may develop in the future and become cheaper,

which may improve corresponding heat contributions. Notably, serious

increases in natural gas (NG) prices provides a scope for geothermal heat

as most of the current built environment (BE) heat demand is met by NG

in Groningen, which can be analyzed either through sensitivity analyses

or creating scenarios within the modeling framework.

Thirdly, ground-based photovoltaics (GBPV) and rooftop photovol-

taics (PVs) contributions are expected to increase due to current policy

instruments [68], which might lead to congestion problems of both the

low voltage and medium voltage (MV) network. Our model does not

consider the low voltage network as it is considered relevant mostly for

modeling at a city or district level. Hence, further analysis on the role of

the low voltage network may either urge for expanding the proposed

modeling framework or linking to other energy system models working

on a local scale. Regarding the MV network, a large number of nodes

were created in our model to represent the MV network with spatial and

technical details to also assess the possible role of grid congestion.

Nonetheless, access to data on current MV capacity was lacking,

implying the model cannot fully incorporate this role. In response, only

additional costs (capital and operational expenditure) of transporting

electricity were imposed so as to represent the need for additional in-

vestments in the grid. Therefore, as a key aspect for improvement, our

Fig. A3. The utilization share of the maximum capacity potential of rooftop PV in the BE within the province of Groningen for every land-use region.

Fig. A4. Various cost components’ comparison between [4] and this paper for the province of Groningen (data in M

€

/year). (A), (B), (C), and (D) are variable O&M,

CAPEX, OPEX, and fuel cost comparison, respectively.

S. Sahoo et al.

Energy Conversion and Management 277 (2023) 116599

framework does allow including a better representation of the technical

characteristics and constraints of an MV network, and can also identify

locations where MV capacity increases signicantly.

Of course, there are other assumptions that may warrant detailing or

further exploration. Examples include the rates of demolishing and

constructing in the build environment, the use of greenhouse area as

proxy for agricultural demand or the 30 % reduction of the district

heating distribution network length network to account for the repeti-

tive nature and bi-directionality of the road network. Rather than dis-

cussing these in detail, it is important to indicate that the modeling

framework allows for such choices to be swiftly changed, represent them

in scenarios or do sensitivity analyses with them. Our framework,

therefore, is robust in allowing these changes quickly. Hence, our

modeling approach of regional integrated energy analysis can also be

easily replicated in other regions, subject to data availability. Tweaks

can be made to the model depending upon the type of data available,

though. As such, our systematic approach is a potentially valuable

addition to the toolbox for performing energy system analysis on a

regional level, next to a range of national and international modeling

tools, along with geographical information system-based tools. Our re-

sults also conrms that such an addition may well be crucial, as it pro-

duced quite different results than our previous study using a more crude

regional representation of the province of Groningen.

Further improvements in modeling framework can include response

to developments regarding the regional policies and incorporate new

features. For example, the model currently follows national emissions

reduction targets. Regional policies incorporate additional regional

emission-related targets, which are not yet integrated into the modeling

framework. Additionally, based on the current discussions in the

Netherlands, the model could be expanded so as to investigate the

applicability of salt caverns for the storage of hydrogen, which is a

spatially-dependent feature and is available within the province. Addi-

tionally, regional infrastructure related to carbon capture and storage is

worth analyzing. Finally, regional policies currently point to becoming

(nearly) self-sufcient regarding energy in the future, while various

stakeholders look quite differently at which technology options and

where they may be used to do so. Inputs from stakeholders on the above-

mentioned aspects can make our modeling framework more robust,

exible, and dynamic. The impact of these multitude of aspects on the

regional energy balances and costs, for example, can be analyzed by

creating scenarios and performing sensitivity analyses. This also shows

that, although a variety of spatially sensitive aspects of the regional

energy system analysis were covered in this paper, more regional energy

system-related topics need to be investigated.

5. Conclusions and future work

This study focused on improving a regional energy system integrated

modeling framework by providing systematic steps to add relevant

spatial detail. The methodology involved creating regions and nodes

within the modeling framework under categories corresponding to dif-

ferences in land use (cities, industry, geothermal doublets, and other

regions), energy supply, and energy infrastructure. Furthermore, the

methodology involved a unidirectional soft linking with geographical

information system-based modeling results. As a result, the modeling

framework allows for regionally allocating spatially sensitive elements,

such as renewable resources or heat demand. A detailed breakdown was

provided for sectoral energy demand, supply options, and energy

infrastructure for electricity and heat, including district heating (DH).

Our regional energy system model (ESM) can translate the impact of

national and international energy-related planning and policy decisions

to a regional level, in addition to implementing regional policies. They

included policies and regulations of, for example, energy infrastructure,

spatial constraints for renewables, and renovations of the built envi-

ronment (BE). Linkages with other models at higher geographical scales

were established.

Important modeling results are as follows:

•Compared to the crude spatial modeling results, our detailed spatial

modeling showed signicant changes in renewable supply mix, such

as, an increased role of biomass, 18.4 PJ increase (+460 %), and

decreased role of solar, 19 PJ decrease (−59 %). The detailed spatial

modeling has nearly 100 regions and nodes categorization,

compared to only two in the crude modeling case for the province of

Groningen.

•The capacity potential of onshore wind was fully utilized in every

remaining part of the municipalities, except for Veendam (a utili-

zation potential of 2.5 GW (GW) from a total of 2.6 GW for the

province of Groningen). However, neither rooftop PVs (1.8 GW from

a total of 6.1 GW) nor ground-based photovoltaics capacity (1.3 GW

from a total of 16.6 GW) potential was fully utilized in any land-use

region. This shows onshore wind has a higher utilization potential

compared to photovoltaics in our regional modeling context.

•Important high voltage (HV) network connections within the prov-

ince of Groningen were between Het Hogeland (offshore connection)

and Eemsdelta municipality and between Het Hogeland (offshore

connection) and Groningen municipality, as demonstrated by the

utilization of their maximum future capacity potential of 8.8 GW and

4 GW, respectively.

•Cities with high heat demand densities and/or compact structures

had high DH penetration, such as, Groningen outer city (20.1 %

penetration), Appingedam (16.7 %), and Winschoten (14.9 %).

•The capital expenditure costs of the regional HV and MV electricity

networks demonstrated that the HV network was dominant in the

Het Hogeland (offshore connection) which has links with abroad and

the North Sea, contributing 74 M

€

/year, and the MV network con-

nected to the Delfzijl industrial cluster, contributing 205 M

€

/year.

The method and results demonstrated that this paper lled a major

research and knowledge gap on the regional level. Adding spatial detail

has several benets, such as a better understanding of renewable energy

supply potentials and mixes due to regional spatial policies and cir-

cumstances, an understanding of regional sectoral demand differences,

e.g., related to the BE, industries, and agriculture, identifying con-

straints and costs related to energy infrastructure, and identifying

possible demand and supply mismatches. While the province of Gro-

ningen was tested, the modeling framework has a wide applicability in

other regions if sufcient data is available. The exibility of our

approach also allows for use on a geographical scale somewhat higher or

lower than in our case. The model has a strong potential for use in

reecting on and providing input to regional policies, both regarding

spatial and energy planning. Our modeling framework is a useful addi-

tion to already existing tools at the national and European levels.

We intend to test our modeling framework using inputs from various

regional stakeholders. Policymakers at various regional scales are ex-

pected to play a major role in deciding the future of a regional energy

system modeling framework, along with energy infrastructure and

environmental experts. These inputs will aid in ne-tuning our model

and produce multiple scenarios in synchronization with stakeholders’

expectations and regional policy reports. Scenarios can be used to

analyze the impacts of increasing citizen participation, sustainability,

circularity, or lowering cost solutions. Scenarios can include signicant

fossil price rises which can occur due to international political de-

velopments as now in the case of natural gas. Similarly, sensitivity an-

alyses can include choices related to the use of hydrogen-related

technology options and increase in DH penetration in population cen-

ters. Our modeling framework has provisions to incorporate all these

changes in future research.

CRediT authorship contribution statement

Somadutta Sahoo: Conceptualization, Methodology, Resources,

S. Sahoo et al.

Energy Conversion and Management 277 (2023) 116599

Formal analysis, Data curation, Visualization, Validation, Writing –

original draft, Writing – review & editing. Joost N.P. van Stralen:

Conceptualization, Methodology, Resources, Software, Validation, Su-

pervision. Christian Zuidema: Writing – review & editing, Supervision.

Jos Sijm: Writing – review & editing, Supervision. Andre Faaij:

Conceptualization, Methodology, Resources, Validation, Writing – re-

view & editing, Project administration, Funding acquisition.

Declaration of Competing Interest

The authors declare that they have no known competing nancial

interests or personal relationships that could have appeared to inuence

the work reported in this paper.

Data availability

Data were collected from Open Sources only and appropriate refer-

ences were made to them throughout the manuscript.

Acknowledgements

We acknowledge the support provided by the ESTRAC Integrated

Energy System Analysis Project nanced by the New Energy Coalition

(nance code: 656039).

Appendix A. Detailed Groningen and other region-specic data and analysis

Sectoral demand

Built environment

For the province of Groningen, municipality level construction and demolition projection was based on 2013–20 historical data from CBS [38] (see

Table 2). For Groningen city, the construction of apartments was distributed by 20 and 80 % between the inner and outer cities, respectively, based on

their current apartment distribution share.

Services buildings were segregated based on activity solely dedicated to ofces, education, industrial halls, and hospitals in GIS, as these are

independent building type activities in OPERA (Table 2). The remaining buildings types were allocated to “other” service buildings. We calculated the

oor area of these buildings based on the effective area dedicated to that activity within a building (the BAG area). Building level aggregation toward

regions was similar to that of household buildings.

Industries

For salt, an additional industry, Nedmag in Veendam, was included compared to the previous study [4], as this industry, even though has less

annual production volume compared to other salt producing industries in the province of Groningen, is important in a detailed spatial analysis due to

high energy demand and industrial waste heat (IWH) supply potential. The procedure for adding a new industrial activity is explained in the previous

study [4] and Table 2. For newly added activities, the model can choose between the current technology option or processes (free of cost) and generic

options available within the subsector to meet future energy demands (including costs).

Energy supply

Wind

KNMI is the meteorological agency of the Netherlands providing wind speed proles at the center of a 2 ×2 km

square mesh for the whole of the

Netherlands for various hub heights and years [45]. We considered wind energy annual prole at location/point which coincided with KNMI data

points and are almost in the central location of region/municipality suitable for wind power production in 2050. For this, we manually matched these

data points with the central locations of each region. For onshore regions outside the province of Groningen, we considered a location closer to the

centroid of these regions. We used a MATLAB script to extract hourly wind speed proles at 150 m hub height for 2015 from a large dataset in KNMI

and used MS Excel spreadsheets for these proles, from where they were uploaded to the OPERA database (Fig. 7).

Solar

Currently, the province of Groningen does not have any examples of oating PVs. Because the data related to the conversion of inland water space

potential into oating PVs potential was not available in OPERA, we used an upcoming project in Ubbena, Drenthe, as an example in our research. In

that project, 28 hectare of inland water space is planned to be utilized in a given space of 50 hectare, with a solar PV capacity of approximately 25 MW

[69]. The inland water PV average capacity is estimated to be higher than that on land [70]. Other regions might have different potentials than what

we calculated.

Geothermal

For each of the geothermal doublets or node, we calculated recoverable heat [41] considering the technical potential by aggregating the potential

values of each 1 ×1 km

cell surrounding the doublet for 2–3 kms. In addition, we estimated the investment costs in technology related to geothermal

heat extraction based on [48,71] and applied a learning curve.

Industrial waste heat

We calculated the IWH potentials of various industrial activities added in [4] and in this study (see Table 2). The method associated with

calculating the IWH potential for most industries is described in [5]. IWH production from various industrial activities is dependent on the energy

input associated with related technology options. The IWH production volumes per unit nal main product were included in the energy balance of the

corresponding technology options for various activities in the OPERA database.

S. Sahoo et al.

Energy Conversion and Management 277 (2023) 116599

Energy infrastructure

Electricity network

Each municipality within the province of Groningen through their respective HV nodes, except Pekela, (Fig. 9 (B)) which has no nodes, as this

municipality is small with no distinct activities allowing reliance on an MV connection. Some cities have additional HV connections because of the

peculiarities of the existing HV network structure, the need to make an additional node to connect the city via an MV network, or the relative

importance of the city. This includes Delfzijl, Groningen, Haren, Hoogezand-Sappemeer, Stadskanaal, and Veendam. There are two additional HV

nodes for establishing connections specically to the Delfzijl industrial cluster and the Het Hogeland offshore connection (responsible for abroad

connections and the North Sea offshore connection) – Fig. 9.

Appendix B. Miscellaneous costs

Tables A1 and A2

Appendix C. Detailed modeling of DH network

As mentioned in Section 2.3.2, the equations below are written in the OPERA model in a general way, i.e., ows are non-sequential and bidi-

rectional. The distinction related to ow direction is made in the database to reduce the number of variables and constraints the model needs to solve.

There are two types of ows: inter-regional/nodal and intra-regional. Inter-regional connection is responsible for the connection between nodes and

intra-regional is responsible for the connection between technology/options within a region/node. Appx. Table A3 provides denitions of variables,

parameters, and indices used in the equations.

The heat ow through a transmission network (from node rhtn to rhtn1) is

FlowHeatInterRg(otdh,rhtn,rhtn1,ts) ≤ AF(otdh,ts)*CapFlow(otdh,rhtn,rhtn1)*C2A(otdh)*Y(ts)(A1)

The above-mentioned equation is used to determine the capacity of the network CapFlow. Readers are directed to refer Joost et al. [34] for a

detailed description of parameters AF, C2A, and Y.

The amount of heat that enters a heat transmission node in rhtn is

FlowIn(otdh,rhtn,ts) = ∑

rhtn1

FlowHeatInterRg(otdh,rhtn,rhtn1,ts)*(1−LossInterRg(otdh,rhtn,rhtn1,ts)) + ∑

ohsdh

FlowHeatIntraRg(otdh,ohsdh,rhtn,ts)*(1

−LossIntraRg(rhtn,otdh,ohsdh,ts))

(A2)

where FlowInterRg(otdh, rhtn, rhtn1, ts) represents heat ow from transmission node rhtn to node rhtn1 in a unidirectional manner (see also Section

2.3.2), within transmission network (otdh). FlowHeatIntraRg(otdh, ohsdh, rhtn, ts) represents heat ow from heat supply source (ohsdh) within the node

rhtn. The parameter LossInterRg is dependent on the pipe length along with capacity (diameter). LossIntraRg corresponds to losses in piping or

transformers, which are not included in heat supply source losses. This loss was considered zero in this paper; however, the model now has the

capability to include this loss in the future.

The amount of heat that exits a heat transmission node is the amount of heat owing to the distribution grid with the same region/node. Heat

transmission node does not have a nal demand. The heat owing out of a transmission node is therefore represented as

FlowOut(rhtn,otdh,ts) = ∑

rhtn1

FlowHeatInterRg(otdh,rhtn,rhtn1,ts) + FlowHeatIntraRg(otdh,oddh,rhtn,ts)*(1−LossIntraRg(rhtn,otdh,oddh,ts)) (A3)

where FlowHeatIntraRg (otdh, oddh, rhtn, ts) represents heat that is transferred from the heat transmission option, otdh, to the heat distribution option,

oddh, at node rhtn. LossIntraRg in this case represents losses in the transformer responsible for stepping down the ow, which is assumed to be zero in

this paper. Then, we have a nodal ow balance equation as

FlowIn(rhtn,otdh,ts) = FlowOut(rhtn,otdh,ts)(A4)

Heat needs to ow from the transmission node to the distribution node (the city), which happens via the distribution grid. The ow of heat from the

heat transmission node, rhtn, to the distribution node, rhdn, via the distribution option/grid, oddh, is (an equation analogous to eq. 1)

FlowHeatInterRg(oddh,rhtn,rhdn,ts) ≤ AF(oddh,ts)*CapFlow(rhtn,rhdn,oddh)*C2A(oddh)*Y(ts)(A5)

The generalized inow equation for distribution DH is as follows:

FlowIn(rhdn,oddh,ts) = ∑

rhtn

FlowHeatInterRg(oddh,rhtn,rhdn,ts)*(1−LossInterRg(oddh,rhtn,rhdn,ts)) (A6)

The heat that enters the heat distribution node is analogous to Eq. 2, only differences being the distribution has a unique ow from one trans-

mission node and no supply of heat sources to the distribution grid within the heat distribution node. The equation is generalized though where

multiple entry points is represented. LossInterRg represents heat losses in the distribution network for connection between transmission node, rhtn,

and distribution node, rhdn.

The heat that leaves the heat distribution grid meets the end-use heat demand in the BE sector. The equation is

FlowOut(rhdn,oddh,ts) = ∑

obe

FlowHeatIntraRg(oddh,obe,rhdn,ts)*(1−LossesIntraRg(oddh,obe,rhdn,ts)) (A7)

S. Sahoo et al.

Energy Conversion and Management 277 (2023) 116599

where LosssesIntraRg corresponds to heat exchanger losses before nal delivery to the end-use options.

Similar to ow balance in transmission, we have the following balance equation at the heat distribution node rhdn as

FlowIn(rhdn,oddh,ts) = FlowOut(rhdn,oddh,ts)(A8)

The total DH infrastructure cost equation is as follows:

CostDH =CostTrans +CostDistri (A9)

where transmission DH cost is

CostTrans =a*∑

otdh,rhtn,rhtn1

(CapFlow(rhtn,rhtn1,otdh)*(InvestmentTrans (otdh,rhtn,rhtn1) ) )

+∑

otdh,rhtn,rhtn1,ts

(FlowHeatInterRg (otdh,rhtn,rhtn1,ts)*VarO&MT rans (otdh,rhtn,rhtn1,ts)) (A10)

where annuity factor a is calculated considering a discount rate of 2.25 % was considered based on the Central Planning Bureau of the Netherlands and

a lifetime of 50 years for both transmission and distribution DHs [56]. InvestmentTrans and VarO&MTrans represent capital expenditure and

variable operation and maintenance cost, respectively, associated with the transmission network.

Distribution DH cost is

CostDistri =a*∑

oddh,rhtn,rhdn(CapFlow (oddh,rhtn,rhdn)*(C1(rhtn,rhdn,oddh) + C2(rhtn,rhdn,oddh)*Dd(rhtn,rhdn,oddh) )

Capad(rhtn,rhdn,oddh))

+∑

otdh,rhtn,rhtn1,ts

(FlowHeatInterRg (oddh,rhtn,rhdn,ts)*VarO&MDistri (oddh,rhtn,rhdn,ts))

(A11)

where parameters C

and C

are construction cost constant and construction cost coefcient, respectively – refer Persson and Werner [7] for a detailed

description on these parameters (also see the description below for the values of these parameters used in this paper). Capa

represents the standard

capacity of the distribution cost as obtained from Appx. Table A4, which also presents standard Investment costs and interregional losses. This table

represents attributes for piping series 1 and considered assumptions mentioned in the catalog from a pipe producer Powerpipe [73]. The distribution

DH capital cost (the rst part of the RHS of the above equation) is a modied version of the capital cost mentioned in [58]. For the DH network, we

only focused on distribution capital cost as this cost constitutes more than half of the total distribution costs of a standard DH network [7]. Var-

O&MDistri represents variable O&M costs associated with distribution network. Calculating operating costs can be uncertain and would require more

detailed modeling of a distribution network, for example including water ow levels, pumping pressure and current, and pressure losses, which are

beyond the scope of our paper. The cost equations provide the possibility of including other operating costs, though. We considered a standard ca-

pacity pipeline of 125 MW and 45 MW for transmission and distribution DH, respectively.

Appx. Table A5 provides data on xed parameters used in the distribution capital cost equation for different region types. We used inner city area

parameters for Groningen inner city and outer city area parameters for other cities. Other region-specic modeling change is that we divided the

Groningen outer city into eight equal parts, along with equal BE demand distribution in each region, because this region is much larger than that of

other cities.

Appendix D. Detailed analysis

Biomass has a variety of supply sources and results in different energy carriers (Appx. Fig. A1), hence we analyzed biomass in detail. Biomass grass

converted to biogas has the highest contribution, i.e., 9.5 PJ, with predominant contributions from Het Hogeland municipality (3.1 PJ) and the

remaining part of Westerkwartier (2.6 PJ) municipality. Industry Delfzijl has the highest biomass supply (4 PJ) due to the presence of a large industrial

cluster utilizing various biomass carriers. Regarding conversion, biogas has the highest contribution of 9.8 PJ, which is lower compared to the 15 PJ

considered in [74], though [74] considered biogas from sources not included in our study, such as agricultural residues, energy crops, and sea algae.

Heat has the second highest contribution of 7 PJ.

Biomass grass and energy crops are either fully utilized within a municipality or completely left out (Appx. Fig. A2). Eemsdelta and Oldambt are

amongst municipalities without these sources. Biomass local wood chips and straw are almost completely unutilized in the province of Groningen –

also see Appx. Fig. A1. For grass and energy crops utilization within the province of Groningen, we see some municipalities utilize their full potential,

whereas neighboring municipalities did not have any utilization as each municipality is a distinct region in our modeling, and the transport between

regions of these regional types of biomass is not included in our method.

The average utilization share of rooftop PV-BE capacity potential is higher in the cities (0.44 average) compared to the rest of the municipalities

(0.35) - Appx. Fig. A3. Neither cities nor municipalities rest have a full utilization share of rooftop PV. Our rooftop PV capacity utilization of 1.8 GW is

higher compared to the range considered in [65] for 2050.

Appx. Fig. A4 represents a cost breakdown for cost of options (energy demand, supply, and infrastructure) and compares between [4] and this

study. Variable O&M costs shows supply options making the largest contribution, even though the difference between [4] and this study is minimal, i.

e., this paper has a 0.3 M

€

/year or 0.3 % more contribution compared to [4]. CAPEX graph shows signicant difference for the energy-demanding

sector BE, for which this study is 1070 M

€

/year (191 % more) higher than that of [4]. The primary reason for this difference is the changes in the

energy labels of the building stocks as mentioned in the main text. CAPEX investment in energy infrastructure is also higher in this paper compared to

[4], i.e., 691 M

€

/year. OPEX investment comparison also shows signicant positive difference related to the BE and energy infrastructure for this

paper compared to [4]. For example, energy infrastructure has a cost difference of 172 M

€

/year. Fuel cost comparison shows the BE and agriculture

have high cost for this paper compared to [4].

S. Sahoo et al.

Energy Conversion and Management 277 (2023) 116599

References

[1] Eurostat. Key gures on Europe; 2019. Doi: 0628.

[2] Ministry of Economic Affairs. Energy Agenda. Den Haag; 2017.

[3] Netbeheer E. TenneT, Netbeheer Nederland. Netimpactrapportage RES 2021;1 (in

Dutch).

[4] Sahoo S, van Stralen JNP, Zuidema C, Sijm J, Yamu C, Faaij APC. Regionalization

of a national integrated energy system model: A case study of the northern

Netherlands. Appl Energy 2022:306.

[5] Sahoo S, Zuidema C, van Stralen JNP, Sijm J, Faaij A. Detailed spatial analysis of

renewables’ potential and heat: A study of Groningen Province in the northern

Netherlands. Appl Energy 2022:318. https://doi.org/10.1016/j.

apenergy.2022.119149.

[6] ¨

Ozdemir ¨

O, Hers JS, Fisher EB, Brunekreeft G, Hobbs BF. A nodal pricing analysis

of the future german electricity market. 2009 6th Int Conf Eur Energy Mark EEM

2009; 2009. Doi: 10.1109/EEM.2009.5207112.

[7] Persson U, Werner S. Heat distribution and the future competitiveness of district

heating. Appl Energy 2011;88:568–76. https://doi.org/10.1016/j.

apenergy.2010.09.020.

[8] Petrovic SN, Karlsson KB. Danish heat atlas as a support tool for energy system

models. Energy Convers Manag 2014;87:1063–76. https://doi.org/10.1016/j.

enconman.2014.04.084.

[9] Strachan N, Balta-Ozkan N, Joffe D, McGeevor K, Hughes N. Soft-linking energy

systems and GIS models to investigate spatial hydrogen infrastructure development

in a low-carbon UK energy system. Int J Hydrogen Energy 2009;34:642–57.

https://doi.org/10.1016/j.ijhydene.2008.10.083.

[10] M¨

oller B, Lund H. Conversion of individual natural gas to district heating:

Geographical studies of supply costs and consequences for the Danish energy

system. Appl Energy 2010;87:1846–57. https://doi.org/10.1016/j.

apenergy.2009.12.001.

[11] Lund H, M¨

oller B, Mathiesen BV, Dyrelund A. The role of district heating in future

renewable energy systems. Energy 2010;35:1381–90. https://doi.org/10.1016/j.

energy.2009.11.023.

[12] Petrovi´

c SN, Karlsson KB. Residential heat pumps in the future Danish energy

system. Energy 2016;114:787–97. https://doi.org/10.1016/j.energy.2016.08.007.

[13] van den Broek M, Brederode E, Ramírez A, Kramers L, van der Kuip M,

Wildenborg T, et al. Designing a cost-effective CO2storage infrastructure using a

GIS based linear optimization energy model. Environ Model Softw 2010;25:

1754–68. https://doi.org/10.1016/j.envsoft.2010.06.015.

[14] van den Broek M, Ramírez A, Groenenberg H, Neele F, Viebahn P, Turkenburg W,

et al. Feasibility of storing CO2in the Utsira formation as part of a long term Dutch

CCS strategy. An evaluation based on a GIS/MARKAL toolbox. Int J Greenh Gas

Control 2010;4:351–66. https://doi.org/10.1016/j.ijggc.2009.09.002.

[15] Alhamwi A, Medjroubi W, Vogt T, Agert C. GIS-based urban energy systems models

and tools: Introducing a model for the optimisation of exibilisation technologies

in urban areas. Appl Energy 2017;191:1–9. https://doi.org/10.1016/j.

apenergy.2017.01.048.

[16] Alhamwi A, Medjroubi W, Vogt T, Agert C. Development of a GIS-based platform

for the allocation and optimisation of distributed storage in urban energy systems.

Appl Energy 2019;251:113360. https://doi.org/10.1016/j.apenergy.2019.113360.

[17] Schiefelbein J, Javadi AP, Diekerhof M, Streblow R. GIS supported city district

energy system modeling. Proc 9th Int Conf Syst Simul Build; 2014. Doi: 10.13140/

RG.2.1.5036.1444.

[18] Pakere I, Kacare M, Gr¯

avelsin¸ ˇ

s A, Freimanis R, Blumberga A. Spatial analyses of

smart energy system implementation through system dynamics and GIS modelling.

Wind power case study in Latvia. Smart Energy 2022:7. https://doi.org/10.1016/j.

segy.2022.100081.

[19] Colbertaldo P, Guandalini G, Campanari S. Modelling the integrated power and

transport energy system: The role of power-to-gas and hydrogen in long-term

scenarios for Italy. Energy 2018;154:592–601. https://doi.org/10.1016/j.

energy.2018.04.089.

[20] Colbertaldo P, Cerniauskas S, Grube T, Robinius M, Stolten D, Campanari S. Clean

mobility infrastructure and sector integration in long-term energy scenarios: The

case of Italy. Renew Sustain Energy Rev 2020;133:110086. https://doi.org/

10.1016/j.rser.2020.110086.

[21] Lombardi F, Pickering B, Colombo E, Pfenninger S. Policy Decision Support for

Renewables Deployment through Spatially Explicit Practically Optimal

Alternatives. Joule 2020;4:2185–207. https://doi.org/10.1016/j.

joule.2020.08.002.

[22] Borasio M, Moret S. Deep decarbonisation of regional energy systems: A novel

modelling approach and its application to the Italian energy transition. Renew

Sustain Energy Rev 2022;153:111730. https://doi.org/10.1016/j.

rser.2021.111730.

[23] Pye S, Sabio N, Strachan N. An integrated systematic analysis of uncertainties in UK

energy transition pathways. Energy Policy 2015;87:673–84. https://doi.org/

10.1016/j.enpol.2014.12.031.

[24] Pfenninger S, Keirstead J. Renewables, nuclear, or fossil fuels? Scenarios for Great

Britain’s power system considering costs, emissions and energy security. Appl

Energy 2015;152:83–93. https://doi.org/10.1016/j.apenergy.2015.04.102.

[25] Moore A, Price J, Zeyringer M. The role of oating offshore wind in a renewable

focused electricity system for Great Britain in 2050. EnergyStrateg Rev 2018;22:

270–8. https://doi.org/10.1016/j.esr.2018.10.002.

[26] Bartholdsen HK, Eidens A, L¨

ofer K, Seehaus F, Wejda F, Burandt T, et al. Pathways

for Germany’s low-carbon energy transformation towards 2050. Energies 2019;14:

1–33. https://doi.org/10.3390/en12152988.

[27] Tigas K, Giannakidis G, Mantzaris J, Lalas D, Sakellaridis N, Nakos C, et al. Wide

scale penetration of renewable electricity in the Greek energy system in view of the

European decarbonization targets for 2050. Renew Sustain Energy Rev 2015;42:

158–69. https://doi.org/10.1016/j.rser.2014.10.007.

[28] Song H, Dotzauer E, Thorin E, Guziana B, Huopana T, Yan J. A dynamic model to

optimize a regional energy system with waste and crops as energy resources for

greenhouse gases mitigation. Energy 2012;46:522–32. https://doi.org/10.1016/j.

energy.2012.07.060.

[29] Limpens G, Moret S, Jeanmart H, Mar´

echal F. EnergyScope TD: A novel open-

source model for regional energy systems. Appl Energy 2019;255:113729. https://

doi.org/10.1016/j.apenergy.2019.113729.

[30] Amirhossein F, Manuel SD, Jos S, Germ´

an ME, Andr´

e F. Measuring accuracy and

computational capacity trade-offs in an hourly integrated energy system model.

Adv Appl Energy 2021;1:100009. https://doi.org/10.1016/j.adapen.2021.100009.

[31] Marques-Perez I, Guaita-Pradas I, Gallego A, Segura B. Territorial planning for

photovoltaic power plants using an outranking approach and GIS. J Clean Prod

2020;257.. https://doi.org/10.1016/j.jclepro.2020.120602.

[32] Watson JJW, Hudson MD. Regional Scale wind farm and solar farm suitability

assessment using GIS-assisted multi-criteria evaluation. Landsc Urban Plan 2015;

138:20–31. https://doi.org/10.1016/j.landurbplan.2015.02.001.

[33] Liu Y, Chen S, von Cossel M, Xu B, Gao H, Jiang R, et al. Evaluating the suitability

of marginal land for a perennial energy crop on the Loess Plateau of China. GCB

Bioenergy 2021;13:1388–406. https://doi.org/10.1111/gcbb.12865.

[34] van Stralen JNP, Dalla Longa F, Dani¨

els B, Smekens K, van der Zwaan B. OPERA: A

New High-Resolution Energy System Model for Sector Integration Research.

Environ Model Assess 2020. https://doi.org/10.1007/s10666-020-09741-7.

[35] AIMMS. AIMMS:: AIMMS Optimization Modeling; 2018. https://aimms.com/e

nglish/developers/resources/manuals/optimization-modeling/ (accessed October

9, 2018).

[36] Open source community. QGIS n.d. https://qgis.org/en/site/ (accessed October 5,

2020).

[37] ESRI. Get started with ArcMap—ArcMap | Documentation n.d. https://desktop.arc

gis.com/en/arcmap/latest/get-started/main/get-started-with-arcmap.htm

(accessed April 11, 2022).

[38] CBS. Dwellings and non-residential stock; changes, utility function, regions; 2021.

https://opendata.cbs.nl/statline/#/CBS/en/dataset/81955ENG/table (accessed

February 9, 2021).

[39] Persson U, Wiechers E, M¨

oller B, Werner S. Heat Roadmap Europe: Heat

distribution costs. Energy 2019;176:604–22. https://doi.org/10.1016/j.

energy.2019.03.189.

[40] Provincie Groningen. Verordening van Provinciale Staten van de provincie

Groningen houdende ruimtelijke ordening Omgevingsverordening Provincie

Groningen 2016 (in Dutch); 2020.

[41] Kramers L, Van Wees JD, Pluymaekers MPD, Kronimus A, Boxem T. Direct heat

resource assessment and subsurface information systems for geothermal aquifers;

The Dutch perspective. Geol En Mijnbouw/Netherlands J Geosci 2012;91:637–49.

https://doi.org/10.1017/S0016774600000421.

[42] Geothermie Nederland. Geothermie locatie van de diverse projecten in Nederland

(in Dutch); 2020. https://geothermie.nl/index.php/nl/geothermie-aardwa

rmte/geothermie-in-nederland/projectoverzicht (accessed April 22, 2022).

[43] TNO. Map Viewer | Thermogis n.d. https://www.thermogis.nl/en/map-viewer

(accessed September 2, 2019).

[44] Rivera Diaz A, Kaya E, Zarrouk SJ. Reinjection in geothermal elds - A worldwide

review update. Renew Sustain Energy Rev 2016;53:105–62. https://doi.org/

10.1016/j.rser.2015.07.151.

[45] KNMI - Koninklijk Nederlands Meteorologisch Instituut. KNW Data | KNW Atlas |

KNMI Projects n.d. https://www.knmiprojects.nl/projects/knw-atlas/knw-data

(accessed April 5, 2022).

[46] Bang C, Vitina A, Gregg JS, Lindboe HH. Analysis of biomass prices; 2013.

[47] van der Hilst F, Dornburg V, Sanders JPM, Elbersen B, Graves A, Turkenburg WC,

et al. Potential, spatial distribution and economic performance of regional biomass

chains: The North of the Netherlands as example. Agric Syst 2010;103:403–17.

https://doi.org/10.1016/j.agsy.2010.03.010.

[48] Lensink S, Schoots K. Eindadvies basisbedragen SDE++2021; 2021.

[49] qpunkt gmbh. Waste heat recovery systems for the Glass Industry; 2014.

[50] TenneT, Elia, ENTSOE. HoogspanningsNet Netkaart 2021. https://webkaart.

hoogspanningsnet.com/index2.php#10/51.3958/4.5483 (accessed November 18,

2021).

[51] TenneT_Nederland. TenneT Assets (hoogspanning) 2020. https://www.arcgis.co

m/home/item.html?id=646a6dee22bf485587bc4daf98da1306 (accessed March

14, 2020).

[52] Matthijsen J, Dammers E, Elzenga H. De toekomst van de Noordzee. De Noordzee

in 2030 en 2050: een scenariostudie; 2018.

[53] ENEXIS. Open data | Enexis - Energie in goede banen (in Dutch) 2021.

https://www.enexis.nl/over-ons/wat-bieden-we/andere-diensten/open-data

(accessed March 23, 2021).

[54] CEER. 2nd CEER Report on Power Losses; 2020.

[55] ETSAP. Pricing in electricity transmission and distribution. vol. E12; 2014. Doi:

10.1109/melcon.1996.550955.

[56] Danish Energy Agency, Energinet. Technology Data for Energy Transport; 2021.

[57] Danish Energy Agency, Energinet. Technology Data Generation of Electricity and

District heating; 2022.

[58] Nielsen S, M¨

oller B. GIS based analysis of future district heating potential in

Denmark. Energy 2013;57:458–68.

[59] Nussbaumer T, Thalmann S, Jenni A, K¨

odel J. Handbook on Planning of District

Heating. Networks 2020.

S. Sahoo et al.

Energy Conversion and Management 277 (2023) 116599

[60] QGIS. Network Analysis; 2022. https://docs.qgis.org/3.22/en/docs/training_man

ual/vector_analysis/network_analysis.html.

[61] Provincie Groningen. RES Naslagwerk regionale structuur Warmte (in Dutch);

2021.

[62] Berenschot, Kalavasta. Klimaatneutrale scenario’s 2050 – Scenariostudie ten

behoeve van de integrale infrastructuurverkenning 2030-2050; 2020 (in Dutch).

[63] Provincie Groningen. Klimaatagenda Provincie Groningen 2030 (in Dutch); 2022.

[64] Provincie Groningen. Regionale Energie Strategie 1.0 Groningen (in Dutch); 2021.

[65] van der Niet S, Rooijers F, van der Veen R, Voulis N, Wirtz A, Lubben M.

Systeemstudie energie-infrastructuur Groningen & Drenthe (in Dutch). Delft; 2019.

[66] CBS. StatLine - Energy consumption private dwellings; type of dwelling and regions

2018. https://opendata.cbs.nl/statline/#/CBS/en/dataset/81528ENG/table?

ts=1550249238313 (accessed February 15, 2019).

[67] ENEXIS. Congestiemanagement onderzoeken | Enexis Netbeheer 2022. https://

www.enexis.nl/zakelijk/aansluitingen/congestie-onderzoeken (accessed April 11,

2022).

[68] Agency NE. Features SDE++ | RVO.nl 2021. https://english.rvo.nl/subsidies-progr

ammes/sde/features (accessed April 24, 2022).

[69] Energiecooperatie duurzaam Assen, Natuur en Milieu Drenthe, Bronen VanOns,

ADAMANT Solar. Bundeling verslagen van en na informatieavonden Zonneplas

Ubbena – Assen; 2020.

[70] Choi YK. A study on power generation analysis of oating PV system considering

environmental impact. Int J Softw Eng Its Appl 2014;8:75–84. https://doi.org/

10.14257/ijseia.2014.8.1.07.

[71] Beckers KF, Performance YKR. Cost, and Financial Parameters of Geothermal

District Heating Systems for Market Penetration Modeling under Various Scenarios.

42nd Work Geotherm Reserv Eng 2017:1–11.

[72] Scheepers M, Palacios SG, Jegu E, Nogueira LP, Rutten L, van Stralen J, et al.

Towards a climate-neutral energy system in the Netherlands. Renew Sustain

Energy Rev 2022;158:112097. https://doi.org/10.1016/j.rser.2022.112097.

[73] Powerpipe. District heating network pipe catalogue; 2016.

[74] Noordelijke Rekenkamer. Energietransitie provincie Groningen (in Dutch); 2016.

[75] European Commission. PVGIS Photovoltaic Geographical Information System.

2022 n.d. https://joint-research-centre.ec.europa.eu/pvgis-photovoltaic-geographi

cal-information-system_en (accessed November 29, 2022).

S. Sahoo et al.

A flexible approach to GIS based modelling of a global hydrogen transport system

Article

Sep 2023
INT J HYDROGEN ENERG

Residual Agroforestry Biomass Supply Chain Simulation Insights and Directions: A Systematic Literature Review

Article

Full-text available

Jun 2023

Residual biomass is a reliable source of energy and hence requires effective supply chain management for optimal performance and sustainability. While there are various studies on this recent trend, a comprehensive review of the literature on simulation-based modeling of the supply chain for residual agroforestry biomass is lacking. This study aims to present a systematic review of relevant literature surrounding residual agroforestry supply chain simulation insights and directions. The systematic literature review was carried out in accordance with PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) 2020 standards and intends to answer the research questions based on (1) Key Performance Indicators (KPI); (2) Simulation techniques; and (3) Efficient supply chain. A search of the Science Direct, SCOPUS, and UA EBSCO databases was conducted using the appropriate keywords combination. The databases were searched, and a total of 1617 papers were appraised automatically. Subsequently, the titles, keywords, and abstracts of 172 papers were examined. Following the full-text analysis, 20 papers in addition to 27 articles taken from other sources matched the requirements for study inclusion. The publications accessed reveals that simulation-based techniques will optimize the supply chain for residual biomass when applied.

A framework to identify offshore spatial trade-offs in different space allocation options for Offshore Wind Farms, as part of the North Sea Offshore Grid

Article

Jun 2024

The role of regional renewable energy integration in electricity decarbonization—A case study of Japan

Article

Jun 2024
APPL ENERG

A review of thermal coupling system of fuel cell-metal hydride tank: Classification, control strategies, and prospect in distributed energy system

Article

May 2023
INT J HYDROGEN ENERG

Hydrogen is of great importance to solve the energy crisis and environment pollution, and multiple hydrogen storage methods have been proposed, among which the metal hydrides (MHs) have attracted wide attentions in recent years. Considering the complementary heat demand of MH and fuel cell (FC) during operation, a thermal coupling system (TCS), where part of the heat generated by FC is transferred to MH tank to facilitate the endothermic hydrogen desorption, has been proposed and investigated in the literatures. However, since the heating of MH tank may retard the evolution of FC towards set operation point and decrease the system efficiency, a control strategy with rapid response and high precision is indispensable for the steady and efficient operation of TCS. This study reviews the research status and limitations of TCS classified according to the heat transfer mode, then the control strategies under various application scenarios are discussed and summarized. Furthermore, three crucial technical issues associated with the application of TCS in distributed energy system are highlighted for future investigations.

Detailed spatial analysis of renewables’ potential and heat: A study of Groningen Province in the northern Netherlands

Article

Full-text available

Jul 2022
APPL ENERG

Spatially sensitive regional renewables’ potentials are greatly influenced by existing land-use claims and related spatial and environmental policies. Similarly, heat particularly related to low-temperature demand applications in the built environment (BE) is highly spatially explicit. This study developed an analytical approach for a detailed spatial analysis of future solar PV, onshore wind, biomass, and geothermal and industrial waste heat potentials at a regional level and applied in the Dutch Province of Groningen. We included spatial policies, various spatial claims, and other land-use constraints in developing renewable scenarios for 2030 and 2050. We simultaneously considered major spatial claims and multiple renewable energy sources. Claims considered are the BE, agriculture, forest, nature, and network and energy infrastructure, with each connected to social, ecological, environmental, technical, economic, and policy-related constraints. Heat demand was further analyzed by creating highly granular demand density maps, comparing them with regional heat supply potential, and identifying the economic feasibility of heat networks. We analyzed the possibilities of combining multiple renewables on the same land. The 2050 renewable scenarios results ranged 2–66 PJ for solar PV and 0–48 PJ for onshore wind and biomass ranged 3.5–25 PJ for both 2030 and 2050. These large ranges of potentials show the significant impact of spatial constraints and underline the need for understanding how they shape future energy policies. The heat demand density map shows that future heat networks are feasible in large population centers. Our approach is pragmatic and replicable in other regions, subject to data availability.

Towards a climate-neutral energy system in the Netherlands

Article

Full-text available

Apr 2022
RENEW SUST ENERG REV

This paper presents two different scenarios for the energy system of the Netherlands that achieve the Dutch government's national target of near net-zero greenhouse gas emissions in 2050. Using the system optimisation model OPERA, the authors have analysed the technology, sector and cost implications of the assumptions underlying these scenarios. While the roles of a number of key energy technology and emission mitigation options are strongly dependent on the scenario and cost assumptions, the analysis yields several common elements that appear in both scenarios and that consistently appear under differing cost assumptions. For example, one of the main options for the decarbonisation of the Dutch energy system is electrification of energy use in end-use sectors and for the production of renewable hydrogen with electrolysers. As a result the level of electricity generation in 2050 will be three to four times higher than present generation levels. Ultimately, renewable energy – particularly from wind turbines and solar panels – is projected to account for the vast majority of electricity generation, around 99% in 2050. Imbalances between supply and demand resulting from this variable renewable electricity production can be managed via flexibility options, including demand response and energy storage. Hydrogen also becomes an important energy carrier, notably for transportation and in industry. If import prices are lower than costs of domestic production from natural gas with CCS or through electrolysis from renewable electricity (2.4–2.7 €/kgH2), the use of hydrogen increases, especially in the built environment.

Deep decarbonisation of regional energy systems: A novel modelling approach and its application to the Italian energy transition

Article

Full-text available

Jan 2022
RENEW SUST ENERG REV

Deep decarbonisation – i.e. the transition towards net-zero emissions energy systems – will be enabled by a high penetration of intermittent renewables, storage and sector-coupling technologies. In this paper, we present a novel modelling approach to capture the increasing complexity of such future energy systems and help policy makers choose among the different possible transition scenarios. Salient features of our model, consisting of an extended and regionalised version of EnergyScope (Limpens et al., 2019 [1]), are a low computational time and a concise formulation which make it suitable for uncertainty and what-if analyses. As a case study, the model is applied to devise scenarios for the Italian energy transition. Specifically, we develop the first open-source whole-energy system model of Italy and assess the feasibility of its decarbonisation strategy with respect to uncertainties in the deployment of carbon capture and storage (CCS) and renewable technologies. Results show that emissions can be cut by 79%–97% vs. 1990 levels thanks to a radical electrification of the energy system coupled to a wide deployment of renewables and efficient energy conversion technologies. Finally, we discuss the synergies, advantages and disadvantages of our proposed approach with respect to alternative modelling approaches used across 88 recent deep decarbonisation studies. The analysis suggests that our model, thanks to its computational efficiency and a snapshot approach (i.e., modelling a target-year in the future), can complement more detailed and established energy models optimising the energy transition pathway (i.e., modelling the pathway from today to the target year).

Evaluating the suitability of marginal land for a perennial energy crop on the Loess Plateau of China

Article

Full-text available

Jun 2021

With a large marginal land area, the Loess Plateau in China holds great potential for biomass production and environmental improvement. Identifying suitable locations for biomass production on marginal land is important for decision makers from the viewpoint of land‐use planning. However, there is limited information on the suitability of marginal land within the Loess Plateau for biomass production. Therefore, this study aims to evaluate the suitability of the promising perennial energy crop switchgrass (Panicum virgatum L.) on marginal land across the Loess Plateau. A fuzzy logical model was developed and validated based on field trials on the Loess Plateau and applied to the marginal land of this region, owing to its ability of dealing with the continuous nature of soil, landscape variations, and uncertainties of the input data. This study identified that approximately 12.8‐20.8 Mha of the Loess Plateau as available marginal land, of which 2.8‐4.7 Mha is theoretically suitable for switchgrass cultivation. These parts of the total marginal land are mainly distributed in northeast and southwest of the Loess Plateau. The potential yield of switchgrass ranges between 44‐77 Tg. This study showed that switchgrass can grow on a large proportion of the marginal land of the Loess Plateau and therefore offers great potential for biomass provision. The spatial suitability maps produced in this study provide information to farmers and policy makers to enable a more sustainable development of biomass production on the Loess Plateau. In addition, the fuzzy theory‐based model developed in this study provided a good framework for evaluating the suitability of marginal land.

Measuring accuracy and computational capacity trade-offs in an hourly integrated energy system model

Article

Full-text available

Feb 2021

Improving energy system modeling capabilities can directly affect the quality of applied studies. However, some modeling trade-offs are necessary as the computational capacity and data availability are constrained. In this paper, we demonstrate modeling trade-offs resulting from the modification in the resolution of four modeling capabilities, namely, transitional scope, European electricity interconnection, hourly demand-side flexibility description, and infrastructure representation. We measure the cost of increasing resolution in each capability in terms of computational time and several energy system modeling indicators, notably, system costs, emission prices, and electricity import and export levels. The analyses are performed in a national-level integrated energy system model with a linear programming approach that includes the hourly electricity dispatch with European nodes. We determined that reducing the transitional scope from seven to two periods can reduce the computational time by 75% while underestimating the objective function by only 4.6%. Modelers can assume a single European Union node that dispatches electricity at an aggregated level, which underestimates the objective function by 1% while halving the computational time. Furthermore, the absence of shedding and storage flexibility options can increase the curtailed electricity by 25% and 8%, respectively. Although neglecting flexibility options can drastically decrease the computational time, it can increase the sub-optimality by 31%. We conclude that an increased resolution in modeling flexibility options can significantly improve the results. While reducing the computational time by half, the lack of electricity and gas infrastructure representation can underestimate the objective function by 4% and 6%, respectively.

OPERA: a New High-Resolution Energy System Model for Sector Integration Research

Article

Full-text available

Dec 2021
ENVIRON MODEL ASSESS

This article introduces and describes OPERA, a new technology-rich bottom-up energy system optimization model for the Netherlands. We give a detailed specification of OPERA’s underlying methodology and approach, as well as a description of its multiple applications. The model has been used extensively to formulate strategic policy advice on energy decarbonization and climate change mitigation for the Dutch government, and to perform exploratory studies on the role of specific low-carbon energy technologies in the energy transition of the Netherlands. Based on a reference scenario established through extensive consultation with industry and the public sector, OPERA allows for examining the impact of autonomous technology diffusion and energy efficiency improvement processes, and investigating a broad range of policy interventions that target greenhouse gas emission abatement or air pollution reduction, amongst others. Two characteristics that render OPERA particularly useful are the fact that (1) it covers the entire energy system and all greenhouse gas emissions of the Netherlands, and (2) it possesses a high temporal resolution, including a module for flexibly handling individual time units. The simulation of the complete Dutch energy system and the ability to represent energy supply and demand on an hourly basis allow for making balanced judgments on how to best accommodate large amounts of variable renewable energy production. OPERA is therefore particularly suited for analyzing subjects in the field of system integration, which makes it an ideal tool for assessing the implementation of the energy transition and the establishment of a low-carbon economy. We outline several near-term developments and opportunities for future improvements and refinements of OPERA. One of OPERA’s attractive features is that its structure can readily be applied to other countries, notably in Europe, and could thus relatively easily be extended to cover the entire European Union.

Spatial analyses of smart energy system implementation through system dynamics and GIS modelling. Wind power case study in Latvia

Article

Jun 2022

Major concern in utilising renewable energy sources, such as solar or wind, is their intermittent nature. Therefore, whether they can be reliable energy sources to provide uninterrupted energy demand when reaching a high share of renewable energy in the system depends on the whole energy supply network. Higher flexibility of RES-based systems can be reached through the development of smart energy concepts. The system dynamics model coupled with a geographical information system platform has been used to analyse space and time dimensions of RES potential using geo-referenced information. This approach allows analysing the whole energy system by determining the best-suited development scenario for each region separately, based on their resource, economic, and technological capabilities. Furthermore, the system dynamics model is complemented with different policies to evaluate their impact on renewable and local energy resource development and economic potential. The results show that the potential for large-scale onshore wind farm capacities with suitable land conditions could be around 5.5 GW, but the forecasted necessary wind power capacity to reach climate neutrality in 2050 is 1.55 GW. Therefore, onshore wind turbines should be promoted to move toward a smart and renewable power system in Latvia.

Regionalization of a national integrated energy system model: A case study of the northern Netherlands

Article

Jan 2022
APPL ENERG

Integrated energy system modeling tools predominantly focus on the (inter)national or local scales. The intermediate level is important from the perspective of regional policy making, particularly for identifying the potentials and constraints of various renewable resources. Additionally, distribution variations of economic and social sectors, such as housing, agriculture, industries, and energy infrastructure, foster regional energy demand differences. We used an existing optimization-based national integrated energy system model, Options Portfolio for Emission Reduction Assessment or OPERA, for our analysis. The modeling framework was subdivided into four major blocks: the economic structure, the built environment and industries, renewable energy potentials, and energy infrastructure, including district heating. Our scenario emphasized extensive use of intermittent renewables to achieve low greenhouse gas emissions. Our multi-node, regionalized model revealed the significant impacts of spatial parameters on the outputs of different technology options. Our case study was the northern region of the Netherlands. The region generated a significant amount of hydrogen (H2) from offshore wind, i.e. 620 Peta Joule (PJ), and transmitted a substantial volume of H2 (390 PJ) to the rest of the Netherlands. Additionally, the total renewable share in the primary energy mix of almost every northern region is ∼90% or more compared to ∼70% for the rest of the Netherlands. The results confirm the added value of regionalized modeling from the perspective of regional policy making as opposed to relying solely on national energy system models. Furthermore, we suggest that the regionalization of national models is an appropriate method to analyze regional energy systems.

Clean mobility infrastructure and sector integration in long-term energy scenarios: The case of Italy

Article

Nov 2020
RENEW SUST ENERG REV

As main contributors to greenhouse gas emissions, power and transportation are crucial sectors for energy system decarbonization. Their interaction is expected to increase significantly: plug-in electric vehicles add a new electric load, increasing grid demand and potentially requiring substantial grid upgrade; hydrogen production for fuel cell electric vehicles or for clean fuels synthesis could exploit the projected massive power overgeneration by intermittent and seasonally-dependent renewable sources via Power-to-Hydrogen. This work investigates the infrastructural needs involved with a broad diffusion of clean mobility, adopting a sector integration perspective at the national scale. The analysis combines a multi-node energy system balance simulation and a techno-economic assessment of the infrastructure to deliver energy vectors for mobility. The article explores the long-term case of Italy, considering a massive increase of renewable power generation capacity and investigating different mobility scenarios, where low-emission vehicles account for 50% of the stock. First, the model solves the energy balances, integrating the consumption related to mobility energy vectors and taking into account power grid constraints. Then, an optimal infrastructure is identified, composed of both a hydrogen delivery network and a widespread installation of charging points. Results show that the infrastructural requirements bring about investment costs in the range of 43–63 G€. Lower specific costs are associated with the exclusive presence of FCEVs, whereas the full reliance on BEVs leads to the most significant costs. Scenarios that combine FCEVs and BEVs lie in between, suggesting that the overall power + mobility system benefits from the presence of both drivetrain options.

Policy Decision Support for Renewables Deployment through Spatially Explicit Practically Optimal Alternatives

Article

Aug 2020

Designing highly renewable power systems involves a number of contested decisions, such as where to locate generation and transmission capacity. Yet, it is common to use a single result from a cost-minimizing energy system model to inform planning. This neglects many more alternative results, which might, for example, avoid problematic concentrations of technology capacity in any one region. To explore such alternatives, we develop a method to generate spatially explicit, practically optimal results (SPORES). Applying SPORES to Italy, we find that only photovoltaic and storage technologies are vital components for decarbonizing the power system by 2050; other decisions, such as locating wind power, allow flexibility of choice. Most alternative configurations are insensitive to cost and demand uncertainty, while dealing with adverse weather requires excess renewable generation and storage capacities. For policymakers, the approach can provide spatially detailed power system transformation options that enable decisions that are socially and politically acceptable.

Regionally integrated energy system detailed spatial analysis: Groningen Province case study in the northern Netherlands

Abstract

Recommended publications

Energy keep us going

University of Groningen in 83rd place on THE ranking list

Origins Center to open at Fundamentals of Life in the Universe symposium

A grand future with small particles. Nanotechnology affects many aspects of our lives

Detailed spatial analysis of renewables’ potential and heat: A study of Groningen Province in the no...

Regionalization of a national integrated energy system model: A case study of the northern Netherlan...

Detail or uncertainty? Applying global sensitivity analysis to strike a balance in energy system mod...

Analyzing the Techno-Economic Role of Nuclear Power in the Transition to the Net-Zero Energy System...