_____________________________________________________________________ 

_____________________________________________________________________ 

Department of Energy and Environment 

Division of Electric Power Engineering 

CHALMERS UNIVERSITY OF TECHNOLOGY 

Gothenburg, Sweden 2017 

 

 

 

 

 

European Electricity Market 
Modelling: Studies on Grid 
Investment and Impacts of 
Renewable Energy Resources 
Master’s thesis in Sustainable Energy Systems 

 

RODOLFO SILVEIRA 

 



 

 

 
 

 

 



 

MASTER’S THESIS 

 

 

 

 

 

European Electricity Market Modelling: Studies on Grid 

Investment and Impacts of Renewable Energy Resources 

  

Master’s Thesis in Sustainable Energy Systems 

RODOLFO SILVEIRA 

 

Supervisor & Examiner: 

TUAN LE 

 

 

 

 

 

 

 

 

 

 

Department of Energy and Environment 

Division of Electric Power Engineering 

CHALMERS UNIVERSITY OF TECHNOLOGY 

Gothenburg, Sweden 2017 



European Electricity Market Modelling: Studies on Grid Investment and Impacts of 

Renewable Energy Resources 

Master’s Thesis in Sustainable Energy Systems 

Rodolfo Silveira 

 

 

© RODOLFO SILVEIRA, 2017.  

 

 

Supervisor and Examiner: Tuan Le, Division of Electric Power Engineering 

 

Department of Energy and Environment 

Division of Electric Power Engineering   

Chalmers University of Technology 

SE-412 96 Gothenburg 

Sweden  

Telephone: + 46 (0)31-772 1000 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Cover: Grid representation of ENTSO-E member countries excluding Iceland and 

Cyprus. Each country is represented by a node. 

 

Printed by Chalmers Reproservice 

Gothenburg, Sweden 2017. 



 

 
I 

European Electricity Market Modelling: Studies on Grid Investment and Impacts of 

Renewable Energy Resources 

Rodolfo Silveira 

 

Abstract 

In order to mitigate greenhouse gas emissions and be less fossil-fuel dependent, 

European countries have been increasing the use of energy from Renewable Energy 

Sources (RES), following ambitious targets defined by the European Commission and 

by national efforts. Consequences emerge in terms of grid investment and in the 

energy market price. The objectives of the thesis are to assess the impacts of large-

scale integration of RES on the European power system on the increased transmission 

capacity requirement and the changes in market prices. The thesis also tries to identify 

best alternatives for investment in transmission interconnection capacity. 

In order to achieve the above objectives, the following were carried out in the thesis. 

Firstly, a base case scenario (2015) is defined by collecting most recent data on the 

existing loads, generation and interconnection capacities, as well as production costs 

for the different countries and different generation technologies. Secondly, future 

scenarios were defined for 2030 and 2050 considering the EU targets for generation 

capacity and future load forecast. Then, an electricity market model was developed 

based on a DC Optimal Power Flow (OPF). The model has been used to analyze the 

European market for future scenarios and interconnection transmission capacities 

between various countries. A Cost-Benefit Analysis (CBA) method is applied to 

identify critical paths (most congested power lines) and rank the best investment 

alternatives in transmission capacity reinforcements between the European Network 

of Transmission System Operators for Electricity (ENTSO-E) member countries. The 

benefits are defined by the avoided congestions costs, and the costs are the annualized 

investment costs. To see if the investment is profitable, a cost-to-benefit index is used. 

Simulations start with the base case scenario and proceeds to the future scenarios. 

Alternatives for transmission capacity expansion for the future scenarios have been 

evaluated and the future development of Locational Marginal Price (LMP) 

considering the different vision scenarios have been analyzed. Reinforcement 

connections in transmission capacity were proposed for many countries depending on 

the future horizon year. In 2030, proposed reinforcement interconnections include  

Norway-Netherlands, Austria-Italy and Switzerland-Germany, of 5784 MW, 4000 

MW and 3540 MW respectfully. For 2050, the number of proposed reinforcement 

interconnections is higher due to the significant increase of the share of RES in the 

generation mix. Yearly investment costs for transmission capacity of the ENTSO-E 

system varies from €4 up to €11 billion of dollars, for 2030 and 2050 alternatives 

respectfully. Countries highly dependent on fossil-fuels in the generation mix such as 

Netherlands, Belgium and United Kingdom present the higher electricity price most 

of the scenarios analyzed, reaching average prices of €90/MWh or higher. Since 

demand for electricity is expected to increase for the future (i.e. due the increase of 

use of electric vehicles and heat pumps), grid reinforcement is expected, also because 

of the high penetration of solar and wind power in the Net Generating Capacity 

(NGC) system. Due to low running costs of RES, LMPs are in most cases affected in 

the sense of bringing electricity prices down when a large time horizon is considered. 

In some scenarios, depending on economic situations assumptions, the running cost of 



 

 
II 

the RES is considered to be higher than that of conventional hydro and nuclear, in 

some countries. 

 

Key words: DC-OPF, Renewable Energy Resources, Locational Marginal Price, Net 

Generating Capacity, Net Transfer Capacity, Investment in Transmission Capacity. 

 



 
1 

Contents 

Abstract I 

Contents 1 

Preface 3 

Symbols 4 

Abbreviations 5 

List of Figures 7 

List of Tables 8 

1  Introduction 9 

1.1 Background 9 

1.2 Motivation 9 

1.3 Objectives 10 

1.4 Scope 10 

1.5 Main contributions 11 

1.6 Thesis organization 12 

2  Literature review 13 

2.1      Renewable Resources 13 

2.2     Previous Studies 13 
2.2.1 Electricity System Model 13 

2.2.2 The market value of variable renewable energy sources 14 
2.2.3 CBA Analysis of transmission network reinforcement 15 

2.2.4 Long-Term transmission planning considering reliability 16 
2.2.5 Ten-Year Network Development Plan (TYNDP) 16 

3  Modelling and Methodology 21 

3.1     Mathematical formulation of the Market Model 21 

     3.1.1 Objective Function 21 
     3.1.2 Active Power Balance 22 

     3.1.3 Maximum Production 22 
     3.1.4 Capacity Factor (CF) 22 
     3.1.5 Transmission Capacity Limits 23 
     3.1.6 Steady-state stability limit 23 

3.2     Locational Marginal Price (LMP) 23 

3.3     Investment in Transmission – Cost-Benefit Analysis (CBA) 24 
     3.3.1 Benefits 24 
     3.3.2 Costs 26 

4  Data Collection and Future Scenarios 29 



 
2 

4.1     Data collection 29 

     4.1.1 The process platform 29 
     4.1.2 About GAMS 30 
     4.1.3 Assumptions 30 

     4.1.4 Limitations consequences 32 

4.2     Definitions of Study Scenarios 32 
     4.2.1 Base Case (2015) 33 
     4.2.2 2030 visions 33 
     4.2.3 2050 visions 34 

5  Case Study: Results and Discussions 37 

5.1     Base Case Scenario 37 

     5.1.1 Energy Dispatch 37 
     5.1.2 Locational Marginal Price (LMP) – Base Case (2015) 39 

5.2     Investment in Transmission Alternatives 41 
     5.2.1 2030 Scenarios proposed alternatives 41 
     5.2.2 Avoided Congestion Costs (ACC) Analysis 44 

     5.2.3 2050 scenarios proposed alternatives 47 
     5.2.4 LMP effect after investment in transmission 51 

6  Conclusions and Future Work 55 

6.1     Conclusions 55 

6.2     Future Work 56 

References 57 

APPENDIX 61 

 

 

 

 

 

 

 

 

 

 



 
3 

Preface 

In this thesis, the project was started in middle of January 2016 with the planning 

phase, consisting of the description of all stages. All tests and simulations have been 

carried out in the laboratory for Master Thesis students in the Division of Electric 

Power Engineering at Chalmers University of Technology. Significant part of the 

work was spent on the data collection, to make them as reliable as possible, and on the 

model development. 

I give special thanks to Dr. Tuan Le, as my supervisor and examiner for all the thesis 

period, and the weekly meetings as a constructive feedback progression. I would also 

like to thank my course colleague Jelte Hipp for his contributions with the GAMS 

code model and important transmission lines data. 

It should be noted that the simulations could never have been conducted without the 

support of high quality and professionalism of the laboratory technical and 

administrative staff. 

Göteborg August 2017 

Rodolfo Silveira 



 
4 

Symbols 

Variables  

T  Power flow transferred from bus i to j. 

P  Active power generation from bus i. 

δ  Voltage angle at bus i. 

TOC Total system Operational Costs. 

Parameters  

AC  Actual transmission capacity reinforcement. 

ACC  Avoided Generation Costs. 

B  Susceptance. 

CF  Capacity Factor. 

CGen  Operational Cost of Generation. 

CRF  Capital Recovery Factor. 

IC  Investment Cost. 

IC0  Initial investment cost. 

ICK  Annualized hourly investment cost. 

PD  Power Demand. 

PMAX  Maximum capacity of generation. 

 Pwind Wind active power output. 

Profile  Wind capacity profile in hourly resolution. 

TMAX  Maximum active power transfer. 

TC  Transmission capacity reinforcement proposed. 

π  Locational Marginal Price. 

Sets  

i  Bus i. 

j  Bus j. 

plant  Generation power plant technology type. 

t  Time in hours. 



 
5 

Abbreviations 

AC  Actual Reinforcement 

ACC  Avoided Congestion Costs 

AEC  Avoided Environmental Costs 

BCI  Benefit-to-Cost Index 

CBA  Cost Benefit Analysis 

CC  Congestion Costs 

CCS  Carbon Capture and Storage 

CF  Capacity Factor 

CRF  Capital Recovery Factor 

CSP  Concentrated Solar Power 

 DG  Distributed Generation 

DSM  Demand Side Management 

EMMA  European Electricity Market Model 

ENTSO-E  European Network of Transmission System Operators for 

  Electricity 

EU  Europe Union 

EVs  Electric Vehicles 

FLH  Full Load Hours 

GAMS  General Algebraic Modelling System 

HP  Heat Pump 

IC  Normalized Investment Cost 

LMP  Locational Marginal Price 

LP  Linear Programming 

n  Number of circuits to expand 

NGC  Net Generating Capacity 

NTC  Net Transfer Capacity 

O&M  Operation and Maintenance 

PV  Photovoltaic 

RES  Renewable Energy Sources 

TC  Reinforcement Capacity proposed 

TC1  Transmission Capacity before expansion 



 
6 

TC2  Transmission Capacity after expansion 

TSO  Transmission System Operator 

TYNDP  Ten-Year Network Development Plan 

UNCCC  United Nations Convention on Climate Change 

VRE  Variable Renewable Energy 

  

 



 
7 

List of Figures 

Figure 1 – ENTSO-E member countries 33 buses network excluding Cyprus and 

Iceland, simplified Grid Model .................................................................................... 11 

Figure 2 – Map of main bottlenecks in the ENTSO-E perimeter [15] ......................... 18 

Figure 3 – Normalized value of wind power profile curve .......................................... 23 

Figure 4 – Representation of the locational marginal price of two different nodes in a 

constrained economic dispatch .................................................................................... 24 

Figure 5 – Flow diagram method for investment alternatives using a CBA ............... 26 

Figure 6  – Flowchart platform of Market Model simulation process ......................... 29 

Figure 7 – Vision Scenarios for 2030 [15] ................................................................... 33 

Figure 8 – Vision Scenarios for 2050 [15] ................................................................... 34 

Figure 9 – Dispatch of energy production January 2015 ENTSO-E countries Base 

Case Scenario ............................................................................................................... 37 

Figure 10 – Dispatch of energy production July 2015 Base Case Scenario ................ 38 

Figure 11 – Swedish Energy Dispatch Base Case Scenario, January 2015 ................. 39 

Figure 12 – Average LMP per country for January and July of 2015, Base Case 

Scenario ........................................................................................................................ 40 

Figure 13 – Average ENTSO-E simulated system price LMP daily curve 21/01/2015, 

Base Case Scenario ...................................................................................................... 40 

Figure 14 – Average NORDPOOL and EEX systems Elspot price daily curve 

21/01/2015 [4],[5], compared with simulated prices ................................................... 41 

Figure 15 – Transmission reinforcement connections proposed for 2030 ................... 46 

Figure 16 – Difference in the BCI for different transmission interconnections 

alternatives for 2030 and 2050 ..................................................................................... 50 

Figure 17 – LMP for 2030 scenarios, Visions 1 and 2 ................................................ 51 

Figure 18 – LMP for 2030 scenarios, Visions 3 and 4 ................................................ 52 

Figure 19 – 2050 LMP (in €/MWh), Vision 5 ............................................................. 53 

Figure 20 – 2050 LMP (in €/MWh), Vision 6 ............................................................. 53 

 
 

 



 
8 

List of Tables 

Table 1 – Comparison between all future vision scenarios with different parameters 35 

Table 2 – Alternative 1 Based on unconstrained case ................................................. 42 

Table 3 – Alternative 2 Based on constrained case considering persisting and new 

congestions from Alternative 1 .................................................................................... 43 

Table 4 – Alternative 3 Based on constrained case considering persisting and new 

congestions from Alternative 2 .................................................................................... 43 

Table 5 – Alternative 4 Based on unconstrained case, ranked by highest ACC values

 ...................................................................................................................................... 44 

Table 6 – Alternative 5 Based on constrained case considering persisting and new 

congestions from Alternative 4, ranked by highest ACC values ................................. 45 

Table 7 – Alternative 6 Based on constrained case considering persisting and new 

congestions from Alternative 5, ranked by highest ACC values ................................. 45 

Table 8 – Comparison between proposed interconnections for 2030 and TYNDP final 

report from 2014 .......................................................................................................... 47 

Table 9 – Alternative 1 for Vision 5 Based on partially constrained values, ranked by 

highest ACC values ...................................................................................................... 48 

Table 10 – Alternative 2 for Vision 6 Based on partially constrained values, ranked by 

highest ACC values ...................................................................................................... 49 

Table 11 – Summary of variable and investment costs, benefits and BCI index for the 

transmission interconnections alternatives for 2030 and 2050 .................................... 50 

 



 
9 

1 Introduction 

 

1.1 Background 

The European Parliament required an aim in 2009 for European Union (EU) countries 

to fulfil at least 20% of its energy demand from renewable sources by the year of 

2020, and at least 60-80% by 2050 [1]. The achievement is to be done through 

individual national efforts. The human contribution for the climate change, through 

the greenhouse gases emissions (especially CO2), is not the only matter to be faced. 

Concern for security of supply and less dependence on fossil fuels resources are 

tendencies of the upcoming years. Solar and wind power capacity are particularly 

growing fast.  The European power system will be expected to face major challenges 

in the future due to intermittent and geographical nature of generation from RES. 

Normally, these challenges include the need for increased transmission capacity 

between countries and within countries in Europe, as well as the need for alternative 

back-up capacity and regulation capacity mechanisms in the power system. A grid 

planning model for the European power system is therefore required to analyse the 

different options for transmission investments and to investigate changes in the 

electricity price in future scenarios. 

1.2 Motivation 

In order to reduce greenhouse gas emissions and be less fossil-fuel dependent, 

complying with the Kyoto Protocol and to the United Nations Convention on Climate 

Change (UNCCC), countries are encouraged to take actions due to the Renewable 

Energy Directive Policy rules. In this thesis, the measure focused is the increased use 

of energy from renewable sources [1] and its consequences in terms of grid 

investment and impacts created by the introduction of the RES in the market energy 

price. In this direction, this thesis demonstrates that there is a need of a new grid 

infrastructure and a value assigned to the investment. The preferences for low carbon 

technologies are driven by the falling costs of the RES options [2] and the fast 

expansion of wind and solar power as renewable alternatives. The introduction of 

solar and wind power and the gradual decrease of conventional generation sources 

affect the electricity prices. It is of interest of this thesis to investigate long-term 

electricity prices changes according to different future scenarios characteristics based 

on economic factors, policies, technology development and load forecast. LMP-based 

market pricing gives a clear market signal for investment in transmission [3]. 

The motivation to focus on the EU power system is that it is a highly interconnected 

system which is expecting a considerable penetration of wind and solar power 

generation for the next decades. Load and generation changes are handled by all units 

inside this system. Despite the EU power grid being highly interconnected, the 

dispatch and market operation of the electricity system are still not centralized. Even 

though not all EU member countries are part of the same common power market at 

the present, known as power pool, there is a future trend for the European power 

system to be more connected and operated in a coordinated manner. As a result, the 

whole system is not operated in an optimal centralized arrangement, although there 

are sub-regions highly integrated such as the NordPool area among Scandinavia and 

Baltic countries [4], and the European Energy Exchange (EEX) area [5] in central 

Europe. 



 
10 

1.3 Objectives 

The objective of this thesis is to assess the impacts of large-scale integration of 

renewable energy resources in the European power system. The thesis will focus on: 

1. Identifying the needs and quantifying the best alternative investments in the 

transmission capacity between European countries; 

2. Analysing the long-term power market impacts (i.e. total system operating 

cost, transmission investment cost and electricity price) in different countries 

in Europe due to the Renewable Energy Directive Policy, for 2030 and 2050 

year scenarios. 

1.4 Scope 

 Grid and market models are limited to the European Network of Transmission 

System Operators for Electricity (ENTSO-E) member countries. There are 35 

member countries [6]. Cyprus and Iceland are excluded because they are 

isolated from the European integrated power system.  

 In the market model, the energy price is represented as a reflection of the 

external costs of the production of energy. The introduction of renewable 

energy resources is analysed for the power system. Other systems such as 

transport system are excluded from the thesis as well as the influence in the 

fuel market.  

 The starting point (base case scenario) considered is 2015, as this year is the 

latest reliable in terms of data on national renewable energy sources shares by 

country, and considering some assumptions that is further detailed in Chapter 

2.  

 The expansion horizon of renewable sources in the power system generation is 

limited to solar and wind power.  

 The renewable penetration is considered at the production side, thus not at 

distribution level, requiring the equivalent investment in transmission which is 

compared with the base scenario existing power grid.  

Figure 1 shows the 33-buses network where each ENTSO-E member country is 

represented by one node. Considering the influence of load size, generation and 

geographical position, Germany (DE) represents the slack bus, where the voltage 

angle is assumed zero, and it is then the reference country. The network is extracted 

from the current power grid, but it is assumed the same interconnections also for the 

future scenarios. It does not mean though that the transmission lines capacities 

necessarily remain the same for the scenarios, but that no extra connections between 

countries are created nor original ones are removed. 

 



 
11 

 

Figure 1 – ENTSO-E member countries 33 buses network excluding Cyprus and 

Iceland, simplified Grid Model 

The full list of the ENTSO-E members and their respective Transmission System 

Operators (TSO) are found in Appendix A. 

1.5 Main contributions 

The main contributions of this thesis are listed below as compared with other relevant 

studies [7], [12], [15].  

 Increased number of nodes (33 nodes resolution) [7]. 

 More updated data, base case scenario updated for 2015 [7], [12]. 

 New RES divided into wind and solar power [7], [12]. 

 Detailed price data for different generation between countries and future 

forecast [12]. 

 More detailed proposed investment alternatives options per scenario/vision. 

Congestion still is not relieved but fewer critical connections remain [15]. 



 
12 

1.6 Thesis organization 

The thesis report consists of 6 Chapters including the Introduction. They are described 

below: 

 Chapter 2 builds a Literature Review offering a technical background of long-

term transmission planning and electricity market modelling. Relevant 

previous studies on market-based transmission planning are discussed. 

 Chapter 3 describes the model and the methodology behind it, including the 

optimization problem and its constraints. The concept of Locational Marginal 

Price (LMP) is presented. Planning process for transmission investment is 

detailed. 

 Chapter 4 describes how the data is collected from different sources and 

countries, the assumptions and their consequences. Future scenarios are also 

described. 

 Chapter 5 presents the simulations and results from the different scenarios. 

Analysis of the energy dispatch, electricity prices and total system running 

costs are presented. Investment alternatives for transmission capacity are 

suggested based on a Cost-Benefit Analysis (CBA). 

 Chapter 6 draws relevant conclusions from the thesis and suggests directions 

for possible future work. 

 

 

 



 
13 

2 Literature review 

In this chapter, relevant technical background is presented as previous studies done 

in the same field that contributed to the development of the thesis. They relate to: 

European electricity model, market value of RES, methodology for investment in 

transmission capacity reinforcement and future planning for the development of the 

European power grid as a common market. 

2.1 Renewable Resources 

Challenges from electricity power systems emerges from the increasing share of RES 

in the power generation sector [7]. RES technologies are intermittent, especially solar 

and wind power. Political decisions and societal changes create targets limiting the 

environmental impacts and following a society each day less dependent on fossil-fuel 

energy sources. But some countries go even further in their generation capacity plan, 

replacing nuclear power as well. For example, Germany represents the largest load 

aggregation node in this model and plans to phase out nuclear power completely by 

2022, and increase the share of solar and wind power significantly to fill the gap. This 

means all future scenarios presented in this thesis, 2030 and 2050, considers no 

nuclear generation for the country. 

An increase of RES if often correlated with a higher electricity demand. Due to the 

intermittent and random nature of wind and solar power, new technology 

development emerges to handle this variation on the electricity supply. Heat Pumps 

(HP) and Electric Vehicles (EVs) are examples of the shift of the power system 

towards systems using electricity. From 2030 to 2050 the final electricity 

consumption is expected to be 33% higher in Europe [8]. Demand Side Management 

(DSM) is an example of how consumers can handle their consumption behaviour and 

the load curve, to avoid higher costs at peak hours for example. 

The deregulation trend of the power system contributes to a larger risk associated with 

decentralized market-based initiatives to invest in transmission separately from the 

generation expansion [9]. It is more complex to achieve optimal power flow solutions 

for a decentralized system, as many externalities such as generation external costs or 

redundant transmission flows are not detected by the model. It is a consequence that 

some misleading price signals from the Locational Marginal Prices (LMP) favour 

investments in some countries nodes over others [10]. During the transmission 

planning process some inputs such load, Net Generating Capacity (NGC) and 

operational production costs are the most uncertain [11]. 

2.2 Previous Studies 

2.2.1 Electricity System Model 

A master’s thesis in [12] models the European electricity market formed by the 

ENTSO-E member countries with focus on Denmark. The work is divided into two 

parts: first part is the electricity system modelling of Europe (low resolution for 

Europe and high resolution for Denmark), and second part studies a thermal electric 

energy storage system. The first part is particularly relevant to this thesis and thus will 

be discussed in this section. 

A multi-node DC-OPF model is constructed as an optimal dispatch model where 

power is transferred by the cheapest generation available. A LP solver is utilised to 

minimize the total system costs. Market and grid data inputs are conducted from 



 
14 

public data available from ENTSO-E and NordPool platform. The susceptance is 

calculated using the line length using the reactance to calculate the admittance. The 

admittance is the inverted impedance of the power line. The reactance constitutes the 

imaginary part of the impedance and the resistance is the real part (the resistance is 

neglected in the model and in this thesis). Virtual line length is computed by 

calculating the distance between geographical centres of neighbour countries buses, 

using Google Maps API query system. The model uses the LMP to represent the 

marginal nodal electricity pricing, as the demand is incremented by one unit. So the 

LMP is a sensitivity measure to represent the regional electricity price. 

To manage system challenges created by the introduction of RES technologies in the 

power system, some possible solutions to soften the problems caused by intermittency 

and market value are divided into long-term and short-term measures. The study 

proposes a large scale electric energy storage situated close to the transmission grid, 

which helps to shave peaks of uncontrolled generation. For example, this is the case 

storage technologies handles the installed capacity of wind power when high winds 

occurs and bus exceeds the load and the net transfer capacity. 

The electricity system model is first simulated without energy storage integration, 

comparing the LMP calculated and the spot price of electricity from NordPool. Values 

are similar and the deviation occurs mainly in the daily high peeks or deep valleys. 

One of the reasons is that NordPool wind power generation data does not include the 

curtailment effect of wind. Prices from the model are also found to be more stable 

compared to the actual market values because the ENTSO-E model act as an 

aggregated model, where all power plants from the same type of a considered country 

having the same cost characteristics. 

2.2.2 The market value of variable renewable energy sources 

The paper from Potsdam-Institute for Climate Impact Research and Vattenfall, 

Germany [13] discusses the market value of Variable Renewable Energy (VRE) 

sources. The value is affected by three technological properties: 

 Supply of VRE is variable. Related to storage limitations and variability in 

supply and demand, electricity price changes with time. The value of 

electricity is affected by the time it is generated and by weather conditions. 

 Output of VRE is uncertain. Forecast errors and uncertainties of VRE 

generation need to be balanced at short notice, which increase the costs thus 

reducing the market value. 

 Primary resource is bound by geographical limitation. Since there are 

transmission constraints between regions, electricity is a heterogeneous good 

across space. Then, the value of electricity depends on where it is generated. 

Sites far from load centres reduced the value of VRE sources. 

The European Electricity Market Model (EMMA) is a calibrated numerical model 

[13] used in this paper to address the optimal or equilibrium yearly generation, 

transmission and storage capacity and hourly market electricity prices for each market 

region. It covers a considerable geographical area (DE, BE, PL, NL and FR), using 

high quality solar and wind data, including technical constraints of the power system. 

EMMA is developed to estimate value factors considering different penetration rates 

of VRE sources under different policies and prices. The model minimizes total system 

costs with respect of investment and production. 



 
15 

Generation technologies are modelled as eleven different technologies with 

continuous capacity and they produce always when the price is above their variable 

costs. VRE generation is limited by hourly generation profiles. Demand is assumed 

price inelastic, assuming perfect and complete markets. Despite of curtailment of 

VRE being possible there are no costs associated to it, meaning that electricity prices 

cannot become lower than zero. Exchange power flow between areas are limited by 

the Net Transfer Capacities (NTCs) and interconnector investments are profitable if 

the social benefits prevail. The EMMA is modelled as a LP problem thus not a unit 

commitment model. Limitations of the model include absence of hydropower 

reservoir modelling and demand response. Technological change is not considered 

such as variation adaptability. Thus, ignoring flexibility aspects makes an 

overestimation of VRE market values and results can be considered conservative. 

The paper finds that if a great share of VRE capacity is installed, electricity prices are 

reduced by the merit-order effect. Thus, the electricity value (i.e. per MWh) of VREs 

decreases as more capacity is installed. For wind power, its market value is slightly 

higher than the value of a constant electricity source at low penetration, but reaches 

values of 0.5-0.8 at a penetration of 30%. Solar power reaches a similar level with 

15% of market penetration, because in this case the generation is concentrated in 

fewer hours. There are a number of integration options that help mitigating the value 

drop of the VRE sources: transmission investments, relaxed constraints on thermal 

generation, change in wind turbine design. Another conclusion is that VREs need mid 

and peak load generation as complementary balancing technologies, such as advanced 

natural gas power plants and biomass. However, base load technologies such as CCS 

or nuclear power do not go well with high penetration of VRE. High carbon prices are 

not enough to make solar and wind power competitive at high penetration shares and 

subsidies may be required for a period beyond 2020, thus decision makers should 

consider a limited role for wind and solar power regarding greenhouse gases 

mitigation. 

2.2.3 CBA Analysis of transmission network reinforcement 

A Cost-Benefit Analysis (CBA) is used in this paper [9] to evaluate an appropriate 

transmission planning strategy, with the costs being the investment in transmission 

capacity lines and the benefits the Avoided Congestion Costs (ACC) and the Avoided 

Environmental Costs (AEC). The attempt is to identify congestion points and propose 

a network reinforcement via investment of new transmission capacity. This study 

deals with the interaction of two projects, “Pathways to Sustainable European Energy 

System” at Chalmers University of Technology and “Towards future electricity 

networks” from the power systems laboratory at ETH Zurich. Increased generation 

from RES technologies and increase of electricity consumption in the transport sector 

imposes challenges. 

The model consists of a 20-bus system based on a DC-OPF model, with focus on the 

interconnections between the nodes rather than connection nodes within the countries. 

Base case scenario considered is December of 2007. Load is considered inelastic, so 

the merit-order curve for the model is considered showing price as a direct response 

of the demand. The methodology for the Benefit-to-Cost Index (BCI) model starts 

with no network reinforcements. To be profitable, the investment needs a BCI higher 

than 1. Permanent congestions are the final candidates for the grid expansion. For the 

investment plan, the BCI is calculated based on the benefits of ACC and AEC. A 

capital recovery factor is used to divide the present value into yearly costs. 



 
16 

The scenarios in this paper represent different proposed transmission investment 

alternatives. The scenario with the higher BCI proposes the following transmission 

investments connections [9]: 

 DE-AT, 3000 MW reinforcement. 

 SL-HR, 2000 MW reinforcement. 

 DE-CZ, 4000 MW reinforcement. 

BCI stands higher than 5 for both low and high CO2 prices. In this study, ACC have 

an insignificant participation, despite the additional transmission capacity proposed, 

thus congestions are not relieved. But regarding environmental costs, the model shows 

that all proposed investment scenarios are profitable, except one with the low CO2 

price horizon. 

2.2.4 Long-Term transmission planning considering reliability 

In previous work, generation and transmission models have been combined in a Cost-

Benefit Analysis (CBA) considering Avoided Congestion Costs (ACC) and 

environmental costs, to compare transmission lines investment costs alternatives. In 

[14] the authors add another indicator, which is the system reliability factor. The steps 

below define the methodology steps to investigate the transmission reinforcement 

assessment. 

 Identification of congested interconnections based on the probability they 

occur. 

 Calculation of the probability of unserved energy for each node for the cases 

when important lines fail. 

 Calculation of probability of unserved energy one more time, but now 

considering the new transmission capacity reinforcement. 

Optimization of the model is based on minimization of the total system costs and a 

DC-OPF [13] constrained by generation capability limits, transmission capacity limits 

and voltage angle limits.  

For the selection of a transmission line proposed alternative the paper includes the 

avoided unserved energy costs. Unserved supply of electricity makes social costs to 

increase when a transmission line is unavailable for a period of time. 

Future scenarios are based on precipitation level (affecting hydro power availability) 

and wind availability for specific countries. Study shows that some lines are 

permanently congested, regardless of the generation mix or load level. The line 

connections are: AT-IT, SL-IT, CH-AT, HR-SL, BG-RO and MK-GR. 

In the final part, the study shows an example of network reinforcement and the impact 

of unserved load is analysed. The transmission line chosen is CH-IT with an initial 

investment of 3890 MW. Line capacity is gradually increased until the unserved 

energy stabilizes at half of additional transmission capacity beyond the initial 

investment. This balance represents a trade-off between investments in transmission 

costs against costs from unserved energy. 

2.2.5 Ten-Year Network Development Plan (TYNDP) 

The TYNDP [15] is a report which gathers information about grid development and 

elaborated by ENTSO-E as a important tool to achieve Europe energy targets, such as 

security of supply across the continent, sustainable development of the energy system 



 
17 

with RES, integration and affordable energy for European consumers via a common 

market integration. The member countries of ENTSO-E submits their transmission 

and storage projects for the upcoming years, including the definition of the affected 

region and the transmission capacity requirements. TYNDP explores different future 

scenarios, and using a CBA analysis by using different future scenarios. A numerical 

quantification of the benefit assessment from all projects according to the CBA 

methodology uses definitions of RES integration, security of supply and socio-

economic indexes. Besides, a stakeholder participation is possible due to the 

accessible results, even for non-ENTSO-E members TSOs or storage project 

supporters. Once the future scenarios have been defined the next step is to 

characterize the investment needs. An investment in transmission capacity refers to 

every concern on the regional grid which is of European significance.  

Main aspects of generation mix in 2030 include [15]: 

 New NGC is mostly RES, especially wind and solar power. The total capacity 

is expected to double or even triple by 2030. 

 These capacities are concentrated mostly in Germany and in regions with 

favourable wind conditions such as Iberia and Italian peninsulas and by the 

North Sea countries. 

 Hydropower is expected to increase between 20% and 40%, the most expected 

areas being the Alps, Iberian Peninsula and Norway. 

 Nuclear phase-out in Germany (by 2022), Belgium (by 2025) and Switzerland 

(by 2034). All present nuclear units in the UK are schedule to be shut down 

and France plans to reduce their share of nuclear to 50% of the power supply 

by the year of 2025. 30-45 GW of nuclear power is expected to be shut down 

in total. However, 20-30 GW is expected to be brought to the system, mainly 

by UK and Finland.  

 Combined shutdown of nuclear and fossil-fired units along central Europe 

increases the distance between generation and load centres, requiring more 

grid infrastructure to transport the electricity. 

After the results of market and network studies, several bottlenecks areas have been 

identified for the European power system and thus requiring new transmission 

investment. Figure 2 shows their locations and their transfer capabilities to 

accommodate the likely power flows between them. 

To understand the causes and effects of bottlenecks, they are categorized into three 

types [15]: 

 Security of supply: when specific areas cannot be supplied according to 

quality standards. 

 Direct connection of generation: introduction of new generation power plants; 

both conventional and RES. 

 Market integration: different prices within and between price zones. 



 
18 

 

Figure 2 – Map of main bottlenecks in the ENTSO-E perimeter [15] 

A pattern of very large power flow is explained by geography. Large RES areas 

emerge from Ireland to Denmark, along the North Sea shores, Iberian Peninsula and 

south of Italy. Densely populated areas such as England and north of Italy, along 

Mediterranean coastline from Spain to Greece, and in the main cities, import most of 

their electricity from neighbour areas. Hydro power pumping and storage in 

Scandinavia and the Alps works as a regulating power capacity with neighbour 

countries. Integration of RES is the main driver for system evolution in Europe. New 

wind power is planned in the North Sea and Baltic Sea regions (offshore) and inlands 

in the North (onshore). They are usually far away from urban centres and highly 

populated areas. 

European transmission projects must match some criteria, including [15]: 

 Main equipment is at least 220 kV for AC lines or 150 kV otherwise. It has to 

be located in one of the ENTSO-E member countries. 

 Investments must contribute to an increase of transmission line capability 

across the ENTSO-E grid network or at its borders. 



 
19 

 NTC additional expansion must fulfil at least one of the minimum 

requirements below: 

o 500 MW or more of additional NTC is a minimum requirement; or 

o Connection of at least 1 GW / 1000 km
2
 of generation capacity; or 

o Load growth is secured for ten years if the area has a consumption 

higher than 3 TWh per year. 

The project profile in the TYNDP report amounts to approximately €150 billion, of 

which one third is for undersea cables. That is approximately equivalent to 2 €/MWh 

of electricity consumption in Europe over a ten year period and about 1% of the 

electricity bill for the final consumer. 

  



 
20 

 

 

 



 
21 

3 Modelling and Methodology 

This chapter describes the mathematical formulation of the EU electricity market 

model, the objective function and its constraints. The concept of LMP is discussed. 

The planning process for transmission investment is presented, based on a cost-

benefit analysis. 

3.1 Mathematical formulation of the Market Model 

From the EU power system perspective, the load is treated in a configuration which 

each country is represented as a node. Due to the high number of member countries, it 

can be considered a low-resolution option. An aggregated model is created to 

represent the electrical transmission network of Europe. The demand is then available 

at all buses individually. At any node, eight different technologies for the production 

of electricity are assigned: hydro, nuclear, wind, solar, biomass, oil, gas and coal 

power. The Net Generation Capacity (NGC) is aggregated and assigned for each 

country (node). The Net Transfer Capacity (NTC) in the model represents the 

capabilities for all interconnections, also represented in an aggregated way. 

In this type of modelling the DC-Optimal Power Flow (DC-OPF) model is suitable. 

The DC power flow method is a linearization of the full AC power flow method, 

which saves computational resources and time when compared with the AC method. 

DC model uses only active power flows, neglecting voltage levels, reactive power and 

transmission losses [17]. The resistance of each bus is negligible compared to its 

reactance and the magnitude of the voltage at every node is equal to its nominal value 

[18]. The method seeks to optimize an objective function and to satisfy all the 

constraints. In a power system, the power flow is described by the physical laws of 

electricity as known as load flow equations. Convergence of the objective function, 

which is the minimization of the total system costs, is checked at every stage of the 

process. This model is also applicable for long-time horizon expansion planning, 

guiding the investment decisions in the transmission system, proper for the 

transmission expansion planning model [16]. 

3.1.1 Objective Function 

The objective function is the Total system Operational Costs (TOC) as presented in 

(1) which will be minimized subject to constraints described in the following sub-

sections. It is the sum over every power plant, at every node i for each time t, 

associated with the cost of generation of each type of power plant for each node 

(country). 

             ∑ ∑ ∑    (       )   (         )

       

  (1) 

where,  P is the active power from power plant production type at bus i and time t; 

CGen is the cost of generation. The CGen includes fuel and O&M (Operations & 

Maintenance) costs. In other words, it represents the running variable costs of each 

power plant type for a specific country. 



 
22 

3.1.2 Active Power Balance 

At each node i (country), the active power balance must fulfil the power demand at 

the same node for every time t (2). The equation is also known as the demand-supply 

balance formula. 

∑  (         )   ∑ (   )   (   )    (   )

      

 (2) 

Where:  P = active power from power plant type at bus i and time t; 

  B = susceptance of the line; 

    = voltage angle of the line; 

  PD = demand at bus i and time t. 

The set plant consists of eight different types of technologies: hydro, nuclear, wind, 

solar, biomass, oil, gas and coal power. Since demand is considered inelastic, the 

model is straightforward responsive to the load demand (PD). Thus, the objective 

function becomes a minimization of system operating costs problem [9]. 

Active power is transferred between two nodes according to: 

 (     )    (   )    (   )   (   ) (3) 

The transferred active power P is constrained by the NTC of the power line 

connecting two nodes. The voltage angle at the slack bus (DE) is the angular 

reference, equal to 0º. Voltage magnitude is assumed 1 pu at all busses in the system. 

3.1.3 Maximum Production 

The power plants do not produce more than the maximum capacity, at every node i 

and every time t. 

 (         )       (       ) (4) 

      

Pmax is known as the NGC value, acquired from the data extraction for each country 

and each power plant type. P is defined as a positive variable. 

3.1.4 Capacity Factor (CF) 

Power plants often do not operate at full rated power capacity all year round. The 

capacity factor is applied as a technical restriction for the RES technologies. Since 

wind and solar power are not always the cheapest technology available, they are 

dispatched according to their marginal costs following the merit order. 

 (         )     (     )      (       ) (5) 

Then, the CF acts as a technical constraint, for each technology, to give the simulation 

more realistic results. A CF of 0.8 is assumed for all hydro power generation. For 

solar power, refer to Appendix C to see the CF correspondent to each country, using 

the Full Load Hours (FLH) method. 



 
23 

For wind power, a wind profile was used in an hourly resolution [19]. 

     (   )          ( )      ( ) (6) 

At every node i, the wind power output Pwind is limited by the wind profile and 

maximum capacity for every hour t. Wind profile is assumed the same for all 

countries. Figure 3 shows the normalized value of wind power in an hourly resolution 

for 1 year. 

 

Figure 3 – Normalized value of wind power profile curve 

3.1.5 Transmission Capacity Limits 

At the transmission level, the power flow is limited by the maximum transmission 

capacity of the lines.          

 (     )       (   ) (7) 

 (     )         (   )  

where, T(i,j,t) is the power transferred from node i to j at period t. 

3.1.6 Steady-state stability limit 

Finally, for steady-state stability limit, the transmission angle differences between two 

connected nodes should be limited to 
 

 
 rad. 

 (   )    (   )        (8) 

 (   )    (   )          

 

3.2 Locational Marginal Price (LMP) 

Due to the different marginal operating costs of the countries analysed and their 

different power plants, the energy is exchanged between the areas in an economic 

way. In reality there are, however, physical and institutional constraints which will not 

permit an optimum economical dispatch as it would be in a free-market situation. 

Still, multi-area joint dispatch method will be used to model the generation 

0
0,1
0,2
0,3
0,4
0,5
0,6
0,7
0,8
0,9

1

1

2
5

2

5
0

3

7
5

4

1
0

0
5

1
2

5
6

1
5

0
7

1
7

5
8

2
0

0
9

2
2

6
0

2
5

1
1

2
7

6
2

3
0

1
3

3
2

6
4

3
5

1
5

3
7

6
6

4
0

1
7

4
2

6
8

4
5

1
9

4
7

7
0

5
0

2
1

5
2

7
2

5
5

2
3

5
7

7
4

6
0

2
5

6
2

7
6

6
5

2
7

6
7

7
8

7
0

2
9

7
2

8
0

7
5

3
1

7
7

8
2

8
0

3
3

8
2

8
4

8
5

3
5



 
24 

coordination which objective function is to minimize total system cost [20]. The 

constraints to be satisfied in this model consists of demand-supply balance equations 

and transmission capacity limitations. The Locational Marginal Price (LMP) or nodal 

price is the cost of supplying an extra unit of load at the node i under consideration by 

the cheapest option available [18]. The LMP is determined by a combination of the 

generation costs of the marginal generation technologies. 

     ( ) (9) 

Where    is the LMP at node i. 

The LMP is calculated as a parameter from the output results of the model as a 

marginal value of dispatching one extra unit of power for a certain hour, associated 

with the active power balance equation, Eq. (2), where supply must fulfil the demand 

for all periods of time and for every node i. The eight generation source types are 

associated with a cost of generation and they are dispatched in a merit order.  The 

LMP is expressed in €/MWh, representing the cost for each hour if the demand in 

node i is increased by 1 MWh. 

Figure 4 below shows a simplified scheme of transferring energy from node i to j, 

constrained by the line capacity Pij(max). 

 

Figure 4 – Representation of the locational marginal price of two different nodes in a 

constrained economic dispatch 

If there is no congestion, LMPi is equal to LMPj. However, if there is a congestion in 

this line, Pij is equal to Pij(max), and the LMP from node i and j are not necessarily 

equals and can be different. In this particular case where the energy is transferred 

from node i to j, if Pij is equal to Pij(max), LMPi will be lower than LMPj. 

Economically counter-intuitive flows occurs when power flows from a higher price to 

a lower price region, and they are common especially in situations with three or more 

buses connected. This is because the laws of physics (Kirchhoff's voltage law) take 

place over the market laws. Increasing the line capacity does not necessarily will 

reduce the LMP of the node [18]. But the system as a whole will be operated more 

efficiently and the total system costs (variable costs) will be lower. 

3.3 Investment in Transmission – Cost-Benefit Analysis 

(CBA) 

The planning process for the decision for the investment in transmission is based on a 

cost-benefit analysis (CBA) that consists of operational (includes maintenance cost) 

and investment cost [7]. 

3.3.1 Benefits 

The benefits of a proposed transmission line investment is connected directly with the 

amount of costs that could be avoided due, for example, congestion between lines. 



 
25 

               (10) 

where ACCk stands for avoided congestion costs when an investment plan k is realized 

for a transmission line between the buses i and j. The congestion costs are calculated 

as the product of the nodal price difference between nodes i and j by the amount of 

power transferred between the two nodes. The method is used for congestion 

management in pool markets [21]. 

The ACC is calculated considering the Congestion Costs (CC) in Eq. (11). 

     ∑ (         )      

     

       

 (11) 

If the nodal prices from the regions i and j are the same, the CC is equal to zero, 

meaning that there is no potential for saving costs investing in transmission for this 

specific time t. From the model, the CC are calculated as output results for every 

existing capacity interconnection between two nodes for every hour. The parameter is 

given in €/MWh and highest interconnections are ranked as candidates for 

reinforcement. Calculating the benefits from a new investment alternative, the ACC of 

an alternative is the sum of all reinforcement connections proposed. It is then 

calculated the hourly value so it can be compared with the hourly investment costs. 

On the other hand, other costs can be included in Eq. (10), for example avoided 

environmental costs (AEC) related with external costs associated with emission of 

CO2. In this thesis, those costs are directly included in the future scenarios, changing 

the CGen, see Eq. (1), increasing the costs of fossil-fuel generation technologies. 

Before proposed CBA methodology is applied, it is necessary to identify the critical 

paths of the network. Critical paths are highly congested lines and lines transmitting 

large amount of power. The lines with high probability of being overloaded are 

promoted for reinforcement. In this thesis reinforcement means new transmission 

capacity for already existing transmission lines and not additional ones. Figure 5 

shows the schematic for the CBA method used for proposing new investment 

alternatives between the countries. 

 



 
26 

 

Figure 5 – Flow diagram method for investment alternatives using a CBA 

For the cases the model cannot find a feasible solution, some constraints are removed 

in order to identify the bottlenecks of highest congestion and their persistence along a 

time period. Sometimes removing the line power flow limits constraints, Eq. (7), is 

not sufficient for the model to find a feasible solution. Then, other constraints can be 

removed, such as Capacity Factor constraints, Eq. (5), and Power transmission angle 

limits, Eq. (8). Then, the proposed interconnections are made based on how much 

lines connections are congested and the Avoided Transmission Costs (ACC) they 

represent. Permanent congestions are candidates for planning expansion. An adjust is 

made to review if some proposed lines are overestimated and suggesting investments 

above the necessary, to reinforce connections where the congestions persist or even to 

suggest new investments where new congestions appear. Proposing new lines based 

on the ACC of congested lines results in the reduction of the Total System Costs, i.e. 

the variable operational costs to run the entire system. If new proposed investments 

can be done keeping the same amount of Investment Costs (IC) or lower, the new 

system is being operated in a less-costly way. 

The best alternative for the new NTC grid is used to run the model for a longer time 

period in order to find a better representation of the LMP effect and how it is affected 

by comparing the different future scenarios. 

3.3.2 Costs 

It is considered that the investment costs are proportional to the transmission capacity 

increase. Considering an initial Investment Cost IC0, and TC1 and TC2 the 

Transmission Capacity before and after the expansion investment [7], the total 

investment cost is calculated as: 



 
27 

           (  
   

   
)             (12) 

The sum of all connection transmission lines gives the Total Investment Cost (TIC). 

An initial cost of 500 m€ per 1000km of transmission line is assumed, and an average 

distance of 1000km between countries. For the assessment of the investment plans a 

Benefit-to-Cost Index (BCI) is used: 

     
         

(   )     
                   (13) 

Where ICk is the hourly cost in the year k of the transmission line investment (IC), 

taking into account the number of years in the transmission project. If the BCI is 

higher than 1, the investment project is considered profitable. The hourly investment 

cost is calculated using a Capital Recovery Factor (CRF) which takes into account a 

number of years y for the investment project and a discount rate r [12].  

    
(      )

    
 (14) 

The CRF distributes a present value to annuities, according to the equation: 

   (   )   
 (   ) 

(   )   
       (15) 



 
28 

 

  



 
29 

4 Data Collection and Future Scenarios 

4.1 Data collection 

Data is collected from different sources mainly based on the ENTSO-E database and 

the specific Transmission System Operators (TSO) of each country. The transparency 

portal from ENTSO-E offers details regarding NGC per production technology type 

and more recent data. 

As mentioned in Section 3.1, the eight types of power production considered in this 

thesis are: hydro, nuclear, wind, solar, biomass, oil, gas and coal power. Appendix D 

shows the NGC and the Cost of Generation (CGen) for each country and each 

technology. Missing values from ENTSO-E for specific technologies are found 

directly from the utility Transmission System Operator (TSO) companies of the 

countries. 

The load data is collected fully for the months of January and July in an hourly 

resolution. Featured days include highest and lowest load days of the year. 

The Net Transfer Capacity (NTC) values are extracted from ENTSO-E NTC matrix 

[6]. Last updated matrix is from winter 2010-2011. More up-to-date values are 

extracted from the ENTSO-E Transparency Platform [6]. Values of NTC for 

Scandinavia and neighbour countries are found at NordPool website platform [4]. 

Cyprus and Iceland are excluded from the data extraction because they are isolated 

from the European power system. The constructed NTC matrix of the power system is 

shown in Appendix B. Values of susceptance are calculated using reactance and line 

length from previous study mentioned in Section 2.2.1 of this thesis. 

4.1.1 The process platform 

The scheme below in Figure 6 shows what type of data is collected, how it is handled 

and transported to the optimizing tool GAMS. 

 

 

Figure 6  – Flowchart platform of Market Model simulation process 



 
30 

Data collected include: 

 Load: in MW, hourly, from each country. 

 NGC: in MW, for each country, for each generation technology type. 

 NTC: in MW, for each transmission line connection between two countries. 

 CGen: in €/MWh, for each country, for each generation technology type. 

The input values are converted from excel tables to GAMS via GDX file, a function 

from the software to execute the data without necessity to include all data inside the 

model code. The results from the simulation are extracted and interpreted using excel 

sheets and tables. Relevant output include: 

 Energy production: in GWh/h, hourly, from each technology type. 

 Energy dispatch: merit order of technologies being used to produce power, 

based on minimization of total operation system costs. 

 ACC: in €/MWh, hourly, for every power flow transfer connection between 

two countries. 

 Line usage: percentage of use of a power line according to the power flow 

transferred and the maximum NTC allowed. 

 LMP: in €/MWh, hourly, for every country. 

4.1.2 About GAMS 

The General Algebraic Modelling System (GAMS) is the chosen software for the 

modelling work in this project. This simulation tool is useful for large and complex 

problems, suitable for high-level modelling systems for programming and 

optimization. It is possible to change the formulation easily from one to another, 

including the constraints and models used [22]. The base code for the models is 

detailed in APPENDIX E. Large number of data input is possible to be added via 

Excel, as well as exported results, using the GDX tool. 

4.1.3 Assumptions 

Since the European power system is in reality very complex, some assumptions are 

made in order to simplify the modelling and still be able to answer the questions 

addressed to this project. 

4.1.3.1 Market Modelling 

The power system is driven by the laws of physics (power flow equations) and by 

human laws (market rules, policy instruments, public opinion etc.). To model the 

European electricity market, only active power is taken into account and there are no 

losses in any level of the grid, from the generation point to the demand. The DC 

power flow method is applied to simplify the power balance equation. Load, 

generation and transmission capacities are represented by aggregators. The problem is 

treated as a Linear Programming (LP) formulation which minimizes the total system 

costs. Market is perfect in the sense that the generation with lower price is always 

preferable first, the demand is inelastic and the system is dispatched in a centralized 

perspective. When congestion is taken into account, electricity price depends on the 

node in which power is being transferred [18] [23]. Thus, the price is the same for all 

consumers connected to the same bus. 



 
31 

4.1.3.2 Marginal Costs of Production 

For the planning process, data collected such as marginal costs of production are not 

very reliable and thus are approximated. There are many methods to account for those 

costs, and this thesis focus on O&M (Operation and Maintenance) and fuel costs [24]. 

Nuclear power fuel costs include waste costs. In solar power, the type of technology 

assumed is residential solar PVs or large-scale concentrated thermal power plants 

(CSP), depending on the country. Wind power is assumed the sum of onshore and 

offshore technologies. O&M costs for onshore wind power is lower compared to 

offshore, due to higher costs including accessing and developing the maintenance of 

the wind turbines. Taking everything into account, O&M costs for wind power are 

expected to be somewhere between €0.027 and €0.054/kWh [25]. In this thesis, the 

cost of wind range varies from €14 to €36 per MWh for the base case scenario of 

2015. Some vision scenarios for the future assume a more favourable economic 

condition and thus costs for solar and wind power are reduced by half for all 

countries. Still looking at the generation costs data, hydro power technology in 

consideration is large-scale. Finally, regarding oil power costs, the energy content of a 

barrel of oil is calculated as 5867946 Btu per Barrel of crude oil [26]. Comparing the 

energy content with the energy conversion to electricity, 10156 Btu/kWh for steam 

electric generators in 2014. This could lead to an overestimation since most of oil 

power plants are older than 2013. Crude oil barrel price is assumed to be €40 [27]. 

Summing up oil fuel price and its O&M costs, each unit of MWh of electricity 

generated will cost €106. The exception is Estonia (EE), which uses shale oil, at a cost 

of €80/MWHel. 

4.1.3.3 Net Generating Capacity 

Considering the Net Generating Capacity (NGC) data extraction, some simplifications 

are done. Total hydro power capacity is calculated as the sum of Hydro Water 

Reservoir, Hydro Pumped Storage and Hydro Run-of-river and poundage (ENTSO-E 

Transparency Platform). Coal power is the sum of fossil hard coal and fossil brown 

coal/lignite. Oil power is the sum of fossil oil and fossil oil shale. Gas power is the 

sum of fossil gas and fossil coal-derived gas. For the nodes Finland (FI) and Lithuania 

(LT), 1460 and 680 MW is added to the NGC. Those values are in reality imports 

from Russia, which is not part of the system representation. Considering that coal is 

on the margin, the values are added to the coal power NGC of the two countries. 

All energy production technologies are assumed to be flexible and to be able to 

operate in part-load. There is no constraint in terms of ramp-up and ramp-down 

activation limits and costs, no must-run units and no part-load costs. Although this 

would sound unrealistic, in the long-term planning modelling perspective it is a 

reasonable assumption. Time step t is set to 1 hour. 

The new RES (wind and solar power) modelling is simplified enough to get a 

seasonal and yearly reasonable outputs. A normalized wind profile is used to represent 

the normalized share of installed capacity of wind. Solar power output is limited by 

the time of utilization method and the respective capacity factor of each country. 

4.1.3.4 Future Scenarios 

For future scenarios (2030 and 2050) the load is considered affected by European 

economic conditions and the degree of integration on the energy roadmap defined by 

the European parliament. Estimations show that the growth of the electricity demand 

is balanced by the decrease in demand due to higher efficiency both in the production 



 
32 

and demand side. No major changes in demand patterns are expected for human 

electricity activity in the next decades [28], but a load factor between 1 and 1.3 is used 

in the visions, depending on the economic conditions, RES penetration and 

technology development. In addition, demand side management is expected in the 

future to smooth out the demand curve thus reducing peak demand [29]. Fuel prices 

are assumed unchanged, following the idea that efficiency increased cancels out the 

increase of the fuel price [15]. 

4.1.3.5 Investment in Transmission Capacity 

The reinforcement transmission lines have a minimum requirement capacity of 500 

MW [15], and a maximum of 10000 MW, considering the whole transmission 

expansion project. Transmission initial investment costs are assumed as €0.5 billion 

for all connections (1000 km average distance for all connections and 500 k€/km). 

The interest rate r is assumed as 0.07 for all scenarios. Individual countries can invest 

a maximum of half of their total NGC in new transmission lines, and also a maximum 

of 10 times the already existing transmission capacity. Unprecedented transmission 

connections are not created. 

4.1.4 Limitations consequences 

Since this is a long-term transmission planning model, the modelling of wind and 

solar power production gathered from weather and geographical data for each specific 

node is not the focus, but may be a topic for future studies. Thus short-term temporal 

effects may not be observed in this analysis, because hourly information data such 

specific local wind and solar profiles are not used. 

The capacity factor of each power plant for each country in reality varies a lot. 

Making assumptions and simplifications tends to underestimate the total system costs, 

mainly due to the cheap dispatch of hydro power. Limitations of water flow and 

weather conditions will affect the full load hours of each hydro power plant directly. 

Cost data is of different generation technologies are estimated and have low accuracy. 

This data is less accessible to get and varies quite a lot from country to country and 

among the different power plant production types. For example, hydro power is 

cheaper when produced in Switzerland (CH) compared to Italy (IT), and coal power is 

cheaper in Germany (DE) when compared to France (FR) [9]. The increasing 

production of wind and solar power expected for the next years makes the cost of 

production to decrease, as a result of economy of scale. However, this effect is not 

detailed in this thesis. 

Nevertheless, due to data uncertainty and unavailability, internal congested 

connections within countries are neglected. The thesis rather focus on inter-country 

connections. So, the detail resolution of the model is considered low for countries 

with diverse power system areas with different parameters. 

4.2 Definitions of Study Scenarios 

The following cases are considered in the thesis: Base case (2015), 2030 and 2050. 

The future scenarios are defined based on the Ten-Year Network Development Plan 

(TYNDP) report from ENTSO-E [15]. NGC tables for all future scenarios are found 

in the Appendix D. The six visions, four from 2030 and two from 2050, represent 

different policies regarding technologies, energy policies, economy and social 

development. The wide range of possibilities gives the report robustness and 



 
33 

neutrality. Table 1 summarizes all six visions regarding their major differences 

compared with each other. 

4.2.1 Base Case (2015) 

The base case scenario consists of a single picture of the ENTSO-E system based on 

all 33 countries. The list of countries, values of NTC, NGC and generation operational 

costs are found in Appendix A, B, D1-A and D1-B respectfully. 

4.2.2 2030 visions 

There are four visions for 2030 which describes possible future scenarios rather than 

trying to make a forecast of the future. To guarantee a high level of certainty in the 

range described for each vision, extreme situations are avoided. In this thesis, the 

parameters taken into consideration include the NGC of each country, generation 

costs and load. Share of RES in the generation mix and CO2 tax on emissions affect 

those parameters. The tax increases the generation costs for the fossil fuel based 

technologies. Economic conditions affect the cost of generation and technology 

development affect the load. 

 Vision 1, Slow Progress: less favourable economic conditions, low share of 

RES in the NGC, low CO2 tax, and low degree of load demand integration. 

 Vision 2, Constrained Progress: less favourable economic conditions, low 

share of RES in the NGC, low CO2 tax, and high degree of load demand 

integration. 

 Vision 3, Green Transition: favourable economic conditions, high share of 

RES in the NGC, high CO2 tax, and low degree of load demand integration. 

 Vision 4, Green Revolution: favourable economic conditions, high share of 

RES in the NGC, high CO2 tax, and high degree of load demand integration. 

Figure 7 below summarized the 4 vision scenarios for 2030 defined by ENTSO-E: 

 

Figure 7 – Vision Scenarios for 2030 [15] 



 
34 

4.2.3 2050 visions 

For 2050, two scenarios visions are presented, also based on the Ten-Year Network 

Development Plan (TYNDP) report from ENTSO-E [15]. The parameters taken into 

account include: electricity demand, RES share in the NGC, electricity exchanges and 

production on fossil fuels. The main differences in the electricity demand difference 

between the two visions are the change in electricity demand caused by electric 

heating, electric vehicles and energy efficiency measures [8]. Also, for both scenarios 

there is an extra exchange of electricity from North Africa, modelled as CSP. 

 Vision 5, Large-scale RES: higher electricity demand, higher share of RES in 

the NGC, high CO2 tax, and favourable economic conditions. 

 Vision 6, Big & Market: lower electricity demand, lower share of RES in the 

NGC, high CO2 tax, and favourable economic conditions. 

The two visions from 2050 are represented in Figure 8 below. 

 

 

 

Figure 8 – Vision Scenarios for 2050 [15] 

The model simulates different future scenarios and a sensitive analysis is done 

varying economic, political and technical factors described in Table 1 (Section 4.2). 

The LMP varies in respect of those factors and also on the time period interval the 

model is simulated. Factors listed in Table 1 affect directly or indirectly the load level, 

the NTC matrix, the NGC values or the cost of generation. 

 

 

 

 

 

 

 



 
35 

Table 1 – Comparison between all future vision scenarios with different parameters 

 

Year 
Vision 

Economic 

Conditions 

Electricity 

Demand 

Share 

of 

RES 

North 

Africa 

Exchange 

CO2 

Tax 

Load 

Integration 

2030 

Vision 1 
less 

favourable 
lowest level 36% no low low 

Vision 2 
less 

favourable 

higher than 

vision 1 
37% no low high 

Vision 3 favourable 
higher than 

vision 2 
44% no high low 

Vision 4 favourable 
higher than 

vision 3 
47% no high high 

2050 

Vision 5 favourable 
highest 

level 
54% 116 GW high high 

Vision 6 favourable 
lower than 

vision 5 
42% 83 GW high high 

In the next chapter simulations are done to investigate how these vision characteristics 

affect the LMP of the countries and the alternatives for investment in transmission 

capacity. 



 
36 

 

  



 
37 

5 Case Study: Results and Discussions 

In this chapter are presented the simulations and results from the different study 

scenarios. Analysis of the energy dispatch, LMPs and total system running costs are 

presented, discussed and validated with real markets values. Investment alternatives 

for transmission capacity are suggested and discussed based on a Cost-Benefit 

Analysis (CBA) and compared with the TYNDP report from ENTSO-E. 

5.1 Base Case Scenario 

5.1.1 Energy Dispatch 

In this section it is shown how the model simulates the energy production dispatch of 

the 33 ENTSO-E member countries for the months January and July of 2015. This is 

an optimum scenario which minimizes total system costs (objective function), 

respecting the constraint inequations (see Section 3.1). The power plants are 

dispatched according to the merit order of each technology, which means cheaper 

technologies (in terms of operational running costs) are run first. Figure 9 shows the 

energy dispatch of all countries for the base case scenario, in January. 

 

Figure 9 – Dispatch of energy production January 2015 ENTSO-E countries Base 

Case Scenario 

Peak capacity production is around 510 GW at hour 400 and all eight generation types 

are dispatched this month for at least some hours. There is a small representation for 

the oil production (green) above the biomass production (blue) in the Figure 06. It can 

0

100

200

300

400

500

600

1

2
3

4
5

6
7

8
9

1
1

1

1
3

3

1
5

5

1
7

7

1
9

9

2
2

1

2
4

3

2
6

5

2
8

7

3
0

9

3
3

1

3
5

3

3
7

5

3
9

7

4
1

9

4
4

1

4
6

3

4
8

5

5
0

7

5
2

9

5
5

1

5
7

3

5
9

5

6
1

7

6
3

9

6
6

1

6
8

3

7
0

5

7
2

7

E
n

er
g

y
 i

n
 G

W
h

 p
er

 h
o
u

r
 

hours 

Hydro Nuclear Wind Solar Biomass Oil Gas Coal



 
38 

be also noticed that according to this model the electricity from coal power is acting 

as regulating power capacity, which is not true in reality. Start-up costs, part-load 

costs, ramp-up and down limitations and other constraints are not part of the model, 

then the technologies are constantly dispatched on a daily basis scale.  

The simulation for the July month is represented by Figure 10.  

 

Figure 10 – Dispatch of energy production July 2015 Base Case Scenario 

Since the load demand is overall lower in July, the oil power is not dispatched. Peak 

capacity is also lower at around 405 GW at the hour 4739. Considering that fossil fuel 

technologies have higher running costs, in July the Locational Marginal Price will be 

lower in average. Base load is predominantly dominated by hydro and nuclear power, 

which can be an overestimation since in reality the capacity factor of those 

technologies are lower than in the model.   

Figure 11 shows how the energy dispatch is for Sweden (SE), base case scenario and 

January of 2015. The peak capacity is 28 GW for just 2 hours, where gas power is 

dispatched (the most expensive technology of SE). Coal and Oil power are not part of 

the Swedish generation mix so only the other 6 technologies are dispatched in the 

month with the highest load. It can be seen that the model is limited in the sense of 

simplifying how some technologies are dispatched. For example, in Figure 9 nuclear 

power (orange) is deviating quite a lot every day, which does not happen in reality. 

This is because ramp-up and down costs are not included in the model as well as part 

load costs operation. Incorporating those elements in the model would require a lot of 

processing and Linear Programing would not be enough to formulate the extra 

constraints. 

0

50

100

150

200

250

300

350

400

450

4
3

4
5

4
3

6
7

4
3

8
9

4
4

1
1

4
4

3
3

4
4

5
5

4
4

7
7

4
4

9
9

4
5

2
1

4
5

4
3

4
5

6
5

4
5

8
7

4
6

0
9

4
6

3
1

4
6

5
3

4
6

7
5

4
6

9
7

4
7

1
9

4
7

4
1

4
7

6
3

4
7

8
5

4
8

0
7

4
8

2
9

4
8

5
1

4
8

7
3

4
8

9
5

4
9

1
7

4
9

3
9

4
9

6
1

4
9

8
3

5
0

0
5

5
0

2
7

5
0

4
9

5
0

7
1

E
n

er
g

y
 i

n
 G

W
h

 p
er

 h
o
u

r 

hours 

Hydro Nuclear Wind Solar Biomass Gas Coal



 
39 

 

Figure 11 – Swedish Energy Dispatch Base Case Scenario, January 2015 

5.1.2 Locational Marginal Price (LMP) – Base Case (2015) 

The monthly average Locational Marginal Price (LMP) is plotted for each ENTSO-E 

country, for the months January and July, as it is shown in Figure 12. The LMP is 

dependent on some parameters such the energy mix of each country, the load level 

and available transmission with neighbour nodes. Countries high dependent on fossil 

fuel technologies (coal, gas and oil power) usually have higher generation production 

costs, thus having a higher LMP, and countries which can fulfil their demand with 

more RES have more chance of having a lower average LMP. Countries with only 

one transmission line connected have a LMP equal or very close from their neighbour 

countries. This is the case of Portugal (PT), Luxemburg (LU) and Lithuania (LT), 

which their only neighbour are Spain (ES), Germany (DE) and Latvia (LV). 

The three highest LMP values for January are: GB (€68/MWh), NL (€63/MWh) and 

BE (€62/MWh). The lowest are: NO (€16/MWh), CH (€20/MWh) and SE 

(€26/MWh). Same simulations are done for the month July of 2015. The maximum 

LMP value is at €53/MWh, during peak hour. It can be noticed that the average value 

is lower for July. The maximum LMP value occurs at €60/MWh. The three highest 

LMP values in this case are: IT (€55/MWh), GB (€53/MWh) and NL (€50/MWh). 

The lowest LMP values are: NO (€15/MWh), SE (€18/MWh) and CH (€20/MWh). 

Figure 12 confirms that there is a considerable decrease in the LMP value in the 

summer month for most countries, except IT, Iberian countries, DE and LU. For CH, 

GR, NI and PL the LMP remains the same. IT becomes the most expensive region for 

July, and the Scandinavian countries NO and SE top the cheapest regions. The price is 

directly related with the electricity demand, which is higher in the winter for northern 

countries (heating demand) and higher in the summer for southern countries (cooling 

demand). Nevertheless, the trend is similar when compared with the winter time, and 

most of the countries present a decrease in the LMP. 

0

5

10

15

20

25

30

35

1

2
4

4
7

7
0

9
3

1
1

6

1
3

9

1
6

2

1
8

5

2
0

8

2
3

1

2
5

4

2
7

7

3
0

0

3
2

3

3
4

6

3
6

9

3
9

2

4
1

5

4
3

8

4
6

1

4
8

4

5
0

7

5
3

0

5
5

3

5
7

6

5
9

9

6
2

2

6
4

5

6
6

8

6
9

1

7
1

4

7
3

7

E
n

er
g

y
 D

is
p

a
tc

h
 i

n
 G

W
h

/h
 

hours 

SE SE SE SE SE SE



 
40 

 

 

Figure 12 – Average LMP per country for January and July of 2015, Base Case 

Scenario 

For a daily variation analysis, two cases are plotted: the day with highest load of the 

year (21/01/2015) and the one with the lowest load (05/07/2015). There is a 

considerable difference in the LMP daily variation (difference between maximum and 

minimum load) for January, 34%, from €35/MWh to €53/MWh. For June the daily 

variation is lower, 19%, from €28/MWh to €35/MWh. Figure 13 shows the LMP 

daily variation curve for January and July. When the system average LMP is 

considered, for all hours, the peak load day price is higher than the valley load day. 

The curve is highly correlated to the daily load demand. 

 

Figure 13 – Average ENTSO-E simulated system price LMP daily curve 21/01/2015, 

Base Case Scenario 

0

10

20

30

40

50

60

70

80

A
T

B
A B
E

B
G

C
H C
Z

D
E

D
K EE ES FI FR G
B

G
R

H
R

H
U IE IT LT LU LV M
E

M
K N
I

N
L

N
O P
L

P
T

R
O R
S SE SI SK

L
M

P
 i

n
 €

/M
W

h
 

ENTSO-E Countries 

January July

0

10

20

30

40

50

60

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

L
M

P
 E

N
T

S
O

-E
 i

n
 €

/M
W

h
 

hours 

21/01/2015 05/07/2015



 
41 

Comparing it with real market values from the NordPool Elspot [4] and the European 

Energy Exchange [5] prices, the trend is similar, as showed in Figure 14. The 

NordPool system consists of the following countries: Norway (NO), Sweden (SE), 

Denmark (DK), Finland (FI), Estonia (EE), Latvia (LV) and Lithuania (LT). The EEX 

system is an international partnership, and the power spot market includes the 

countries: Germany (DE), Austria (AT), France (FR) and Switzerland (CH). 

 

Figure 14 – Average NORDPOOL and EEX systems Elspot price daily curve 

21/01/2015 [4],[5], compared with simulated prices 

Important to notice that Elspot area price values for the NordPool and EEX systems 

represent market clearing price, based on the bids of producers and retailers or 

consumers. At the same time, the values extracted from the simulation of the base 

case scenario represent the operational costs in an optimized system which reduced 

total system running costs. Since the power market in EU is not completely central-

dispatched, a good reference to see trends based on real market trade is the NordPool 

power market system. The spot market defines the short-term optimal trade between 

the market actors. The NordPool system Elspot average price for January 2015 is 

€34/MWh, and €11/MWh for July 2015 [4]. Monthly averages in the simulations tend 

to be underestimated due to the European centralized and optimum dispatch degree 

which does not occur in reality. 

5.2 Investment in Transmission Alternatives 

In Sections 5.2.1 and 5.2.2, the strategy for investing in new alternatives is discussed. 

Three first alternatives use a frequency of congestion method, and alternatives four, 

five and six incorporates ACC analysis in the process to rank the candidates for 

enforcement in transmission capacity. All transmission interconnection alternatives 

for 2030 use the Vision 1 scenario (see Section 4.2.1) for simplicity. In Section 5.2.3, 

both Visions 5 and 6 are tested, and the methodology for investing in transmission 

alternatives incorporate ACC analysis. 

5.2.1 2030 Scenarios proposed alternatives 

The first step is to remove the power flow transmission capacity constraint, eq. (7), to 

run the model with the minimum requirements. Then, some alternatives 

interconnection reinforcements are proposed based on the most congested lines and 

0

10

20

30

40

50

60

70

1 3 5 7 9 11 13 15 17 19 21 23

A
v

er
a

g
e 

L
M

P
 i

n
 €

/M
W

h
 

hours 21/01/2015 

Actual EEX Simulated EEX

0

10

20

30

40

50

60

1 3 5 7 9 11 13 15 17 19 21 23

A
v

er
a

g
e 

L
M

P
 i

n
 €

/M
W

h
 

hours 21/01/2015 

Actual Nordpool Simulated Nordpool



 
42 

how often they are congested. For those lines, ACC analysis shows the connections 

where costs could be avoided if there was a higher capacity for the transmission line. 

For the 2030 scenarios, ACC analysis is done separately, after first three alternatives 

are proposed. Then the benefit of the investment is the ACC.  The model runs in the 

time horizon of 24 hours for highest peak day of the year (21/01/2015). 

Table 2 – Alternative 1 Based on unconstrained case 

Alternative 1 - 2030 Unconstrained case 

Interconnection TC1 TC AC n TC2 IC 

AT-IT 295 1475 1500 2 1795 2.8057 

CH-FR 1100 1100 1000 1 2100 1.6466 

CZ-PL 600 1800 2000 2 2600 2.4663 

DE-CZ 1000 1000 1000 1 2000 1.6931 

GR-IT 500 1000 1000 1 1500 2.0986 

NO-NL 723 1446 1500 2 2223 2.1232 

RO-RS 700 700 500 1 1200 1.5389 

SE-DE 615 1230 1000 1 1615 1.9654 

SI-IT 620 620 500 1 1120 1.5913 

SK-PL 500 2000 2000 2 2500 2.6094 

Where: 

TC1: Transmission Capacity before expansion, in MW; 

TC2: Transmission Capacity after expansion, in MW; 

TC: reinforcement capacity proposed, in MW; 

AC: actual reinforcement, in MW; 

n: number of circuits to expand; 

IC: normalized Investment Cost, see equation (12). 

The first alternative (k=1) presents a high number of investment lines, a second one an 

intermediate option and a third alternative with fewer proposed investment line 

circuits. After first alternative is presented, adjust is based on the previous state 

alternative, and then a new one is suggested. Running the model with the NTC 

proposed in Alternative 1, the value of the Total Operational Costs (TOC) is €0.32145 

billion. Using Eq. (13), (14) and (15) the BCI of 0.82 is calculated, for r=0.05 and 

y=30, meaning that the alternative is not profitable (lower than 1). 

Alternative 2 is proposed based on a constrained case using the NTC proposed in the 

previous alternative. Now some congestions are eliminated, new congestion are 



 
43 

created and some persist. Adjust is done in order to find a better cost-benefit 

alternative, selecting new candidates or resizing old ones. 

Table 3 – Alternative 2 Based on constrained case considering persisting and new 

congestions from Alternative 1  

Alternative 2 - 2030 
Constrained case considering persisting/new 

congestions 

Interconnection TC1 TC AC n TC2 IC 

AT-IT 295 2950 3000 3 3295 3.4131 

AT-CZ 800 800 1000 1 1800 1.8109 

CH-AT 1200 1200 1000 1 2200 1.6061 

FR-BE 3400 3400 3000 3 6400 1.6325 

GB-IE 530 530 500 1 1030 1.6644 

NO-NL 723 4338 4000 4 4723 2.8767 

SE-PL 600 600 500 1 1100 1.6061 

SK-PL 500 500 500 1 1000 1.6931 

Running the model with the NTC proposed in Alternative 2, the value of the Total 

Operational Costs (TOC) is €0.32100 billion and a BCI of 1.49, making it a profitable 

alternative. Alternative 3 is proposed following the same idea and it is based on 

Alternative 2. Again some congestions are eliminated, new congestion are created and 

some persist. Adjust is done in order to find a better cost-benefit alternative. 

Table 4 – Alternative 3 Based on constrained case considering persisting and new 

congestions from Alternative 2 

Alternative 3 - 2030 
Constrained case considering persisting/new 

congestions 

Interconnection TC1 TC AC n TC2 IC 

AT-IT 295 3540 3500 4 3795 3.5544 

CH-AT 1200 1200 1000 1 2200 1.6061 

CH-DE 4000 4000 4000 4 8000 1.6931 

FR-BE 3400 3400 3000 3 6400 1.6325 

NO-NL 723 5784 6000 6 6723 3.2298 

SK-PL 500 500 500 1 1000 1.6931 



 
44 

The TOC is €0.32161 billion and the BCI lowers to 1.73, indicating that the 

alternative is profitable. A new strategy is used in the next section, and consists of 

using the ACC analysis to rank the highest congestion costs associated with the 

existing transfer interconnections. 

5.2.2 Avoided Congestion Costs (ACC) Analysis 

Coming back to the unconstrained 2030 case scenario, now the highest ten sum of 

ACC of all connections are listed. The proposed interconnections are based on the 

value of the ACC for the connections where the line usage is being constrained by the 

power flow maximum limit.  

The values of the transmission capacities are listed in the Table 5 below. 

Table 5 – Alternative 4 Based on unconstrained case, ranked by highest ACC values 

Alternative 4 - 2030 Unconstrained case 

Interconnection TC1 TC AC n TC2 IC 

DE-NL 3850 3850 4000 4 7850 1.7124 

FR-BE 3400 3400 3000 3 6400 1.6325 

NO-NL 723 1446 1500 2 2223 2.1232 

CH-IT 4165 4165 4000 4 8165 1.6731 

FR-GB 2000 2000 1000 2 3000 1.4054 

SI-IT 620 620 500 1 1120 1.5913 

AT-IT 295 1475 1500 2 1795 2.8057 

CH-FR 1100 1100 1000 1 2100 1.6466 

CH-DE 4000 4000 4000 4 8000 1.6931 

GR-IT 500 500 500 1 1000 1.6931 

For alternative 4, the TOC calculated by GAMS is €0.32045 billion and the 

economical parameter BCI is 1.53 (profitable). Keeping the NTC proposed by 

alternative 4, new congestion network is formed and adjust are made until alternative 

5 is proposed. 

 

 

 

 

 

 



 
45 

Table 6 – Alternative 5 Based on constrained case considering persisting and new 

congestions from Alternative 4, ranked by highest ACC values 

Alternative 5 - 2030 
Constrained case considering persisting/new 

congestions 

Interconnection TC1 TC AC n TC2 IC 

CH-AT 1200 1200 1000 1 2200 1.6061 

FR-BE 3400 3400 3000 3 6400 1.6325 

DE-SE 615 615 500 1 1115 1.5949 

NO-NL 723 4338 4000 4 4723 2.8767 

SE-DE 615 1230 1000 1 1615 1.9654 

SK-PL 500 500 500 1 1000 1.6931 

SE-PL 600 600 500 1 1100 1.6061 

RO-BG 200 200 500 1 700 2.2527 

TOC are reduced to €0.32095 billion and BCI in this case is 1.46. 

 

Table 7 – Alternative 6 Based on constrained case considering persisting and new 

congestions from Alternative 5, ranked by highest ACC values 

Alternative 6 Constrained case considering persisting/new congestions 

Interconnection TC1 TC AC n TC2 IC 

CH-AT 1200 1200 1000 1 2200 1.6061 

FR-BE 3400 3400 3000 3 6400 1.6325 

DE-DK 2100 1050 1000 1 3100 1.3894 

AT-IT 295 3540 3500 4 3795 3.5544 

NL-NO 723 5784 6000 6 6723 3.2298 

SK-PL 500 1000 1000 1 1500 2.0986 

In the attempt to reduce the number of investment interconnections, alternative 6 end 

up increasing the TOC to €0.32095 and increasing the BCI to 1.61. Then, for the 

2030’s future scenarios sensitivity analysis, alternative 3 is assumed as the NTC 

matrix, adding connection DE-DK from alternative 6 for feasibility purposes since 

both countries present high penetration of RES in all scenarios, so the supply must 



 
46 

meet the demand for all periods of time. Combine these alternatives to create a 

proposed investment for 2030, as detailed in Figure 15 and Table 8. 

 

Figure 15 – Transmission reinforcement connections proposed for 2030 

The green lines are the capacity reinforcement proposed for 2030. They are an 

increase of capacity of already existing interconnections (in this thesis no new extra 

connection is created). Some investment can be compared and validated with real 

future projects found at the TYNDP report (12) from 2014. Overall all the suggestions 

are plausible and realistic, and most of them are already addressed in the TYNDP 

report.  

For the connection CH-AT and CH-DE, the project is called Swiss Roof, project 90, 

and its plan connects CH with both DE and AT. The expected date of commissioning 

is between 2017 and 2022 and the contribution is 1200 MW for the connection 

DE/AT (border area) with CH (Rüthi substation). FR-BE connection (project 23) 

presents a future project between Avelin/Mastaing (FR) substation and Horta (BE) 

substation. The expected date of commissioning is 2021 and since the project is in the 

planning phase there is no proposed NTC value presented in the report. Project 179 is 

a project between DE-DK and it is still under consideration phase, with an expected 

date of commissioning for 2030. No NTC value reinforcement is presented yet. For 

AT-IT, project 26 describe several connections between both countries. The project 

focus on supporting the power balancing between RES in Italy and pump storage 



 
47 

hydropower plants in the Austrian Alps. The total transmission capacity proposed is 

1770 MW, with the expected date of commissioning between 2018 and 2023. Projects 

varies from consideration phase, to planning and permitting. North Sea offshore grid 

infrastructure scheme (project 230) describes a complex and large project including 

several countries forming a ring: NO, GB, NI, IE, DK, DE, NL, BE, LU and FR. NO 

connects directly with GB and DE in this scheme. The path to DE is in permitting 

phase with a NTC of 1400 MW and the further connection to NL is still in planning 

phase. Due to this complexity of connections, not a simply direct connection between 

NO-NL, the value proposed for this connection is much higher than the one presented 

in the TYNDP. Also, in the thesis the lines are considered to be the same costs for all 

connections, while in reality deep subsea cables are more expensive. There is no 

planned interconnection between SK and PL directly according to the TYNDP report. 

The market model is run for different future scenarios firstly operated without line 

transmission constraints. Then, using a cost-benefit analysis transmission capacity 

reinforcements are proposed, compared and validated with the TYNDP report from 

ENTSO-E. The Table 8 below comparing the proposed alternative with the actual 

planning projects of the TYNDP. 

Table 8 – Comparison between proposed interconnections for 2030 and TYNDP final 

report from 2014 

Interconnections - 2030 Proposed (MW) TYNDP (MW) 

AT-IT 3540 1170 

CH-AT 1200 1200 

CH-DE 4000 1200 

FR-BE 3400 Planning phase 

DE-DK 1050 Planning phase 

NO-NL 5784 1400 

SK-PL 500/1000* No planned project 

*difference between alternatives 3 and 6 

Total investment costs for the proposed alternative is €4.89 billion per year 

(annualized cost using CRF). 

5.2.3 2050 scenarios proposed alternatives 

For 2050 scenarios, two alternatives for Visions 5 and 6 are proposed (one for each 

vision) using the same strategy as used for 2030. The base case scenario now is the 

base NTC from 2015 adding the proposed alternative from 2030. It is called partially 

constrained case because the power flow constraints are removed but the new 

proposed connections from 2030 remain. Feasibility is still not reached since now 

there is a significant share of RES in the NGC mix in both visions, which implicates 

in larger reinforcements, both in terms of capacity and number of interconnections. 

Transmission flow constraints are removed so critical lines can be identified. Adjust is 

done and alternatives 1 and 2 for 2050 are proposed in the Tables 9 and 10 below, 



 
48 

including the interconnections suggestions and requirements. Then, the total running 

costs and the economic index BCI is calculated for both cases. 

 

Table 9 – Alternative 1 for Vision 5 Based on partially constrained values, ranked by 

highest ACC values 

Alternative 1 - 2050 Based on partially constrained values from 2030 

Interconnection TC1 TC AC n TC2 IC 

AT-IT 295 1475 1500 3 1795 2.8057 

BG-RS 450 2250 2000 4 2450 2.6945 

CH-DE 1500 1500 1500 3 3000 1.6931 

CZ-DE 3650 3650 5000 5 8650 1.8628 

ES-FR 1300 5200 5000 5 6300 2.5781 

FR-BE 2300 2300 2500 3 4800 1.7357 

FR-GB 2000 4000 4000 4 6000 2.0986 

FR-IT 995 995 1000 1 1995 1.6956 

GR-IT 500 2500 2500 3 3000 2.7917 

HU-AT 500 2500 2500 3 3000 2.7917 

HU-SK 600 3000 3000 3 3600 2.7917 

NO-DK 1632 4896 5000 5 6632 2.4021 

NO-NL 723 3615 3500 4 4223 2.7648 

SE-DE 615 3075 3000 3 3615 2.7712 

SI-IT 620 1860 2000 2 2620 2.4412 

For Vision 5, TOC of €0.35164 billion is calculated by GAMS after implementing the 

suggested interconnections, and the BCI is 3.39. 

Finally, the interconnections for the Vision 6 is presented in the table below, with a 

correspondent BCI of 2.33. For both scenarios the interest rate r used is 5% and the 

life-time of the transmission y connection is 30 years, giving a CRF value of 0.0650 

(see Eq. 15). 

 

 



 
49 

Table 10 – Alternative 2 for Vision 6 Based on partially constrained values, ranked by 

highest ACC values 

Alternative 2 - 2050 Based on partially constrained values from 2030 

Interconnection TC1 TC AC n TC2 IC 

AT-IT 295 2360 2500 3 2795 3.2486 

BG-MK 400 400 500 1 900 1.8109 

BG-RS 450 900 1000 1 1450 2.1700 

CH-DE 4000 2000 2000 4 6000 1.4054 

CZ-DE 3650 1825 2000 4 5650 1.4369 

DK-DE 2365 2365 2000 2 4365 1.6128 

EE-FI 1016 3048 3000 3 4016 2.3744 

ES-FR 1300 2600 2500 3 3800 2.0726 

GR-IT 500 2500 2500 3 3000 2.7917 

HU-SK 600 1800 1500 2 2100 2.2527 

LV-EE 879 2637 2500 3 3379 2.3465 

NO-DK 1632 1632 1500 2 3132 1.6518 

NO-NL 723 3615 3500 4 4223 2.7648 

RO-RS 550 550 500 1 1050 1.6466 

SE-DE 615 2460 2500 3 3115 2.6223 

SE-PL 600 2400 2500 3 3100 2.6422 

SI-IT 620 2480 2500 3 3120 2.6158 

 

Total investment costs for alternatives 1 and 2 for 2050 are €11.68 and €11.34 billion 

per year. 

Table 11 and Figure 16 summarizes all interconnections suggested for 2030 and 2050 

and their respective BCI value. 

 

 

 



 
50 

Table 11 – Summary of variable and investment costs, benefits and BCI index for the 

transmission interconnections alternatives for 2030 and 2050 

Year Alternative 

TOC (in 

billions euros, 

1 day) 

IC (in billions 

euros, per 

year) 

Benefits (in 

millions 

euros, 1 day) 

BCI 

2030 

1 0.32145 6.684 0.1347601 0.8177 

2 0.32100 5.300 0.1954410 1.4940 

3 0.32161 4.362 0.1860698 1.7294 

4 0.32162 5.847 0.4414692 1.5303 

5 0.32045 4.953 0.1782042 1.4584 

6 0.32095 4.394 0.1745896 1.6104 

2050 
1 0.35164 11.68 0.9781289 3.3938 

2 0.32061 11.34 0.6525149 2.3334 

 

Figure 16 – Difference in the BCI for different transmission interconnections 

alternatives for 2030 and 2050 

As it is showed in Table 11, the total operational system costs (TOC) calculated by 

GAMS are similar for all alternatives, except alternative 1 from 2050 presenting a 

value of 0.35164 billons euros. This is mainly related to the fact that electricity 

demand for this scenario (Vision 5) is the highest among all visions, and the CO2 tax 

is at the highest value. Since generation is directly connected with the demand for 

electricity, the cost of operation increased as more power plants need to be dispatched. 

Investment Cost (IC) is annualized using the CRF expression, Eq. (14). In this case 

the highest values are the alternatives from 2050, since more transmission lines are 

proposed, both because of the increase of the electricity demand and the increase of 

0

1

2

3

4

Alternative
1 (2030)

Alternative
2 (2030)

Alternative
3 (2030)

Alternative
4 (2030)

Alternative
5 (2030)

Alternative
6 (2030)

Alternative
1 (2050)

Alternative
2 (2050)

B
C

I 

Proposed Interconnections Alternatives 



 
51 

RES in the generation mix. More intermittent energy sources require more 

transmission reinforcement, both for importing and exporting. The benefits calculated 

for each alternative is the sum of the avoided congestion costs (ACC) of all 

connections proposed, for the peak day simulated. The value of ACC is directly 

related to the degree of congestion and the nodal price difference between the two 

countries. The highest benefit value comes from alternative 1 from 2050, representing 

that 0.978 millions of euros could be saved on the peak day, if the proposed line 

reinforcements were applied. A high benefit affects positively the BCI. 

According to the graph in Figure 16, BCI for Alternative 1 is the only not to reach 1, 

thus not being profitable. 2050’s alternatives shows a higher BCI of approximately, 

making them very profitable. This is due the high number of suggested 

interconnections for 2050’s alternatives (16 and 17 interconnections, respectively) and 

their high NTC individual value. 

5.2.4 LMP effect after investment in transmission 

Now that suggested NTC values are calculated, the LMP effect can be studied for the 

different vision scenarios, for 2030 and 2050. Starting with Vision 1, Figure 17 shows 

that regional prices are somewhat similar when compared with base case scenario 

from 2015, with peak values from BE, GB and NL reaching €80/MWh. Vision 2 

presents similar peak values for LMP but many countries presents a lower price when 

compared to Vision 1. This is due technological development effect on the load 

integration, which in the model is translated as a decrease of the load electricity 

demand. 

For Visions 3 and 4, the comparison is also made between them as showed in Figure 

18. Vision 3 shows a significant increase of electricity prices on fossil-fuel based 

countries due the high tax on CO2 implemented. FR reaches €106/MWh and