Optimization and techno-economic study of a PV Battery system for a vacation home in Sweden

Master Level Thesis

Master thesis 30 credits, 2020 Solar Energy Engineering Author:

Joaquin Coll Matas Supervisor:

Frank Fiedler Examiner:

Ewa Wäckelgård Course Code: EG3011 Examination date: 11/09/2020

K

Dalarna University Solar Energy

Engineering

European Solar Engineering School No. 268, September 2020

Optimization and

Techno-Economic Study of a

PV Battery System for a

Vacation Home in Sweden

Abstract

Currently, Sälen area in Sweden is finding issues in the power grid due to an irregular load profile with high peak power demand and an infrastructure that is becoming undersized.

Distributed PV-battery systems are considered a possible solution to solve this problem.

A PV-battery system for a typical vacation home in this area is designed and optimized to give the best economical solution to the homeowner. Then, a techno-economic evaluation of the system is performed. A photovoltaic system and an only grid connected system are also simulated and compared. Finally, a sensitivity analysis is performed on different simulation inputs.

HOMER Grid software is used to simulate and size the system. Firstly, a pre-sized system is modelled using average or typical market prices and component characteristics.

Afterwards, real market components that fit into the pre-sized model are modelled to get a real system design. The optimized design includes a PV system of 13 kW, a BYD lithium ion battery of 5.1 kWh capacity and a Sungrow hybrid inverter of 10 kW.

The economic evaluation of the system indicates that, with current market prices and subsidies, the optimized system is the most economical solution for the homeowner compared to the other systems. In the sensitivity analysis, a significant risk for the

profitability of the system is found on the compensation from selling electricity to the grid.

The technical evaluation of the system indicates that the battery provides a significant

peak-shaving effect that can benefit the power grid. However, large solar energy sales to

the grid with high power peaks that could cause instability issues are observed.

Contents

1 Introduction ... 1

Aims ... 1

Method ... 2

2 Technology background and literature review ... 5

Photovoltaic technologies ... 5

Battery technologies ... 5

Battery storage system applications ... 6

System design ... 7

Sizing and optimization ... 8

3 Method ... 10

Boundary conditions ... 10

3.1.1. General inputs and assumptions ... 10

3.1.2. Solar resource ... 11

3.1.3. Load profile ... 13

3.1.4. Capital subsidies ... 15

3.1.5. Fixed capital costs ... 16

3.1.6. PV Modules ... 16

3.1.7. Inverters ... 17

3.1.8. Batteries ... 19

3.1.9. Energy tariff ... 21

3.1.10. Demand tariff ... 22

3.1.11. Feed-in compensation ... 23

Pre-sizing ... 24

Multi-year simulation ... 25

Result analysis ... 26

3.4.1. Economic results ... 26

3.4.1.1.Cumulative discounted cashflow ... 26

3.4.2. Technical results ... 27

Sensitivity analysis ... 28

4 Results ... 29

Pre-sizing ... 29

Multi-year simulation ... 29

4.2.1. Economic results ... 30

4.2.1.1.Cumulative discounted cashflow ... 32

4.2.1.2.Annualized costs by type ... 33

4.2.2. Technical results ... 34

4.2.2.1.Energy fluxes ... 35

4.2.2.2. Peak power ... 38

Sensitivity analysis ... 39

4.3.1. Energy tariffs ... 39

4.3.2. Load profiles ... 40

4.3.3. Feed-in compensations ... 40

5 Discussion ... 42

Assumptions ... 43

Previous studies ... 44

6 Conclusions ... 46

Future work ... 46

Appendix A Monthly energy bill ... 51

Appendix B Malungs Elnät demand tariff ... 55

Appendix C Pre-sizing results ... 56

Appendix D Component datasheets ... 58

Appendix E Cumulative discounted cashflow scenarios ... 63

Abbreviations

Abbreviation Description

BESS Battery energy storage system O&M Operating and maintenance Li-ion Lithium ion

PV Photovoltaic

PV-BESS Photovoltaic battery energy storage system

SEK Swedish krona

Nomenclature

Symbol Description Unit

DoD Depth of discharge %

DoD

Maximum depth of discharge used during operation % DoD

Maximum depth of discharge using during battery testing %

E Energy throughput of a battery kWh

E

Energy throughput of a battery using test conditions kWh

NPC Net present cost SEK

1 Introduction

Fighting global warming by reducing greenhouse gas emissions is one of the main

challenges of our era. Renewables will play a crucial role and, amongst them, solar energy is having very important development. According to the International Energy Agency [1], between year 2019 and 2024 renewable capacity is expected to grow around 50 % with photovoltaics (PV) leading with approximately 60 % of these additions driven by important cost reductions and policy efforts. However, renewable sources have an

stochastic nature and it is challenging to integrate them into the power grid where demand and supply must be balanced at all time.

Energy storage is one solution to shift the renewable production to match the electricity demand. Battery Energy Storage Systems (BESS) are becoming one of the most promising technologies thanks to its fast response time, ease of control and location flexibility and low Operating and Maintenance (O&M) costs [2]. Nowadays, their main disadvantage is its high cost but it has dropped significantly during last years, and this trend is expected to continue in the coming years with cost reductions [3]. Distributed solar PV systems are growing exponentially, with a total of 41 GW additions in 2018, supported by strong policy support mainly in Europe, United States and Japan [4]. They can combine PV systems and BESS (PV-BESS) improving significantly their benefits becoming a promising technology [5].

Currently, the power grid in Sälen area, in Malung-Sälen municipality in Dalarna, Sweden, is finding some issues. It is a remote vacation area, with no great grid infrastructure, where many electricity customers have their demand at the same time (e.g. saunas after skiing, breakfast time, etc.) and an irregular load profile is generated with high peaks. Also, electricity demand in this region is increasing and the power grid is becoming undersized.

This has a negative impact on both customers and grid operators [6]. Electricity customers have high energy tariff (cost per kWh of energy use) and demand tariffs (cost per kW peak power demand). On the other side, grid operators have difficulties managing grid stability and need to increase the grid power capacity to supply these peak power loads[7].

A possible solution that grid operators are studying is to improve this situation by installing distributed PV-BESS in several homes in Dalarna. Multiple benefits could be achieved by both customers and grid operators [5],[6]. Customers can cut down their energy and demand charges by reducing energy use from the grid, shifting energy use from expensive to cheap hours and reducing their peak power demand [7]. System operators can benefit too by having a smoother electricity load profile with lower peak power demand (peak-shaving) that makes frequency regulation easier. Other benefits are grid and peak capacity investment deferral [7] and power losses reduction having on-site production [8].

Aims

The objective of this thesis is to design a PV-BESS for a typical vacation home located in Sälen area in Sweden. The aim is to optimize the system size and component selection to get the most economical solution for the electricity customer to satisfy the home energy use. Considering the current problems of the electricity system, the hypothesis is that it is possible to find a PV-BESS that, with the current electricity market situation and subsidies, can provide benefits to both electricity customers and grid operators. A techno-economic assessment of the performance of the system needs to be done.

The main requirements of this system are:

• Include a PV system to produce solar energy.

• Include a BESS to store energy.

• Include a smart energy management system.

The objectives and therefore the boundary conditions are focused exclusively on a system located in Sälen area in Sweden. The assessment is performed for a typical vacation home, not a particular case. In addition, the legislation framework (e.g. operating licenses,

permits, registration fees…) is not fully studied.

Technical results are presented for the PV-BESS operation and its interaction with the grid in terms of energy and power exchange. Economic results are presented only for the electricity customer who is expected to be the homeowner. From a grid operator perspective, changes on the house load profile are studied and its possible benefits and drawbacks on the grid are presented in a qualitative way.

One of the future projections of this system is to work in a microgrid. However, the optimization and design are focused to operate the system individually. How to control and optimize the performance of this system when working in a microgrid is out of the scope of the thesis.

Method

The method followed is divided into 5 phases. Firstly, the boundary conditions for the simulation are established. Then, HOMER Grid software is used in second and third phases which are shown in Figure 1.1. The second phase is a pre-sizing of the system done by an analytical iterating method. The third phase is a multi-year simulation based on the pre-sizing results. The fourth phase consists in a technical and economic analysis of multi-year simulation results. Finally, the fifth phase is a sensitivity analysis is performed on different variables.

Figure 1.1: Optimization analytical method

A brief explanation of each phase is presented below:

1. Boundary conditions. The first part of the method consists in obtaining all data necessary to run the simulation. An online research is done to know which PV-BESS components are currently available on the market and their characteristics and prices. Also, information on government subsidies, solar resource, load profiles, electricity price trends and more is gathered. This data defines the input values and boundary conditions of the simulations.

2. Pre-sizing. HOMER Grid software is used to simulate and optimize a

preliminary system. This is done through an iterating analytical simulation using typical and average values for the variables found in the market analysis. This iterating part is composed by several simulations that allow HOMER to use its auto-sizing feature. A base case system is firstly simulated and then different variables are changed until a satisfactory pre-sizing is obtained. This simulation provides a preliminary result because only generic components are used with average values. Also, when the auto-sizing feature of HOMER is used, the

software does not allow to simulate the time evolution of a system through several years. This means that only one year is simulated and therefore PV and battery degradation, price changes and other time-dependent parameters cannot be simulated. Other advantages of this presizing are that simulations can be run in a short time and they allow to find the best way to input the boundary conditions and to discard some assumptions.

3. Multi-year simulation. Based on results from the pre-sizing, a multi-year simulation is run. In this case, real market components with their own

characteristics and price are used instead of generic components with average values. The selection of these components is based on real market products that are as similar as possible to the optimized pre-sized model. Here, a multi-year simulation is used for a period of 25 years and time-dependent variables are included. Different systems (only grid connected system, PV system and

PV-BESS) are simulated to be able to compare results. As this is a simulation with future time perspective, six future scenarios are simulated combining 2 %, 3 % and 4 % electricity price increase per year and 100 % and 80 % of capital costs found in the market analysis.

4. Result analysis. An economic and technical analysis of results is performed. Raw data from the simulation outputs is obtain and sometimes treated to show the most important results and get an easier view of the system performance. The economic analysis is performed from an electricity customer perspective and the technical analysis is performed from a utility grid company perspective.

5. Sensitivity analysis. A sensitivity analysis is performed to different variables to

understand how and in what extent they change the simulation results. This allows

to understand results better and improve its reliability. The sensitivity analysis is

performed on electricity tariff, load profile and capital costs. The sensitivity of

electricity price and capital cost is also analyzed in the different future scenarios

and therefore it is not included in this chapter.

The optimization and simulation of the PV-BESS is done using HOMER software [9]. It

is a widely recognized software originally created by the National Renewable Energy

Laboratory. Its main advantage is that it allows to simulate hybrid systems that include

different energy sources and energy storage and at the same time optimize the size and

operational strategy for each component. Furthermore, it allows to input sensitivity

analysis to find the different factors that affect results the most. These are features that

similar tools such as PVSyst [10], Polysun [11] and Trnsys [12] cannot do. HOMER

software is divided in HOMER Pro and HOMER Grid. Homer Pro is designed to

simulate isolated microgrids, unreliable grids and rural electrification. HOMER Grid is

designed to simulate behind-the-meter distributed generation and commercial scale solar

and storage. The main advantages of using HOMER Grid are the availability to create

complex electricity tariffs, apply incentives and optimize a system by peak-shaving

dispatch, features that HOMER Pro cannot do. For simplicity, HOMER Grid will be

named as HOMER in advance.

2 Technology background and literature review

The next sections present an overview and literature review about PV and BESS technologies, PV-BESS system applications, design, sizing and optimization tools for energy systems and market costs..

The International Renewable Energy Agency [6] published recently an electricity storage guide to assess the system value and ensure project viability of renewable power systems with electricity storage. Listed below there is a summary of the phases that they

recommend following to design a renewable power system with electricity storage:

1. Identify services and other benefits that BESS can provide to integrate renewables energy customers to the grid.

2. Study the suitability of different technologies for this case.

3. Optimize battery capacity and renewable system size. In this case, modeling of the system is often necessary. Iterating methods are often required to optimize

benefits.

4. Optimize energy dispatch tool and storage operation for maximum benefits.

Hourly electricity costs and weather data is often required.

5. Develop a project feasibility model to study total life-time costs and revenues.

Sometimes, monetizable revenues are not enough to overcome costs but other benefits can make the system feasible.

Photovoltaic technologies

Currently, the solar energy market is dominated by crystalline silicon technology with 90 % share of the market in 2017. There are two types of silicon modules with mono-crystalline technology with higher cost but also higher efficiency and poly-crystalline with lower cost but also lower efficiency. The other 10 % of the market is composed mainly by thin-film technologies [2]. The main advantages of crystalline silicon are its high efficiency and reliability, with low degradation rates. Also, it is a proven and mature technology. Thin film technology however can become a cheaper solution in some cases and have a lower energy payback time [13].

Battery technologies

Currently, the energy storage market is dominated by pump hydro storage which in 2019 accounted for 153 GW compared to only 4 GW for BESS [1]. Pumped hydro storage accounts for 96 % of storage capacity, followed by thermal storage (1.9 %), batteries (1.1 %) and other mechanical energy storages (0.9 %) [14]. However, the current situation of BESS is in an exceptional development and increase of worldwide installed capacity, driven by its main advantages: fast response time, ease of control, modularity and geographical flexibility [2].

Batteries such as lead-acid, lithium-ion (li-ion), zinc-bromide and nickel cadmium technologies stand out because its wide range of capacity and power rating and having a fast time response in the order of minutes or even seconds.

Zubi et al. [15] show an extensive description of the advantages and disadvantages of the main commercial available BESS technologies: lead-acid, li-ion, NiMH and NiCd. Table 2.1 summarizes the most important advantages and disadvantages of each technology

according to this study.

Table 2.1: Advantage and disadvantages of main battery storage technologies [15].

Technology Advantages Disadvantages

Lead-acid

Relative low cost Mature technology

Large number of manufactures No memory effects

Low self-discharge

Low specific energy and power Short life cycle

High O&M costs Temperature sensitive Limited reliability

Pb and gas release concerns

Li-ion

High specific energy and power High roundtrip efficiency Long lifetime

High reliability

Usually eco-friendly materials High R&D investments

High initial cost

Advanced monitoring required Safety concern (heat explosive) Low recovery and recycling

NiMH

High reliability Eco-friendly materials Good safety

Fast recharge

High self-discharge Memory effect Short life cycle

Low recovery and recycling

NiCd

Low initial cost Mature technology High reliability

High operative temperature range

Low specific energy and power Memory effect

Low roundtrip efficiency Hazardous cadmium

Battery storage system applications

Applications for BESS are varied. Depending on the objectives, the optimization of the system is different. According to Hesse et al. [5], a summary of the five main applications that exist for BESS in power grids is presented.

• Ancillary service: Batteries can be used as an ancillary service to an existing power grid to keep balance between energy generation and demand. When power sources are variable (e.g. renewables) or electricity load is unstable, battery storage can balance the grid by storing or releasing electricity. This application will likely become more important because of the higher renewable share in modern power grids and batteries have shown to be capable of regulating frequency with very fast response time, up to milliseconds [16].

• Behind-the-meter: It is the use of BESS for residential customers together with another energy source, often renewables such as PV or wind power. This allows to increase self-consumption and other benefits such as lower tariffs due to low peak power demand. Also, ramping control and quality increase of voltage frequency can be achieved.

• Energy trade: Electricity market price fluctuates with time in a range from minutes to even years. Usually, peak power demand in the evening lead to high prices and low power demand overnight lead to low prices. BESS allow to buy electricity when prices are low to sell when they are higher to make profit.

Difference between bought and sold electricity must be high in other to

compensate the investment costs of the BESS.

• Grid support and investment deferral: BESS can be used for supporting the grid to adapt power variations, control ramp-up events or reduce power flow in some lines. Without BESS systems, power lines, transformers, switches, etc. between production and end use of energy must be designed for peak power demand. By using BESS it is possible to smooth peak demands and reduce costs of the grid infrastructure. Also, voltage fluctuations and active/reactive power can be controlled

• Combined applications: Multi-purpose applications are gaining importance recently where a BESS is used to do multiple tasks and have multiple benefits. For example, a BESS can be used both for increasing renewable self-consumption and for peak-shaving.

System design

There are three main connection topologies available in the market for PV-BESS [5], [17], which are represented in Figure 2.1. A brief description with advantages and disadvantages for each topology is listed below [5].

• AC coupled: Figure 2.1 (a). The connection between the PV system, the BESS and the loads is done through the power alternating current (AC) lines of the house instead of a direct current (DC) connection between PV system and BESS. The advantages of this topology are the ease of adding a BESS to an existing PV system and the flexibility of the implemented system design, size and technology. Also, the battery can be charged and discharged directly from/to the grid with low losses.

The main disadvantages are a higher cost because two inverters are required and the higher conversion losses for PV generation because DC/AC/DC/AC conversions are required to store and discharge energy from PV. Furthermore, smart control and communication between inverters is required.

• DC link coupled: Figure 2.1 (b). Both systems are connected after its DC/DC converter. The main advantages are low charging losses because AC conversion is not necessary, cheaper cost because only one inverter is required and simpler battery charging control. The main disadvantages are that the inverter size limits the maximum power that can be supplied to the loads and usually it is not possible to install to an existing PV system. If the inverter it is rated for peak power

demand, when supplying lower loads efficiency will be lower.

• MPPT coupled: Figure 2.1 (c). In this case the inverter is connected before the

Maximum Power Point Tracker (MPPT). The main advantages are a cheaper cost

because only one inverter is necessary, its simplicity and the high efficiency due to

no AC conversion required for charging. The main disadvantages are difficulty to

add BESS to already existing PV systems, charging from AC side is not possible

and difficulty to find products because it is not a very common technology.

Figure 2.1: System connection topologies PV-BESS: (a) AC-coupled; (b) DC-link coupled; (c) MPPT-coupled. Used with CC 4.0 permission [18].

Sizing and optimization

Appropriate sizing of BESS for distributed renewables is a key factor to achieve the optimum system and maximize benefits. Yang et al.. [2] present a review of the main battery sizing criteria, methods and applications for renewable systems based on a great number of studies. For distributed renewable energy systems, analytical methods clearly stand out from other methods such as probabilistic methods, mathematical optimization methods or heuristic methods.

Several modeling and optimization tools that are designed or can be used for PV-BESS exist. Different studies have been published analyzing these tools with their advantages and disadvantages [5], [19]. Ringkjøb et al. [19] present a review of 75 modelling tools that can be used for energy and electricity power systems. They serve different purposes such as investment or operation decision support, power system analysis, forecasting…and have different interfaces, code availability, cost, etc. Some of the most common for PV-BESS system design are HOMER, SAM and Trnsys, however, many other tools can be found.

Cao et al. [20] study the impact of the time resolution used when simulating energy

systems on the self-consumption result. A comparison between simulations using the same

PV system and different time-steps is performed. Results show that a one-hour time-step

can have a mismatch 80 % higher than a one-minute timestep. Results also show that by

installing a battery of 0.25 kWh in the system, the error can be reduced from around 80 %

to 10 % in some cases with 1-hour time-step resolution thanks to the peak shaving that

cuts intermittent peaks that appear at higher resolutions. Luthander et al. [21], based on

Cao et al. [20] and other literature, recommend using a time-step shorter than one hour for

PV systems.

Talent et al. [22] researched about the optimal sizing for PV-BESS in Australia under two different electricity tariffs using a mixed integer linear program. The first tariff includes only an energy charge in Swedish krona (SEK) per kWh and the second one is a demand tariff that includes an energy charge (SEK/kWh) plus a demand charge (SEK/kW). The system size is optimized to minimize the total expenses of the electricity customer over the project lifetime. The study case is a residential home with an annual energy use of 10.8 MWh and a residential load profile. The economic boundary conditions are 58600 SEK for a 5 kW PV system and 42300 SEK for a 7kWh battery system. The results show that for both tariffs the best system architecture is a PV size of 5 kW and not to install a battery. When forcing the optimization to install a battery, an optimum size of 5 kW PV system and 7 kWh battery (the lowest limit) is found. The optimum energy dispatch strategy for the energy tariff is mainly following the load and, for the demand tariff, is peak-shaving. With the demand tariff, the maximum annual peak power demand goes from 6 kWh/h to 2.25 kW/h. With the energy tariff, the maximum peak is almost no reduced. The addition of a battery is significantly more profitable with a demand charge for customers. However, the amount of energy used from the grid, battery and PV system is very similar.

Wu et al. [23] analyze the optimum size for different scenarios of behind-the-meter battery storage using a linear programming algorithm. The battery cost modelled is divided in two parts: cost per battery capacity (3969 SEK/kW) and cost per maximum battery discharge power (880 SEK/kW). Results show that the optimum battery size has to provide a battery autonomy between 3.2 h and 4.1 h depending on location (Chicago, Houston, New York City and San Francisco) and load profile (office, retail, school, hotel, hospital, warehouse).

Despite the fact that a residential building load profile is not used, one of the conclusions is that the load profile does not influence significantly the optimum size. Therefore, this optimization might be also useful to compare with this thesis. Their results show that the savings in energy charge by the battery are very small compared to the savings of the demand charge.

Weniger et al. [24] analyze the optimum size of PV-BESS for residential homes in Lindenberg, Germany. The optimization is based on having the lowest minimum electricity price including all costs during the project lifetime. System prices of

15000 SEK/kW for PV system, 15000 SEK/kWh battery storage capacity, and a feed-in tariff of 1.1 SEK/kWh are used. The optimum size is based on the total annual load, giving a result of 1.1 kW PV power and 0.5 kWh battery storage per MWh of annual energy use. The demand charges, energy charges and capital costs play a crucial role in the sizing of the system. However, a trend is found that self-consumption is a key factor for sizing because energy sales to grid will likely be reduced in the future.

Ollas et al. [25] study the impact of three different energy dispatch algorithms (day-ahead,, day-behind and a Target-Zero) in order to optimize peak-shaving to reduce demand charges and maximizing self-consumption in a single family house in Sweden. The system has a 3.6 kW PV system and a low energy use (not specified but close to nearly-zero energy building). Results show that the optimum size of the battery to increase

self-consumption is 7.2 kWh. For larger capacities, the increase of self-consumption is low

due to the limited PV production. Depending on the energy dispatch algorithm, they get a

self-consumption between 30 % and 50 % and an average peak power between 0.5 kW

and 2.0 kW.

3 Method

In this chapter the method followed is explained step by step. There are 5 subchapters, each one explaining one phase of the method:

1. Boundary conditions: What data is necessary to establish the boundary

conditions and run the simulations. An explanation of data sources, calculations, assumptions and simplifications is provided for every HOMER Grid simulation input.

2. Pre-sizing simulation: Description of the iterating analytical method for the pre-sizing of the system.

3. Multi-year simulation: Description of the multi-year simulation for the final sizing of the system and the different future scenarios.

4. Result analysis: How the economic and technical analysis of simulation results is performed: description of the outputs, parameters, calculations and comparisons.

5. Sensitivity analysis: Description of the variables studied in the sensitivity analysis:

energy tariff, load profile and feed-in benefits.

Boundary conditions

The boundary conditions of the simulations are the input parameters that HOMER needs to run to calculate the results. An extensive analysis needs to be done to input the most reliable inputs and minimize uncertainties related with assumptions.

All market components with their characteristics and prices are taken from two retailers that sell their products in Sweden: Europe-solarstore [26] and Solelgrossisten [27]. The research period is between July and August 2020. Unless specified, the boundary data used is from year 2019. In the case of subsidies, only official websites are used.

In a list, these are the eleven categories of inputs that have been used:

• General inputs and assumptions.

• Solar resource.

• Load profile.

• Capital subsidies.

• Fixed capital costs.

• PV modules.

• Inverters

• Batteries.

• Energy tariff

• Demand tariff.

• Feed-in benefits.

All these inputs are shown in detail in the next sections describing data sources, calculations, assumptions, and simplifications.

3.1.1. General inputs and assumptions

The general assumptions that affect the overall project are 25-year lifetime (typical PV module lifetime) and discount rate of 8 % and an inflation rate of 2 %, values

recommended by HOMER Grid. Currency used is Swedish krona (SEK) and the equivalence with euro is considered to be 1 euro equal to 10 SEK. For simplicity, unless specified, all inputs include 25 % VAT. Electricity prices have an additional energy tax rate analyzed in its own section.

The time-step resolution of the simulation is limited by the load profile provided by

Dalakraft and the solar data downloaded from different sources to 1 h hour timestep.

The O&M costs are very low for PV and li-ion batteries. For the simulations, each

component is included with 0 SEK/year O&M cost and a small fixed amount of 100 SEK per year is included for the entire system. The replacement costs are modelled separately.

3.1.2. Solar resource

One of the main factors when designing a PV-BSS system is the solar irradiance available in the location. Unfortunately, no ground data is available for Sälen and therefore different tools that use satellite data have been used to compare and choose the most reliable resource. PolySun [11], PVGIS [9], Strång [28], PVSyst [10], HOMER [9] and SMHI [29]

tools and databases are used. Listed below there is a description of each profile generated:

• PolySun: It is a simulation software for solar energy systems. It imports solar data from Meteonorm.

• PVGIS: It is the Photovoltaic Geographical Information System created by the European Commission provides a database of meteorological data. Three different satellites are available: SARAH, COSMO and ERA-5. The three satellites are compared. COSMO and ERA-5 are designed for high latitudes (greater than 60º) which is the case for Sälen located at around 61º north. Furthermore, this database allows to include horizon profiles to consider shadowing from topographic

obstacles.

• STRÅNG: This Swedish tool financed by the Swedish Meteorological and Hydrological Institute, the Swedish Radiation Safety Authority and the Swedish Environmental Protection Agency allows to download solar data for Sweden from previous years.

• PVSyst: This simulation software also imports solar data from Meteonorm. It allows to download a solar profile based on NASA satellite data (PVSyst 100 % NASA) or a synthetic file created with 90 % NASA satellite data and 10 % data from nearby ground stations (PVSyst 90 % NASA).

• HOMER: This simulation tool allows to download solar data from Meteonorm using NASA satellite data and it can also create an hourly profile based on monthly averages. Two profiles are created: one importing the hourly profile from NASA (HOMER NASA) and one using the monthly average from NASA to generate a hourly profile in HOMER (HOMER monthly).

Figure 3.1 shows the monthly irradiance based on the different data sources. January, February, November and December have almost no irradiance and the months with higher solar resource are May, June, July and August as should be expected for the northern hemisphere. The irradiance profiles are all very similar for all resources with higher peaks in summer which never reach more than 900 W/m².

Figure 3.1: Monthly solar irradiance in Sälen based on different sources.

The solar irradiance of each source is shown in Table 3.1. Most results show an annual value of around 900 kWh/m² which is an expected value. Data from STRÅNG is significantly lower and therefore it is discarded. Comparing PVGIS results, the most reliable sources should be ERA5 and COSMO satellites which are specially designed for higher latitudes. As Table 3.1 shows, both provide the same data which gives a higher solar resource than SARAH satellite, designed for lower latitudes. Both PVSyst results are slightly higher than the others with annual irradiance of 932 kWh/m².

PVSyst data provides two irradiance profiles with the same annual solar irradiance. The PVSyst 100 % NASA gives a higher peak in June with 172 kWh/m² compared to the 155 kWh/m

from PVSYST 90 % NASA. However, this second one has higher values for the other months of summer. PolySun gives a lower annual irradiance than average, with higher values in summer but lower values in the other seasons. Finally, HOMER NASA gives the same output than PVSyst with with 90 % satellite data and HOMER monthly data looks slightly higher than average.

Table 3.1: Monthly and annually solar irradiance by source [kWh/m²].

Month Poly- Sun

PVGIS COSMO

PVGIS SARAH

PVGIS

ERA5 Strång PVSyst 100%

NASA

PVSYST 90 % NASA

HOMER

NASA HOMER monthly

Jan 7.8 8.7 6.1 8.7 7.7 8.1 8.1 9.3 8.0

Feb 24.1 22.4 12.8 22.4 19.7 26.6 26.6 19.6 26.6

Mar 63.5 72.7 58.3 72.7 50.1 69.8 69.8 73.5 69.7

Apr 106.0 132.2 107.9 132.2 71.3 95.8 112.5 119.9 112.4 May 146.6 133.7 122.0 133.7 129.1 151.3 158.4 133.8 158.3 Jun 156.1 169.6 160.2 169.6 125.5 172.1 155.7 165.7 155.4 Jul 151.5 124.6 121.7 124.6 125.6 139.7 151.9 126.3 151.9 Aug 109.4 117.0 111.2 117.0 107.8 135.2 120.3 122.9 120.1

Sep 68.1 70.1 70.2 70.1 45.2 74.3 74.1 69.8 74.1

Oct 30.2 37.9 34.4 37.9 30.3 38.9 36.9 35.2 36.7

Nov 9.0 12.5 8.8 12.5 12.3 14.8 13.2 10.1 11.1

Dec 3.9 4.3 2.7 4.3 3.6 5.2 4.3 4.4 4.4

YEAR 876 906 816 906 728 932 932 891 928

The final decision is to use PVGIS SARAH/COSMO. This is because they provide average values and because these satellites are specially designed for higher latitudes.

Snow soiling losses are also considered. In this region, close to a skiing resort, snow can

cover solar modules for several weeks or even months. However, this can vary a lot from

year to year. An assumption is made assuming that in winter most of the time solar

modules will be covered by snow except for a few days when the snow is melt. The

snow-covered periods considered are from 1

to 25

of January, from 1

to 15

of

February, from 1-5

of March and from 1

to 25

of December. This is included in

HOMER simulation making irradiance equal to 0 W/m² in these periods.

3.1.3. Load profile

The load profile is one of the most important factors when designing and optimizing a PV-BESS. An hourly load profile for a vacation home in Sweden is provided by Dalakraft for the month of March shown in Figure 3.2Figure 3.2: Hourly load profile for a house in Sweden in March provided by Dalakraft. The data input for HOMER is an average hourly value calculated as kWh/h.

Figure 3.2: Hourly load profile for a house in Sweden in March provided by Dalakraft.

To be able to find any weekly trends, all four full weeks of March are shown in Figure 3.3.

The energy use of per day of every week is shown on top of each other. No significant difference is found between weekdays and weekends which is something expected because this is a vacation home without a routine of working days. Also, every week looks different and there is no clear trend.

Figure 3.3: Weekly load profile comparison.

The daily load profile is also analyzed in Figure 3.4 by adding on top of each other the four Tuesdays and four Wednesdays of the load profile. The variability between days is very high, for example comparing Tuesday 4 and the Wednesday 1. However, the average of all days shows a trend with low energy use during night and peaks which are usually around 10 h and 17 h.

Figure 3.4: Hourly load profile for Tuesdays and Wednesdays of March 2017.

Also, monthly total energy use is provided for 14 months, from January 2019 to February 2020 by Malungs Elnät company. Data from January until December 2019 (see Appendix A) is used to get the monthly total energy use for an entire year. As shown in Figure 3.5, it is clear that months with the highest energy use are in winter and the lowest energy use is in summer. The total load for an entire year is 11219 kWh. With this data it is possible to generate an hourly load profile for an entire year.

As the hourly profile provided by Malungs Elnät is for only one month (March 2019). An entire year load profile is generated by taking the hourly profile for March as a reference.

The hourly load profile of March is divided by its total monthly energy use. Then, this profile is used every month multiplied by the total energy use of each month. This is an assumption that makes all months have the same load profile, which might be not the reality. The hourly load profile generated (repetitive month profile) is also shown in Figure 3.5. There, it is possible to see that every month has the same shape but different size.

Figure 3.5: Generated daily load profile and monthly load profile provided by Dalakraft.

0 200 400 600 800 1000 1200 1400 1600

0 1 2 3 4 5 6 7 8

J F M A M J J A S O N D

Energy use [kWh/ month]

Energy use [kWh/ h]

Month

Hourly load profile Monthly load profile

As the load profile includes several assumptions, another load profile is simulated to reduce uncertainties and be able to compare. This load profile (HOMER profile) is generated by HOMER. By introducing the total load for an entire year, HOMER can generate an hourly load profile for an entire year. The specifications chosen to generate this profile are “Residential Home Profile” with higher energy use in winter and daily variability of 10 %. These two load profiles compared have the same annual energy use but different profile.

Figure 3.6 shows a comparison of both profiles using the average load of every hour in a year. An average is used because it is difficult to compare hourly values because there is high variability between days and weeks. With this figure, it is easy to see that HOMER profile has lower energy use during night. Furthermore, it has three different peaks in a day, one early in the morning around 5 h, one around 13 h and one around 19 h. In comparison, the repetitive month profile has a higher energy use during night and only two peaks, one in the morning around 9 h and another one around 18 h. The maximum peak of HOMER profile is found to be much higher on average than the repetitive month profile.

Figure 3.6: Repetitive month profile and HOMER profile comparison.

3.1.4. Capital subsidies

Capital subsidies are subsidies that can be obtained for the capital cost a system or part of it. In Sweden, capital subsidies can be obtained from the government for both PV systems and BESS with some restrictions:

• PV system subsidy:[30]–[32]. A 20 % cost subsidy can be obtained for design, workforce, and material costs. This includes modules, wires, frames, clamping structures, monitoring systems, switches, inverters, measurement equipment and protection equipment. The minimum self-consumption of PV energy must be 20 %. A limitation exists for a maximum of 1.2 million SEK per system and a maximum of 37.000 SEK per kW installed.

• BESS subsidy: [32]–[34]. A 60 % cost subsidy can be obtained for design, workforce and material costs. This includes batteries, wiring, control systems and smart energy management systems. The battery system must be used for

self-consumption with a renewable energy source. There is a maximum amount of 50.000 SEK per system.

Unfortunately, HOMER only allows to introduce subsidies for two components: solar modules and batteries. The other capital costs of the system (fixed capital costs and

0 1 2 3 4

0 5 10 15 20 25

Energy use [kWh /h]

Hour of the day [h]

Repetitive month profile HOMER profile

inverter capital costs) cannot be subsidized. To solve this problem, the subsidies that are applicable to them are introduced directly in the component cost.

3.1.5. Fixed capital costs

Fixed capital costs are defined as capital costs which are not PV modules, inverter, and battery costs. In HOMER, these costs are included in fixed project costs. According to an International Energy Agency report [35], costs for PV systems in Sweden (systems without battery) were distributed as follows: 29 % PV modules, 13 % inverter, 12 % installation, 12 % project margin, 14 % other costs and 20 % VAT. The same report two years after, in 2018 [8], concluded that the average price for roof-mounted PV systems without energy storage for single family houses (around 5 kW) was approximately 14400 SEK/kW installed. With this data, Table 3.2 is calculated:

Table 3.2: Assumed fixed capital costs per kW for PV systems in Sweden based on data from 2016 and 2018.

Cost type Cost

[%] SEK/kW

(VAT excl) SEK/kW

(VAT apart) SEK/kW

(VAT incl.) Fixed capital cost (VAT incl.)

PV 29 5220 5011 6264 -

Inverter 13 2340 2246 2808 -

Installation 12 2160 2073 2592 2592

Project margin 12 2160 2073 2592 2592

Other 14 2520 2419 3024 3024

VAT 20 - 3456 - -

TOTAL 14400 17280 17280 8208

The total fixed capital costs for PV systems are 8208 SEK per kW installed of PV. As HOMER does not allow to introduce fixed costs per kW (only total amount), the fixed capital costs are adjusted every time the size of the PV is changed. As PV systems can get a 20 % subsidy from the government, a 20 % reduction is applied to fixed capital costs giving a final cost of 6566 SEK/kW.

Fixed capital costs for PV-BESS systems are calculated similarly. Unfortunately, no data for Sweden was found, and therefore an assumption is made that fixed capital costs for PV-BESS are 15 % higher than for PV systems. In this case, two different subsidies can be applied to capital costs: 60 % coming from battery subsidy and 20 % coming from PV subsidy. As an assumption, the battery subsidy is applied to 30 % of the fixed capital costs and the PV subsidy is applied to the other 70 % of fixed capital costs. The total fixed capital costs are 9439 SEK/kW (6418 SEK/kW including subsidies).

3.1.6. PV Modules

In Sweden, end-consumer prices of crystalline silicon modules reported by Swedish installers ranged in 2019 from 3.2 SEK/W to 6.6 SEK/W, with an average of 4.5 SEK/W (excluding VAT) [8]. A value of 5.625 SEK/W including VAT (4.5 SEK/W excluding VAT) is considered for pre-sizing and multi-year simulation. Note that there is a lower cost for module prices compared to the value calculated in previous Table 3.2 because it mixes data between 2016 and 2018 and prices have declined since then.

HOMER cannot define the number of modules per string, module orientation and tilt,

wire losses, soiling losses, mismatch losses, etc. To include all these losses in the software,

an assumed derating factor of 85 % is included in the PV module component (that means

that the power output is reduced by 15 % all time).

The PV modules used for the simulation are generic flat plate PV from HOMER database with an assumed lifetime of 25 years, the same than the project lifetime. The current PV module market is already a mature market with hundreds of modules with very similar characteristics. Typical values are power ranging between 250 W and 350 W, efficiencies around 18 % and degradation rates with a warranty of at least 85 % relative efficiency after 25 years. Amongst the best quality-price modules, only silicon modules have been found.

Therefore, thin film technologies are discarded. No significant benefits have been found between mono and poly crystalline. Mono crystalline have higher efficiencies but they are also more expensive.

3.1.7. Inverters

Many different inverters can be found in the market, some of them are specific for PV, some are specific for batteries and some are hybrid inverters that can work with both at the same time. The selection of the inverter is very important because it will determine the system topology as previously shown in Figure 2.1.

The factors that are considered for the inverter selection are: price, peak power, maximum PV power input, battery compatibility, power range, efficiency and warranty. After

research, it is decided that the best option for PV-BESS is to use a hybrid inverter. This means that the system topology is a DC-linked topology as shown previously in Figure 2.1 (b).

This solution is cheaper than purchasing two inverters separately for battery and PV (AC-linked topology). In addition, it has a simpler scheme which is expected to have lower installation costs. Also, hybrid inverters have lower compatibility constraints and they already include smart energy management systems compatible with certain battery models.

The MPPT coupled topology has not been found. Also, unless it had a very competitive price, it would not be considered because it does not allow battery charging with AC power from the grid, a feature that might be useful. The three brands of hybrid inverters selected are:

• Fronius Symo Hybrid inverter.

• Sungrow Hybrid Residential Three Phase Inverter

• Huawei Sun 2000L.

Each brand has different inverter sizes. Its main characteristics are shown in Table 3.3. For all of them, 12.5 years of lifetime is considered (half time of the project).

Table 3.3: Hybrid inverter characteristics and battery compatibility. *3000 SEK added for CheckBox compatibility component. Data from https://www.europe-solarstore.com/

Inverter AC Power

[kW]

Price [SEK]

Price (VAT incl.)

[SEK]

Cost per kW [SEK/kW]

Max. PV power input

[kW] Battery compatibility Fronius

Symo Hybrid

3.0 17850 22312 7437 5.0 BYD Battery Box HVS

LG RESU 7H and 10H*

Fronius Solar Battery

4.0 19780 24725 6181 6.5

5.0 21680 27100 5420 8.0

Sungrow Hybrid Residential

5.0 16510 20638 4128 7.5

BYD Battery Box HVM BYD Battery Box HVS LG RESU 7H and 10H

6.0 17340 21675 3613 9.0

8.0 18110 22638 2830 12.0

10 19380 24225 2423 15.0

HUAWEI SUN2000L

2.0 7750 9687 4843 3.0

LG RESU 7H and 10H

3.0 10040 12550 4183 4.5

3.7 10890 13612 3699 5.5

4.0 11290 14112 3528 6.0

4.6 11790 14737 3203 6.9

5.0 12200 15250 3050 7.5

All three inverters have two maximum power point trackers and an european weighted efficiency greater than 97 %. Huawei and Sungrow inverters have a warranty of 10 years and Fronius has a warranty of 5 years plus another optional 5 years upon payment. As expected, the higher the power, the lower the price per kW. All three inverters have a very similar scheme with the previously mentioned DC-link coupled topology. Also, they all include a smart energy management system.

In the pre-sizing simulation, HOMER needs an input of the inverter cost per kW. In the multi-year simulation, the selected inverters are introduced with its respective cost. Figure 3.7 shows the price and kW of each of the inverters. There, it is possible to see that Fronius inverters are more expensive per kW than the others, however, they do not offer better characteristics. Therefore, they are discarded from the component selection.

Figure 3.7: Inverter power and price comparison and price trend excluding Fronius inverters To input the inverter costs per kW in HOMER, a linear fit of the points in Figure 3.7 excluding Fronius inverters is done (Equation 3.1). Then, two points are calculated for the expected range of inverter power. These two points are 2 kW and 12 kW with cost of 10937 SEK and 3040 SEK respectively. When these two points are introduced in

HOMER, it automatically produces a linear price trend per kW that follows Equation 3.1.

As said before, HOMER cannot apply subsidies to inverter costs (only to PV modules and batteries) and therefore they are applied directly to the inverter cost. In this case, a 60 % subsidy is applied to the capital cost of the inverter. The battery subsidy is greater than the PV subsidy and it can be applied to the inverter because it includes a smart energy

management system. Replacement costs are introduced with the assumption of 80 % of capital initial costs assuming that some components such as cables will probably be reused and prices will very likely drop in following years.

For PV systems without energy storage, an inverter cost of 2808 SEK/kW is used in HOMER based on values from previous Table 3.2 and a 20 % reduction of cost has been applied as PV system subsidy. Again, an assumption of 20 % reduction in replacement capital cost is assumed.

0 5000 10000 15000 20000 25000 30000 35000

0 2 4 6 8 10 12

Pr ice [SE K]

Power [kW]

Fronius Symo Hybrid Sungrow Hybrid Residential

Huawei SUN 2000L Lineal (Price trend without Fronius)

𝑦 = 1910.3𝑥 + 7116.5 Equation 3.1

3.1.8. Batteries

Different batteries have been found that fit into the purpose of this project and are compatible with the three previous inverters. Other parameters that have been considered for the selection are: capacity between 1 and 22 kWh, peak power, cost, lifetime, maximum depth of discharge (DoD) and system compatibility. Only li-ion batteries have been found that are compatible with the previous inverters. The selected batteries are:

• BYD Battery BOX Premium HVS

• BYD Battery BOX Premium HVM

• LG RESU 7H and 10H.

• Fronius Solar Batteries.

Their main characteristics are shown in Table 3.4. The HVS and HVM models from BYD are very similar, being the HVM the newer version of the product. This is the reason why, for example, the HVM 8.3 kWh model has better characteristics than the HVS 7.7 kWh with almost the same price.

Table 3.4: Selected battery models and their main characteristics. Prices from https://www.europe-solarstore.com/.

Battery

model Cap.

[kWh]

Cost with VAT [SEK]

Cost per kWh [SEK/kWh]

Nom.

Volt.

[V]

Roundt.

Eff. [%]

Min.

SOC [%]

Max.

Charge curr. [A]

Max. Disch.

Curr.

[A]

Lifetime [kWh]

BYD Battery Box Premium HVM

8.3 49975 6021 153 96 0 43.4 50 25620

11.0 64875 5898 204 96 0 43.1 50 34150

13.8 77875 5643 256 96 0 43.1 50 42690

16.6 90750 5467 307 96 0 43.3 50 51230

19.3 103625 5369 358 96 0 43.1 50 59770

22.1 116875 5288 409 96 0 43.2 50 68310

BYD Battery Box Premium HVS

5.1 37250 7304 205 96 0 19.9 25 15410

7.7 49625 6445 256 96 0 24.1 25 23120

10.2 66125 6483 307 96 0 26.6 25 30820

12.8 79875 6240 358 96 0 28.6 25 38530

LG RESU 7H and 10H

7.0 53425 7632 450 95 5.7 8.5 10 39600

9.8 62250 6352 450 93.5 5.1 11.7 10 55800

Fronius Solar Battery

4.5 59675 13261 170 93.5 20 16.0 16 28800

6.0 69975 11663 230 93.5 20 16.0 16 38400

7.5 80938 10792 290 93.5 20 16.0 16 48000

9.0 91050 10117 345 93.5 20 16.0 16 57600

10.5 101838 9699 400 93.5 20 16.0 16 67200

12 111788 9316 460 93.5 20 16.0 16 76800

Because HOMER load is input in average hourly values and not in terms of minutes or seconds, peak power demand can be underestimated. Because of this, the maximum discharge current input in HOMER is the current that the battery can discharge for a long time. For example. BYD Battery Box Premium HVS models can supply 50 A for 5 seconds, but only 25 A is considered for the simulation.

Similarly to inverters, component prices and power are compared in Figure 3.8. Again,

Fronius batteries are also discarded from component selection because they have a higher

price without providing better characteristics.

Figure 3.8: Battery capacity and price comparison with price trend excluding Fronius inverters A linear fit is done excluding Fronius batteries to get Equation 3.2. In this case, the maximum and minimum points introduced in HOMER are 5.1 kWh and 22.1 kWh

according to the lowest and highest capacities from the component selection. Their price is of 39813 SEK and 117038 SEK respectively. Introducing these two points, HOMER automatically generates a price trend following Equation 3.2. Replacement costs are introduced as 70 % as capital initial costs. The same assumptions than for inverters are made but expecting larger cost reduction because batteries are a more immature technology than inverters.

In the pre-sizing simulation, the generic HOMER li-ion battery is used with a size ranging between 5.1 kWh and 22.1 kWh (the smallest and biggest batteries available). The battery lifetime is determined as 12.5 years (half of the project). In the multi-year simulation the specific characteristics of each battery are introduced.

The Tesla Powerwall battery is also included in the simulations. However, this battery is already including an inverter and the simulation is done in a different way than the others.

It has 13.5 kWh of useful capacity and the simulation is run with an inverter of 7 kW without any cost as it is already included in the cost of the battery. Its price is 87650 SEK including VAT, fixed capital costs are the same than for PV-BESS and the subsidies are applied in the same way.

In the multi-year simulation, Tesla Powerwall battery already has its own defined component in the HOMER Grid Library but the other batteries do not. These

components are created as a new HOMER library component introducing its respective characteristics shown in previous Table 3.4. Regarding battery degradation, no data was found to define an energy throughput curve depending on the maximum operating DoD of the battery. The energy throughput is the amount of energy that a battery can store and deliver over its lifetime. This energy throughput depends on degradation of the battery during operation and especially on the maximum DoD at which is discharged. The only data found for these batteries is a unique test point with the expected energy throughput using a fixed maximum DoD. As one unique point of data is not enough to input a

y = 4542,6x + 16646

0 20000 40000 60000 80000 100000 120000 140000

0 5 10 15 20 25

Price [SE K]

Capacity [kWh]

BYD Battery Box Premium HVM BYD Battery Box Premium HVS

LG RESU 7H and 10H Fronius Solar Battery

Lineal (Price trend excluding fronius)

𝑦 = 1910.3𝑥 + 7116.5 Equation 3.2

degradation curve in HOMER, an assumption is done so that all batteries follow Equation 3.3 for degradation.

A calculated point is introduced from 30 % until 100 % DoD in steps of 10 %. An example for BYD Battery Box HVS 5.1 and HVM 8.3 is shown in Figure 3.9 where the 100 % DoD energy throughput is provided in the datasheet and the other points are calculated using Equation 3.3.

Figure 3.9: Assumed battery lifetime in kWh by maximum DoD for BYD Battery Box HVS 5.1 and HVM 8.3

3.1.9. Energy tariff

Electricity prices in Sweden are based on the Nord Pool hourly market. However, energy companies buy this energy and add their costs and benefit margin on top of these prices.

The amount of money that the electricity customer must pay for the electricity tariff is the energy charge.

To input into HOMER an energy tariff (named monthly tariff), Borlänge Energi company tariff is followed. A fixed monthly tariff per kWh is introduced which is the monthly price from Borlänge Energi in 2019 shown in Table 3.5. Then, a 20 SEK/month fixed tariff is also added. On top of both, 25 % VAT is added. From 1

of January 2020 there is an energy tax of 35.3 öre/kWh which including VAT makes a total of 44.13 öre/kWh [36]. A fixed charge of 44.13 öre/kWh for purchased electricity is included in HOMER to

consider this energy tax.

A second energy tariff is used in the sensitivity analysis to be able to compare. While in the monthly tariff the electricity price is independent from the hour when the electricity is purchased, a second tariff (hourly tariff) is used where the electricity price changes every hour based on the Nord Pool price. Hourly electricity prices can be downloaded from the Nord Pool webpage [37]. Sweden is divided in four different regions with different prices, and Sälen is located in region S3. Data from year 2017 is used.

𝐸 = 𝐷𝑜𝐷

𝐷𝑜𝐷

· 𝐸

Equation 3.3

Datasheet

test point

To generate prices for this hourly tariff, electricity prices from Borlänge Energi company have been compared [38] with Nord Pool prices to see how much money customers pay for their electricity compared to the Nord Pool price. Table 3.5 shows the electricity price comparison (excluding fixed costs, electricity fees and VAT) for every month in 2019 and the annual average. From this table, it can be seen that, on average, customer electricity prices are around 22.5 % more expensive than Nord Pool prices.

Table 3.5: Electricity price comparison between Nord Pool and Borlänge Energi electricity prices.

Month Monthly average price Nord Pool

[SEK/kWh]

Monthly price Borlänge Energi

[SEK/kWh]

Price increase [%]

Jan 0.56 0.65 17.2

Feb 0.48 0.57 19.8

Mar 0.41 0.51 22.5

Apr 0.42 0.51 21.9

May 0.37 0.46 23.3

Jun 0.26 0.35 34.6

Jul 0.37 0.45 22.6

Aug 0.40 0.48 21.1

Sep 0.37 0.46 23.2

Oct 0.41 0.49 21.3

Nov 0.45 0.53 19.2

Dec 0.38 0.47 23.4

Average 0.41 0.49 22.5

To input the hourly tariff into HOMER, hourly prices for 2019 are downloaded from Nordpool webpage [37]. Then, an assumed 23 % higher price is input to consider the higher costs that electricity customers have to pay to their energy company compared to Nordpool prices. Also, a 20 SEK/month fixed fee is considered and, on top of both, 25 % VAT is added. Electricity tax is also included with VAT.

In order to run the multi-year simulation, price changes over time have to be considered.

An analysis of the electricity price evolution [8] shows that there has been a significant increase in recent years. Between year 2000 and year 2018 electricity prices went from around 0.7 SEK/kWh to 1.3 SEK/kWh including all costs and taxes. This makes an average year increase of 3.5 % per year. To project three future electricity price scenarios, a price increase of 2 %, 3 %, and 4 % per year is considered.

3.1.10. Demand tariff

Apart from electricity prices, costs related to the grid utility must be paid by the consumer.

The amount that must be paid is the demand charge. This cost is based not on the energy use but on the power demand for the grid. In this case, Malungs Elnät prices used as reference [39] (See Appendix B). The demand charge has a fixed tariff and a variable tariff.

The variable tariff changes depending on summer period (April, May, June, July, August, September and October) and winter period (November, December, January, February and March).

The fixed part depends on the grid connection (measured in Amperes) which for

households is usually 16 A. Its price is 1772 SEK/year or 2215 SEK/year including VAT

independently from summer or winter period.