SECOND CYCLE, 30 CREDITS STOCKHOLM SWEDEN 2020,
Techno-economic analysis of solar powered hydrogen production in vicinity of Swedish steel industries
MAX FRIMAN
KTH ROYAL INSTITUTE OF TECHNOLOGY
SCHOOL OF INDUSTRIAL ENGINEERING AND MANAGEMENT
Master of Science Thesis Department of Energy Technology
KTH 2020
Techno-economic analysis of solar powered hydrogen
production in vicinity of Swedish steel industries TRITA: TRITA-ITM-EX 2020:517
Max Friman
Approved
2020-09-11
Examiner
Hatef Madani Larijani
Supervisor
Nelson Sommerfledt
Industrial Supervisor
Hugo Larsson
Contact person Hugo Larsson
solar powered hydrogen production in vicinity of Swedish steel industries
MAX FRIMAN
Master in Sustainable Energy Engineering Date: September 23, 2020
Supervisor: Nelson Sommerfeldt, Hugo Larsson, Filip Enblom Examiner: Hatef Madani Larijani
KTH, Royal Institute of Technology Host company: Save by Solar AB
Swedish title: Analys av tekniska och ekonomiska parametrar för solkraft och vätgas production vid Svenska stålkraftverk
Abstract
The world’s steel industries account for 7% of the global CO2 emissions and the demand for steel is estimated to increase in the near future. Swedish steel manufactures are investigating processes to produce steel in an environmen- tal friendly way. One attractive option is to include hydrogen gas as much as possible in the processes. This thesis investigates hydrogen production from solar power in vicinity of 7 different steel producing facilities in Sweden. The objective is to construct systems with electrolyzers, PV and storage to possibly lower the cost of operation as well as lowering the CO2 emissions. The types of industries investigated are divided into two groups. The first group consists of industries with no current hydrogen demand. These industries are evalu- ated with a system which produces hydrogen from PV during the summer and storing it until winter when electricity prices increase. Then the hydrogen is converted back to electricity with a fuel cell. The second group consists of industries with a existing demand of hydrogen. These industries are evaluated with a system similar to the first, but the hydrogen is used in gasous form in melt-shops and furnaces, not as a source for electricity. The systems are eval- uated on two parameters; the change in CO2emissions and NPV. The system for group one showed a negative NPV for all industries, as well as an increase of CO2 emissions. The system for group two showed both positive NPV and a decrease in CO2 emissions compared to operations as it is done today. As a conclusion of this thesis, industries would benefit from producing their own hydrogen gas at the facilities if they use it as a raw product. If it shall be used as storage for electricity usage, the storage cost today are too high.
Sammanfattning
Världens stålindustrier bidrar till 7% av de totala CO2utsläppen, och efterfrå- gan av stål i världen antas öka. I Sverige undersöker ståltillverkare nya mer miljövänliga alternativ för att producera stål. Ett av dessa alternativ är att i större utsträckning använda använda vätgas i stålproduktione. Syftet för den- na uppsats var att undersöka hur vätgasproduktion från solkraft i närhet av 7 svenska stålindustrier påverkar CO2utsläpp samt driftkostnader. Detta gjordes genom att bygga system bestående av elektrolysörer, PV samt förvaring. Det 7 industrierna blev uppdelade i två grupper. Den första gruppen består av in- dustrier som inte använder sig utav vätgas idag. Dessa industrier undersöktes med ett system där vätgas produceras från solkraft under sommarhalvåret för att lagras till vintern, då priset på elektricitet ökar. Vätgasen blir då omväxlad till elektricitet via en bränslecell. Den andra gruppen består av industrier som redan använder sig utav vätgas. Dessa industrier undersöktes med ett system som liknar det första, men här används vätgasen som råprodukt i ugnar, och omvandlas därav inte till elektricitet. Båda systemen mättes på två paramet- rar, skillnad i CO2 utsläpp och NPV. Alla industrier som utvärderades med första systemet visade en ökning i CO2 utsläpp och negativa NPV. Industri- erna i grupp två som utvärderades med det andra systemet visade båden en förminskning i CO2 utsläpp samt possitiva NPV. Slutsatsen av denna uppsats är att industrier som använder sig utav vätgas som råvara skulle gynnas av att producera vätgasen själva. Om vätgasen ska användas till elektricitet vid ett senare tillfälle är kostnaden för förvaring idag alldeles för hög.
I would like to start of by saying thank you to my supervisor at KTH Nelson Sommerfeldt who has help me in both academical guidance and constructive feedback which increased my motivation.
Thank you Hugo Larsson at Save by Solar AB for your expertise within the economics of solar power in Sweden, and thank you Filip Enblom for your help providing contacts within the Swedish steel industry.
I want to thank Valentin Vogl for his help and our conversations regarding using hydrogen in steel production.
Last but not least I want to thank Krista Winkler and the people closest to me who have been giving me support and increased my motivation when it was needed the most.
v
1 Introduction 1
1.1 Purpose . . . 3
1.2 Deliminations . . . 3
1.3 Methodology . . . 4
2 Technological Review 5 2.1 Steel Production . . . 5
2.1.1 Blast Furnace . . . 8
2.1.2 Direct Reduction using Hydrogen . . . 9
2.1.3 Electric Arc Furnace . . . 11
2.2 Swedens Electricity & Hydrogen Market . . . 13
2.2.1 Electricity Market . . . 13
2.2.2 Hydrogen Market . . . 15
2.3 Solar Photovoltaics . . . 17
2.4 Hydrogen Production . . . 19
2.4.1 Electrolysis . . . 19
2.5 Storage . . . 23
2.5.1 Compressed Hydrogen . . . 23
3 Method 26 3.1 Problem Identification . . . 26
3.2 Facilities . . . 27
3.3 System Presentation . . . 29
3.4 Components . . . 33
3.4.1 PV . . . 33
3.4.2 Electrolyzer . . . 34
3.4.3 Compressor . . . 34
3.4.4 Storage . . . 35
3.4.5 Fuel Cell . . . 37
vi
3.5 Purchasing from the Grid . . . 38
3.6 Economics . . . 40
3.7 CO2 Emissions . . . 42
3.8 Sensitivity Analysis . . . 43
4 Results 44 4.1 System One . . . 44
4.1.1 Sandvik - Surahammar . . . 46
4.1.2 TPC - Hallstahammar . . . 47
4.1.3 Outokumpu - Avesta . . . 48
4.2 System Two . . . 49
4.2.1 Sandvik - Hallstahammar . . . 51
4.2.2 Sandvik - Sandviken . . . 52
4.2.3 SSAB - Luleå . . . 53
4.2.4 SSAB - Oxelösund . . . 54
5 Discussion & Conclusion 55 5.1 Discussion . . . 55
5.2 Conclusion . . . 56
A Something Extra 64 A.1 Sensitivity Analysis . . . 65
A.2 System One . . . 70
A.3 System Two . . . 85
A.4 Sensitivity Analysis . . . 101
Acronyms and Abbrevations
• AEL - Alkaline Electrolyzer
• AFC - Alkaline Fuel Cell
• AsGa - Gallium Arsenide
• BF - Blast Furnace
• BOF - Basic Oxygen Furnace
• CAPEX - Capital Expenditure
• CO - Carbon Monoxid
• COPV - Composite Overwrapped Pressure Vessel
• CO2 - Carbon Dioxide
• DC - Direct Current
• DR - Direct Reduction
• EAF - Electric Arc Furnace
• FC - Fuel Cell
• Ge - Germanium
• Gt - Giga tonne
• H-DR Direct Reduction using Hydrogen
• HHV - Higher Heating Value
• H2 - Hydrogen
• IEA - International Energy Agency
• kg - Kilogramme
• kW - Kilo Watt
• kWh - Kilo Watt Hour
• LCOE - Levelized Cost of Electricity
• LCOH - Levelized Cost of Hydrogen
• LHV - Lower Heating Value
• MPa - Mega Pascal
• Mt - Million tonne
• MW - Mega Watt
• MWh - Mega Watt Hour
• m3- Cubic Meter
• NPV - Net Present Value
• Nm3- Normal Cubic meter
• OPEX - Operational Expenditure
• O2- Oxygen
• O&M - Operation and Maintenance
• PEM - Proton Electrolyte Membrane
• PEMFC - Proton Electrolyte Membrane Fuel Cell
• PV - Photovoltaic
• R&D - Research and Development
• SAM - System Advisor Model
• SEA - Swedish Energy Agency
• SEC - Specific Energy Consumption
• Se - Selenium
• SOEC Solid Oxide Electrolyzer
• TWh - Terra Watt Hour
• WACC - Weighted Average Cost of Capital
2.1 Energy Intensity for primary and secondary steel production (Seetharaman 2014a) . . . . 5 2.2 Energy usage of Swedish iron and steel industries (Jernkon-
toret 2020) . . . 6 2.3 Energy usage of Swedish iron and steel industries (Jernkon-
toret 2020) . . . 7 2.4 Blast Furnace & Basic Oxygen Furnace process compared to
Direct Reduction process (Karakaya et al. 2018) . . . 7 2.5 Components of a standard Blast Furnace (Seetharaman 2014a) 8 2.6 Schematics of a MIDREX process (Seetharaman 2014b) . . . 9 2.7 Schematics of H-DR EAF system with integrated Electrolyzer
(Vogl et al. 2018) . . . 10 2.8 Improvements of EAF since 1960 (Seetharaman 2014c) . . . . 11 2.9 Standard EAF design: 1) Transformer, 2) Cable connections,
3) Electrode arms, 4) Electrode clampings, 5) Arms, 6) Cooled off-gas duct, 7) Cooled panels, 8) Structure, 9) Basculating structure, 10) Rack, 11) Cooled roof, 12) Basculating device, 13) Hydraulic group. (Seetharaman 2014c) . . . 12 2.10 The yearly average of SPOT prices in sweden since 1996, with
the linearly trend. Data acquired from (SCB 2020b) . . . 14 2.11 Estimated renewable capacity growth between 2019 and 2024
by technology (IEA 2019b) . . . 18 2.12 Capacity of Swedish PV by size (IEA 2018) . . . 18 2.13 Monopolar and Bipolar electrolyzers (Krishnan et al. 2020) . . 20 2.14 Showing the different types of COPV (Barthelemy et al. 2017) 24 3.1 Layout of System One . . . 30 3.2 Layout of System Two . . . 32
x
3.3 Figure of how one year of charge and discharge from storage was evaluated . . . 36 3.4 Orange: Vattenfalls Distribution areas (Vattenfall 2020b), Blue
dots indicate the locations of facilities . . . 38 4.1 How the costs affects the NPV of the system for Sandvik in
Surahammar . . . 45 4.2 How WACC and Grid cost affects the NPV of the system for
Sandvik in Surahammar . . . 45 4.3 How the price of components affects the NPV of the system
at the Sandvik facility in Sandviken . . . 50 4.4 How WACC and Grid cost affects the NPV of the system at
the Sandvik facility in Sandviken . . . 50 A.1 How the costs affects the NPV of the system at TPC in Hall-
stahammar . . . 65 A.2 How WACC and Grid cost affects the NPV of the system at
TPC in Hallstahammar . . . 65 A.3 How the price of components affects the NPV of the system
for Outokumpu in Avesta . . . 66 A.4 How WACC and Grid cost affects the NPV of the system for
Outokumpu in Avesta . . . 66 A.5 How the price of components affects the NPV of the system
at the Sanvik facility in Hallstahammar . . . 67 A.6 How WACC and Grid cost affects the NPV of the system at
the Sanvik facility in Hallstahammar . . . 67 A.7 How the price of components affects the NPV of the system
at the SSAB facility in Luleå . . . 68 A.8 How WACC and Grid cost affects the NPV of the system at
the SSAB facility in Luleå . . . 68 A.9 How the price of components affects the NPV of the system . 69 A.10 How WACC and Grid cost affects the NPV of the system at
the SSAB facility in Oxelösund . . . 69
2.1 Example subscriptions from Vattenfalls northen distribution
area (Vattenfall 2020b), *Nov-Mar Mon-Fri 06:00-22.00 . . . 13
2.2 Capital cost in millions of SEK for electrolyzers depending on size (Parra & Patel 2016) . . . 21
3.1 Yearly demands acquired from the different facilities. *With one out of two BF running, **No available data . . . . 27
3.2 Calculated yearly demands for the different facilities. . . 28
3.3 Technical data is aqcuired from SAM software, economic data was supplied by Save by Solar, and confirmed by ETIP-PV (2017) Walker (2017). . . 33
3.4 Technical data is aqcuired from Buttler & Spliethoff (2018), Mayyas & Mann (2019) and Parra & Patel (2016) . . . 34
3.5 Data required for sizing and cost calculations (Gökçek & Kale 2018) . . . 35
3.6 Data required for sizing and cost calculations (Preuster et al. 2017), (Bangoura 2020) . . . 35
3.7 Technical data is aqcuired from Ruth et al. (2019), Mayyas & Mann (2019), . . . 37
3.8 The three evaluated subscriptions with regard to distribution area (Vattenfall 2020a), *Nov-Mar Mon-Fri 06:00-22.00 . . . 39
3.9 Variables used in the sensitivity analysis. * Initial value of cost is calculated depending on the size of the electrolyzer . . 43
4.1 System design for Sandvik in Surahammar . . . 46
4.2 System design for TPC in Hallstahammar . . . 47
4.3 System design for Outokumpu in Avesta . . . 48
4.4 System design for Sandvik in Hallstahammar . . . 51
4.5 System design for Sandvik in Sandviken . . . 52
4.6 System design for SSAB in Luleå . . . 53
xii
4.7 System design for SSAB in Oxelösund . . . 54
The Swedish Climate Policy Council has set a target of achieving zero net greenhouse gas emissions by 2045. The portion of renewable energy has reached almost 50%, and Sweden is improving rapidly in the transport sec- tor (SEA 2019b). However, to achieve net zero, they will have to restructure their industrial sector. According to IEA (2019c) the iron and steel industries account for 7% of the global CO2emissions (2.24 Gt) , which is mainly related to the use of coal and coke in blast furnaces (Sundqvist Ökvist et al. 2017).
Sweden’s steel sector is one of the most energy efficient in the world, but even they produce a lot of emissions due to the use of coal.The Swedish steel in- dustries created 5.8 Mt CO2 emissions in 2016, 85% of these emissions are from the use of coal in the blast furnace’s that produces molten iron (Axelsson 2019). In the future there must be a shift towards decarbonizing technologies globally, and for Sweden to reach their goals.
Three companies SSAB, LKAB and Vattenfall have created HYBRIT. The HYBRIT initiative seeks to find a fossil free way to produce steel, by using green electricity and hydrogen instead of coke and coal. Recently these three companies together with the Swedish Energy Agency have invested 200 mil- lion SEK into a large scale hydrogen storage in Luleå (HYBRIT 2019).
Currently 95% of the Hydrogen produced globally is from fossil fuels, 48%
is made by steam methane reforming that produces high emissions (Ngoh &
Njomo 2012). Therefore, hydrogen production facilities should evolve from using fossil fuels and shift towards the technologies available using renewable sources.
In Sweden, the steel industries consumes around 4 TWh of electricity yearly (Jernkontoret 2019). If SSAB’s facility in Luleå would exchange their blast furnaces for direct reduction with hydrogen, and produce their own hydro- gen through electrolysis, their electricity consumption would reach roughly 15 TWh (Enocsson 2020). The need for electricity from renewable sources
1
will increase drastically as the steel industry shifts towards a CO2neutral pro- duction.
When it comes to industrial applications of solar in Sweden, there are some papers researching the potential of integrating solar-thermal and PV into Swe- den’s heating and energy network, among these are Joly et al. (2017) who investigates the possibility for solar thermal integration to the district heating networks (Joly et al. 2017). Siyal et al. (2015) researched the Swedish poten- tial of stand alone wind-powered hydrogen production for refueling stations at three different locations. Their conclusion was that the systems modelled for each location would provide enough hydrogen to meet the hydrogen load at a competitive price (Siyal et al. 2015). Lindblad (2019) performed a eco- nomic feasibility study of electrolysis connected to an off-shore wind farm in Oxelösund. Some of the systems evaluated in this study was created to match a potential hydrogen load from the steel manufacturer SSAB located in Ox- elösund. Vogl et al. (2018) have done research regarding direct reduction us- ing hydrogen for steel production, in collaboration with HYBRIT (Vogl et al.
2018). Nilsson (2015) has built a demo site of an off grid house, using solar power and hydrogen as sources for energy. This is however on a quite small scale (Nilsson 2015).
Wind and Hydro power are the dominant resources of renewable energy in Sweden (SEA 2019a), and solar power is often overlooked due to the absence of the sun during Nordic winter periods. At the moment, solar power has a very small role in Swedens renewable fraction with 0.2 TWh produced in 2017 (Electricity generation for private use is not included). From 2017 to 2018 the amount of PV connected to the grid increased with 67% (SEA 2019a).
This paper will conduct a research of the potential of solar powered hydro- gen production systems for direct or indirect use in different steel production facilities.
1.1 Purpose
The main purpose of this thesis is to construct systems which utilizes solar power and hydrogen production with the current cost of components, then evaluate the effects they have on operating the facilities compared to the present systems. The two main parameters evaluated are CO2 reduction, and the de- crease or increase of cost. With the results of this thesis, one can get a clearer picture of how hydrogen and solar can be utilized in steel industries today and in the future.
1.2 Deliminations
In this study some boundary conditions are taken from seven steel manufactur- ing sites around Sweden. Therefore, the thesis contains seven case studies with regards to technical demands. However, it does not contain economic data for each specific facility. The cost of electricity is calculated from the different subscriptions offered by Vattenfall for industries, and therefore the exact cost is not acquired from the industries themself. Collecting price information of electrolyzers and hydrogen storage from companies was proven difficult since the ranges of sizes was so great. The price data was instead collected from various articles researching the different technologies.
1.3 Methodology
Creating this techno-economic analysis for solar-hydrogen system to support steel industries will consist of these four main steps.
• Collecting relevant data regarding the different industries that will be studied. Having interviews with employees discussing their process lay- out.
• A technological and economical study regarding hydrogen. This will include:
– How hydrogen can be produced using solar power as the energy source.
– Different storage technologies for hydrogen and oxygen – How hydrogen and oxygen is utilized in the steel industry
• Identify what causes highest cost of operation regarding power and hy- drogen for these industries. To do this, understanding the Swedish en- ergy market and laws with regards to larger industries is very important.
Gaining knowledge of how the price of electricity changes through out the year is also relevant.
• Modelling different systems which utilizes technologies for hydrogen production and power generation with goals to lower the cost of hydro- gen and/or electricity for steel manufacturing facilities. With these mod- els, calculations of the Levelized Cost of Electricity (LCOE), Levelized Cost of Hydrogen (LCOH) and Net Present Value (NPV) for each system can be done. Lastly, with a sensitivity analysis one can identify which factors affects the cost more than others, one can also see tipping points of when a system is feasible or not.
The work process will begin with a technological study together with inter- views to gather the relevant data, then the modelling will be conducted in SAM and Matlab.
The purpose of this technological review was to get an overview of the current situation in the relevant technologies. The theoretical information for each subcategory is presented.
2.1 Steel Production
The process to produce steel can vary, depending on what type of input ma- terial you use and what type of steel you want to produce. There are three steps in the primary steel production process, which are; preparation of the raw material, creating iron and then producing steel. Depending on how the iron is produced the raw material needs to be prepared differently. The most common technology for iron making is the Blast Furnace (BF).
Figure 2.1: Energy Intensity for primary and secondary steel production (Seetharaman 2014a)
5
To produce steel in a Basic Oxygen Furnace (BOF) with iron made in a BF, the energy required is 19.8-31.2 GJ per tonne of crude steel (Seetharaman 2014a).
If compared to secondary steel production where only recycled steel scraps are used, it requires 2.3 times more energy. In figure 2.1 the different production lines with connected energy intensity is presented.
In Sweden there are 10 secondary steel production facilities using scraps as their main resource and two primary steel production facilities using raw iron as their main resource (Jernkontoret 2019). Even though there are only two facilities having blast furnaces, two thirds of Sweden’s total steel production is from the iron they produce. The total production of crude steel in Sweden in 2018 was estimated at 4,7 million tonne (Jernkontoret 2020).
The Swedish steel industries are specialized in different areas of steel pro- duction, and they have a reputation of producing the best quality possible.
Figure 2.2: Energy usage of Swedish iron and steel industries (Jernkontoret 2020)
Most of the steel produced in Sweden is being ex- ported, mainly to coun- tries within the EU. Ger- many is the largest im- porter of Swedish steel, a large amount is also ex- ported to the USA. In recent years China has become a vital costumer of Swedish steel. Since China is one of the largest producers of steel in the world, this is a good in- dicator of Sweden’s high quality steel.
In 2018 the export of Swedish steel amounted for 53 billion SEK, and the im- ports of steel was 43 billion SEK. However, Sweden imports about the same amount of steel as they export measured in weight (Jernkontoret 2019). The 10 billion difference in exports and imports even though the amount of steel is the same is because Sweden exports high quality steel and import standard quality.
The energy usage changes from year to year as it depends on the demand of steel. The main energy carries used in Sweden are coke, oil, electricity and natural gases. The usage of these carriers in the steel industry from 1970 until 2017 is shown in figure 2.3.
Figure 2.3: Energy usage of Swedish iron and steel industries (Jernkontoret 2020)
Due to the two BF facilities, coke is accounting for the major part of energy carriers used in Sweden. SSAB, the company owning the two BF facilities, are currently investigating on how they can change their production to ex- clude coke. The most promising technology is direct reduction with hydrogen gas. The technologies used in Sweden today are BF and Electric Arc Furnace (EAF), but the shift towards fossil free steel production will exchange the BF with Direct Reduction using Hydrogen gas (H-DR). Therefore the process coal and coke will be exchanged for hydrogen gas and electricity. Figure 2.4 shows both the BF-BOF and DR-EAF route.
Figure 2.4: Blast Furnace & Basic Oxygen Furnace process compared to Di- rect Reduction process (Karakaya et al. 2018)
2.1.1 Blast Furnace
Blast furnaces have been a central part of iron making for a very long time.
The early furnaces was about 7 meters high and built out of stone blocks and timber. Today the heights of a normal furnace can reach around 19 meters, measured up to the top of the throat of the furnace. This technology have been refined through out the years and become a much cleaner process that it used to be. Such improvements are; the change from charcoal to coke as the furnace fuel and preheating of the blast. Both these changes improved the efficiency of the BF’s significantly and also reduced the emissions (Seetharaman 2014a).
The raw material used in BF’s are lump ore, pellets or sinter, these forms of iron are put together with coke inside the furnace in alternating layers. These furnaces are heated up to temperatures around 1,500°C to create molten iron (Seetharaman 2014a). The main product from BF’s is the molten iron also called hot metal, which is used for BOF steel making, the main bi-product is called molten slag. The slag will form on top of the hot metal because of the lower density (Seetharaman 2014a). Both the slag and hot metal are being tapped regularly. Once the slag is solidified, it is sold and can be used for road aggregate and cement production (Peacey & Davenport 2016). Figure 2.5 shows how the standard BF is composed.
Figure 2.5: Components of a standard Blast Furnace (Seetharaman 2014a) The steel industry accounts for 5% of the global energy consumption and about 6% of the anthropogenic CO2emissions (Zhao et al. 2020). This is mainly due to the most common steel making route of BF-BOF. According to Zhao et al.
(2018) this route emits 2084.37 kg of CO2per tonne of carbon steel produced.
The process that creates the most emission is the iron making with BF, which accounts for 79% of the emissions (Zhao et al. 2018).
2.1.2 Direct Reduction using Hydrogen
As the BF, direct reduction is a method for processing iron ore’s into man- ageable products for the steel production. The last 100 years BF has been the dominant technology for ironmaking (Seetharaman 2014b). The beginning of DR started in Mexico in the 50’s. Here the HYL process was developed. In the 70’s the MIDREX process was invented and now there are hundreds of DR plants around the world, mostly using natural gas as the reducing agent.
The total production of iron from DR is around 70 million tonnes per year (Seetharaman 2014b).
In this process the reduction of the iron occurs in a solid state, compared to in a BF where the iron is melted. As an effect of this, the energy demand is lower than for a BF process. However, it requires that the iron pellets is of higher quality. It is very difficult to remove impurities from the solid iron, which could result in steel makers receiving Direct Reduced Iron (DRI) with impurities.
Figure 2.6: Schematics of a MIDREX process (Seetharaman 2014b)
Therefore DR processes requires DR-grade pellets to be able to provide good quality DRI (Seetharaman 2014b). The final prod- uct is DRI or sponge iron, which is solid compared to the liquid hot metal pro- duced from BF. This prod- uct is unstable, and not so dense. If the product requires transportation, it can be processed further into Hot Briquette Iron (HBI), which is more com- pact and stable.
The DR process uses reducing agent in gaseous state. Most commonly used is natural gas. The natural gas must be prepared in a reformer first, be- for being able to act as a reducing agent. It has to be reformed into Hydrogen (H2) and Carbon Monoxide (CO). To be able to reform the natural gas usually steam and/or carbon dioxide (CO2) is used. A simplified schematic picture of the standard MIDREX process is shown in figure 2.6.
Instead of reforming natural gas, there is the possibility to use pure hydrogen.
According to Vogl et al. (2018) there is one plant in Trinidad producing DRI using fluidised bed reactors and hydrogen from steam reforming. In Vogl et al.
(2018) research, they constructed a H-DR and EAF system for steel produc- tion. According to their calculation, if the system would operate continuously on 100% HBI, the production of one tonne steel would require 1504 kg of pel- lets and 51 kg of hydrogen (without regards to hydrogen losses). If the EAF would be loaded with equal parts pellets and scraps instead, one tonne steel would demand 738 kg pellets, 536 kg scraps and 25 kg of hydrogen (Vogl et al.
2018). A schematics of Vogl et al. (2018) system is shown in figure 2.7. Since the process is entirely electrified, the emissions will mainly depend on the mix of the power grid. They conclude that a switch from BF-BOF to this type of system would reduce the CO2 emissions in most of Europe.
Figure 2.7: Schematics of H-DR EAF system with integrated Electrolyzer (Vogl et al. 2018)
2.1.3 Electric Arc Furnace
The history of EAF is quite short with only 100 years from the first experi- ments using electricity to melt iron to today. The capital cost of small scale EAF steel mills were relatively low after world war II, therefore it became a common choice for new steel mills in post war Europe (Hani 2019). The im- provements of the EAF have been rapidly developed. In figure 2.8 it shows how the electric consumption has decreased from 630 to 345 kWh/t. The tap- to-tap time refers to the time between taping out the hot metal.
Figure 2.8: Improvements of EAF since 1960 (Seetharaman 2014c) The main components in a EAF is the furnace and electrodes. The elec- trodes are most commonly made out of graphite. These electrodes will be consumed during operation, this consumption has decreased from 6.5 to 1.1 kg of graphite consumed per tonne of steel produces(Seetharaman 2014c). The material that shall be melted is put inside the furnace, then the electrodes are introduced. When connected to a power source an arc of electricity will be created between the electrodes, generating heat and melting the material. A general design of a EAF is shown in figure 2.9
For secondary steel production, the input for EAF is mostly scrap metal, which can be categorized into three different types: Obsolete scraps (appliances, ma- chinery, old cars), Industrial scrap and Internal scraps (quality rejections, steel recovery from slag). These scraps must be prepared before introduced into the EAF, and the most dominant technology is shredding (Seetharaman 2014c).
Another input for EAF is DRI, also called sponge iron.
Melting DRI in an EAF is more energy demanding than melting scraps. How- ever, EAF using DRI requires 3.48 MWh per tonne liquid steel, while a BF- BOF would consume 3.68 MWh/t (Vogl et al. 2018). This comparison is more interesting since simply using EAF for scrap does not compete with primary steel production such as BF-BOF. Even though BOF’s can be replaced with this technology, oxygen is still very useful in a system with EAF, since it can be added to heat up "cold spots" in the EAF, creating a more uniform heating of the scraps. This will decrease electricity consumption of the EAF (Pardo
& Moya 2013).
An EAF primary use of energy is electricity, therefore the emissions depends greatly on how the electricity is produced. India is one of the countries with most DR-EAF steel plants, this is because of their cheap electricity and abun- dance of natural gas (Seetharaman 2014c). Since their energy mix is about 50% coal and 20% renewables (IEA 2019a), the emissions of an EAF in India is remarkably higher than one in Sweden.
Figure 2.9: Standard EAF design: 1) Transformer, 2) Cable connections, 3) Electrode arms, 4) Electrode clampings, 5) Arms, 6) Cooled off-gas duct, 7) Cooled panels, 8) Structure, 9) Basculating structure, 10) Rack, 11) Cooled roof, 12) Basculating device, 13) Hydraulic group. (Seetharaman 2014c)
2.2 Swedens Electricity & Hydrogen Market
To be able to estimate the electicity cost for different steel industries and un- derstanding the laws and restrictions of producing power from solar energy, a deeper knowledge of the electricity market in Sweden was required.
2.2.1 Electricity Market
The total cost of your energy use can be divided into different segments. One part goes to the network operator for the Operation and Maintenance (O&M) of the power grid. This is made up of a fixed charge and a flexible charge SEK/kW. Furthermore you will also pay the supplier of the energy, this can be done either with fixed or flexible cost depending on the subscription you have chosen. Different subscriptions are shown in table 2.1. In addition to these cost, you will have to pay tax on you energy use as well. Normally the tax contributes to about 40% of the total cost of your energy (Vattenfall 2020b).
However, there are laws regarding larger industries that lowers this tax. Ac- cording to Vattenfall (2020b) steel and iron producing companies can get a tax refund, resulting in a total tax cost of 0.005 SEK/kW. The cost of electricity varies each hour depending on demand and supply, in addition to these cost there are fees depending on when you use the electricity as shown in table 2.1.
High Voltage Low Voltage
Subscription N2 N2T N3 N3T N4
Fixed Cost
SEK/month 220 400 23 500 2 500 3 300 385 Monthly Peak
SEK/kW, month 10 28 28 28 48
High Demand*
SEK/kW, month 20 56 71 90 0
High Demand Transfer*
ÖRE/kWh 2.8 5.7 18.0 20.7 52.0
Regular Transfer
ÖRE/kWh 1.7 3.4 5.6 7.8 14.4
Table 2.1: Example subscriptions from Vattenfalls northen distribution area (Vattenfall 2020b), *Nov-Mar Mon-Fri 06:00-22.00
There are two monthly tariffs, the first one is a cost for each months maxi-
mum hourly average consumption, also called monthly peak cost. The second one is the same type of cost but applies only for the high demand months, which are November to Mars. During these months the hourly price is in- creased Monday-Friday between 06.00 and 22.00.
In 2003 Sweden introduced power certificates of renewable energy to increase Sweden’s share of said energy. As a power producer you receive a certificate for each MWh of renewable energy produced. These certificates can then be sold on an open market, where the price is decided between the buyer and the seller (Energimyndigheterna 2020). As a producer, there is a ratio of 30.5%
of renewables needed. If you do not produce enough renewable energy, you can buy certificates from a producer that has more than they need. When this policy was implemented, the prices for certificates were quite high. However, as more producers use more renewable sources, the prices of the certificates has dropped from around 180 to 19 SEK in the last 5 years (SKM 2020). This shows that the incentive has worked properly, since producers do not see value in buying others certificate anymore.
When receiving the certificate, one also receives a "Ursprungsgaranti" per pro- duced MWh. This gives you the right to state to the public that you are pro- ducing renewable energy. This can also be sold, and then the buyer owns the right to state that they produce renewable energy.
Figure 2.10: The yearly average of SPOT prices in sweden since 1996, with the linearly trend. Data acquired from (SCB 2020b)
It is very difficult to predict the future prices for the electricity, basically be- cause there are to many depending variables. In figure 2.10 the average yearly SPOT prices from 1996 are shown. From figure 2.10, two conclusions can be made; firstly the price varies a lot from year to year. So predicting the cost
within just a few years in almost impossible. Secondly, the trend shows that overall the prices are increasing. However, to predict the future of electricity prices, the past is not always the best source as there are so many factors play- ing vital roles. Information that could contradict this increase of cost is that Sweden has as a goal to increase both solar and wind farms (IEA 2018). Since the LCOE from these sources of electricity is steadily decreasing, this could have a lowering effect of the future SPOT price. According to Porsö (2018) Swedens electricity grid produced 62.9 g-CO2/kWh in 2016.
2.2.2 Hydrogen Market
The European Commissioner for Climate Action and Energy stated March 2015: "Energy markets and grids have to be fit for renewables, not vice versa.”
(Nastasi 2019). This statement implies that there has to come improvements in how to utilize the renewable energy. The technologies for producing re- newable energies exist, but the existing infrastructure is built during the era of fossil fuels. Hydrogen can play a vital role in the future civilisation. As this energy carrier is included in more decarbonization strategics and is gain- ing more public awareness, it attracts investments and research from industries (Nastasi 2019)
During standard atmospheric conditions hydrogen exists in gaseous form. It has a density of 0.0899 kg/m3 at 25°C and 1 atm (Normal condition). This means that one kg of hydrogen would take up 11.12 m3 during most naturally occurring conditions. Compared to petrol, which density is 748.9 kg/m3, it takes up roughly 8000 times more space. However, one kg of hydrogen con- tains almost 4 times the energy of petrol, 9.1 and 33.33 kW h/kg respectively.
To be able to compete with petrol as a energy carrier, the cost of storing hy- drogen at a reasonable volume must be feasible.
Hydrogen has gained more attention as renewable energy power plants are in- creasing. This is partly because hydrogen has the properties suitable for long time storage. The dominant technology today for storing excess electricity from solar and wind are batteries. They have a great efficiency if the elec- tricity will be used within a short period of time. Storing electricity through- out seasons in a battery will generate great losses (Zhang et al. 2016). If the electricity is instead converted into hydrogen it can be stored from summer to
winter without any significant losses. However, there will be losses while con- verting electricity to hydrogen and again when converting it back to electricity.
Due to these losses, this type of storage is only suitable for long term storage, when the losses of other types of storage are greater. Therefore, these two stor- age options for intermittent power sources does not really compete with each other as of now (Robles et al. 2018).
The industry sector has 90% of the hydrogen market share and is the also the largest hydrogen producer in the EU (Fraile et al. 2015). The steel and iron industry is the third largest consumer of hydrogen with 0.41 Mt per year.
Ammonia production alone stands for 50% of the industry share with a con- sumption of 3.6 Mton of hydrogen. Most of the Chemical sector uses large on-site hydrogen production or pipelines for transportation. The steel and iron industry are mostly using cylinders or tube trailers as means of transporting the hydrogen to them. The price of hydrogen is very different, one parameter affecting the cost is how far it needs to be transported. According to Fraile et al. (2015) prices are estimated to be between 100-600 SEK/kg. If the steel and iron industry would increase their demand of hydrogen, it would be more economical to produce their own or install a pipeline to a hydrogen production facility, such as it is done by the chemical sector.
2.3 Solar Photovoltaics
The basic principle behind a Photovoltaic (PV) panel is based on the photo- voltaic effect, which is the conversion of light energy from the sun into elec- tricity. A panel is constructed of multiple solar cells, constituted by semicon- ductor materials, most common is silicon. There are other material that are better suited for solar cells, such as Gallium Arsenide (AsGa), Selenium (Se) and Germanium (Ge) to name a few. However, silicone is one of the cheapest option and also the most commonly used. Within these materials a artificial electric field has been constructed (pn junction). The electric field causes the negatively charged particles to move in one direction and the positively charge particles to move in the opposite direction. Light, which is composed of pho- tons, will transfer energy to an atom of the semiconducting material. More specifically it transfer energy to a electron in the p-n junction, which will be excited to a higher energy state, called conduction band. When a electron makes this "jump" it leaves behind a so called "hole" in the valence band. The electron will now travel through a connected load and back to the hole, thus generating electricity (Bayod-Rújula 2019).
There are multiple types of solar cells, such as, Monocrystalline, Polycrys- talline and Thin-film, to name the most common ones. How these types of cell differ lies within the technology of the material. A Monocrystalline cell is made out of a single silicone crystal. The process of manufacturing Monocrys- talline cells is expensive and have a large amount of waste material (40-50%) (Bayod-Rújula 2019). Polycrystalline cells is made up out of multiple silicone crystals. These cells are cheaper to produce with the cost of lower efficiency than a Monocrystalline cell. Thin-film is a technology which aims to reduce the cost even further. Thin-film cells are constructed as Polycrystalline cells, however much thinner, between tenths and several microns (Bayod-Rújula 2019). By reducing the thickness of the cell, the amount of material used is decreased and with it the cost. This technology also opens up the possibility for a bendable panel. Save by Solar almost exclusively uses Monocrystalline technology for their projects. Therefore it will be used in this project as well.
The Monocrystalline panels on the market can have an efficiency between 12- 20 % (Peng et al. 2017). The total cost of installing a ground mounted PV project is between 6 000 to 10 000 SEK/kW (IEA 2018). This cost is esti- mated to decline, and in 2050 the price could be between 1615-4708 SEK/kW (IRENA 2019)
According to IEA (2019b) renewable power capacity is set to expand 50%
from 2019 to 2024 globally. Solar PV alone will stand for 60% of this capac- ity increase as seen in figure 2.11. In 2018 the installed capacity was 512,3 GW globally (IEA 2019d). In Sweden the capacity of grid connected PV was 411 MW in 2018 and it is growing steadily. In 2018 156.36 MW was installed, that is an increase of 89% compared to the amount installed in 2017 (IEA 2018).
As seen in figure 2.12 the PV plants of larger scale are both few nor does it contribute with a large capacity. The main part of Swedish grid-connected PV comes from the commercial sector (47%) and the private sector (33%) (IEA 2018). Most of these systems are rooftop installed.
Figure 2.11: Estimated renewable capacity growth between 2019 and 2024 by technology (IEA 2019b)
Figure 2.12: Capacity of Swedish PV by size (IEA 2018)
2.4 Hydrogen Production
There are several methods to produce hydrogen gas. As mentioned in the in- troduction, the most common method is steam reforming using methane gas.
Most of the technologies used today cannot be categorized as renewable meth- ods of hydrogen production. The energy source this paper focuses on is solar radiation, therefore the focus will be on technologies using this source of en- ergy. Recently there has been multiple research conducted on biological pro- cesses of producing hydrogen, such as Nikolaidis & Poullikkas (2017). The main focus are mostly on direct and indirect bio-photolysis. Nikolaidis & Poul- likkas (2017) are estimating the LCOH for direct bio-photolysis to be 21.42 SEK/kg and for indirect 14.28 SEK/kg. Since these processes are not commer- cially available they will not be included in the system modelling. Hydrogen production through water electrolysis is the most advanced and available re- newable production technology today (Krishnan et al. 2020). The two most common electrolyzers are Alkaline and Proton Electrolyte Membrane (PEM).
Both of these technologies will be investigated to understand which applica- tions they are most suitable for.
2.4.1 Electrolysis
Water electrolysis, also called water splitting, is a method to produce hydro- gen and oxygen in gaseous form from water with the use of electricity. This method has been around since the 1800 and the first large scale water electrol- ysis plant producing 10,000 Nm3of H2 per hour was built in 1939.
The most common medium for electrolyzers is Alkaline, even though the dis- covery of electrolytical water splitting was through experiments with acidic water. The reason for this were because the materials became cheaper for an Alkaline electrolyzer (AEL), and also the corrosion can more easily be dealt with(Kreuter & Hofmann 1998).
The basic electrolyzer system consist of an cathode, a anode, an electrolyte and a power supply. By introducing a Direct Current (DC) to the system, elec- trons will travel from the anode to the cathode creating oxygen and hydrogen respectively at the electrodes. The reactions occurring are shown in equations 2.1.
Anode : 2OH− → 1
2O2+ H2O + 2e Cathode : 2H++ 2e → H2
Overall : H2O → 1
2O2+ H2
(2.1)
During the 20th century there were two main constructions, monopolar and bipolar electrolyzers. In the monopolar construct, each cell has an indi- vidual tank of electrolyte, and are then connected to each other externally. The Bipolar electrolyzers have their cells assembled back to back, with each side of the cell being a anode on one side and a cathode on the other, thus being bipolar. An illustration of monopolar and bipolar electrolyzers are shown in figure 2.13. Today the bipolar construct is the most available, partly due to size reduction (Krishnan et al. 2020).
Figure 2.13: Monopolar and Bipolar electrolyzers (Krishnan et al. 2020) According to Mayyas et al. (2019) the three most advanced technologies for water electrolysis are Alkaline, PEM and Solid Oxide (SOEC) electrolyzers.
AEL is an commercial technology at large scale. PEM electrolyzers are being used but can still be improved through additional R&D, therefore in a early commercial stage (Mayyas et al. 2019). SOEC shows great promise of good efficiencies, but need more work before reaching a commercial stage and will not be investigated further (Krishnan et al. 2020). Both Alkaline and PEM electrolyzers follows the same principle as described above, with the differ- ence of the PEM using a solid membrane as the electrolyte. PEM is preferred for vehicle uses and also for intermittent power supplies, since the variable load spectra and the start up time is suitable for these purposes (Krishnan et al.
2020).
Alkaline & PEM Electrolyzer cost estimations
The data availability of investment and O&M cost for electrolyzers in general are very limited. This is mostly because the suppliers will not reveal their prices due to rivalry. Therefore the cost presented in this sections are estima- tion, and the real price can vary.
Alkaline is the most mature technology when it comes to electrolyzers. The investment cost of AEL lies within the range of 4700 - 12 800 SEK/kW (Kr- ishnan et al. 2020). PEM electrolyzers generally have a higher costs than Al- kaline and they are still being produced in small quantities. Companies such as NEL and Giner are starting to produce more PEM electrolyzers because of the recent decline in manufacturing cost and because PEM electrolyzers shows promise of a higher efficiency than Alkaline (Mayyas et al. 2019).
According to Mayyas et al. (2019) manufactures can lower the production cost by increasing amount of electrolyzers produced from 10 to 1000 by 58%. The CAPEX cost of a 1MW PEM electrolyzer would approximatly be 8.46 mil- lion SEK, which is 8 460 SEK/kW (Parra & Patel 2016). Efficiencies for this type of electrolyzer lies within the range of 69-82% according to (Saba et al.
2018) and they estimate that by 2030 CAPEX cost can land at 3675 SEK/kW with efficiencies up to 84%. In tabel 2.2 capital expenditures (CAPEX) for electrolyzers are presented.
Electrolyzer 25 kW 100 kW 1 MW 10 MW 100 MW 1 GW
Alkaline 0.50 1.33 6.69 41.43 285 4240
PEM 0.64 1.69 8.46 59.64 464 3993
Table 2.2: Capital cost in millions of SEK for electrolyzers depending on size (Parra & Patel 2016)
Fuel Cells
Even though fuel cells does not produce hydrogen, the technology is very sim- ilar to that of a electrolyzer. The purpose of a fuel cell is to produce electricity from hydrogen, simply put, a reverse electrolyzer. The components of a PEM fuel cell (PEMFC) and a PEM electrolyzer are mostly the same, therefore the their production cost are correlating (Mayyas & Mann 2019). The efficien- cies of a PEMFC is around 40-50 % (Taner 2018). While Alkaline Fuel Cells
(AFC) can have an efficiency up to 70%, they are not as relevant for intermit- tent use (Salameh 2014). According to (Abdon et al. 2017) the CAPEX of an Alkaline and PEMFC are 4600 and 9750 SEK/kW respectively. The lifetime of fuel cells differ immensely depending on sources. According to de Bruijn (2011) a stationary PEMFC can reach 40 000 operating hours, while Abderez- zak (2018) states that PEMFC have a lifetime of 4 000 to 8 000 operating hours.
A company called Nedstack offers PEMFC for stationary purposes, and they have had a PEMFC operating in Delfzijl, the Netherlands, which has reached above 65 000 operating hours with the original stacks. Unfortunately no data regarding price could be gathered from Nedstack. Two AFC companies, Gen- Cell and AFCEnergy both states that their AFCs have extreme longevity, Gen- Cell even states that their AFC have an "Unlimited Runtime".
2.5 Storage
An investigation in current technologies regarding hydrogen and oxygen stor- age was necessary to see if any of the options could be economically feasible to implement in the system. The industries analysed in this study have more or less constant demand of hydrogen. Without storage, the systems would need an hourly production of hydrogen equal to the hourly demand. Since the hydro- gen production will be powered by solar PV, an intermittent source of power, storage could help to utilize the abundance of power during the summer time in Sweden.
Studies shows that the most relevant storage technologies for hydrogen are liquefied storage, metal-hydride storage, clathrate-hydrate storage and com- pressed hydrogen storage (Low-pressure, High-pressure, Underground) (Ozaki et al. 2014). By liquefying hydrogen, the density is increased from 0.089 to 70.99 kg/m3, which results in lower storage volumes. This technique is very energy intensive since the gas must be cooled down to temperatures of -253°C.
It is also problematic to store liquid hydrogen for longer periods of time due to product loss by evaporation (Barthelemy et al. 2017). The most common uti- lization of this technique is hydrogen transportation via trucks that can exceed a capacity of 60 000 L. From a stationary aspect, there are more economical options than liquefied storage.
Metal-hydride is a promising solid state storage, and could be part of hydro- gen heavy duty utility vehicles (Lototskyy et al. 2020). Some Metal-hydride storage options exist on the market, suitable for storing less than 200 grams of hydrogen (Bangoura 2020).
Clathrate’s can be describes as cage-like lattices of water molecules. In these lattices, hydrogen can be encapsulated and stored (Schüth 2005). Both Metal- hydride and clathrate-hydrate storage needs more development before they will reach a commercial level (Barthelemy et al. 2017). In this thesis, different types of compressed hydrogen will be further investigated due to their avail- ability in commercial use.
2.5.1 Compressed Hydrogen
Storing Hydrogen gas in compressed vessels is the oldest form of storing hy- drogen, dating back to the 1880’s (Barthelemy et al. 2017). There are 4 types of Composite overwrapped pressure vessels (COPV), as shown in figure 2.14.
Figure 2.14: Showing the different types of COPV (Barthelemy et al. 2017) The different types of vessels are suitable for different uses due to manufac- turing cost and pressure. Type I and type II vessels are cheapest to produce (Barthelemy et al. 2017). Pressures above 50 Mpa are not suitable for type I vessels. For stationary uses this type of storage could be suitable depending on the amount of land area available. The most common hydrogen vessel is of 50 L and can withstand pressures of 30 Mpa, which is the standard for the last 15 years according to Preuster et al. (2017). These vessels are often of type IV.
Oxygen, which has a higher density than hydrogen in gaseous form 1.42 kg/m3, will demand less space when compressed to the same pressure. The standard cylinders for oxygen storage can withstand pressures of up to 30 MPa (Linde 2020). The price of a 50 L 30 MPa cylinder ranges from 1000 - 2000 SEK (Preuster et al. 2017).
Low-Pressure
Hydrogen stored at low pressures will demand larger volume. If there is suf- ficient space and the application does not demand pressurized hydrogen, this could be a valid option. According to Bangoura (2020) storing up to 100kg of hydrogen should preferably be done with low pressurized vessels, if there is sufficient space. This is mainly because of the compressor cost reduction compared to high pressurized storage. The CAPEX is usually higher for low pressure storage compared to high pressure, however, the operating expendi- tures (OPEX) is lower. For a 100 kg storage facility using 6 MPa vessels, the CAPEX can be 1 432 000 SEK, or 14 320 SEK/kg (Bangoura 2020).
High-Pressure
Higher pressure is equivalent to lower volume, which is highly sought after when looking of ways to store hydrogen. The main drawback with increasing pressure is both the higher investment cost of the compressor and increasing
electricity demand. The OPEX of high pressurized storage is mostly cost of compression (Bangoura 2020). The CAPEX for 100 kg of 50 MPa vessels are 848 000 SEK, or 8480 SEK/kg (Bangoura 2020).
Underground
Underground hydrogen storage is a very economic option for storing large quantities of gases. Even if storing hydrogen does not differ significantly from storing natural gas underground (which is commercially available), it is not yet a commercial technology and it will not be in the near future according to Tarkowski (2019). However, there are three existing underground hydrogen storage today according to Tarkowski (2019). The most recent one (2007) is located in Moss Bluff, eastern Texas. This salt cavern has a volume of 566 000 m3. The construction cost of this project was roughly 160 million SEK, which amount to a CAPEX of 83 SEK/kg of hydrogen (Preuster et al. 2017).
Compressor
In addition to the cost of storage, the compressor must also be included. It is commonly assumed that, depending on the output pressure, the compressor amounts to 10-15% of the hydrogen energy content (HHV) (IEA 2014). Ac- cording to IEA (2014) a 30 MPa compressor handling a flow rate of 1.79 kg/h has a CAPEX range of 475 000 - 950 000 SEK. While according to Ozaki et al.
(2014) a 35 MPa compressor with cooling unit would have a CAPEX of 131 SEK/W. The latter of the two cost estimation is more detailed in their calcula- tions and more applicable on different sizes of compressors. The compressors evaluated in IEA (2014) and Ozaki et al. (2014) have efficiencies between 65- 75%.
In this section the work process of the thesis is described in detail. Firstly the problem identification will be presented, then the data provided from facilities together with relevant information regarding said facilities. Two systems will be presented and after, the method of how they are evaluated together with economics will be presented.
3.1 Problem Identification
Normally the purpose of installing solar power for an industry is to decrease the demand peaks during summertime. If the peaks are decreased the cost of highest peak per month will also decrease. A perfect load profile for solar PV is one that has high peaks during daytime and during summer. Such load profiles can be found at large grocery stores, who has a high base load due to a large amount of refrigerators and added electricity demand for cooling during hot summer days. Unfortunately the load profile for steel and iron pro- cessing industries can be assumed constant throughout the year (Vogl 2020).
This means that there is no significant peaks. Even though the load is more of less constant, the cost of electricity is not. As presented in section 2.2.1 there is an increase in cost during weekdays 06:00 - 22:00 from November through March. This is definitely not the prime time for solar power in the Nordic hemisphere. However, with PV and hydrogen storage, the energy can be stored and used in this period of high cost. To address this cost problem, two systems will be presented to lower the cost of high demand season, using solar power and hydrogen production. The systems will suit different types of industries depending on their demands, which are presented in section 3.2.
26
3.2 Facilities
The companies willing to contribute with data were Outokumpu, Sandvik Group, TPC AB and SSAB. From their facilities, data regarding electricity consumption, yearly steel production, hydrogen and oxygen demand was col- lected. This data is summarized in table 3.1.
SSAB has two steel producing plants, one in Luleå and one in Oxelösund.
The facility in Luleå is producing steel plates from raw material using blast furnace technology. Then the steel plates are transported by cargo trains to SSAB’s facility in Borlänge where they produce products out of the plates.
Facility H2 O2 Electricity Steel
Hallstahammar
Sandvik 300 000 500 000 50 11 000
Surahammar
Sandvik 0 100 000 4 700
Sandviken
Sandvik 1 500 000 25 000 000 600 250 000 Luleå
SSAB 0 –** 527 1 893 008
Oxelösund
SSAB 0 –** 564 1 144 000*
Hallstahammar
TPC 0 0 6 132
Avesta
Outokumpu 0 900 000 413 413 000
[Nm3] [Nm3] [GWh] [tonne]
Table 3.1: Yearly demands acquired from the different facilities.
*With one out of two BF running, **No available data
The facility in Oxelösund is a combination of the Luleå and Borlänge op- eration, so it is larger in size but does not produce as much steel as the facility in Luleå. Oxelösund also adds scraps to their raw iron when producing steel, so their electricity consumption is slightly higher. According to Mikaelsson (2020) Oxelösund produces 1,1 tonne steel with 1 tonne raw iron due to the ad- dition of scraps. SSAB’s two facilities are the only ones producing steel from raw iron, both using BF technology. SSAB have plans to transform both of these facilities from BF-BOF to H-DR and EAF systems (Vogl 2020). There-
fore, they will be investigated as though they are using H-DR and EAF. Ac- cording to Vogl et al. (2018), 51 kg of hydrogen is needed per tonne produced steel. Vogl et al. (2018) also states that using H-DR and EAF system, a specific energy consumption (SEC) of 3.48 MWh/t of liquid steel is demanded. In this consumption, they include an electrlolyzer, which stands for two thirds of the SEC. The EAF and ore heating processes are the other large consumers of en- ergy. To calculate the new energy demand of SSAB’s facilites, the electrolyzer will not be included, since this demand will be calculated in this projects pro- posed systems. The energy demand of a BF-BOF system is mainly from coke and coal, therefore it is assumed that the existing electricity demand will re- main. Assuming the facilities will produce the same amount of steel, the new demands were calculated with 51 kg of hydrogen and 1.16 MWh/t of steel pro- duced. Table 3.2 shows the updated demands assuming SSAB using H-DR and EAF technology.
Facility H2 O2 Electricity Steel
Hallstahammar
Sandvik 3∗105 5∗105 50 11 000
Surahammar
Sandvik 0 1∗105 4 700
Sandviken
Sandvik 1.5∗106 2.5∗106 600 250 000 Luleå
SSAB 1.07∗109 – 2 722 1 893 008
Oxelösund
SSAB 6.49∗108 – 1 891 1 144 000
Hallstahammar
TPC 0 0 6 132
Avesta
Outokumpu 0 9∗105 413 413 000
[Nm3] [Nm3] [GWh] [tonne]
Table 3.2: Calculated yearly demands for the different facilities.
Outokumpu’s facility in Avesta produces only steel from scrap metal, then the product is transported further to their other facilities where they do refining of the material. Sandvik provided data from multiple facilities where they both produce steel from scrap and do refining of the material. The oxygen demand from these facilities are mainly used in melt-shops and furnaces. Sandvik uses hydrogen at Hallstahammar and Sandviken, mainly for heat treatment. TPC
is one out of two lost-wax casting facility in Sweden, using scrap to produce high alloy products.
3.3 System Presentation
As mentioned in section 3.1, two systems utilizing solar power and hydrogen will be constructed to suit the different demands presented in table 3.2. Sys- tem one will address the facilities that does not have any hydrogen demand, which are; TPC, Outokumpu and Sandvik in Surahammar. System two will be constructed to meet the hydrogen demand of SSAB’s facilities and Sandviks’s in Hallstahammar and Sandviken.
System one: Electricity demand
The purpose of this system is to utilize hydrogen as a long term storage and convert it into electricity at a time when the grid prices are high. As shown in figures 3.1 the facilities will have an electricity demand and two out of three will also have an oxygen demand. The electricity demand will be satisfied by purchasing power from the grid and producing power from PV panels. The electricity from the PV will be converted into hydrogen and oxygen by water splitting using a PEM electrolyzer. When the PV produces more electricity than the electrolyzer can handle, the electricity will instead be used to lower the amount of electricity bought from the grid. The oxygen is assumed to be utilized directly in furnaces. The hydrogen will be compressed and stored until it is high demand season. Then the hydrogen will be converted into electricity by a fuel cell. Only high-pressure storage is evaluated, the reason for this is presented in section 3.4.4. The compressor is sized according to what pressure is needed for the storage as well as the flow rate from the electrolyzer. The fuel cell capacity is calculated to have an output in kWh to meet the reduced hourly demand of the facility. The storage is sized to hold enough hydrogen to reduce the amount of power bought from the grid through out the high demand hours in a year. The sizes of the PEM electrolyzer and PV’s are being altered to find a system that can produce enough hydrogen so that the fuel cell can lower the electricity demand. The amount of demand to be reduced will be altered to find a system which yields the highest NPV.
Figure 3.1: Layout of System One
System two: Hydrogen demand
The purpose of this system is to supply hydrogen and oxygen that is being pro- duced on site to meet the demands instead of buying the gases from a third- party company. The hydrogen demand is assumed to be constant throughout the year. Since the cost of hydrogen (100-600 SEK/kg) is much higher than the cost of oxygen (0.5 SEK/kg), the system will be designed to meet the hydro- gen demand. This will result in a reduced amount of oxygen purchased from a third-party company. As figure 3.2 shows, this system contains two types of electrolyzers which produces hydrogen and oxygen. The AEL is sized to pro- duce the hourly hydrogen demand. This electrolyzer utilizes electricity from the grid and will therefore increase the electricity demand of the facility. As section 3.1 states, the systems are suppose to reduce the cost during high de- mand season. To achieve this, the system uses a PEM electrolyzer that utilizes electricity from PV panels to store hydrogen in high-pressurized vessels so it can be used in November-March and therefore reduce the amount of electric- ity purchased from the grid during this period. The storage capacity is sized to meet the desired reduction of hydrogen from the AEL. This reduction is dependent on the variable load range of the AEL, which is 15-100% of the nominal load. Therefore the storage will be sized to contain enough hydrogen to meet 85% of the hourly demand during the high cost hours. As in system one, the compressor is sized according to what pressure is needed for the stor- age as well as the flow rate from the electrolyzer. The PEM electrolyzer and PV’s sizes are varying to find a system which yields the highest NPV possible, while producing enough hydrogen so that the storage can meet the demand.
Figure 3.2: Layout of System Two
3.4 Components
In this section all technical and economical parameters for each component are presented.
3.4.1 PV
A 1.05 kW solar system was constructed in SAM to create a vector with the hourly output for a year, for each location. This vector can then be multiplied to evaluate different sizes of PV systems. The PV panel used in the systems are from Jinko Solar Co and the model is called "JKM350M-72-J4". This panel was chosen since it has similar properties as the ones most commonly used by Save by Solar AB. The PV system in SAM consist of 3 panels and one inverter
"SMA America: SB1100U-SBD [240V]". The CAPEX and OPEX cost is for the whole PV system, including inverters, cables, transformers etc (ETIP-PV 2017). All technical and economical data used in the modelling is presented in table 3.3.
JKM350M-72-J4
Efficiency 19.07%
Maximum Power 350 Wdc
Lifetime 25 years
CAPEX 8 000 SEK/kW
OPEX 25 SEK/kW
Degradation 0.36% /year SMA America: SB1100U-SBD [240V]
Efficiency 90.551%
Maximum DC Power 1215 Wdc Maximum AC Power 1100 Wac
Lifetime 15 years
Replacement Cost 1 820 SEK/kW
Table 3.3: Technical data is aqcuired from SAM software, economic data was supplied by Save by Solar, and confirmed by ETIP-PV (2017) Walker (2017).
All systems in the southern distribution area were optimized in SAM with regards to maximum AC output, and resulted in a tilt angle of 40°C and an azimuth angle of 160°C. SSAB facility in Luleå was optimized in the same manner and achieved a system design with a tilt angle of 50°C and an azimuth angle of 140°C.
3.4.2 Electrolyzer
The AEL is exclusively used in system two, where a constant demand of hy- drogen is required. The CAPEX for the electrolyzers are calculated from table 2.2 by interpolation, depending on required size for the system. The technical and economical data used in the modelling is presented in table 3.4.
AEL PEM Unit
Efficiency 70 63 %
Load Flexibility 15 - 100 0 - 160 %
Lifetime 80 000 45 000 Operating hours
OPEX 2 3 % of CAPEX
Replacement Cost 3 500 4 200 SEK/kW
Output Pressure 0.5 3 MPa
Table 3.4: Technical data is aqcuired from Buttler & Spliethoff (2018), Mayyas
& Mann (2019) and Parra & Patel (2016)
The AEL is operating 8760 hours per year, therefore it will be replaced every 9 years over the lifetime of the system, which is 25 years. The PEM electrolysers operating hours are fewer since it only operates when the PV system produces electricity. For the southern facilities, the PV produces electricity 3543 hours per year. For the facility in Luleå, it will operate 3602 hours per year. This results in a replacement every 12 years for both the southern and northern fa- cilities. Both electrolyzers will be replaced twice throughout the lifetime of the system.
3.4.3 Compressor
The compressor size will be depending on the desired output pressure and hydrogen flow rate from the PEM electrolyzer. There will be no compressor connected to the AEL since it is assumed that the hydrogen produces from the AEL will be utilized directly. The size of the compressor is calculated with equation 3.1
Wcomp = CpT1 ηc
"
P2 P1
r−1r
− 1
#
mc (3.1)
Cp is the specific heat of hydrogen at constant pressure, T1 is the inlet tem- perature of the hydrogen gas in Kelvin, ηcis the efficiency of the compressor, P2 and P1 are the outlet and inlet pressures respectively, r is the isentropic exponent of hydrogen gas and mcis the flow rate of the gas when entering the
