Examensarbete 30 hp
Energy storage solutions for
electric bus fast charging stations
Cost optimization of grid connection and
grid reinforcements
Teknisk- naturvetenskaplig fakultet UTH-enheten Besöksadress: Ångströmlaboratoriet Lägerhyddsvägen 1 Hus 4, Plan 0 Postadress: Box 536 751 21 Uppsala Telefon: 018 – 471 30 03 Telefax: 018 – 471 30 00 Hemsida: http://www.teknat.uu.se/student
Abstract
Energy storage solutions for electric bus fast charging
stations
Malin Andersson
This study investigates the economic benefits of installing a
lithium-ion battery storage (lithium iron phosphate, LFP and lithium titanate, LTO) at an electric bus fast charging station. It is
conducted on a potential electric bus system in the Swedish city Västerås, and based on the existing bus schedules and routes as well as the local distribution system. The size of the energy storage as well as the maximum power outtake from the grid is optimized in order to minimize the total annual cost of the connection. The assessment of the distribution system shows that implementing an electric bus system based on opportunity charging in Västerås does not cause over-capacity in the 10 kV grid during normal feeding mode. However, grid reinforcement might become necessary to guarantee potential backup feeding modes. Batteries are not a cost effective option to decrease grid owner investments in new transformers. However, battery energy storage have the possibility to decrease the annual cost of
connecting a fast charging station to the low-voltage grid. The main advantage of the storage system is to decrease the fees to the grid owner. Of the studied batteries, LTO is the most cost effective solution because of its larger possible depth-of-discharge for a given cycle life. The most important characteristics, that determine if a fast charging station could benefit economically from an energy storage, is the bus frequency. The longer the time in between buses and the higher the power demand, the more advantageous is the energy storage.
ISSN: 1650-8300, UPTEC ES 17 002 Examinator: Petra Jönsson
Ämnesgranskare: Juan de Santiago Handledare: Kenny Granath
Executive summary
In Västerås, the possibility to implement electric buses is investigated. The system that is considered is based on high power battery chargers at the end stations. This type of charging would cause a high load on the distribution system and potentially demand expensive reinforcements. In addition, the grid connec-tion constitute a high cost for the charigng staconnec-tion owner. In the investigaconnec-tion, the economic benefits of connecting a lithium-ion battery (lithium titanate, LTO and lithium iron phosphate, LFP) to these charging stations are investigated.
The assessment of Mälarenergi’s distribution grid shows that there is enough available capacity in the 10 kV-grid during normal feeding to support the potential electric bus system. However, the transformer substations that are located close to the bus end stations Bjurhovda, Björnögården, Hälla and Hacksta lacks the capacity to connect the fast charging station in question. The study shows that the most cost effective solution would be to invest in a new transformer substation instead of decreasing the power demand by the use of an energy storage system.
For the charging station owner, the most cost effective solution in all studied cases would be to become high-voltage customers because of the lower fees. Considering a high-voltage connection, lithium-ion batteries of today’s price can not decrease the annual cost of connection. However, at some stations an LTO battery storage can be used to decrease the annual costs of a low-voltage connection, mainly by decreasing the fees. The characteristics of such a station are a low bus frequency and high power demand (Hälla is one example).
The LTO battery is superior to the LFP battery because it can discharge to a deeper level for the same cycle life. The lifetime is the main limiting factor for the battery storage since it is approximately 8 years whereas an grid investment would last for 40 years. The price of the LTO battery is assumed to be 8500 SEK/kWh in this study, but according to the trend, the cost of lithium-ion batteries is decreasing. A decrease in price would mean that more stations benefits from using an energy storage.
Acknowledgment
This thesis covers 30 credits (hp) and completes my master’s degree in Energy Systems Engineering at Uppsala University and the Swedish University of Agricultural Sciences (SLU). The project was conducted on behalf of Mälarenergi Elnät in Västerås.
I would like to thank my supervisors at Mälarenergi Elnät, Kenny Granath and Johanna Rosenlind, for your guidance throughout the project. A big thank you also to everyone at the grid planning department for creating a welcoming environment and for answering all my questions and helping me out in various ways. In addition, I want to thank Juan de Santiago at the Division of Electricity at Uppsala University for your involvement and help.
Others that have contributed to this paper include Peter Norrman at Hybricon Bus Systems, Fredrik Persson at Göteborg Energi and David Steen at Chalmers. Thank you for taking the time to discuss and answer my questions.
Lastly, I would like to thank my parents and Sebastian for all your input, advice and encouragement and for reading the whole thesis multiple times without even complaining.
Sammanfattning
I takt med att befolkningen i städerna växer, ökar kraven på en hållbar och effektiv kollektivtrafik. Flera svenska städer, däribland Umeå, Stockholm och Göteborg, har implementerat snabbladdande elbussar på några av sina busslinjer. Även i Västerås utreds möjligheten att införa den här sortens bussystem.
Systemet bygger på något som kallas Opportunity charging, vilket innebär att bussarna laddas på hög effekt vid ändhållplatserna under den tid som bussen normalt står stilla där. Den här typen av laddning möjliggör mindre bussbatterier och färre bussar än nattlig depåladdning, men den höga laddeffekten in-nebär också en påfrestning på distributionsnätet. Om snabbladdningsstationerna leder till överbelastning på det befintliga distributionsnätet blir det nödvändigt att göra förstärkningar. Laddoperatören, som äger och driver laddstationerna, betalar en avgift till nätägaren för att vara ansluten till nätet och ta ut effekt. Dessutom bekostar laddoperatören själva anslutningen. Denna måste dimensioneras med avseende på termisk kapacitet, spänningsfall och utlösningsvillkor. Laddstationer med högt effektuttag och/eller stort avstånd till en nätstation kan bli kostsamma att ansluta till nätet på grund av höga avgifter och väl tilltagna kablar.
Den här rapporten syftar till att undersöka huruvida ett energilager skulle kunna sänka nätinvester-ingskostnader för nätägaren och/eller nätanslutnnätinvester-ingskostnader för laddoperatören genom att sänka och jämna ut effektuttaget från laddstationerna. Utredningen är gjort med grund i det elbussytem som övervägs i Västerås samt det lokala distributionsnätet.
Lastprofilerna från de olika tänkta laddstationerna uppskattas med hjälp av bussarnas tidtabeller, körsträckor samt antagen förbrukning om 2.3 kWh/km. Distributionsnätets tillgängliga kapacitet under normalmatning utvärderas med hjälp av data över alla transformatorers högsta effektflöde samt kablarnas, transformatorernas och nätstationernas belastningsförmåga. Det visar sig att existerande 10kV-nät inte löper någon risk att överbelastas vid implementering av snabbladdningsstationer vid någon av bussänd-hållplatserna i Västerås. Däremot räcker inte märkeffekten till hos en del av de transformatorer och nätstationer som ligger närmast den tänkta laddstationen. Investeringskalkylen visar att den lägsta årliga kosntaden, annuiteten, erhålls av att investera i nya transformatorstationer istället för att sänka effektuttaget med hjälp av ett batterilager. En bidragande orsak till detta är batteriernas antagna livslängd på 8 år jämfört med nätkomponenters 40-åriga livslängd.
Nätanslutningskostnaderna minimeras genom att optimera systemet, bestående av laddare, energilager och nätanslutningskomponenter (kablar och säkringar). Nätets kapacitet och storleken på energilagret väljs för minsta möjliga årliga kostnad. Detta görs genom linjärprogrammering i MATLAB på 4 olika platser med olika egenskaper, så som avstånd mellan laddstation och nätstation, bussfrekvenser och laddeffekter. De kostnader som minimeras är dels annuiteten av investeringen och dels de årliga avgifterna. Det visar sig att energilager i vissa fall kan minska kostnaderna för en lågspänningsanslutning då den stora investeringen i ett energilager vägs upp av de minskade årliga avgifterna. Den främsta bidragande faktorn som gör det ekonomiskt fördelaktigt är en låg bussfrekvens. Det lager som visar sig lönsamt är LTO-batteriet (litium-titanat), som klarar djupast urladdning och högst effekt av de studerade batterierna. Högspän-ninganslutning har dock lägst årlig kosntad i alla de studerade fallen, och det är inte kostnadseffektivt att investera i batterilager för att minska effektuttaget vid en sådan anslutning.
Teknisk ordlista
balance responsible party: balansansvarig cable loadability: kabelbelastning
current collector: strömavtagare electricity supplier: elhandlare feeder: matarkabel
flywheel: svänghjul
grid concession: nätkosession harmonics: övertoner
looped grid structure: maskat nät national grid: stamnät
network operator: nätföretag pantograph: pantograf
primary substation: fördelningsstation
Transmission System Operator, TSO: Systemansvarig trolley bus: linjebus
secondary substation: nätstation trigger condition: utlösningsvillkor type curve: typkurva
Contents
1 Introduction 12
1.1 Problem and aim of study . . . 12
1.1.1 Aim . . . 13
1.1.2 Limitations and assumptions . . . 14
1.2 Related studies and projects . . . 14
1.3 Structure of report . . . 15
2 Theory 16 2.1 The Swedish power system . . . 16
2.1.1 Actors on the electricity market . . . 16
2.1.2 Grid network . . . 16
2.1.3 Substations . . . 17
2.1.4 Power lines and cables . . . 18
2.2 Electric buses . . . 20
2.2.1 Energy supply and demand . . . 20
2.2.2 Charging of electric buses . . . 20
2.2.3 Grid connection . . . 23
2.2.4 Swedish examples . . . 25
2.3 Energy storage . . . 25
2.3.1 Applications and technologies . . . 25
2.3.2 Battery Energy Storage (BES) . . . 27
3 Method and data 30 3.1 Technical feasibility . . . 30
3.1.1 Electric bus system . . . 30
3.1.2 Evaluation of grid capacity . . . 31
3.1.3 Cable sizing . . . 31
3.1.4 Energy storage system . . . 32
3.2 Cost estimates . . . 33
3.3 Comparison of costs . . . 34
3.4 Optimization of annual grid connection cost . . . 34
3.4.1 Linear programming problem . . . 35
3.4.2 Parameters . . . 35
4 Results 39 4.1 Annual grid connection cost . . . 39
4.1.1 Hällagatan, Hälla . . . 39
4.1.2 Flisavägen, Bjurhovda . . . 44
4.1.3 Forntidsgatan, Bjurhovda . . . 46
4.1.4 Björnögården . . . 47
4.2 Grid owner investments . . . 49
4.2.1 Hälla . . . 51
4.2.2 Forntidsgatan, Bjurhovda . . . 52
4.2.3 Björnögården . . . 53
5 Discussion 54 5.1 Annual grid connection cost . . . 54
5.2 Suggested design approach . . . 56
5.3 Grid owner investments . . . 56
5.4 Criticism of the sources . . . 56
5.5 Criticism of the methods . . . 57
5.6 Suggested further studies . . . 57
6 Conclusions 58
List of Figures
1 Map of buss lines in Västerås 12
2 Structure of local electricity grid in Västerås 13
3 Network structures 17
4 Substation 18
5 Underground cable 18
6 Bus energy consumption 20
7 (a) Trolley bus conection (b) Catenary-free dynamic charging 22
8 (a) Hight power pantograph charging station. (b) Ground-based conductive fast charging
station 22
9 Wireless charging station principle. 23
10 System without ESS 24
11 System layout with energy storage system. 24
12 Applications for energy storage systems 26
13 Charging characteristics of batteries 27
14 Available battery capacity for different C-rates 27
15 Past and estimated future cost of li-ion batteries 29
16 Relationship between battery cycle-life and depth-of-discharge 29
17 Calculation of available capacity during normal feeding 31
18 Calculation of trigger condition 32
19 Suitable of energy storage technologies 33
20 Model with system parameters. 36
22 Annual cost for the Hälla station connection. 40
23 The expected power consumption from the potential charging station at Hälla. 41
24 Optimization of 5000 randomized loadprofiles at Hälla, low-voltage, LTO. 42
25 The annual connection costs achieved by certain grid capacities and storage sizes at Hälla. 43
26 Battery sizes obtained when optimizing 5000 randomized load profiles with a fixed grid
capacity of 554 kW. 43
27 The substations at Bjurhovda. 44
28 The expected power consumption from the potential charging station at Bjurhovda. 45
29 Optimization of 5000 randomized load profiles at Flisavägen, low-voltage, LTO. 46
30 The substations at Björnön. 47
31 Optimal battery size in relation to battery cost. 48
32 The expected power consumption from the potential charging station at Björnögården. 49
33 Relationship between battery cycle-life and depth-of-discharge for LTO battery. 64
34 Sensitivity analysis Björnögården, LTO 68
35 Optimization of 5000 randomized load profiles at Björnögården, low-voltage, LTO. 68
36 Optimization of 5000 randomized load profiles at Björnögården, high-voltage, LTO. 68
37 Sensitivity analysis, Björnögården, LFP. 69
38 Optimization of 5000 randomized load profiles at Hälla, high-voltage, LTO. 69
39 Sensitivity analysis, Hälla, LFP. 69
40 Sensitivity analysis, Flisavägen, LTO. 70
41 Optimization of 5000 randomized load profiles at Flisavägen, low-voltage, LTO. 70
42 Sensitivity analysis, Flisavägen, LFP. 70
43 Sensitivity analysis, Forntidsgatan, LTO. 71
44 Optimization of 5000 randomized load profiles at Forntidsgatan, low-voltage, LTO. 71
45 Optimization of 5000 randomized load profiles at Forntidsgatan, high-voltage, LTO. 71
List of Tables
1 Fast charging stations for electric buses, operated in Sweden 25
2 Battery characteristics 28
3 Details about bus routes in the study 30
4 Battery model parameters 33
5 Optimization of the connection cost at the Hälla station. 39
6 Sensitivity analysis for LTO battery at Hälla. 42
7 Optimization of the connection cost at the Flisavägen, Bjurhovda station. 44
8 Optimization of the connection cost at the Forntidsgatan, Bjurhovda station. 46
9 Optimization of the connection cost at the Björnögården station. 47
10 Available capacity in primary substations. 50
11 Available capacity for connection at suitable secondary substations. 51
12 Transformer investments at the Hälla station. 52
13 Transformer investments at the Forntidsgatan, Bjurhovda station. 52
14 Transformer investments at the Björnögården station. 53
15 Normal values of cables, substations and other grid components as estimated by Ei. 63
16 Maximum cable loading. 63
17 Low-voltage connection prices. 64
18 High-voltage connection prices. 64
19 Flywheel characteristics. 65
20 SMES characteristics. 65
List of Symbols and Acronyms
APS Aesthic Power SupplyBEB Battery Electric Bus BES Battery Energy Storage
CAES Compressed Air Energy Storage CPT Capacitive Power Transfer
DoD Depth-of-discharge EAC Equivalent Annual Cost
Ei Swedish Energy Markets Inspectorate (Energimarknadsinspektionen) ERS Electric Road Systems
EV Electric Vehicle
FCEB Fuel Cell Electric Bus FES Flywheel Energy Storage HEB Hybrid Electric Bus
LFP Lithium iron phosphate battery LTO Lithium titanate battery NPV Net Present Value
PHS Pumped Hydroelectric Storage RIPT Resonant Inductive Power Transfer RPT Resonant Power Transfer
SMES Superconducting Magnetic Energy Storage SRS technology by Alstom
TSO Transmission System Operator WPT Wireless Power Transfer
AArea (m2)
C LMaximum power drawn from the grid
E Energy
E BSize of the energy storage I Current (A)
φ Phase shift
kE SS Battery interest rate
kgGrid interest rate
LLenght (m)
nLifetime η Efficiency
Pc(t) Charging power of the storage
Pd(t) Discharge power of the storage
Pg(t) Power drawn from the grid
Ploss Resistive power loss
r rate of interest RResistance ρ Resistivity () SSalvage value
U Voltage
YE SS Battery life time
YgGrid life time
1 Introduction
In 2014, 54 % 1 of the world’s population lived in urban areas [1]. For Europe, the number was 73 % [1] and for Sweden 85 % (2010) [2]. The urbanization is expected to continue in all continents during the first half of the century and the urban residents are estimated to increase with 3 billion people between 2014 and 2050 [3]. Sweden is no exception from the trend and about 70 % of the country’s population increase is expected to take place in the three largest cities [4]. Stockholm county, with 2.2 million inhabitants (2016), grows with 35,000 new residents each year [5]. The increasing urbanization seen in the world puts strains on infrastructure as well as the environment [1, 5]. To decrease the traffic density and air pollution experienced in many urban areas, an efficient and sustainable public transport system is one important puzzle piece [6]. Electrification of buses shows promise to reduce the fossil fuel dependency of the transportation sector as well as create a healthier urban environment [7].
Electric buses are nothing new, but have been around for over 100 years in the shape of trolley buses [8]. In recent years, there has been a major development in battery electric buses that are charged either overnight in the depot or in fast charging stations at selected bus stops. In Sweden, several bus routes are being electrified. Volvo, Siemens and Vattenfall together carry out the project ElectriCity in Gothenburg and since 2015 run a 7.5 km long bus route with electric and hybrid buses charged in high power charging stations at the end stops [9]. In Umeå, the company Hybricon has developed electric buses that charge in only three minutes and can run for about an hour [8]. Scania and Vattenfall are conducting trials with wireless charging of electric buses and in December of 2016 their first buses were taken into operation in Södertälje [10].
1.1 Problem and aim of study
Figure 1: Västerås with its seven urban bus lines that
all pass through the central station (Centralen) [11].
In Västerås, Sweden, the local public trans-portation company Västmanlands Länstrafik is investigating the possibility to convert part of the bus fleet from biogas buses to battery
electric buses. All seven urban bus lines,
viewed in figure 1, are considered. The
pro-posed concept is to charge the bus batter-ies at the end stations during idling, a strat-egy that requires high power fast charging sta-tions.
Depending on the bus line, the scheduled time at the end station is 1-10 minutes and the bus frequency 2-8 buses/hour. The bus routes have one-way distances of 6-14 km and the es-timated power demand is in the range 250-1000 kW.
The distribution network operator in Västerås is Mälarenergi Elnät,which are the owners of the 10 kV and 400 V grid. The general structure of the local grid is presented in figure 2. It consists of medium- and low-voltage networks as well as primary and secondary substations. The main flow of power is from the overlying transmission grid via the primary substations and out to the customers and the end of the 400 V/230 V grid. The grid is dimensioned for a high power flow close to the primary substations but have
less capacity further away. The electric bus fast charging stations can be connected to the grid either on the 10 kV-side or on the 400 V-side of a secondary substation.
Figure 2: The local grid in Västerås is owned by Mälarenergi Elnät and consist of the 10 kV and the
400 V networks. Power from the overlying grid is transformed into medium-voltage level in the primary substations. It is distributed in the 10 kV grid that consists mainly of underground cables. In secondary substations the voltage is transformed to 3-phase 400 V and in the cable cabinets the phases, being 230 V each, are separated. Dotted lines illustrate customer owned feeders that connects the customers to the grid. They can be connected at different voltage levels.
There is a limit to the amount of power you can input or output to a point in the grid, constrained by the required voltage quality and the thermal capacity of the grid components [12]. In addition, it is of impor-tance to fulfill electrical installation criteria. All these factors must be taken into account when designing a grid connection based on the estimated power demand. Every customer that connects to the electric grid pay the cost price of the connection as well as fees to the grid owner that are based on the subscribed power. Fast charging stations are expensive to connect to the grid due to the short and high power peaks that occur only when a bus is there to charge. Because of these load characteristics, the implementation of electric vehicles also gives rise to some concern about the load on the power distribution network [6]. If the power outtake exceeds the available grid capacity, costly reinforcements must be made by the grid owner. An energy storage system is a promising solution to make fast charging stations more cost effective since it can decrease and even out its power demand [13]. A large storage can affect the cost both positively and negatively since it enables a weaker grid and lower fees but is costly in itself. Because of this, there is a trade-off between the size of the energy storage and the strength of the grid. For each system, there exists an ideal combination of storage and grid capacity that minimizes the costs for the charging station owner or the grid owner.
1.1.1 Aim
The aim of this study is to determine whether an energy storage system connected to an electric bus fast charging station can decrease the grid connection costs for the charging station owner or grid reinforcement costs for the grid owner. In addition, the objective is to propose a design approach for the system that consists of the charging station, the energy storage and the grid connection components.
The study is conducted on the electric bus system that is considered in Västerås and four grid connection cases, with different characteristics, are analyzed in more detail. In each case, the cost of the connection is evaluated in terms of the annuity and minimized by optimally sizing the energy storage system in relation to the grid components. Different lithium-ion batteries are considered. The optimization is done by use of
1.1.2 Limitations and assumptions
This study focuses only on the Västerås case, and the bus lines and substations that exist there. Only fast charging at the end stations is considered. It is assumed that there will be no changes to the present bus schedules nor the routes. The charging station design load profile uses the weekday schedule and a bus energy consumption of 2.3 kWh/km (as estimated by VL).
In the investigation of potential grid reinforcements, only the design (normal) feeding mode as well as primary and secondary substations are considered. The available power is evaluated based on a worst case scenario. Costs considered are the investment costs of 10 kV- cables, transformers and substations.
When analyzing the grid connection, the costs taken into account are the investment costs of the energy storage and the grid connection cables as well as the fees that are payed by the customer to the grid owner. The cost of the charger and the bus is not considered. It is assumed that the price of electricity does not change enough with time for the yearly electricity cost to change if the consumption pattern changes.
It is assumed that the lifetime of the electric buses as well as the battery storage system is 8 years. The grid investment has an expected lifetime of 40 years, even though it is not certain that it will be used for the same purpose, or at all, during that whole time.
No regard has been taken to the question of ownership of the potential energy storage. Grid owners are prohibited from producing and trading electricity and there is a disagreement whether operating an energy storage falls into that category.
1.2 Related studies and projects
In the recent years, several Swedish master’s thesis projects have been conducted that concerns electric buses. Zisimopoulos [14] investigates the electrification of internal buses at Arlanda and evaluates charging systems, costs and CO2emissions. Lindberg [15] studies the power quality of the two fast charging stations
in Umeå and concludes that the stations emit harmonics that might affect grid connected consumers nearby. Karlsson [16] evaluates the total cost of electrifying a whole bus system and compares various charging strategies. The thesis concludes that end-station fast charging combined with depot charging is most cost effective and that the grid connection cost is a main component of the total cost.
At Chalmers in Gothenburg a project is currently being conducted that aims to analyze the effect that electric bus fast charging stations have on the distribution grid [17]. Energy storage solutions are being investigated as well as the possibility to use the charging stations to stabilize the grid voltage. Another Chalmers study, conducted by Grauers et. al. [18], evaluates the cost effectiveness of various charging systems for electric buses. The study concludes that opportunity charging (fast charging during scheduled idle time) is the least costly option.
Fusco et. al. [6] has modeled average energy consumption of electric buses as a function of average speed during the route. The paper proposes a methodology for optimal design of a transit system that includes electric buses and charging stations. Ding et. al. [13] suggests an optimal control strategy for fast charging stations combined with energy storage systems as well as the optimal storage sizing. The article also evaluates how much the investment cost and the charging cost can be decreased by the use of a lithium-ion battery. In the study, the costs considered are the investment in cables, transformer and battery as well as the cost of the electricity, that varies substantially throughout the day. It is evaluated on load data from a real system where several buses are charged by the same charger. The conclusion of the study is that the two lithium based batteries LFP and LTO can decrease the costs associated with a fast charging
station, mostly due to the varying electricity price. The model in the study does not take into account the difference in lifetime between the grid components and the energy storage.
1.3 Structure of report
The report is structured as follows. Section 2 presents the background theory necessary to understand the study and its results. It gives a general overview of the Swedish power system and also provides a more thorough description of the distribution grid and distribution grid planning. Furthermore, section 2 gives insight in the area of technology that is electric buses and go over possible charging strategies, present manufacturers and ongoing projects. Section 2 also describes various energy storage technologies and compare their characteristics.
In section 3, the general methods of the study are described and the data presented. In addition, the section describes the model that was created to optimize the system, consisting of the bus, the charger, the grid connection and the energy storage.
The results of the optimization are compiled in section 4, discussed in section 5 and in section 6 the conclusions drawn from the study are presented.
2 Theory
2.1 The Swedish power system
2.1.1 Actors on the electricity market
The Nordic countries share a common electricity market that was deregulated in 1996 [19]. Players on the market are producers, consumers, suppliers, network operators, a Transmission System Operator as well as the authorities Swedish Energy Markets Inspectorate (Ei) and Swedish Energy Agency.
The Transmission System Operator (TSO) is Svenska Kraftnät, a state-owned authority that has the overall system responsibility. They own and operate the national grid, maintain power balance in the system and procures a power reserve before each winter [19].
Network operators own the grid and are responsible for the transport of electricity. There are about 160 network operators in Sweden and Mälarenergi Enlät AB is one of them [20]. Because of natural mo-nopolies there can be only one actor providing electric grid at each location. According to the electricity law(2013:207) 2 chap. §1, a network operator must have permission, so called grid concession, from Ei to build and operate a grid in a certain area [21]. To ensure a well-functioning market as well as reasonable prices, competition is simulated through rules and regulations managed by Ei [20]. As established in the electricity market reform in 1996, Swedish network operators are not allowed to produce or trade electricity other than to cover up for their grid losses or secure operation in case of faults (law(2011:712) 3 chap. §1) [21]. They are obliged to maintain a high power quality and make sure the right amount of power is delivered to the consumers despite losses in the grid (law(2005:1110) 3 chap. §9 [21]). Any producer or consumer that so wishes, must be connected to the local network, although the connection is payed for by the connecting actor (law(2005:404) 3 chap. §6) [22].
Electricity suppliers buy electricity, commonly on the market Nord Pool Spot, and sell it to their cus-tomers. As opposed to grid owners, they compete with each other, which gives the consumer the option to choose supplier company and deal. Every electricity supplier is obliged to, at every instant, deliver the same amount of power as their customers consume, which makes them the Balance responsible party. In case of imbalance, the TSO trades electricity on the balance market with short notice and charges the Balance responsible party that failed to keep the balance. This motivates all actors responsible for the balance to perform good estimates of their customer’s consumption [20].
Producers sell electricity directly to customers or to electricity suppliers on the spot market. They pay a fee to the network operator to connect to the grid and feed it with power. Consumers are charged for the used power and electricity both by the local network operator and the electricity supplier as well as pay electricity tax [20].
2.1.2 Grid network
The electric grid consists of a high-voltage transmission system and a low-voltage distribution system. Traditionally, power is generated at large plants connected to the national grid, transported further through the regional grid and lastly distributed to consumers in the local low-voltage grid. With the implementation of renewable power production, such as solar and wind power, electricity is to a larger extent produced and connected on the consumer side of the grid [23].
The stem in the Swedish transmission network is the national grid that consists of AC power lines at 400 kV and 220 kV and stretches from north to south to connect the large producers with the consumer
areas [19]. HVDC cables link Sweden to neighboring countries Finland, Lithuania, Poland, Germany and Denmark [19]. The national grid connects to the local distribution network via regional grids with voltages of 20−130 kV [19]. The regional networks are mainly owned by Vattenfall, E.ON and Ellevio [24]. Locally, power is distributed to consumers through either medium-voltage networks (10-20 kV) or low-voltage networks (400/230 V) [23]. When faults occur in the power system it is most often in the local grids [19]. Networks can have either radial or looped structure, as illustrated in figure 3. A radial grid is the least costly network structure and the easiest to protect but in the event of failure in a line, all nodes connected behind the fault will be affected [25]. This structure is mostly used on the countryside. Loop and multi-loop structures are used in urban areas [25]. They are more expensive to construct but have higher resilience since each point can be fed in several ways. During normal feeding, these networks are operated as radials but have the possibility to open or close loops so that the power can be alternatively fed in case of faults or maintenance on a line.
Figure 3: Network structures can be either radial, looped or multi-looped. The dots represents substations
or consumers. Dotted lines mark where the loop can be closed with switches to enable an alternative operation [25]. (a) Radial network with single-point feeding. (b) Loop network with single-point feeding.
(c) Multi-loop network with single-point feeding. (d) Multi-loop network with multiple-point feeding. Distribution network planning involves planning long-term and short-term investments to meet demand changes or maintain quality as well as construction design of network structure, cables, power lines, trans-formers and other components. The main goal is to achieve safe and reliable power transfer to the lowest possible lifetime cost [26]. When major installations are made in a node, both the capacity of the normal feeding mode as well as the back-up feeding should be analyzed so that it is not exceeded.
The load patterns of various electricity customers can be described by a type curve that graphs the consumed power during a day as mean and standard deviation. Type curves are an important resource that aid in grid planning activities such as grid dimensioning, load forecasting, investments, reinforcement planning etc. A load forecasting model was developed by Svenska Elverksföreningen in the 90’s and although updated, it is still used today. The model is based on substantial measurements and covers many load types. Several type curves can be superimposed to create load behaviors closely linked to reality. Inputs to the model are degree-day, month, weekday/weekend, mean temperature (24h), yearly energy consumption and likelihood that the load does not overstep the forecast. For a more detailed description of the model, the reader is referred to [27].
2.1.3 Substations
A substation is a node in the grid network where power lines can be divided and current or voltage levels changed. It is also where the system protection is located and where the current can be stopped. Substations
Figure 4: Layout of a typical 10/0.4 kV
distribution substation in the Västerås grid.
can be equipped with transformers, protective relays, break-ers and disconnectors as well as metbreak-ers and devises for reactive compensation [25]. Relays measure currents and voltages, detect abnormalities and control the breakers. A breaker stops the current, most commonly with the insu-lating SF6 − gas, while the disconnector physically
sepa-rates two conductors as a visual confirmation of the bro-ken current. Fuses and other surge protection devises pro-tects the components from over-voltages [28]. Substations can be located in open air, when space is enough, or in-side a metal enclosed construction isolated with SF6− gas,
when space is restricted [25]. Inside the substation incom-ing and outgoincom-ing feeders are connected to one or several common bus bars. Figure 4 depicts a typical substation lay-out.
2.1.4 Power lines and cables
Over-head transmission lines and under-ground cables transport electricity from producers to consumers. Burying cables under ground increases the system reliability due to avoided exposure to weather, lightning, falling trees etc., but the investment is more expensive than equally rated over-head lines [25]. Power can be transferred either through direct current (DC) or alternating current (AC). A DC current uses the whole cross-section area of a conductor while an AC current, due to the skin effect, flows only on the conductor surface. This makes DC transfer more efficient [25]. Despite this, AC is dominating the distribution system. Over-head lines generally uses light aluminum conductors combined with a steel core to improve the strength [25]. The transmission lines are hung on towers separated from each other and can be either uninsulated or insulated [29]. Since it is the cheapest method, it is the most common choice for long-distance power transfer [25]. However, over 97 % of the existing kilometers of low-voltage power conductors in Sweden are under-ground cables [23]. A cable and its components are illustrated in figure 5.
Figure 5: A cable consists of several layers, each with its special function. Conductors are made out of
copper or aluminum. Copper has a lower resistivity than aluminum, but it is heavier and more expensive [25]. Copper is used for cross section areas of 0.5 - 2500 mm2while aluminum only is preferred for areas of 50 mm2 and above [29]. To isolate the conductor either plastic or rubber is used. For high voltage cables, PEX-isolation is dominating. This is a cross linked polyethylene with thermal and mechanical characteristics very suitable for electric isolation applications [29]. Separated from the conductor with insulation is the concentric neutral conductor, made out of copper or aluminum wire or tape. To give the cable a circular cross section, a filling material is often used. As protection, several layers of plastic, sometimes with metal reinforcement, are added on the surface [29].
2.1.4.1 Cable selection
The cable load capacity specifies the maximum current or power a certain cable can transport and is limited by the temperature level the cable materials can withstand. When a current runs in a cable, the temperature rises due to heat power losses in the conductor. These losses are determined by the current and the resistance according to [28]:
Ploss = R · I2 (1)
where Ploss is the resistive losses in the wire. The current depends on the transferred power and nominal
voltage. To minimize the losses during high power transfer (large I) and long distances (large R) the voltage is increased.
The percentage voltage drop in a cable is described by equation 2. It is seen that it depends on the transmitted power (P), the resistance (R), the reactance (X) and phase shift (φ). The reactance can be neglected and the resistance described by the cable resistivity (ρ), length (L) and cross-section area (A). In a customer facility, the voltage drop should not exceed 4 % according to the Swedish electricity standardization Svensk elstandard SS 436 40 00 [30]. This requirement affects the feasible cable geometry.
∆U Unom = √ 3 · I(Rcosφ + X sinφ) Unom ≈ √ 3 · I · ρ · L · cosφ A · Unom (2)
In addition to these two requirements, the cable must be short enough for the service fuse to release if there is an overcurrent. In case of a fault that creates a current higher than the fuse’s breaking current, the fuse releases and protects the circuit from thermal and mechanical stress. The short circuit current should be completely isolated within a specified time (often 5s). The let-through energy can be expressed as I2t,
where I is the RMS short circuit current and t the breaking time [30]. This is however not a measure of the energy, but of a quantity that is proportional to the energy transmission. Each cable can withstand a certain level of high fault currents, expressed in k A2s, and the fuse must be chosen so that its specific let-through
energy does not exceed the cable limits.
The high current regulations (Starkströmsföreskrifterna) states that a faulted facility must be discon-nected rapidly from the grid [30]. To meet these requirements both the fuses and the cables must be correctly dimensioned with respect to the highest as well as the lowest possible fault-current. It is the lowest fault current that limits cable length. The lowest fault current in directly earthed systems occurs when there is a 1-phase ground fault at the furthest distance from the fuse [30]. To ensure that the fuse breaks within its given time the short circuit current Is must be at least the size of the fuse’s breaking
current Ib[30]:
Is ≥ k · Ib (3)
kis a tolerance factor. This is called the trigger condition. The short circuit current is described by Is =
U
z · L (4)
where, in the case of a 1-phase earth fault, U is the phase-to-neutral voltage and z equals the combined impedance of one phase and the cable neutral expressed in W/m. L is the distance from the fault to the fuse. To assure that the fuse releases as it should, the following must hold [30]:
Ln=
U Is · z
≤ U
k · Ib· z (5)
The above equation is obtained by combining equation 3 and 4 and assumes that the short circuit power at the start of the cable is infinite. Usually it is important to take into account also the impedance at the
start, that consist of the impedances of the feeding grid (ZQ), the transformer (ZT) and, if applicable, the
feeder before the fuse (ZF). The maximum allowed cable length is thus described as
Lmax = Ln(1 −
ZQ + ZT + ZF
Zmax
) (6)
where Zmaxequals the total cable impedance in W at length Ln[30].
2.2 Electric buses
2.2.1 Energy supply and demand
Figure 6: The energy demand of electric buses
de-pend on factors such as climate, speed and topog-raphy. The graphics illustrates a consumption esti-mation made by Volvo, where the electric energy use in best case is 0.8 kWh/km and in worst case 2.82 kWh/km [31].
Electric vehicles are driven by various types of traction motors, such as the brush-less DC motor, the switched reluctance motor and
the induction motor [32]. The engine
effi-ciency is around 25 % for regular combus-tion engines while electric engines can
op-erate at efficiencies of 80-90 % [6]. The
energy demand of an electric bus depends on many factors, such as speed, route dis-tance, number of passengers (weight), temper-ature, topography, road quality and driver
be-havior [31]. As shown in figure 6, the
consumption varies between 0.8 kWh/km in the best case to 2.82 kWh/km in the worst case.
The difference between the various types of electric buses (grid bounded, hybrid, fuel cell and battery electric buses) is the electricity sources they use to power their motors [33]. Battery Electric Buses (BEB) can be equipped with smaller batteries that are fully charged in 5-10 minutes or larger batteries that charges over-night and lasts throughout the day [33]. Range is a limiting factor for electric vehicles since storing energy in large long-lasting batteries increases the weight of the vehicle substantially [6]. Battery types used in EV:s today are Lead-acid, Ni-Fe, Ni-Cd, NiMH (Nickel-Metal hydride), Sodium-metal chloride, Na-S and various Lithium based [32].
A hybrid bus commonly combines an electric motor with a conventional internal combustion engine (ICE). If the engines are connected in parallel the traction power can be delivered from both engines simultaneously, or either of them separately. When series connection is used, the ICE functions as a generator that provides electricity to the EM, sometimes via a battery [32]. Plug-in hybrid vehicles have series connected motors, but also the ability to charge the battery through an external source, thus enabling fully electric operation [33].
2.2.2 Charging of electric buses
Charging can take place either while the vehicle is moving or when it is at rest. Furthermore, energy can be supplied to electric vehicles by means of conduction or by wireless coupling (Wireless power transfer, WPT). Four main charging technologies can thus be identified; static conductive, dynamic conductive, static wireless and dynamic wireless.
International standards for electric bus charging are in the making and expected to be finalized by ISO/IEC in 2020 but presently a broad variety of solutions can be seen throughout the world [34]. Because of this uncertainty, bus manufacturers commonly produce buses that allows for the customer to specify either conductive or wireless charging systems rather than focusing on one specific technology [35]. Sev-eral European electric bus manufacturers (Irizar, Solaris, VDL and Volvo) are cooperating with charging system developers (ABB, Siemens and Heliox) around a common charging interface [34].
Electric Road Systems (ERS) are roads that dynamically provides vehicles with power [36]. This dynamic charging is a strategy that secures the range of an EV without relying on large energy storage systems [36]. However, this type of charging requires a lot of infrastructure. Static and dynamic charging can be combined by letting the vehicle drive outside the ERS on a combustion engine (hybrid vehicles) or on electricity stored in batteries (fully electric vehicles) [36].
In cases where the bus is to be exclusively charged statically, there are several options on when to charge. For city buses, opportunity charging has been identified as a suitable operational strategy [13, 37, 18]. It entails charging the bus during its scheduled idle time at a few stations along the route (usually end stations). Frequent charging allows for a smaller battery to be used but it also requires high power, which increases the price of the charger [38, 6, 39]. The alternative to opportunity charging would be slow charging overnight of a large, long-lasting battery, fast charging at every bus stop along the route of a small battery or battery swapping one or a few times a day [39]. A comparison made by Grauers et. al. [18] shows that opportunity charging at end stations have a lower total cost than night charging and bus stop charging (cost of battery, charger and electricity included).
A few other charging strategies has been proposed, for example charging fast and very frequent though built-in structures on the road while storing the energy in super capacitors [40]. Musavi et. al. [38] recommends a solution for public transport that consists of a few fast DC charging stations combined with wireless chargers at bus stops or traffic lights. The choice of charging strategy depends on the route distance as well as on project economy and available charing station area. For a more detailed analysis of different charging methods the reader is referred to [39] and [18].
2.2.2.1 Conductive charging
During conductive charging energy is transferred to the vehicle from a voltage source through an electric conductor. Dynamic conductive charging through overhead lines is a well-established technology used for trains, trams and trolley buses. Electric trolley buses draw a current from overhead wires through trolley poles as seen in Figure 7a [7]. The poles are dragged behind the vehicle and allow for lateral and vertical movement, although disconnection sometimes occurs in sharp turns [41]. In their project eHighway, Siemens together with Scania has developed an ERS concept for trucks connected to over-head catenary wires that allows for higher speed than regular trolley buses [42]. Advantages of trolley buses include low noise, no emissions, easy maintenance, similarity to diesel buses as well as the possibility to operate on existing infrastructure [7].
In order to avoid the visual impact from overhead wires, solutions for ground-based charging systems has been developed. Alstom is the largest supplier of catenary-free charging systems for trams. Their Aesthetic Power Supply (APS) technology includes a third middle rail that provides electricity through several current collector shoes mounted at the bottom of the tram [43]. Alstom and Volvo are both participating in a project to implement the APS technology in an ERS that can power all sorts of vehicles [36], see Figure 7b. Ground-based charging systems have the advantages of easy extension of the lines and no overhead-wires but electric rails on the road might affect the safety of humans and animals due to changed driver behavior, road friction and magnetic fields [43, 36].
(a) (b)
Figure 7: (a) Electric buses connected through trolley poles are seen all over the world. This particular
one is operated in Vancouver, Canada. The overhead wire has two lines; one with a voltage of around 600-700 VDC and one at ground potential [41]. This is required since the vehicle is rubber tired and thus not grounded. The control system of the trolley poles enables automatic connection to the catenary [7]. The connecting material in the trolley shoe is usually carbon-based, such as graphite [44]. Photograph by Steve Morgan [45]. (b) Pilot ERS that uses APS technology developed by Alstom and trucks with conductive pick-ups constructed by Volvo GTT. The truck is provided 750 VDC from ground-based rails through the collector shoe at the rear of the vehicle [36]. Image from Volvo [46].
(a) (b)
Figure 8: (a) Station for high power opportunity-charging,
constructed by ABB. The system consists of a current col-lector, a pantograph, that automatically connects to con-tacts at the bus roof to provide the battery with 150 kw, 300 kW or 450 kW DC [47]. The utility AC power must be rectified at the charging station. Image from ABB [47].
(b) Illustration of the SRS technology, a conductive
ground-based fast charging station developed by Alstom. The bus is supposed to charge in a few minutes while idling at a bus stop. Image from Alstom [48].
A pantograph is a current collect-ing devise traditionally used to power
trains and trams. Today, the
panto-graph has been reused in static conduc-tive fast charging stations for electrical
buses, as shown in figure 8a.
Com-panies that develop this technology are ABB, Hybricon, Siemens and Proterra,
amongst others. The power provided by
these stations is in the range of
150-1000 kW [49, 50, 51]. Both ABB and
Siemens provide charging stations of 150 kW, 300 kW and 450kW DC, which en-ables a charging time of 4-6 minutes
for a regular city bus [47, 50]. The
pantograph can be placed either on the roof of the bus or at the station
(re-versed pantograph). A pantograph at
each vehicle increases the system resilience but it adds weight and cost to the bus [47].
In addition to the APS system, Alstom has also developed a static conductive ground-based fast charging system for electric vehicles called SRS. The technology is very similar to APS but the idea is for trams and buses to charge while idling at bus stops [48]. The charging concept is viewed in Figure 8b.
2.2.2.2 Wireless charging
Wireless power transfer can be divided into two categories; Inductive power transfer (IPT) and Resonant power transfer (RPT) [52]. The two technologies are similar since they are both based on electromagnetic coupling. Coupling can also be achieved using capacitors (Capacitive power transfer, CPT) but this so-lution is only suitable at short distances [38]. For a more thorough review of different wireless charging
technologies the reader is referred to [38].
In a wireless charging station, like the one outlined in figure 9, energy is transfered in the electromagnetic field between a coil buried under ground, the track, and another coil in the vehicle, the pickup-coil [53]. In order to achieve efficient energy transfer an AC current of high frequency is required (80-500 kHz) [35]. This is obtained by the use of high frequency inverters at the charging station. Power electronic devises in the vehicle converts the AC current to DC before charging the battery, as seen in figure 9.
Figure 9: In a wireless charging station AC power from the distribution grid is converted to a higher
frequency before being transmitted from the connector under ground to the pick-up coil inside the vehicle. In the vehicle the high frequency AC power is rectified and the battery charged. Controls in the charging station as well as in the vehicle ensures the battery is fully charged.
Wireless charging poses many advantages over conventional conductive charging; it is convenient for the user, there are no issues with charging in wet weather and the size and weight of the charger can be reduced [38]. Other advantages of inductive charging are the low visual impact as well as the possibility to use the same standard for all types of vehicles [35]. The grid interaction is a major drawback as well as the cost and safety issues concerning human exposure to high frequency radiation [35]. In addition, the power transfer efficiency is sensitive to misalignment [53]
Several pilot projects with static inductive charging of buses is or has been conducted; Scania (Sweden, 2016), Flaunder DRIVE (Belgium, 2011), City of Den Bosch (Netherlands, 2012), Bombardier (Germany, 2013), Dong Won Olev (South Korea, 2013) and Wrightbus (UK, 2014) to mention a few [9, 35]. As of 2014 there were seven companies providing inductive charging systems worldwide, Bombardier and Conductix-Wampfler being the only two with solutions for buses. Examples of companies producing buses with inductive charging systems are VCL (Netherlands), BYD (China), VanHool (Belgium) and Solaris (Poland) [35].
2.2.3 Grid connection
The grid connection is an important difference between conductive and inductive charging stations. While regular conductive charging points are located at the grid connection point, inductive charging points are separated from the grid connecting point since the coils are placed under ground [35]. Underground grid connection is being investigated to avoid separate installation [35].
Loads that use power electronics, such as chargers for electric buses, have a dynamic behavior. The internal control systems in the power electronic devises will make sure a constant power is supplied to the charger, which leads to increased current if the voltage drops. Locations in the grid with a lot of power electronics therefore have a higher risk of developing resonance issues. [54]. Lindberg [15] concludes that the fast charging stations for electric buses in Umeå emits harmonics that might affect the voltage quality for nearby customers.
A simplified image of the grid connection of a reversed pantograph fast charging station is shown in figure 10. The system consists of the primary substation, the 10/0.4 kV substation, service cables, the charging station with its rectifier and pantograph and the bus containing a battery. In addition, there are
systems for safety, control and communication between bus and charger. The input to the fast charger should be 3-phase 400 VAC and the output a DC voltage of 400-850 V.
The charging station owner can be either a high-voltage or a low voltage-customer, depending on if the connection point is 10 kV or 400 V, respectively. At the connection point, the ownership of the grid shifts from being the network operator’s to being the connecting customer’s. In a low voltage connection, the customer is responsible for the facility itself and the 400 V service. In a high voltage connection, on the other hand, the customer is also the owner of a 10/0.4 kV substation and a connecting 10 kV-cable in addition to a 400 V service. The connection fees to the network operator are different, depending on the connection voltage. They include the cost price of the necessary cables, a starting fee that depends on the subscribed current or power as well as monthly and yearly power and energy transfer fees.
Figure 10: The system, that consists of a primary and a secondary substation, the service cables, the fast
charger and the bus as well as control system, safety system and wireless communication. Input to the charger is 400 VAC and output is 400-850 VDC.
An energy storage system can be connected to the charging station as shown in figure 11. The idea of using an energy storage is to decrease the power demand from the buses. This is done by charging the storage when no bus is at the station, and later, when a bus arrives, help provide charge power. This way, the energy outtake from the grid is spread out over a longer time, which decreases the required power. The storage can be sized to provide the full power demand or only a portion. It can be designed to recharge fully in between each bus or make it so large that it does not have to recharge fully during the busiest hours.
2.2.4 Swedish examples
In Sweden, there are at present four different bus lines that are operated with opportunity charging. Fast charging stations are not owned or managed by the bus companies. Instead the charging service is procured by the municipality or local public transport authority and the stations owned and operated by a charging station operator. Table 1 below summarizes the opportunity charged Swedish bus lines.
Table 1: A summary of Swedish bus lines operated electrically with fast chargers at end stops. Most of
the information was gathered through interviews with grid companies, charging suppliers and charging station operators. Location Operator: network/ charging station Supplier: charger/ bus Charging
technology Voltage:input/output (V)
Power
(kW) Comment
Gothenburg Göteborg
Energi/? Siemens/Volvo Static conduc-tive, reversed pantograph
400/ 600 2x120 and 2x300
All fed from previously existing distribu-tion substadistribu-tions. One stadistribu-tion shares service with other consumers, the others have their own. [55]
Umeå Umeå Energi/ Hybricon
Hybricon/
Hybricon Static conduc-tive, reversed pantograph
400 / 700 300 and 650
The 300 kW station has its own transformer in a shared substation and is combined with a BESS of 120 kWh to decrease the load on the weak grid. The 650 kW station is fed by its own substation. [15, 56]
Södertälje Vattenfall/
Vattenfall Bombardier/Scania Static induc-tive, ground based
?/? 200 The bus line is a research project in coop-eration with KTH and was implemented in december 2016 [10].
Stockholm,
Ropsten Vattenfall/Vattenfall Siemens/Volvo Static conduc-tive, reversed pantograph
400/ 600 150 The bus line is evaluated between 2015 and 2017 [57].
2.3 Energy storage
2.3.1 Applications and technologies
The following grid benefits from energy storage has been identified:
• Balance the capacity Energy storage can be used to cover the few short periods when line capacity is exceeded, as an alternative to reinforcement of the grid [58].
• Level out load To even out the load and relieve the grid, energy storage can be charged from the grid when the load is low and help meet the demand when it is high [54].
• Improve voltage quality Energy storage can help mitigate harmonics, flicker, voltage oscillations, voltage drops and other voltage quality issues [54].
• Balance production from intermittent sources. [54]
• Increase hosting capacity The hosting capacity is a measurement of the amount of power from renewable resources that can be inputed to the grid. It is determined by evaluating various limiting factors such as grid thermal capacity and voltage drop. Energy storage can help increasing the level of renewables in the grid by improving its characteristics [54].
• Participate in frequency regulation The purpose of frequency regulation is to make sure the system frequency is kept at 50 Hz by instantaneously matching production and consumption. Since energy storage systems can be both producers and consumers they can assist in regulating the frequency [54].
• Back-up power Back-up energy storage increases the resilience in the grid [54].
• Minimize losses In each energy storage system there are internal losses, but energy storage can partially compensate these losses by reducing the grid losses that are due to variations in the transmitted power [54].
• Decrease grid fees A network operator can decrease its fees to superjacent grid owners by reducing grid losses and evening out the power outtake. Consumers can reduce their fees by decreasing their power consumption. This can be achieved by applying energy storage in the grid [59].
Various applications for energy storage are presented in figure 12 according to their respective time scale. The application determines the required time frame, power and energy rating, life-time etc. and thus the energy storage technology. According to a study conducted by Ei, the profitability of energy storage for network operators strongly depends on the amount of system benefits the storage system has as well as the alternative investments that are avoided [60].
Figure 12: The image collected from Elforsk [54] points out various grid issues and energy storage
applications. The logarithmic time scale indicates within what time frame each grid phenomenon or application take place.
Energy can be stored using a large variety of methods, classified according to the following [61, 62]: • Electrochemical: Batteries
• Mechanical: Pumped Hydroelectric Storage, Compressed Air Energy Storage, Flywheels • Electric: Superconducting Magnetic Energy Storage, Capacitors and Supercapacitors • Chemical: Hydrogen and other fuels
• Thermal: Low temperature thermal storage, High temperature thermal storage
Chemical and thermal storage will not be further considered in this work. Thermal storage has few power system applications and chemical storage is primarily suitable for long term storage [61]. Batteries will be described detail since they are the most common energy storage devices [62] and already are combined with EV charging stations. Characteristics of the remaining energy storage technologies are found in Appendix A5.
2.3.2 Battery Energy Storage (BES)
A battery has the ability to perform conversion between electrical energy and chemical energy through redox reactions at the electrode/ electrolyte interface [63]. The battery cell consists of two electrodes, the anode and the cathode, separated by an ionic solution, the electrolyte. When the battery is discharged, electrons flow from the anode to the cathode via a load while positively charged ions flow from the anode to the cathode through the electrolyte [62]. The voltage between the two terminals of a battery, when no current flows, is denoted open-circuit voltage (OCV) and it is the chemical potential difference between the two electrodes. When a current is drawn from the battery the voltage between the terminals is called closed-circuit voltage or terminal voltage, and it varies during the discharging process. Every battery has an internal resistance to current flow [63].
2.3.2.1 Battery charging characteristics
Figure 13: The image shows the charging process
of a lithium-ion battery [64]. The current stays con-stant until the nominal voltage is reached, and then decreases exponentially.
A primary battery is discharged once and then
disposed of whereas a secondary battery can be charged again by applying a current in the oppo-site direction [63]. The charing characteristics of a lithium-ion battery is shown in figure 13. The pro-cess of first charging and then discharging an en-ergy storage is called cycling. The state-of-charge (SOC) is a measure of how many percent of the total capacity a battery contains at the moment. 2.3.2.2 Battery capacity
The battery capacity is measured in Ampere-hours (Ah) and indicates the amount of charge that is stored. It can be described by Peukerts law [61]
Figure 14: Depending on the C-rate during
dis-charge, the battery delivers different capacities. The larger the C-rate (and thus the discharge current), the smaller the available capacity. The image shows the discharge characteristics of a Kokam 17-Ah/3.7-V lithium-ion battery [61].
Capacit y= Ik · td (7)
where the discharge current I is assumed to be constant throughout the whole course, and the dis-charge time tdis expressed in hours. k is the Peukert
constant, that is 1 for an ideal battery. The rate, at which the battery is discharged (and charged), may vary. It can be related to the maximum battery ca-pacity by the C-rate. At 1C, the current required to discharge the full capacity in an hour is used. At 2C the current is doubled and at 0.5C the current is halved. In an ideal battery, the discharge time would according to equation 7 be 1/C −rate hours. However, that is not the case. The bigger the C-rate, the smaller the available battery capacity, as illustrated in figure 14.
2.3.2.3 Battery chemistries
Characteristics for several common batteries are presented in table 2. Lead-acid is the most used battery chemistry and also the cheapest available [61]. Batteries in the lithium-ion family (for example lithium
iron phosphate, LFP, and lithium titanate, LTO) all have high energy as well as power densities and are thus favored in portable and vehicle applications. The NaS battery has the highest rated capacity out of all batteries. The materials are common and non-toxic, but require a high temperature (575-624 K) and is therefore expensive to operate [62]. Although NiCd batteries are robust and cheap to operate and maintain, the compounds are toxic. Also, the capacity of the NiCd battery is reduced substantially if repeatedly recharged without being fully discharged (so called memory effect) [62]. Less common battery chemistries used in vehicle applications are Nickel-metalhydride (NiMH) and sodium-nickel chloride (ZEBRA) [61]. Flow batteries differ from regular batteries in the sense that the energy is stored in the electrolytes instead of the electrodes. The electrolyte solutions are pumped to and from flow compartments and are oxidized and reduced at the electrodes [62]. Common flow batteries include vanadium-redox flow batteries (VRB), zinc-bromine (ZnBr) and polysulfide-bromine (PSB). Also flow battery characteristics are found in table 2.
Table 2: Characteristics for common batteries. The best performing battery for each category is marked
in bold. The cycle life implies 100 % DoD. When nothing else is noted, the numbers are collected from a review by Lou et.al. [62] and the conversion 1 SEK = 8.5 USD is used.
Chem. Density (Wh/L; W/L) Power rating (MW) Energy rating (MWh) Efficiency (cycle; charge; discharge) Life-time (cycle; years) Self-discharge (%/day) Discharge time @ rated power Cost (Invest SEK/kWh; Oper. SEK/kW/y)
Lead-acid 50-90;10-400 0.05-40 0.0005-40 63-90; 85; 85 500-1800; 5-15 0.1-0.3 seconds -hours 425-3,400;425 Li-ion 150-500; 1,500-10,000 0.1-100 0.004-10 75-90;85; 85 1000-20,000;5-15 0.1-5 minutes -hours 5,100-32,300(3,485 [65]); -NaS 150-300; 140-180 <34 0.4-244.8 75-90;85;85 2500-4500; 10-20 0 seconds -hours 2,550-4,250;680 NiCd 60-150; 80-600 <40 6.75 60-85;85;85 2000-3500;3-20 0.03-0.6 seconds -hours 3,400-20,400; 20 VRB 16-35; <2 0.03-3 <60 65-85; - ;75-82 12,000+; 5-20 small seconds -24+ h 1,275-8,500; 595 ZnBr 30-65; <25 0.05-10 0.05-4 65-80; - ;60-70 1500-2000+;5-10 small seconds -10+ h 1,2758,500; -PSB 20-30; <2 0.004-15 <0.06 60-75; - ; - - ;10-15 0 seconds -10+ h 1,2758,500;
-Battery storage systems have a broad variety of applications, such as power quality improvement, peak shaving, energy management, back-up power and transportation [62]. Applications most suited for batteries have discharge times from minutes to hours [54]. Flow batteries typically operate as daily storage, with longer discharge times than regular batteries [61]. It is important to notice that each battery chemistry has its own characteristics and therefore are suited for its specific applications.
2.3.2.4 Lithium-ion battery cost
A review by Nykvist and Nilsson [65] shows li-ion battery cost has decreased by 14 % each year between 2007-2014. According to the market study, the price was around 410 USD/kWh in 2014. Figure 15 shows past as well as estimated future cost of li-ion batteries (all different types).
Figure 15: The graph illustrates the cost development of li-ion batteries for BEV applications, based
on a review of publications, journals, reports, the market trends, expert statements and manufacturer’s estimates [65]. 150 USD/kWh is considered the point of commercialization of BEV.
2.3.2.5 Battery aging
Figure 16: The figure illustrates the relationship between
cycle-life and depth-of-discharge for a typical lithium-ion battery [66]. The number of cycles indicates how many cycles the battery can perform before the capacity is decreased to 80 % of the original. Shallow charge cycles significantly increases the battery life.
. A battery’s aging is determined by
oper-ating and storing temperature, depth-of-discharge (DoD), number of cycles and the charge/discharge current. High tem-perature, deep discharge, a large number of cycles and a high discharge current all contribute to capacity fading and de-crease of battery life [67]. The DoD is defined as the amount (in percent) of the total battery capacity that has been dis-charged by the time charging starts again. At 100 % DoD, the battery is completely emptied at every discharge. Figure 16 illustrates the relationship between the DoD and the number of cycles a lithium-ion battery can perform before 20 % of its capacity is lost. It is in accordance with similar graphs presented in [18] and [68].
3 Method and data
When performing a distribution system design assessment, Lakervi and Holmes [26] suggests the following three main tasks, that are the basis for the execution of this study:
• Determination of a technically feasible solution. • Estimation of cost per unit for all components. • Conversion of costs to make them comparable.
3.1 Technical feasibility
3.1.1 Electric bus system
The public transportation company VL has provided information on the following design parameters: maximum bus energy consumption, bus frequencies, time at the end stations, total drive time on the routes and route distances (table 3). The power and energy requirement at each end station is based on the route distance and the scheduled stops as well as the energy consumption for the buses. Each time the bus arrives to the charging station it is assumed to require the amount of energy it consumes during the route (one way). This is calculated through equation 8 where k is the energy demand in kWh/km and l the route distance in kilometers.
E = k · l (8)
VL estimates the dimensioning energy demand to be 1.6 kWh/km and 2.3 kWh/km for a 12 m and a 18 m bus, respectively. In this study, the figure that refers to the largest bus type is used. It is assumed that the charging efficiency of the bus battery is 85 %. Table 3 lists the demanded charging power that is based on the charge energy and scheduled time at the end station.
Table 3: The table shows the power demand during one charge for each bus (18 m) as well as route
characteristics.
Bus End station Distance
(m) Total routetime (min) Bus frequency(buses/hour) Time at endstation (min) Required power(kW)
1 Bjurhovda Skälby 1276512727 44 8 62 3451032 2 Björnön NorraGryta 1410814019 44 5 46 573379 3 Erikslund Airport 1393014293 45 6 108 226291 4 Brottberga Finnslätten 1216112264 39 3.5 47 494285 5 Tunbytorp Hälla 1281512848 40 4 63 347695 6 Rönnby Finnslätten 1192011876 36 3 710 276285 7 Centralen Hacksta 60066018 14 2 11 975976
