Benchmarking of Smart Grid Concepts in Low-Voltage Distribution GridsODILIA BERTETTI

(1)

IN

DEGREE PROJECT TECHNOLOGY AND MANAGEMENT, SECOND CYCLE, 30 CREDITS

STOCKHOLM SWEDEN 2017,

Benchmarking of Smart Grid Concepts in Low-Voltage Distribution Grids

ODILIA BERTETTI

KTH ROYAL INSTITUTE OF TECHNOLOGY

(2)

Benchmarking of Smart Grid Concepts in Low-Voltage Distribution Grids

Odilia Bertetti

Supervisor and Examiner: Anders Malmquist

Master of Science Thesis EGI_2017-0076-MSC EKV1202 MSc in Sustainable Energy Engineering

Division of Energy Technology

KTH Royal Institute of Technology SE-100 44 STOCKHOLM

20

^th

September 2017

(3)

Abstract

Due to increasing penetration of decentralized variable renewable energy generators and the increasing demand of electrical power due to the electrification of the heat and transport sectors, low voltage grids are facing critical problems. Deviation of the permitted voltage range and local overloads of the grid equipment, are the two main issues that are compromising a smooth distribution grid operation. An intelligent integration of distributed generators, heat-pumps and electric vehicles into a Smart Grid, allows the flexibility that they intrinsically provide, to be used by distribution system operators to avoid critical grid conditions.

Smart grid suppliers currently available on the market, have been categorized into Local, Decentralized and Centralized Smart Grid Concepts. Their main difference is represented by the level of control, communication and coordination that they make use of. The aim of the thesis was to evaluate the effectiveness of solution of the Smart Grid Concepts implementation in specific low voltage grids, especially in term of voltages and loadings mitigation capabilities, to be used as a decision making tool for future smart grid imple- mentations.

A control architecture that emulates the way the analyzed Smart Grid Concepts operate, has been implemented in Python and tested on three different low voltage distribution networks in DigSILENT PowerFactory. The control architecture is an algorithm that communicates to DigSILENT PowerFactory how the Smart Grid needs to operate in response to detected critical grid conditions. The flexibility that the Smart Grid Concepts make use of, are battery storage, active power curtailment and reactive power compensation from photovoltaic inverters and demand side management by means of electric vehicles and heat pumps. In particular, in order to make most use of the available flexibility, an intelligent electric vehicles charging strategy has been implemented as well as an intelligent heat pump operation.

Both static worst-case simulations and time-dependent simulations, over a winter and a summer day, for different penetration scenarios, have been carried out. The summary of the simulation results showed that while the Decentralized Smart Grid Concept, if the flexibility is available, is always able to keep voltages and loadings between their critical values, the Local Smart Grid Concept is not able to do the same for the loadings.

Keywords: Smart Grid, Grid Integration, Distributed Generation, Electric Vehicles, Photovoltaic, Battery Storage, Heat Pump, Flexibility, DigSILENT PowerFactory.

(4)

Sammanfattning

På grund av ökad penetration av decentraliserade variabla förnybara energikällor och den ökande efterfrågan på elkraft på grund av elektrifiering av värme- och transportsektorn, står lågspänningsnätet inför kritiska problem. Avvikelse av det tillåtna spänningsområdet och lokala överbelastningar av nätutrustningen är de två huvudproblemen som äventyrar en smidig nätdrift. En intelligent integration av distribuerade generatorer, värmepumpar och elektriska fordon i ett smart nät, tillåter flexibiliteten som de egentligen tillhanda- håller, för att undvika kritiska rutnätförhållanden.

Smartnätleverantörer som för närvarande är tillgängliga på marknaden har system som kategoriserats som lokalt, decentraliserat och centralt Smart Grid Concepts. Deras hu- vudsakliga skillnad representeras av den nivå av kontroll, kommunikation och samordning som de utnyttjar. Syftet med avhandlingen var att utvärdera effektiviteten av lösningen av implementeringen av Smart Grid Concepts i specifika lågspänningsnät, särskilt när det gäller spänningar och belastningsreducerande förmågor, som ska användas som be- slutsverktyg för framtida smarta nätverksimplementeringar.

En reglerarkitektur som emulerar hur ett analyserat Smart Grid Concepts fungerar, har implementerats i Python och testats på tre olika lågspänningsdistributionsnä i DigSI- LENT PowerFactory. Kontrollarkitekturen är en algoritm som kommunicerar med DigSI- LENT PowerFactory hur Smart Grid bör fungera som svar på detekterade kritiska grid- förhållanden. Den flexibilitet som Smart Grid Concepts använder sig av är batterilagring, aktiv strömavbrott och reaktiv effektkompensation från fotovoltaiska omvandlare och ef- terfrågesidan hantering med elbilar och värmepumpar. I synnerhet för att på bästa sätt utnyttja den tillgängliga flexibiliteten har en intelligent laddningsstrategi för elfordon implementerats liksom en intelligent värmepumpsoperation.

Både statiska wärsta fall simuleringar och tidsberoende simuleringar, över en vinter och en sommardag, för olika penetrationsscenarier har utförts. Sammanfattningen av simu- leringsresultaten visade att medan det decentraliserade Smart Grid Conceptet, om flexibiliteten är tillgänglig, alltid kan hålla spänningar och belastningar mellan sina kritiska värden, kan det lokala Smart Grid Concepts inte göra samma för belastningarna.

Nyckelord: Smart Grid, Nätintegration, Distribuerad Generation, Elfordon, Solceller, Batterilagring, Värmepump, Flexibilitet, DigSILENT PowerFactory.

(5)

Acknowledgments

This thesis is the conclusion of my Master Programme education in Sustainable Energy Engineering at KTH Royal Institute of Technology. I am very thankful to the professors and coordinators that made it possible. KTH not only gave me a strong theoretical and practical background, but also it constantly provided external stimuli from industries and professionals to link the bridge between academic and professional life.

The thesis project has been carried out in Energynautics GmbH, a company located in Darmstadt, Germany, specialized in variable renewable energy grid integration studies.

I would like to thank Eckehard Tröster for his valuable guidance and support throughout these six months and for ideating the topic of the thesis. I would also like to thank Thomas Ackermann for trusting in me and giving me the great opportunity to work on my thesis in Energynautics. Finally I would like to thank all my colleagues in the company for their support, friendship and for creating an enjoyable working atmosphere. Working on the thesis in Energynautics has been a memorable working experience through which I have aquired important tools which will be essential for my working life ahead.

Finally I would like to thank my supervisor at KTH, Anders Malmquist, for his time and patience in supervising and guiding my thesis report.

(6)

List of Symbols and Abbreviations

aBox Actuator Box

ADN Active Distribution Network AMI Advanced Metering Infrastructure ASHP Air Source Heat Pump

AVR Automatic Voltage Regulator BESS Battery Energy Storage System BEV Battery Electric Vehicle

CHP Combined Heat and Power CIS Customer Information System COP Coefficient of Performance cos ϕ Power Factor

Cp Specific Heat db Deadband

DER Distributed Energy Resources DSM Demand Side Management DSO Distribution Sustem Operator EU European Union

EV Electric Vehicle

FCEV Fuel Cell Electric Vehicle

(7)

GHG Greenhouse Gases G2V Grid to Vehicle

GSHP Ground Source Heat Pump H_T Transmission Losses

H_V Ventilation Losses

HEV Hybrid Electric Vehicle HP Heat Pump

HV High Voltage I Current

ICE Internal Combustion Engine

ICT Information and Communication Technology IEC International Electro-technical Commission kW kilowatt

kWh kilowatt-hour LV Low Voltage

mBox Measurement Box MV Medium Voltage

OLTC On Load Tap Changer OMS Outage Management System p.u. Per Unit

PCC Point of Common Coupling PF Power Factor

PHEV Plug-in electric Vehicle PV Photovoltaic

Q_HL Heat Losses Q Thermal Energy˙ R Resistance

RES Renewable Energy Sources

(8)

S Apparent Power [kVA]

sBox Intelligent Substation

SCADA Supervisor Control and Data Acquisition SE State Estimation

SG Smart Grid SOC State of Charge

STATCOM Static Compensator T Temperature

TES Thermal Energy Storage U_n Nominal Voltage

V Voltage

V2G Vehicle to Grid

VRES Variable Renewable Energy Sources Wp Watt Peak

WAMS Wide Area Measurement System X Inductance

Z Impedance

η_ch Charging efficiency

(9)

CHAPTER 1 Introduction

1.1 Background

To combat global warming and to tackle the fossil fuel scarcity, a shift to less carbon intensive resources in energy generation and supply is essential. The power sector plays a central role in the decarbonization of the energy sector, being it the largest source of energy related green houses emission (GHG). Today in fact, in Europe, 40% of the electricity demand is met through fossil fuel-based power generation such as coal, oil and gas [53]. The power sector offers the possibility to efficiently use renewable energies and therefore provide carbon-free electricity from renewable resources. As part of the effort to reduce GHG emissions from the energy sector, the European Union’s (EU) energy strategy envisions a wide scale deployment of low carbon electricity generation from renewable energy sources (RES) [29]. According to the current EU Climate and Energy Package the share of electricity generation from RES should increase from 20% in 2010 to 30-45% in 2020 with a target of 55% in 2050 [31]. Beside the increase of renewable energy generation, other targets include a 20% reduction of greenhouse gas emission and a 20% improvement in energy efficiency [31].

Today the main RES contributor to electricity generation is hydro electrical generation [30]. However in Europe its potential has already being well exploited and there are limited opportunities for a further development [30]. Meeting European RES targets will require the development of alternative sources of RES generation such as wind, solar and biomass [31]. In this context, the European RES strategy will be largely based on the development of wind and solar generation, which are defined as Variable Renewable Energy Sources (VRES). Starting from an energy share of 10% of electricity demand in 2014, the share of VRES in the European Mix is expected to reach 20% in 2020 and around 30% in 2030 [2].

The political will to realize such a change can be observed in several countries: in particular Germany has witnessed an unprecedented increase in volatile renewable generation

(15)

capacity during the last decade, incentivized by the Erneuerbare-Energien-Gesetz (EEG) that has heavily subsidized renewable generation[23]. Due to the incentives designed to accelerate investment in renewable energy, the installed capacity of wind power and solar photovoltaic (PV) surpassed a combined installed capacity of 88 GW [58] (Figure 1.1).

In particular, small and medium-scale systems of less than 30 kilowatt-peak (kWp) have rapidly emerged during the last decade. As a result, about 70% of the installed PV capacity is connected to the low-voltage (LV) grid (Figure 1.2). Germany’s goal is to transform its electrical energy supply to one that is based on a renewable energy share of more than 80% by 2050 [35]. The so called Energiewende, or energy transition, which among other points mandates the phase out nuclear power by 2022, will be accompanied by high PV and wind power penetration in certain distribution grids [35]. These high-penetration scenarios will create challenges for existing grids and thus bring a demand for advanced control concepts to guarantee reliable and cost-efficient future grid operation.

Figure 1.1: Wind and Solar Total Installed Ca- pacity in Germany[58]

Figure 1.2: Voltage Level Con- nection of RES Capacity in Ger- many[12]

As generation shifts to increased RES, electrification creates further environmental bene- fits by shifting many end users of electricity such as transportation and heating, that now contribute to the 6% and 22% respectively of the total GHG emission, away from fossil fuel sources [53]. In the residential sector, heat pump (HP) technology is becoming more adopted due to the high efficiency through which the process is carried out, which will increasingly replace commonly used gas-fire boilers to satisfy the heating requirement of the buildings [17]. In the transportation sector hybrid and electric vehicles (EV) car will replace traditional oil fired internal combustion engine (ICE) [46]. Moreover if a large share of electricity generation in the national power system comes from RES, the electricity used to power the heat pumps and electric vehicles is almost carbon free, allowing the decarbonization of their respective sectors.

It is therefore clear that the energy system is undergoing a wide transformation. Firstly the decentralization of the power system which has been spurred by policies and by a sharp decrease in manufacturing cost. Secondly the electrification of large sectors of the economy such as heating and transport. Those two transformation make customers active elements of the system and require significant coordination, which is leading towards the electrification of the power system with smart metering, smart sensors, automation and other digital network technologies [46].

(16)

Figure 1.3: Three trends of the grid edge transformation [46]

1.2 Motivation

The changes of the future distribution grid, which include enormous numbers of decentralized power generation units and increasing demand of electrical power due to heat pumps, electric vehicles and similar high power loads, poses challenges to the low voltage grids in term of [87][14]:

• Deviations of the permitted voltage range (±10% of the rated voltage according to EN 50160)

• Local inner overloads of the grid equipment, especially of the power cables and transformers

The power system was originally designed to transport energy from large power plants to the consumers, in a top down approach [86]. Now large power plants are being replaced by a multitude of small generators widespread over the whole country which produce electricity based on local meteorological conditions [78]. As a result of the changes in the distribution grid, critical situations like a complete inversion of power flow started being more common especially in moments of the day when there is a peak electricity production from the PV systems and low requirement of electricity from loads [78].

As long as only few and small DG units are connected to the distribution grid, the load is still prevailing and the power injection from the DG units only reduces the total network load, decreasing low voltage issues. However, when the penetration of DG is large and the power flow is reversed, over-voltages and over-loadings become an issues for distribution grids. In the German power system around 41 GWp of PV have been connected to the grid until the beginning of 2017 [3]. The largest part of these generation units (70%) are connected to the low voltage distribution system, where voltage rise and overloading of grid components, caused by the injection of active power, are already an issue [12]. A total installed capacity of 80 GWp is expected by 2030, which will increasingly accentuate those issues [3].

(17)

Due to the electrification of the residential and transportation sector moreover, new high-power loads such as electric vehicles and heat pump will play an important role at household level [15]. They will draw high amount of power and create issues to the grid in term of under-voltages and overloading of grid components, especially is some particular moment of the day when the coincidence factor of utilization of such loads is particularly high [15] with simultaneous no generation.

For those future requirements the distribution grid have never been planned and con- structed for. The way the distribution system operators (DSOs) have distributed energy and designed their grids up until recently, has been according to a top-down approach [75]. Their primary role was to deliver energy flowing in one direction, from the transmission substation down to end users [75]. This approach makes use of very few monitoring tools and is suitable for distribution networks with predictable flows. DSO’s are responsible for assuring security of operation of the power system which include maintaining the voltage fluctuations on the system and the loading of grid components within given limits [75]. As a result, DSOs with high shares of DG in their grids already face challenges in meeting some of their responsibilities and these challenges are expected to become more frequent. To integrate safely increasing amount of decentralized and often fluctuating generation and for allowing a further electrification of the heat and transportation sector, to guarantee a stable operation of the grid, two solutions are available [55]:

• Massive investment in additional low voltage equipment, such as new transformer and cables

• Moderate investment to introduce a higher level of automation in distribution networks, by means of Smart Grids.

Investments in grid equipment are cost-intensive, especially if underground cables are used [55]. Such networks have in fact to be designed to tackle worst cases situations such as maximum load/minimum generation and minimum load/maximum generation by dimensioning the lines and other network equipment to fulfill the grid requirements.

However it is difficult to forecast the necessary level of grid enhancement due to the increasing and not totally predictable integration of decentralized generation units [55].

In addition, the regulatory requirements for DSO necessitate more detailed documenta- tion of reliability of supply and power quality of the generation units in their networks [42]. Therefore, DSOs will need more information and control of the state of their distribution grids especially on the LV and MV level where the usage of monitoring and control devices is currently very limited [44]. This represents a big challenge for DSOs as most of the PV and Wind Farms are connected to these voltage levels.

Therefore the solution is being adopted especially by German DSO, for integrating fluctuating renewable energy and high-power loads into the distribution grid and prevent voltage congestion issues, is the introduction of a higher level of automation in their network by means of Smart Grids [38]. This allows the usage of flexibility provided by the distributed energy resources (DERs) by integrating information and communication

(18)

technology with the electricity infrastructure into a Smart Grid (SG).

From a technical perspective, three types of Smart Grids can be found on the market based on their level of control, communication and measurement infrastructure and con- sequently on the way they react to critical grid status [12]:

• Local Smart Grid Concept do not require a communication infrastructure therefore it can be easily integrated into the overall grid operation. The control is done based on the local measurement sensed at the point of common coupling (PCC) where the control intelligence is installed, according predefined parameters and droop functions. The local control intelligence has no information about the status of the grid in other parts of the network and about how the other control intelli- gences are reacting to the grid status.

• Decentralized Smart Grid Concept is a communication based control where the exchange of information is a crucial element. The control intelligence located at the secondary substation, gets information from independent sensors installed in critical points of the LV network. The control is achieved via the coordination of several active system components clustered into different areas.

• Centralized Smart Grid Concept has its control intelligence located at the primary substation: all the functions provided by the distribution management system (DMS) are performed in one central location. It is also a communication based concept however its communication and control are over all the LV, MV and HV network in the area where the control intelligence in installed. It makes use of wide area measurement system (WAMS) or advanced measurement infrastructure (AMI) to enable the control intelligence to continuously monitor and optimize the grid operation.

Figure 1.4: Overview of Smart Grid Concepts[12]

(19)

The three Smart Grid Concepts have different ways of reacting to a critical grid status, based on different level of information available, due to the different extent of measurement and communication infrastructures they make use of. Local, Decentralized and Centralized Concepts have respectively a low, medium and high level of information about the grid conditions. The more information are available to the control intelligence, the more it will be able to detect and respond to critical grid situations [13]. The Cen- tralized Concept, with its WAMS and related outage management capabilities, represents the solution with the highest degree of network security. The Local Concept, instead is the weakest to provide network security due to its inability to detect critical gird conditions, with the exception of the terminal where the local control intelligence is installed.

However, the wider the measurement and communication infrastructure are, the more costly and complicated to implement the the Smart Grid solution gets [13]. The Local Smart Grid Concept can be implemented quickly as a plug-and-play solution and only requires a small initial investment. The Centralized Smart Grid Concept instead, due to its wide area control and communication, is more costly and complicated to integrate into existing distribution grids. The selection of the Smart Grid concept to invest on in a particular network topology, is therefore a compromise between effectiveness of solution provided by the Smart Grid concept in term of ability to relieve the grid from critical condition, and its related investment

As the main objective of the thesis is to analyze if any Smart Grid Concept presents some weaknesses in term of ability to react and mitigate a critical grid condition, the focus of the research is the Local and Decentralized Concepts. The Centralized Smart Grid Concept has been kept out from the research because due to its wide-area monitoring and communication infrastructure, it does not present weaknesses in term of detection and therefore mitigation of critical grid conditions.

In particular in this research, the measures that the Smart Grid Concepts make use of to react to critical grid status, are the flexibility means available at household level. They include battery energy storage systems (BESS), active power curtailment and reactive power compensation from PV inverters and demand side management (DSM) by means of EVs and HPs.

A variation of the Local Concept, the Local Reactive Concept, which has additional reactive power capabilities has been examined, as well as a variation of the Decentralized Concept, with additional on-load tap changer (OLTC) installed at distribution transformer, the Decentralized OLTC Concept. The two main concepts and their variations will be further discussed in Chapter 4.

1.3 Research Question

The aim of the thesis is to evaluate the effectiveness of solution of various Smart Grid Concepts, especially in term of voltages and loading mitigation capabilities. It is intended to be used as a decision making tool for selecting the most suitable type of Smart Grid

(20)

to invest on in a particular network topology.

The thesis will:

• Investigate the voltage and loading issues in future distribution networks.

• Analyze the behavior of the analyzed Smart Grid Concepts in response to critical grid condition.

More specific research questions being addressed are:

• If the measurement infrastructure that the local Smart Grid Concept has, are enough to detect a critical grid status and if the response of the local control based on local measurements is always "Smart".

• If there is a limit of PV/HP/EV penetration after which the control from the Smart Grid concepts do not mitigate the network conditions anymore and grid reinforcement need to be done.

• In which cases the Decentralized Smart Grid Concept works better than the Local Smart Grid Concept.

• If a local control with the additional reactive power capabilities brings any substan- tial advantage/disadvantage.

• In which cases is worth the investment of an OLTC in the Decentralized Smart Grid Concept.

1.4 Literature Review

In the studies [40] and [68] carried out by DG demo Netz active voltage control strategies to increase the hosting capacity of distribution grid have been implemented and tested.

They have been divided into "Local Voltage Control", "Distributed Voltage Control"

and "Coordinated Voltage Control". The objective was to estimate the hosting capacity as a result of each analyzed control strategy. The result was that coordinated voltage control (the equivalent of the "Centralized Smart Grid Concept" in this thesis) had the same effectiveness as grid reinforcement, allowing a penetration of distributed generation capacity equal to 90% of the peak load . Distributed voltage control (the equivalent of the "Decentralized Smart Grid Concept" in this thesis) instead allows for 80% while local control for 68% of the peak load.

Source [49] has compared local and coordinated control of reactive power compensation to solve voltage issues: while in the first case PVs responds to local measurements, in the second the reactive power provision happens in a decentralized manner, both the strategies were successful for keeping voltages between limits. In the study carried out by [1] and [49] the local control strategy based on several reactive power controls for voltage support has been implemented. The outcome of the studies was that local voltage support by means of reactive power from PV has the ability to maintain the voltage within

(21)

operating limits in different distribution grids.

Sources [67], [26], [36] and [110] have analyzed the decentralized voltage control by means of PV inverter based on optimization parameters varying from line losses minimization to active power curtailment minimization and voltage profile flattering maximization. In all the cases the control strategies were able to keep the voltages between acceptable limits.

Several studies have investigated the optimal battery energy storage integration to solve voltage issue mitigation in LV grids. Sources[102] and [9] have investigated voltage control strategies using central storage systems while [109]and [108] analyzed a decentralized energy storage system by means of local voltage control.

Voltage support by means of heat pumps’ thermal energy storage (TES) system in distribution grids with high PV penetration,through demand side management (DSM) have been analyzed in [25], [62] and [15]. Results show that an effective integration can reduce the PV-induced voltage rise issues.

A number of studies have analyzed the impact that large penetration of EVs have at low voltage level, in term of overloading and undervoltages and have implemented different types of control strategies. Several papers ([50][98])have focused on centralized scheduling and control. These studies found out that the feeder load profile can be flattened, voltage violations can be reduced and transformer loading can be reduced. Other sources ([47],[5]) have focused on decentralized optimization of EV charging which require reduced communication infrastructure. Some others ([7], [92],[89])) have implemented a local voltage control strategy that regulates the charging process based on local information to avoid local voltage violations.

1.5 Methodology

The methodology that has been adopted to address the research questions is as follow:

1. An analysis of the Smart Grid providers currently available on the market has been carried out. They have then been divided into Local, Decentralized and Centralized Smart Grid Concepts.

2. Thee low voltage network topologies have been built in DigSILENT PowerFactory and scenarios in term of PV, HP and EV penetration, in the range between 20%

and 100%, have been created.

3. Worst Case scenarios in term of simultaneous combination of PV infeed, load and voltage at the secondary substation - due to either a wind farm or an energy- intensive industry located in the MV in the proximity of the distribution transformer - have been created.

4. The Local, Local Reactive, Decentralized and Decentralized OLTC Smart Grid Concepts have been simulated in Python through algorithms that emulate as close as possible to reality how the Smart Grid Concept would operate. The algorithms

(22)

communicate with DigSILENT PowerFactory when control from the Smart Grid is needed.

5. The analyzed Smart Grid Concepts has been applied to the three network topologies for 8 worst case scenarios, which study the most critical conditions that the grid could have to face and react to.

6. The flexibility the Smart Grid Concepts make use of when control is needed in the time-dependent analysis, has been modelled: Electric vehicle intelligent charging strategy and control, Heat Pump intelligent operation and control, reactive power (Q) control from PV inverter and active power (P) control from Battery Energy Storage Systems (BESS).

7. Time-dependent simulations of the analyzed Smart Grid Concepts for 24 hours period in a Winter and a summer day in the three network topologies, has been carried out.

8. The analyzed Smart Grid Concepts have been compared and graded in term of ability to mitigate voltage and loading violation, of active and reactive power losses and necessary PV curtailment.

The presented methodology has been selected based on the analysis of publications that concern the topics of this thesis, which have been presented in Section 1.4. Due to the fact that the thesis has dealt with a wide range of topics (i.e. smart grid architectures, battery storage, reactive power compensation, flexibility from HPs and EVs), and given the lack of established methodology to address the research question of the thesis, separate methodologies for each area of the research has been adopted. Each part that has been modelled in the thesis, follows the methodology of certain publications in that particular research area (i.e. how heat pump flexibility have been modelled is based on previous studies which have modelled in detail the flexibility from heat pumps), which has then been included in the main methodology.

1.6 Thesis Structure

Chapter 1 gives an overall introduction to the thesis. The research motivation and ob- jectives are presented, as well as the adopted methodology and the thesis structure.

Chapter 2 starts by giving a description of the distribution grid architecture, then the new actors (DG, HP and EV) in the distribution grid are presented, together with the issues related to their high penetration in the LV network. Current regulations are then described.

Chapter 3 describes the solutions than can be adopted to mitigate the impacts that the new actors have at distribution grid level with a focus on flexibility means in Active Dis- tribution Network (ADN).

(23)

Chapter 4 describes the three Smart Grid Concepts and their representative smart grid supplier. The description if the implementation of the smart Grid Concepts is then presented.

Chapter 5 describes how the flexibility from EV, HP and EV has been modelled and integrated into the each Smart Grid Concept.

Chapter 6 describes the network topologies where different penetration scenarios have been applied, which have been used as testbeds for the implementation of the Smart Grid Concepts.

Chapter 7 presents the results for the static worst-case and time-dependent simulations as well as the final comparison between control strategies in the static and dynamic case.

Chapter 8 presents the summary, key findings, recommendation and main conclusions of the thesis. The topics for future work are also discussed.

(24)

CHAPTER 2 Impact of Distributed Generation, Heat Pumps and Electric Vehicles on Distribution Networks

2.1 Introduction

Power systems were traditionally planned and designed by assuming unidirectional power flows from power stations to loads [86]. Nowadays, several factors have led to a situa- tion where significant small-to-medium scale renewable generation capacity have been already connected to the distribution networks, giving rise to technical issues such as bi-directional power flows.

Meanwhile, also the loads are changing: new loads like EVs and HPs, deriving from the electrification of the residential and transport sector, are appearing in the network and they are transforming the electricity consumption pattern and power [74].

In this chapter the architecture of the traditional power system, with the focus of MV and LV distribution grid is presented, together with the challenges that new actors in the power system create and the regulatory boundary condition that applies to them.

2.2 The architecture of the Modern Power System

The purpose of the electric power system as a whole is to generate electric power to be delivered to the users. This rather complex system is formed by different sub-systems which can be divided into electric power generation, transmission, distribution and utilization [19]. As these sub-systems operate at different voltage levels, transformers or transformation substations are fundamental network components where the voltage level is adjusted accordingly to their required value [19]. Figure 2.1 shows a schematic representation of the electric power system with its different sub-systems connected by different step-up and step-down transformers. The distribution subsystem and in particular the

(25)

LV network including the distribution transformer, is the object of this thesis and represents the boundary of the study. Moreover, the impact that the MV grid has on the LV grid in term of voltage deviation, has been taken into consideration. This depends on the distance between the primary and secondary substation, and it also depends on the the proximity of generation plants or energy-intensive industries to the distribution transformer. Medium Voltage and Low Voltage grids are described below.

Figure 2.1: Overview of the modern power system and boundary of the study definition

2.2.1 Medium Voltage Distribution Network

Medium Voltage distribution networks are the connection between the high voltage network (10 kV or 20 kV but it can vary up to 110 kV depending on the country) and the low voltage network (0.4 kV) [19]. Distributed generation (DG) up to 10 MW, such as wind turbines, combined heat and power (CHP) and large scale PV installations are connected here, as well as large users such as industrial utilities [20].

For the lines in MV distribution networks, several different line types are being used, namely overhead lines and underground cables, both common in MV networks. Usually overhead lines are adopted in rural areas, while underground cables are more common in urban areas [20]. The voltage drop characteristic of overheads line is usually much higher than in underground cables due to the fact that for the same power delivery their diameter is smaller [19].

Transformers at the primary substation are usually equipped with an automatic OLTC to compensate the voltage deviation in the MV networks. The secondary substations are

(26)

sometimes equipped with an off-load tap changer compensate for voltage deviations in the LV networks, however these tap changers have only limited positions and have to be adjusted manually [106].

2.2.2 Low Voltage Distribution Network

Low voltage distribution networks receives the power from the MV/LV distribution transformer, which steps down the voltage to values low enough to deliver electricity safely to small consumers and households [19]. LV grids in Europe usually consists on a three- phase connection with a voltage of 400 V. Small scale DG units are in general connected to low voltage distribution networks and their size ranges from very small units of some hundreds of watts, up to some hundreds of kilowatts [52]. Typical types of DG connected to the LV network are usually installed at a residential level and include photovoltaic plants, small scale CHP and small scale wind turbine.

Traditionally, also the low voltage distribution network was planned and built for an unidirectional power flow from the secondary substation to the customers. By now, low voltage distribution networks are quite passive which means that there is usually no voltage control or measurements behind the primary substation. They are dimensioned to deal with maximum load and still maintaining a sufficient voltage level at the customer connection point.

Urban and rural distribution networks have different characteristics over which distributed generation and loads have different impacts [77][95]:

• Rural networks are characterized by small transformers, overhead lines with limited cross sections and long distances. They usually have long feeders, which experience a higher voltage drop across the lines compared to urban feeders. These features therefore result in a limited hosting capacity.

• Urban networks instead, are characterized by large transformers, underground cables with large cross section and short distances. The impact that DG and load have on the voltage profile is therefore less severe than in rural distribution networks.

In this thesis, the impact in urban, suburban and rural distribution grid has be analysed, however the focus has been on suburban and rural networks as they represent the network topologies which experience the most severe voltage and loadings violations.

Another distinction that can be done regarding LV distribution grids is between radial and meshed topologies configuration [61][41][111]:

• Radial topology is the simplest and less expensive, and is often used in rural distribution networks. In this configuration there is only one way for electricity to flow from the transformer to all nodes of the grid. Its main drawback is that, in case of failure in any point of the network involving cables, transformers or generators, all the customers connected to the branch would be subjected to a power outage.

Automatic switches can be used along the network in order to reduce the duration

(27)

of the interruptions and to allow unaffected sections to remain in service in case of temporary failure. A simple radial topology is shown in Figure 2.2.

• Meshed topology provides a higher level reliability, however it requires the use of redundant network equipment, which make this solution more expensive to implement. In its most basic form, two feeders form a closed loop, open at one point.

In case of a failure, by opening or closing the switches the configuration of the grid can be changed. The benefit of this structure is the additional flexibility provided in load flow; if a fault occurs, there will always be a way to supply part or all the consumers by isolating the fault area. During normal operation if the switch is closed moreover, the electricity can decide in which branch to flow, depending on where it meets lower resistivity.

Figure 2.2: Radial vs Meshed Network Topologies

Radial topologies correspond to the most common LV network configuration in rural and suburban grid, therefore they have been analyzed in this thesis.

2.3 Challenges of the Future Grid

The increasing amount of DG and high-power loads installed in the LV and MV grid poses difficult challenges to DSOs that need to guarantee security of operation of the distribution networks they are responsible of. In distribution networks, voltage issues and network equipment loading, are the main challenges concerning the integration of distributed generators and high power-loads [90]. Dynamic issues derived by the integration of DG and new loads in the distribution grid, such as network frequency, transients, flickers, harmonics and unbalances with a dynamic nature, also affect the power system, but they are outside the scope of this thesis.

Hereafter the impacts that DG and high power loads have at distribution level are presented.

2.3.1 Violation of Voltage Operational Limits

All equipment connected to the power utility system is designed to be used within a certain voltage range. Since all consumers have appliances of the same standard of service

(28)

voltage, it is necessary to provide them almost the same service voltages. In case the voltage limits are violated, customer facilities and network components are jeopardized [90].

Because voltage drop exists in each part of the system, the consumer who is electrically furthest from the substation always receives the lowest voltage. In case of no DG connected, the voltage along the feeders in fact varies from a maximum value at the customer closest to the substation, to a minimum value at the customer at the end of the feeder [90].

When integrating high-power loads in the LV grid, the extent of the voltage variation between the secondary substation and the last customer in the feeder will be more significant than in the case where just traditional loads are connected, which may lead to severe under-voltage situations [74]. The extent of the voltage violations depends on the coincidence factor of the high-power loads, which is the ratio of the simultaneous demand of the loads in the system to the sum of their maximum demand. The larger the load coincidence factor is, the larger the under-voltage violations in the network will be [73].

When integrating distributed generation to the LV grid instead, opposite voltage violations may occur. The injection of active power increases the voltage at the point of connection where the PV is installed if there is not enough local demand to consume the generated electricity [69]. In this case the voltage will locally increase and may lead to violations of the upper voltage operational limits [69]. The highest voltages in the case of high PV infeed, are always experienced at the end of the feeder; the extent of the over- voltages therefore depends on the relative position of the PV in the feeder. This voltage violations jeopardize customer facilities and network components and violate obligatory standards on power quality[103].

Figure 2.3 shows the voltage profiles in the two opposite cases of combination of maximum power generation and minimum loads vice versa.

Figure 2.3: Voltage profiles along a feeder in two combinations of load and generation.

(29)

2.3.1.1 Demonstration of Steady State Voltage Variations in Distribution Networks

The amount of voltage drop in a distribution network can be calculated from the analysis of the single-line diagram of Figure 2.4 and is given by:

U₂ = U₁+ I(R + jX) (2.1)

where I is the phasor representation of the current flowing through the feeder, R and X are respectively the line resistance and reactance, U₁ is the voltage at the secondary substation and U2 is the voltage where the load is installed.

Figure 2.4: Single-line diagram for voltage drop representation in a distribution system without DG.

The power supplied from the grid can be written as:

P_L+ jQ_L = U₂I (2.2)

where P_L and Q_L are respectively the active and reactive power consumption of the load.

The current flowing through the feeder can therefore be written as:

I = P_L− jQ_L U2

(2.3) By using the value of I, the voltage U₂ can be expressed as:

U₂ = U₁ +RPL+ XQL

U₂ + jXPL− RQ_L

U₂ (2.4)

In practice, the voltage angle between U1 and U2 is actually small, the voltage change can therefore be approximated as:

∆U = U₂− U₁ ≈ RPL+ XQL

U₂ (2.5)

If the voltage U₁ is considered as the base voltage, it can be assumed as unity. Therefore the voltage variation can be written as:

∆U ≈ RP_L+ XQ_L (2.6)

In LV distribution networks, the X/R ratio of the branch is small as R is usually higher than X. Therefore neither RPL nor XQL are negligible. The voltage where the load is

(30)

installed therefore depends on both the network resistance and reactance, on the real and reactive powers of the load and the voltage at the secondary substation.

The amount of voltage variation in a distribution system where also a DG is installed, can be calculated from the analysis the single-line diagram in Figure 2.5. A distributed generator is connected and P_Gand Q_Gare respectively its generated active and reactive power.

Figure 2.5: Single line diagram for voltage drop representation in a distribution system with DG.

The voltage variation in the feeder in this case can be approximated as:

∆U ≈ R(P_G− P_L) + X(±Q_G− Q_L) (2.7) Loads consume both active (−P_L) and reactive (−Q_L) power, whereas generators always supply active power (+P_G) and may inject or draw reactive power(±Q_G).

From Equation Equation (2.7), can be derived that when a DG injects active power into the grid, the voltage drop along the feeder is decreased. However, if the active power injected by the DG is larger than the feeder load, the power direction is reversed which therefore flows towards the secondary substation. As a consequence, the voltage at the point of connection of the generator U₂ rises above the voltage at the secondary side of the transformer U1. Also, from Equation (2.7) can be derived that if the distributed generator injects or consumes reactive power, the voltage variation can respectively be increased or decreased, which is one of the measures for voltage mitigation adopted in this thesis.

2.3.2 Reverse Power Flow

When the total electricity generation from PVs exceeds the total electricity demand in the LV network, reverse power flow occurs and the local LV grid exports the residual load towards the MV grid [103]. The frequency of occurrence of this phenomenon depends on the penetration level of the PV systems installed in the LV grid. When the penetration is low, the loads will consume the electricity generation locally and no inversion of the power flow occurs. With higher shares of of PV penetration instead, when there is peak electricity generation, due to mismatch between peak load and peak PV generation, the load is not high enough to consume the electricity locally generated and reverse power flow is likely to occur [103].

(31)

Figure 2.6 shows the increased amount of reverse active power flow across the distribution transformer with an increased PV penetration level in the LV grid. The figure also shows the increase of reactive power flow across the transformer with the same increasing PV penetration, due to reactive power requirement from PV inverters.

As there is a technical limitation on the actual power that can flow across a transformer, which is determined by its ampacity, reverse power changes the design of distribution networks which have to be designed to comply with peak generation and peak load situations [79].

Figure 2.6: Reverse Power Flow at dis- tributuon transformer with increasing penetration of PV systems.

Figure 2.7: Active and Reactive Power Losses with increasing penetration of PV systems.

2.3.3 Overloading of Grid Components

Distributed generation and high-power loads increase the loading on distribution grids:

transformers and cables originally, have not been designed for handling such large power flows [90]. Power system congestion occurs when flows across a system component, such as lines, cables and transformers exceed their safe design capacity which as discussed in Section 2.5 is 100% of their rated capacity [79]. Current DSO’s experience shows that wide integration of PVs located LV grids cause unexpected congestions especially in rural areas [75]. Similar problems occur in cases of massive connection of new loads in the network due to sector electrification, such as EVs and HPs. Figure 2.8 shows the overloading of lines and transformer as a result of the reversed power flow from the feeder where the PVs are connected, towards the distribution transformer. The most severe overloading issues are usually experienced in parts of the network where the reverse power flow from different feeders, is agglomerated into one line, which is usually close to the distribution transformer. Figure 2.9 shows the overloading issues experienced in network topologies with high penetration of EV and HPs drawing power from the grid with a high coincidence factor.

Normally the heat losses and thus the transformer over-heating is the limiting factor which determines the maximum allowed power flow. The lifetime of a transformer is in fact

(32)

determined by the deterioration of the insulation, which increases with high temperatures [64]. It is therefore important not to overload the transformer during long time periods.

Figure 2.8: Overloading of grid components due to PV reverse power flow.

Figure 2.9: Overloading of grid components due high-power loads.

2.3.4 Grid Losses

Power losses are the difference between the power that leaves a distribution substation and the power delivered to the customers. The technical losses are losses due to the finite conductivity of cables, lines, transformers and other network equipment [66]. A line can be modelled as the combination of a resistance and an inductance; the higher the ratio between resistance and inductance is, the lower the conductivity, which in a conductor results in higher losses. The length of the lines in distribution networks are much shorter than in transmission networks, therefore the ratio between the resistance and the inductance is higher than in transmissive networks, which lead to higher losses [66].

Power losses represent generated energy that is not sold to the final customer leading to lower revenues to the DSO. There is therefore the necessity to minimize the power losses in the network. In order to do so, the solution is usually to supply the power demand in a distribution network as locally as possible, zero losses happens when the load exactly matches the generation. The reduction in current flow leads to reduced power losses in the system [94].

The effect of PVs on grid losses depends on two factors: the amount of injected power and the location of the DG in the grid [63]. When PVs in situated close to the demand, the injected power is consumed locally and the power losses are reduced compared to the case without PVs. Figure 2.7 shows the line losses for varying generation size connected at bus of fixed load. The functions that express the active and reactive power losses depending on line resistance and current flowing in the lines, is shown in Equation (2.8) and Equation (2.9) respectively [63]. The higher the current flowing in the lines and the higher the resistance in the network, the higher the active power losses will be.

P_loss= 3I²R = P V

2

R + Q V

2

R (2.8)

(33)

Qloss = 3I²X = P V

2

X + Q V

2

X (2.9)

2.4 Main changes in the Distribution Network

2.4.1 Distributed Generation

Integrating distributed generation into distribution network represents a challenge for the DSOs due to the DG generation profiles and location [43]. Electricity generation from PV systems is in fact not dispatchable, as it is strictly dependent on meteorological conditions. Unlike traditional power plants they are in fact not able to vary their power output based on utility’s demand. Moreover peak generation from PV systems most often do not coincide with the peak demand. Therefore distribution system operator are facing difficulties to respect relevant grid codes and guidelines regarding the safe integration of DG in distribution networks [43].

5 10 15 20 25 30

Terminal 0.98

1.00 1.02 1.04 1.06 1.08 1.10 1.12

Voltage [p.u.]

0 kW150 kW 450 kW

750 kW 1050 kW

1350 kW 1650kW

1950 kW

0 kW 150 kW 450 kW 750 kW 1050 kW1350 kW1560 kW1950 kW 0

50 100 150 200 250 300 350 400

Transformer Loading [%]

Impact of PVs on Voltages and Transformer Loading

Figure 2.10: Voltages Profile along the feeder in different PV penetration scenarios in distribution grid.

Figure 2.10 illustrates the impact that an increasing amount of distributed generation installed at residential level, have on voltages and loadings in the network. When no DGs are connected, voltages along the feeder are typical of load prevailing profile, where lowest voltages are experienced at the end of the feeder. When the power generation from PVs matches the load requirement, it mitigates the low voltages in the network and a flat voltage profile is experienced. In this combination of load and generation, also the transformer and lines loading experience their minimum value. As soon as the generation exceeds the demand, voltages and loadings in the network start increasing; in this case the highest voltages will always be experienced at the end of the feeder. Violations of operational limit due to high penetration of distributed generation, happens when either the highest voltage in the feeder or the loadings of lines and transformers, are above their maximum allowable limit. Depending on the network topology voltages or loading issues are reached first. In Figure 2.10 the transformer is overloaded before over-voltages are

(34)

experienced in the feeders (if an upper voltage limit of 1.1 p.u. is considered), and this happens for an installed capacity between 450 and 750 kW. Over-voltages are instead experienced starting from an installed capacity of 1950 kW.

The intrinsic limitations of distributed variable renewable generators make their grid integration not as simple as for traditional dispatchable generators. In order to facilitate the integration, their operation needs to be shifted from a fit-and-forget approach, which has been commonly adopted by grid operators until recently, to a more active approach [43].

2.4.2 Heat Pumps

Environmental concerns not only have influenced the appearance of DG, but are also transforming the way the heat requirements in the residential sector have traditionally been supplied [74]. An electrification of the residential sector is on the way. Electrically driven heat pumps can in fact save primary energy compared to standard heat generators such as gas boilers, this is especially true with a high share of renewable electricity generation in the national energy mix [33].

In order to increase heat pump market share, the German government has subsidized them with grants for heating systems that use renewable energy [57]. Electric heat pumps experienced significant growth rates: the share of installation in new houses rose from 0.8% in 2000 up to 32.2% in 2017 [4]. This trend, which is pushed by the German En- ergiewende will undergo a steady increase. According to the study "Wärmewende 2030"

in fact, heat pumps must become a major pillar of the german heating system, supplying around twenty times more than they do today [57].

Moderate climates represent the ideal conditions for effective operations of HPs, whose efficiency is inversely proportional to the temperature difference between the environ- ment and the heated space [32]. Ground source heat pumps (GSHP) are usually installed in new, detached dwellings whereas air-source heat pumps (ASHP) are generally more suited for retrofit applications [82]. The coefficient of performance (COP), which is the heat produced per unit of electrical energy consumed is typically in the range between 3 and 4 for the GSHP and between 2 and 3 for the ASHP, which make this technology highly-efficient [32]. GSHP can achieve higher efficiencies given the relatively constant temperature of the ground throughout the year, however they have a higher capital cost for terrain drilling [82]. The most common HP type in Germany is ASHP, which has been considered over the course of this thesis.

Typical size of heat pumps ranges from 5 kWth to 20 kWth depending on the square meters of the house where the HP is installed [81][104]. A Typical rule of thumb for sizing HP system is 0.12kWth for each m² of room [104]. Considering an average COP of 2.5 for ASHP, the electrical power of the HPs ranges between 2 and 8 kWe.

The growth of heat pump load however, poses a challenge for distribution engineers [25].

The steering of heat pumps can cause problems in local distribution grids since the assets

(35)

of the distribution grid may not be dimensioned for large consumption peaks. This may lead to possible violations of grid technical constraints, such as voltage and loading levels, which will require adjustive measures to be taken [15]. The challenge of a huge penetration of heat pumps in the distribution grid is particularly true when the coincidence factor of heat pump utilization located in the same feeder is high and there is no local generation from PV systems.

Figure 2.11 shows the impact that different level of heat pump penetration in a LV grid, have on voltages and on transformer loadings in one winter day in Germany. Five HPs penetration scenarios, which represent the percentage of residential units equipped with a heat pump, varying from 20% to 100%, have been analyzed. To investigate the impact that the electrical load from the heat pumps have on the distribution grid, the traditional part of the household load has been kept constant throughout the whole day. The heat pumps analyzed are ASHP with an average size of 12.5 kWth (5 kWe), operating in combination of a thermal energy storage (TES).

000102030405060708091011121314151617181920212223 0.92

0.93 0.94 0.95 0.96 0.97 0.98

Voltage [p.u.]

100%80%

60%40%

20%No HP

000102030405060708091011121314151617181920212223 0

20 40 60 80 100 120 140 160

Transformer Loading [%]

100%80%

60%40%

20%No HP

Impact of HPs on Voltages and Transformer Loading

Figure 2.11: Voltages and Transformer Loading profiles with increasing share of HP in the LV grid.

With a heat pump penetration of 40% the transformer resulted to be overloaded in some moments of the day and to be continuously overloaded for higher penetration scenarios, reaching peaks of 150% when all the residential units have a heat pump activated. Volt- ages below 0.95 per unit [p.u.] are reached for heat pumps penetration scenarios of 60%

and the lowest voltage level that is reached in the 100% penetration scenario is below 0.93 p.u.

On one hand heat pumps poses challenges in the grid in term of violation of technical constraints, but on the other, they play an important role as a flexibility provider. The thermal energy storage in fact allows to decouple heat demand and heat generation, which make demand side management measures possible [15].

(36)

Other problems caused by a large HPs integration in the distribution grid are the very high starting currents and the switching transients, however they are outside the scope of the thesis.

2.4.3 Electric Vehicles

Electric vehicles have emerged as one of most attractive and promising solutions to decrease the level of GHGs in the transportation sector. With rapid and ongoing development of high-capacity batteries, high-efficiency motors and power electronics in fact, EVs have entered the large-scale commercialization stage [113].

Germany has set itself the goal of becoming the lead market and provider for electric mobility by 2020 as part of its long-term zero emission mobility vision [48]. From an EV stock of 49.2 thousand vehicles in 2015, it has sets its target of 1 million electric vehicles on the road by 2020 and 6 millions EV in 2030 [48]. The electric vehicles required to realize Germany’s 2020 vision, can be classified in terms of the following categories:

• Battery electric vehicle (BEV)

• Plug-in hybrid vehicle (PHEV)

• Fuel cell electric vehicle (FCEV)

• Internal combustion engine (ICE) including hybridization

PHEVs and BEVs vehicles require the use of batteries with high-energy storage capacity and with charging infrastructure connected to the electrical grid. Hybrid Electric Vehi- cles (HEVs) and Fuel Cell Vehicles (FCV) are not considered in this thesis because these technologies do not represent a load for the power system. In the thesis, the term EV only refers to PHEVs and BEVs.

High-capacity and high-efficiency charging infrastructures are mandatory to sustain the growing charging demands and to improve pure electric driving mileages and operational economy of EVs. Charging the battery of an electric vehicle depends on the combination of:

• Charging Power (i.e voltage and the number of phases of the power supply)

• Battery characteristics (i.e battery capacity)

Standard EV’s battery capacity depends on the EV supplier and on the distinction between full-electric vehicles and plug in electric vehicles. It varies from 4.5 kWh to 85kWh with an average size for the first category being round 8 kWh and for the second category around 35 kWh [96] [80]. It is expected that in the future the EV’s battery capacity range size will increase given the advancement on material technology that improves the energy density of the batteries, which make them increasingly lighter and smaller.

(37)

Nowadays classical charging power are 3.7 kW (3 phase) which is classified as slow charging technology, as to charge a battery of 30 kWh, around 8 hours are required [85]. In order to make the charging process faster and thus to make the technology more attractive from a potential user point of view, higher charging power, from 11 kW to 120 kW are becoming more common [85]. Fast charging stations, which can use multiple chargers working in parallel to deliver up to 120 kW of charging power, are already implemented and are becoming more widespread. A classification of the most common charging power is shown in table (Table 2.1.Regulations and standardization of the charging process are still ongoing charging infrastructures represent a relatively new technology. The In- ternational Electrotechnical Commision (IEC) has recently published International EV charging Standards (IEC 62196-1 and IEC 62196-2) which offer clarification by standard- izing the plugs and sockets which can be used in different electricity infrastructures in term of maximum current, voltage and protection required [85].

Power Connection Charging Power Recharge range per hour

Low Power 1-phase AC connectio <= 3.7 <20 km

Medium Power 1 or 3 phases AC connection 3.7 - 22 kW 20 - 110 km

High Power 3-phase AC connection >22 kW >110 km

High Power DC connection >22 kW >110 km

Table 2.1: Overview of EV charging power

EV charging stations can be divided into two typical classes: home-based charging and dedicated parking loads charging. This thesis has focused on EV charging stations of the first type (Figure 2.12).

Figure 2.12: Home based EV charging station.

Widespread home-based EV charging stations introduce large and intermittent load demands with new temporal and spatial characteristics [74]. EVs will absorb and store

Benchmarking of Smart Grid Concepts in Low-Voltage Distribution GridsODILIA BERTETTI

Benchmarking of Smart Grid Concepts in Low-Voltage Distribution Grids

ODILIA BERTETTI

Benchmarking of Smart Grid Concepts in Low-Voltage Distribution Grids

Odilia Bertetti

Supervisor and Examiner: Anders Malmquist

Master of Science Thesis EGI_2017-0076-MSC EKV1202 MSc in Sustainable Energy Engineering

Division of Energy Technology

KTH Royal Institute of Technology SE-100 44 STOCKHOLM

20

September 2017

Abstract

Sammanfattning

Acknowledgments

List of Symbols and Abbreviations

Contents

CHAPTER 1

Introduction

1.1 Background

1.2 Motivation

1.3 Research Question

1.4 Literature Review

1.5 Methodology

1.6 Thesis Structure

CHAPTER 2

Impact of Distributed Generation, Heat Pumps and Electric Vehicles on Distribution Networks

2.1 Introduction

2.2 The architecture of the Modern Power System

2.2.1 Medium Voltage Distribution Network

2.2.2 Low Voltage Distribution Network

2.3 Challenges of the Future Grid

2.3.1 Violation of Voltage Operational Limits

2.3.2 Reverse Power Flow

2.3.3 Overloading of Grid Components

2.3.4 Grid Losses

2.4 Main changes in the Distribution Network

2.4.1 Distributed Generation

Impact of PVs on Voltages and Transformer Loading

2.4.2 Heat Pumps

Impact of HPs on Voltages and Transformer Loading

2.4.3 Electric Vehicles