Assessing the peak shaving ability of Vehicle-to-Grid technologies on a city-level

(1)

Assessing the peak shaving ability of Vehicle-to-Grid

technologies on a city-level

Pre-master Thesis

Written by: Cas Wieggers S4114639 C.H.J.Wieggers@student.rug.nl 06-15388380 June 2020 University of Groningen Faculty of Economics and Business

(2)

- 1 -

Abstract

This paper addresses the use of vehicle-to-grid technologies to increase grid flexibility for coping with the increasing use of renewable energy sources. By creating a bi-directional power supply between electric vehicles (EVs) and the energy grid, peak power can be stored in the batteries and fed back to the grid at a later point of time. This research simulated a fleet of grid-connected EVs which is used for peak shaving services for a supply of photovoltaic energy. The results of these simulations showed that implementing this technology in current conditions would reduce the curtailment rate by approximately 90%.

Keywords: Peak Shaving; Electric Vehicles; State of Charge (SOC); Photovoltaic Energy; curtailment level

Supervisor: J.E. Fokkema

(3)

- 2 -

1. Introduction

1.1 General introduction

Increasing importance is placed on reducing emissions worldwide. In 2007 80 percent of the emissions were energy related (Bellekom et al., 2012). This emphasizes the need for a transition of the energy market towards renewable energy sources. However, one of the challenges of using renewable energy sources is the fact that these are weather-related. The energy supply is unpredictable (Tan et al., 2016). The current energy grid is unable to cope with these levels of unpredictability. One solution is to increase grid flexibility and better match supply and demand. Research suggests that the current energy grid, which carriers power from a small number of generators to many consumers, must be converted into smart grid, which enables two-way flow of electricity and information to create distributed

automated energy delivery network (Fang et al., 2012). A smart grid creates an automated energy network that transports energy in a two-way flow, rather than solely supplying power to the user. One component of the smart grid are vehicle-to-grid technologies.

1.2 Vehicle-to-grid technologies

Vehicle-to-grid (V2G) systems can increase the flexibility by enabling bi-directional power flow between the energy grid and electrical vehicle’s (EV) batteries. The EV batteries can be used to store excess energy during high-supply periods and feed back energy in period of high demand (Wang & Wang, 2013). In their research Lund, & Kempton proved that

implementing the technology and integrating it with wind power in the Danish national grid, can increase the efficiency of the grid, lower emissions and improve the ability to integrate wind energy in the power grid (Lund & Kempton, 2008). This shows that aggregating EVs on a national level can increase grid flexibility, which is the ability of the power system to maintain balance between generation and demand of energy. In their research Lauinger, Vuille & Kuhn even suggest that the technology can support a global doubling of the renewable energy share by 2030 in comparison to 2015 (Lauinger et al., 2017).

In other research the combination was made between V2G-systems and photovoltaic energy. Van der Kam and Van Sark implemented these ideas into their research and examined

implementing V2G systems in a microgrid with PV Installation in Lombok, a neighbourhood in the city of Utrecht (van der Kam & van Sark, 2015). Their results showed that the self-consumption of PV-power increased, and the losses of excess energy decreased due to the storage capabilities of EVs. This sample principle applies to the provincial level, in which a system with 500.000 North-Italian citizens and 1.000 EVs was simulated (Fattori et al., 2014). 1.3 Research objective

Previous research showed that the integration of V2G-technologies and photovoltaic energy supply on a provincial (North-Italian) and neighbourhood-level (Lombok - Utrecht) will increase the flexibility of the energy grid and enable a better match between energy demand and supply. These studies included simulations based on penetration levels of Electric Vehicles. Secondly was concluded that V2G technologies improve the efficiency of the energy grid in Denmark when integrated with wind energy.

(5)

- 4 - When supply is high, energy is transported to the battery of the EV. When demand is high, the energy can be transported to the energy grid.

Figure 1 The proposed V2G-system.

Research has not focused on connecting the supply of PV-energy to the batteries of EVs on a city-level. Aggregating a bigger amount of EVs, in comparison to Lombok, and connecting them to a V2G-system allows for a bigger battery capacity to collect previously curtailed PV-energy. This would hypothetically improve the flexibility of the energy grid, as through peak shaving a bigger percentage of the energy supplier will be used. Secondly, based on the assumption that the EVs are unused during the middle of the day, when PV-energy supply is at its peak, a combination between V2G-technologies and PV-energy is optimal. The

combination of both factors leads to the following research question;

To what extent can an aggregated EV-fleet, when connected to the energy grid of a city via V2G-systems to enable a two-way flow of energy, lower the curtailment levels of PV-energy supply?

As suggested by the research question above, this paper will address the effect that V2G-technologies have on the efficiency of the electricity grid, measured in the curtailment levels. Based on the idea that V2G-technologies improve the peak shaving capabilities of the grid and subsequently lower the amount of curtailed energy, the curtailment levels provide sufficient insights into the efficiency of the grid.

1.4 structure of this paper

Chapter two will provide the theoretical background of this research, based on a review of previous research. In chapter three the conceptual model and the experimental setup of the simulation study will be described. Chapter four will state the results of the simulation study and will discuss these insights in comparison to previous research. Lastly, chapter five will give answer to the research question as given above.

2. Literature Review

Vehicle-to-grid (V2G) is the establishment of a bidirectional power transfer between electric vehicles and the energy grid, which could offer several services to the grid. Services as power regulation, load balancing and reactive power support can improve the penetration of

renewable energy sources in the energy grid by increasing grid flexibility (Lauinger et al., 2017). Since the publication of fundamental articles by Kempton and Tomić (2005), the research interest has taken off, but the realisation of V2G systems is still in its infancy. Due to the uncertainty of the economic viability of the technology and limitations regarding the aggregation of multiple vehicles, the technology is still in the pilot stage. A number of

(6)

- 5 - a maximum of 100 charging stations and are hindered by economic viability and the lack of competitive hardware (EVConsult, 2018).

The fundamental article of Kempton and Tomić (2005) covers implementation, business models and transition of V2G in the energy grid at that time. They concluded that the automobile fleet was surprisingly complementary to the energy grid and could function as storage for peak power. Further research supplemented on this by analysing the effect V2G will have on national energy systems. National power systems build up of 100% electric vehicles and fuelled by wind power were simulated (Lund & Kempton, 2008). Results

showed that increasing the intelligence of the energy system, would improve the efficiency of the energy system. This model assumes that the entire vehicle fleet consists of EVs, which is not a valid representation of the real-world situation. This model can be improved by using real time data of the number of EVs in the system. Secondly the model simulated only wind energy. As the share of energy has increased since, the importance of simulating PV-energy supply in combination with V2G-technologies, has increased.

Additional research did include real time information, such as the number of EVs and battery capacity. In 2013, Wang & Wang performed simulations to asses the peak shaving

capabilities of V2G (2013). This study assumed that only five percent of the automobiles were fully electric. The results of this study showed that an increase in the connectedness of EVs would increase the efficiency of V2G’s peak shaving abilities. Depending on the number of EVs, V2G could replace other peak shaving services. However, this study did not incorporate the supply of renewable energy into the simulation, which is less consistent in comparison to conventional energy sources. Therefore, the research in this paper will research the effects of using V2G for peak-shaving of renewable energy sources and gather new insights.

(7)

- 6 - As suggested previous research has proven that V2G has sufficient capabilities to increase self-sufficiency in a micro grid. However, for a larger scale the results were marginal. This can be incorporated with the fact that previous research mainly focused on the combination between V2G and wind energy. As the share of PV-energy in the grid is increasing, further research should focus on conducting research regarding the combination between V2G and PV-energy. Lastly, previous research did not include specific parameters in the simulations that improve the representation of the peak shaving abilities of V2G. By including

(dis)charging facilities at non-residential areas, real time information of the electricity grid and the supply of renewable energy, the simulation will be able to increase the understanding of the peak shaving capabilities of V2G. This paper will remove the above-mentioned

insufficiencies and will simulate an electricity grid on a city level. The goal of these

simulations will be to assess the peak shaving capabilities of V2G for PV-energy by analysing the curtailment levels. A more flexible grid will be able to apply peak shaving and make greater use of the supplied renewable energy, thus having lower curtailment levels.

3. Methodology

3.1 Problem description

This paper aims to understand the effect of V2G-technologies on energy grid flexibility through peak shaving in combination with photovoltaic energy. Flexibility will be measured by analysing the curtailment levels in the simulation.

In the simulations we consider a PV-energy supply per hour St (kWh) which is connected to a city-level energy, with local electricity demand per hour Dt (kWh). A fleet of electric vehicles will be connected to the energy grid through aggregators and can be used as battery storage with capacity Bt (MWh) for peak shaving services. Dt will be calculated based on the number of households H, the average electricity demand per household per year A and the user profile of electricity, which will be acquired from Nedu (Verbruikersprofielen, 2020). St is

determined by the average PV-system losses L and the total capacity of the solar panels C. Bt is dependent on the number of EVs connected to the gridvia aggregators N, the maximum battery capacity per vehicle M and the usage profile of an EV during the day. The user profile depicts the use of EVs by their owners in terms of available (dis-)charging hours U and the minimum required state of charge (SOC) R.

The goal of this research is to study the viability of V2G-systems in reducing curtailment by simulating model in which a fleet of EVs is used as battery storage for excess energy. When supply is higher than demand, excess energy will be stored in the batteries if storage space is available. When supply is lower than demand the energy in the battery will be fed back to the system. This increases the percentage of energy used by consumers and lowers the

(8)

- 7 - 3.2 Conceptual Model

3.2.1 Parameters and variables:

Parameter Symbol Description In/exclude State

Local electricity demand

Dt Demand for electricity per hour in kWh.

Include Changes every hour.

Number of households

H Total number of households Include Fixed

Average electricity demand

A Average electricity demand per household per year

Include Fixed PV-energy supply St Energy generated by a solar

park in MWh per hour.

Include Calculated

PV-capacity C PV peak power supply in

kWh.

Include Fixed

Average system loss

L Average losses of PV-system Include Fixed

Wind-energy supply Wt Energy generated by a wind

park in kWh per hour.

Exclude

-EV battery capacity

Bt Total available EV battery capacity at a certain time.

Include Calculated

Number of EVs N Total number of EVs in the

research area.

Include Fixed

Battery capacity per electric vehicle

M Maximum battery capacity per EV, based on average of all EV-models.

Include Fixed

Available (dis-) charging hours

U Hours that an EV is available for (dis-) charging.

Include Fixed per hour

Minimum required SOC

R Minimal energy level of an EV-battery that must remain in the battery.

Include Fixed

Other storage options

X Total available battery capacity from other sources.

Exclude -

Seasonal changes in EV user pattern

Z User pattern data changes due to seasonality

Exclude

-Table 1 The list of parameters of this simulation study.

Variable Symbol Description State

Curtailment level c Percentage of curtailed PV-energy out

of the total supply.

Calculated

The average SOC u Average capacity utilisation of the

EV-battery.

Calculated

Curtailment reduction

r Reduction of the curtailment level in comparison to the real world.

Calculated

Table 2 The list of variables of this simulation study.

3.2.2 Assumptions and simplifications:

In this research some assumptions will be made. Firstly, it will be assumed that the necessary infrastructure for a bidirectional flow of energy is in place. The costs to set up this

(9)

- 8 - In addition, the only available battery capacity is generated through EVs, no other batteries are incorporated in this study. The reason for this is that this study aims to understand the effect of using EVs as batteries. Additionally, the study will not differentiate the battery capacity per EV within a single experiment. The number of different EV models and battery capacities will add complexity to the simulation model, while not significantly contributing to a better understanding. Lastly, no seasonal pattern in the available charging hours will be simulated. It is assumed that by simulating a large number of EVs, the usage pattern will approach an average.

3.2.3 Scope and level of detail:

This research will entail a simulation of a city-level energy grid and will incorporate

photovoltaic energy as a renewable energy source. This scope was chosen as it is expected to provide a level of EV-aggregation which can provide an adequate battery capacity for peak-shaving services. Secondly, based on the literature review, the city level has not been researched intensively so far. The experiments in this simulation study will be based on hourly data and will entail an entire year. The reason for this is that the available data also has this level of detail and based on this level of detail the daily fluctuations regarding the EV-use and electricity demand are incorporated in this study.

3.2.4 Logic flow diagram:

The logic behind the simulation is presented in a logic flow diagram in appendix A. This diagram summarizes the decisions that are made regarding the flow of energy between demand, the batteries, and the energy grid. The simulation starts with the generation of PV-energy which can be used to satisfy local PV-energy demand. When no PV-energy is generated, the available energy in the EV-batteries, up to the minimum required SOC, will be used to satisfy demand if the EVs are available for charging during specific hours. At all times, the minimum required SOC will be left, as the EV-owners need energy to use the EV for their daily

activities. In the case that no energy is available in the batteries, the electricity of the other energy sources will be used. Additionally, when the energy in the EV-batteries can only partially satisfy demand, additional energy sources are used. In the case that the PV-energy supply is higher than local demand, excess energy will be stored in the batteries. This excess energy can only be sent to the battery up to the maximum available battery capacity and during available charging hours. If no storage space is available or the EVs are not available for (dis)charging, the energy will be lost (curtailment).

3.3 Experimental setup

The simulation study in this paper will consists of a multitude of different experiments to gain a deeper understanding in the peak shaving ability of V2G. Each experiments will include a solar energy supply, generated by 230.000 solar panels that were present in Groningen in October 2018 (Grunneger-Power, 2018), an EV-fleet and a local demand generated by households. The setting of these experiments is the city of Groningen.

(10)

- 9 - & Oda in their research regarding EVs (2020)(2016). The available charging hours are set at 8:00 – 17:00 and 22:00 – 7:00, which was also assumed in other research. The results of the base case experiment will be compared with a real-world experiment. The difference between the two is solely that in the real world experiment no EVs are connected to the grid, N is 0. To gather insights in the peak shaving ability of V2G, several experiments will be performed next to the base case. The first simulation will include several experiments in which

differentiate the number of EVs (N) and the minimum required SOC (R). Both factors directly influence the available battery capacity. N will range between zero (real world scenario) and 2.500 EVs, which roughly account for double the amount of EVs in comparison to current day. Additionally, R will range between ten and fifty percent, which states that in between fifty and ninety percent of the battery capacity can be used for peak shaving. This simulation will show the effect that the number of EVs and the minimum required SOC have on the reduction of the curtailment level. The second simulation will involve one experiment in which all hours are available for charging instead of the limited hours. Based on this

simulation can be concluded whether this parameter had a significant influence on the results. Thirdly, a set of experiments will be performed in which the battery size (M) and minimum required SOC (R) will be differentiated. As in the first simulation, the battery size also directly influences the available battery capacity. M will range between twenty-five and two hundred kWh as this is the battery capacity range for the available EV models (Elektrische-Voertuigen-Database, 2020). Lastly, a sensitivity analysis will be performed in which the supply data of 2014 and 2015 will be compared with the data of 2016. This sensitivity analysis will show whether these years significantly differentiate from the base case. The table below summarizes the different sets of experiments that will be performed in this simulation study. Each experiment will consist of simulation an energy grid for one year. The bold printed text are the parameters that have changed in that specific set of experiments.

Set of experiments # of experiments EV-battery capacity for V2G (kWh) Number of EVs (dis-)charging hours Min. required SOC

Real world 1 0 1.3751 _{8:00-17:00 &}

22:00-7:00

20%

Base case 1 57.12 _1.375 _{8:00-17:00 &}

22:00-7:00

20% # of EVs &

min. req. SOC

50 57.1 0 – 2500 in steps of 250 8:00-17:00 & 22:00-7:00 10% - 50% in steps of 10% Available (dis-) charging hours 1 57.1 1.375 All hours 20%

Battery size & min. req. SOC

40 25 – 200 in steps of 25 1375 8:00-17:00 & 22:00-7:00 10% - 50% in steps of 10% Sensitivity analysis 3 57.1 1.375 8:00-17:00 & 22:00-7:00 20%

4. Results

Multiple experiments have been performed to gather insight in the peak shaving capabilities of vehicle-to-grid technologies. This chapter will describe the various experiments and the conclusions that can be based on them. Appendix B shows the results of the simulation study.

1_{The number of EV’s in Groningen. Calculated by multiplying percentage of EVs out of the national vehicle} fleet with the number of cars in Groningen.

2_{The average battery capacity of EVs based on EV database (Elektrische-Voertuigen-Database, 2020).}

(11)

- 10 - 4.1 Base case

The base case shows the effect of implementing V2G in the current day situation, as it assumes the current day number of EVs. By comparing these results with the real-world scenario, the effects of implementing V2G can be assessed. The figure blow displays the monthly statistics of the base case. The X-axis states the months of the year and the Y-axis the amount of energy in kWh. The graph shows that on average the supply of PV-energy is

significantly lower than the average demand, which is caused by the fact that the generation of energy is limited so day hours only. Secondly, the figure shows that during the summer months electricity demand is lower and energy supply is higher than during the winter months. This signifies an increasing energy surplus during the day hours, which is stored in the batteries of EVs. This can also be seen in the higher SOC.

Figure 2 Base case statistics – average monthly demand Dt, supply St, and SOC at end of period

In the table 4 the results of both the real-world scenario and the base case are shown. The results show that the in a scenario without V2G 7.52% of the PV energy is curtailed. This suggest that out of the entire energy supply 7.52% of the

energy is supplied at periods in which demand is lower than supply. This excess energy can not be stored in batteries, as no battery capacity is available.

In the base case an available energy storage of 62.810 kWh is generated by the EVs. By connecting this battery capacity to the energy grid, the curtailment level is reduced to 0.58%. In comparison to the real world, the curtailment level is reduced by 92.31%. Based on these results can be concluded that if the current number of EVs was connected to the grid through V2G, a significant curtailment reduction can be achieved. Additionally, in this scenario only 23.31% of the battery capacity is used on average. Approximately 80% of the battery capacity remains unused, which could sign to overcapacity. However, in their research Wikner & Thiringer conclude that keeping the SOC at approximately 15% for a few days a week, the battery life will be prolonged by at least half a year (2018). For the base case can be

concluded that although a large share of the battery capacity remains unused, it is beneficial the battery lifetime.

Variable Result

Real world Curtailment Level 7.52% Base case

Curtailment Level 0.58%

Average SOC 23.31%

Curtailment Reduction 92.31%

(12)

- 11 - 4.2 Number of EVs & Minimum required SOC

In figure 3, the curtailment levels of the second simulation are shown. The corresponding data can be found in appendix B. The X-axis shows the number of EVs that are connected to the grid through V2G, whereas the Y-axis depict the curtailment levels in percentages. Each line depicts the relation between the number of EVs and the curtailment levels for that specific minimum required SOC. This relation is shown to be a decreasing curve, in which each increase in the number of EVs has a smaller effect on the curtailment levels. This can be explained as the smaller peaks in supply can be absorbed by a smaller number of EVs, whereas a higher available battery capacity is needed to absorb the higher peak supply. Secondly, the figure also shows that the minimum required SOC is affecting the curtailment levels. By increasing this parameter, more EVs are needed to cushion the peak supply. For example, zero curtailment can be achieved with 2.000 EVs if the minimal required SOC is 10%. However, for a minimal required SOC of 50% the curtailment level would still be 0.82% with 2.000 EVs.

Figure 3 The curtailment levels for simulation 2 (# of EVs and min. required SOC)

Based on the results of this simulation can be concluded that by increasing the number of EVs, curtailment levels can be reduced. Secondly, depending on the minimum required SOC, zero curtailment can be achieved when approximately 3% of the entire vehicle fleet of 70.000 cars would consists of grid connected EVs, showing that implementing the technology on a smaller scale would already pay off. However, if EV owners would increase the minimum required SOC to 50%, double the amount of EVs is needed to achieve zero curtailment. 4.3 Available charging hours

The third simulation showed that limiting the available charging hours does not influence the curtailment level. The base case and the experiment in which all hours are available for (dis-)charging both show a curtailment level of 0.58%. This suggest that during the hours in which EVs are unavailable in the base case, no surpluses of PV energy are visible, and no extra energy is curtailed. However, the average SOC is affected by the parameter. When all hours are available for charging or discharging, the EVs can feed back energy to the grid directly when demand is higher than the supply. In the base case the energy remained in the battery until EVs were connected to the grid at 22:00, raising the average SOC from 22.66% to 23.31%. This would mean that if (dis-)charging hours are solely limited during non-peak

0,00% 1,00% 2,00% 3,00% 4,00% 5,00% 6,00% 7,00% 8,00% 0 250 500 750 1000 1250 1500 1750 2000 2250 2500 C ur tai lm ent le ve l # of EV's

Curtailment levels for simulation 2 (# of EV's / min. Req SOC)

(13)

- 12 - supply; curtailment levels are not affected. In the case that EVs are used by their owners midday, in which energy supply is at its peak, curtailment levels could be affected as less battery capacity is available for peak shaving.

4.4 Battery size & Minimum Required SOC

The battery size of each EV also directly influences the available battery capacity for peak shaving. Figure 4 shows the curtailment levels of each experiment in this simulation. The X-axis shows the battery capacity per EV and the Y-X-axis depicts the curtailment level. In the figure can be seen that the relation between the battery capacity and a minimum required SOC follows the same pattern as with the number of EVs, a decreasing curve. As the battery

capacity increases, the available battery capacity increases and curtailment levels decrease. Each increase in battery capacity will have a lesser effect on the curtailment. Secondly, the results also show that no significant changes in the average SOC in comparison to the base case are seen. In all experiments the average SOC was only 1% to 3% above the minimum required SOC, never approaching the maximum battery capacity.

The figure also shows that when the 1.375 EVs in the experiments have a battery capacity of 125 kWh, the curtailment levels approach zero, regardless of the minimum required SOC. If the battery technologies of EV continue to advance, the average battery capacity of an EV will increase further. This increases the available battery capacity for peak shaving and thus improves the peak shaving ability of V2G.

Figure 4 The curtailment levels of simulation 3 (Battery Capacity & Minimum Required SOC).

4.5 Sensitivity analysis

The goal of the sensitivity analysis is to determine whether the variables in the experiments are affects by changes in the data input of parameters. Therefore, two experiments were performed in which solar data from 2014 and 2015 was used, as opposed to 2016 for the base case. These results show that the curtailment reduction of each year falls in the range of 91% - 97%, which implies that the simulations shows approximately the same results. The

difference in curtailment reduction can be explained by the differences in produced PV energy as solar intensity differs each year. This also applies for the average SOC as every year on

0,00% 0,50% 1,00% 1,50% 2,00% 2,50% 3,00% 3,50% 4,00% 25 50 75 100 125 150 175 200 Curt ailm en t Perc en ta ge Battery Capacity (kWh)

Curtailment levels simulation 3 (Battery Capacity & Min. Required SOC)

(14)

- 13 - average 23% of the maximum battery capacity is used. Based on these results can be

concluded that no significant changes in the results become visible by using different years of solar data.

4.6 Discussion

By comparing the results of this study with that of existing literature, some similarities become clear. In their research Wang and Wang (2013) showed that EVs are a viable option for peak shaving and could replace other methods. This corresponds with the results of this study, as both suggest that by solely using EVs to provide battery capacity, a significant share of curtailment can be absorbed. Wand and Wang also pointed out that the ability for peak shaving is highly dependent on the number of EVs, which directly links back to the total available battery capacity. Richardson adds on this and points out that only a small part of the vehicle fleet has to consists of grid-connected EVs to be able to store all peak energy and feed it back to the grid at a later point of time. This corresponds with the results of this research in which using only a small proportion of the vehicle fleet of Groningen would reduce the curtailment rate by 92%. However, in comparison to the research of Fattori et. All there was no similarity in the results of the research. According to Fattori et. all, the use of V2G in combination with PV-energy led to no significant benefits for the grid. Their research

however limited the (dis-)charging possibilities for EVs to charging at home. By limiting the available (dis-)charging time, the results would show less significant results. This research did include the charging possibilities at work, which significantly increased the validity of the technology.

5. Conclusion

The problem that is addressed in this paper is the peak in energy supply that puts a strain on the energy grid. By connecting EV’s to the energy grid through V2G, a bi-directional energy flow between the EV and the energy grid is created. This enables the energy grid to store energy in the EVs in time of peak supply and use this energy at a later moment. To asses the peak shaving ability of V2G, the following research question is constructed; To what extent can an aggregated EV-fleet, when connected to the energy grid of a city via V2G-systems to enable a two-way flow of energy, lower the curtailment levels of PV-energy supply?

To answer this question several simulations were performed in which a fleet of EVs is used to store energy surpluses from a PV-energy supply. These simulations varied in terms of the number of EVs, the battery capacity per EV, the available charging hours and the minimum required SOC of an EV. Multiple conclusions can be drawn based on the results of these simulations. Firstly, the base case showed that if the current number of EVs (1375) would be connected to the energy grid, a reduction of approximately 90% of the curtailment level could be achieved as 62.000 kWh of available battery capacity is created. This implies that the implementation of V2G can achieve results by solely using the resources that are currently available, without the need for further investment in new EVs. Not only does this increase the technical viability of the technology in terms of peak shaving abilities, it also increases the economic viability.

(15)

- 14 - the technology parties could assess these factors to gain insight in the potential curtailment reduction of the specific scenario. Secondly, the results show that on average a relatively small percentage of the battery capacity above the minimum required SOC is used, ranging from 1% to 3%. This SOC is beneficial to prolonging battery life as it decreases battery degradation.

5.1 Limitations

This study did include limitations which could affects the results. Firstly, the simulations did not include the depletion of the battery through usage. The simulations solely simulated a minimum SOC that makes sure that energy is left in the battery for use. Adding this parameter to the study could possible lower the average SOC and increase the available battery capacity for peak shaving. Secondly, this study simulated a PV-energy supply and left out wind energy. The results showed that a large share of the battery capacity remains unused.

Incorporating wind energy supply in the simulations could increase the average SOC and give insights in the effectiveness of V2G systems in the peak shaving of wind and solar energy. Lastly, the simulations have set the fleet of EVs as a kind of solid battery in which all EVs behave in the same way, in terms of available charging hours. If the simulations would

randomize the number of EVs that is connected at certain times, the simulation would become more representative for the reality.

5.2 Future research

Further research could focus on adding parameters or changing the current parameters to eliminate the limitations. This would involve adding additional sources of energy and changing the simulation of the EV fleet to increase realism. For the latter, information regarding the behaviour of EV-owners and the use of the EV could be necessary.

Bibliography

Aziz, M., & Oda, T. (2016). Load Leveling Utilizing Electric Vehicles and their Used Batteries. Modeling and Simulation for Electric Vehicle Applications, October. https://doi.org/10.5772/64432

Bellekom, S., Benders, R., Pelgröm, S., & Moll, H. (2012). Electric cars and wind energy: Two problems, one solution? A study to combine wind energy and electric cars in 2020 in The Netherlands. Energy, 45(1), 859–866.

https://doi.org/10.1016/j.energy.2012.07.003

Elektrische-Voertuigen-Database. (2020). Accu capaciteit van elektrische auto’s. https://ev-database.nl/cheatsheet/accu-capaciteit-elektrische-auto

EVConsult. (2018). V2G - A Global Roadtrip - Around the World in 50 Projects. October, 1– 38. https://www.evconsult.nl/wp-content/uploads/2018/10/Final-Report-UKPN001-S-01-I-V2G-global-review.pdf

Fang, X., Misra, S., Xue, G., & Yang, D. (2012). Smart grid - The new and improved power grid: A survey. IEEE Communications Surveys and Tutorials, 14(4), 944–980.

https://doi.org/10.1109/SURV.2011.101911.00087

(16)

- 15 - https://doi.org/10.1016/j.solener.2014.09.034

Grunneger-Power. (2018). Aantal zonnepanelen in de stad verdubbeld tot 230.000 stuks. https://grunnegerpower.nl/in-een-jaar-verdubbeling-aantal-zonnepanelen-in-stad/ Kempton, W., & Tomić, J. (2005). Vehicle-to-grid power implementation: From stabilizing

the grid to supporting large-scale renewable energy. Journal of Power Sources, 144(1), 280–294. https://doi.org/10.1016/j.jpowsour.2004.12.022

Lauinger, D., Vuille, F., & Kuhn, D. (2017). European Battery , Hybrid and Fuel Cell Electric Vehicle Congress A review of the state of research on vehicle - to - grid ( V2G ):

Progress and barriers to deployment. European Battery , Hybrid and Fuel Cell Electric Vehicle Congress, March, 1–8.

Lund, H., & Kempton, W. (2008). Integration of renewable energy into the transport and electricity sectors through V2G. Energy Policy, 36(9), 3578–3587.

https://doi.org/10.1016/j.enpol.2008.06.007

Richardson, D. B. (2013). Electric vehicles and the electric grid: A review of modeling approaches, Impacts, and renewable energy integration. In Renewable and Sustainable Energy Reviews. https://doi.org/10.1016/j.rser.2012.11.042

Tan, K. M., Ramachandaramurthy, V. K., & Yong, J. Y. (2016). Integration of electric vehicles in smart grid: A review on vehicle to grid technologies and optimization techniques. Renewable and Sustainable Energy Reviews, 53, 720–732.

https://doi.org/10.1016/j.rser.2015.09.012

van der Kam, M., & van Sark, W. (2015). Smart charging of electric vehicles with

photovoltaic power and vehicle-to-grid technology in a microgrid; a case study. Applied Energy, 152(August), 20–30. https://doi.org/10.1016/j.apenergy.2015.04.092

Wang, Z., & Wang, S. (2013). Grid power peak shaving and valley filling using vehicle-to-grid systems. IEEE Transactions on Power Delivery, 28(3), 1822–1829.

https://doi.org/10.1109/TPWRD.2013.2264497

Wikner, E., & Thiringer, T. (2018). Extending battery lifetime by avoiding high SOC. Applied Sciences (Switzerland), 8(10). https://doi.org/10.3390/app8101825

Zima-Bockarjova, M., Sauhats, A., Petrichenko, L., & Petrichenko, R. (2020). Charging and discharging scheduling for electrical vehicles using a shapley-value approach †.

Energies, 13(5), 1–21. https://doi.org/10.3390/en13051160

(17)

- 16 -

Appendixes

(18)

- 17 - B. Simulation results Base Case: Parameter _Result Cur. Level 0,58% Av. SOC 23,31% Curtailment reduction 92,31%

Simulation 1 (# of EVs & Min. Req. SOC)

Minimal required SOC

(19)

- 18 - Simulation 2 (Available charging hours)

Hours Limited hours All hours

Cur. Level 0,58% 0,58%

Av. SOC 23,31% 22,66%

Curtailment reduction 92,31% 92,31%

Simulation 3 (Battery capacity & Minimum Required SOC)

Minimal required SOC

Battery capacity Variable 10% 20% 30% 40% 50% 25 Curtailment Level 2,31% 2,59% 2,89% 3,21% 3,56% 25 Average SOC 13,04% 22,80% 32,54% 42,26% 51,95% 25 Curtailment reduction 69,31% 65,59% 61,60% 57,35% 52,65% 50 Curtailment Level 0,64% 0,88% 1,20% 1,58% 2,04% 50 Average SOC 12,70% 22,49% 32,21% 41,93% 51,64% 50 Curtailment reduction 91,55% 88,36% 83,98% 78,93% 72,92% 75 Curtailment Level 0,30% 0,58% 0,92% 1,34% 1,93% 75 Average SOC 13,53% 23,31% 33,01% 42,67% 52,25% 75 Curtailment reduction 96,02% 92,31% 87,71% 82,14% 74,31% 100 Curtailment Level 0,04% 0,16% 0,37% 0,64% 1,04% 100 Average SOC 12,17% 22,10% 31,97% 41,80% 51,57% 100 Curtailment reduction 99,52% 97,84% 95,12% 91,55% 86,23% 125 Curtailment Level 0,00% 0,00% 0,01% 0,16% 0,45% 125 Average SOC 11,65% 21,65% 31,64% 41,57% 51,44% 125 Curtailment reduction 100,00% 100,00% 99,82% 97,84% 94,03% 150 Curtailment Level 0,00% 0,00% 0,00% 0,00% 0,11% 150 Average SOC 11,33% 21,32% 31,32% 41,32% 51,28% 150 Curtailment reduction 100,00% 100,00% 100,00% 100,00% 98,56% 175 Curtailment Level 0,00% 0,00% 0,00% 0,00% 0,00% 175 Average SOC 11,11% 21,11% 31,10% 41,10% 51,10% 175 Curtailment reduction 100,00% 100,00% 100,00% 100,00% 100,00% 200 Curtailment Level 0,00% 0,00% 0,00% 0,00% 0,00% 200 Average SOC 10,96% 20,95% 30,95% 40,95% 50,94% 200 Curtailment reduction 100,00% 100,00% 100,00% 100,00% 100,00%

Simulation 4 (Sensitivity analysis)

Assessing the peak shaving ability of Vehicle-to-Grid technologies on a city-level