Juni 2019
Development of a Smart Charging Algorithm for Electric Vehicles at Home
Johanna Lundblad
Rebecca Segelsjö Duvernoy
Teknisk- naturvetenskaplig fakultet UTH-enheten
Besöksadress:
Ångströmlaboratoriet Lägerhyddsvägen 1 Hus 4, Plan 0
Postadress:
Box 536 751 21 Uppsala
Telefon:
018 – 471 30 03
Telefax:
018 – 471 30 00
Hemsida:
http://www.teknat.uu.se/student
Development of a Smart Charging Algorithm for Electric Vehicles at Home
Johanna Lundblad and Rebecca Segelsjö Duvernoy
The purpose of this bachelor thesis is to develop a smart charging algorithm for electric vehicles (EVs) and examine the potential of the smart charging scheme, compared to the uncontrolled charging scheme at residential houses with an installed photovoltaic (PV) system. The
thesis examines if the smart charging can increase the photovoltaic self-consumption and self-sufficiency of houses. Also, the thesis will evaluate if the smart charging scheme can reduce the household peak loads compared to the uncontrolled charging scheme.
The presented results show that the implementation of the proposed algorithm will reduce the household peak load on average by 38.64 percent at a house with an installed PV system. The self-consumption and self-sufficiency increased by 4.69 percent and 4.97 percent when the smart charging algorithm was applied. To increase the credibility of the developed model a sensitivity analysis considering number of houses and vehicles was done.
From the results, it can be concluded that the proposed smart charging algorithm could be an option to reduce the household peak load and increase the usage of renewable energy sources.
ISSN: 1650-8319, TVE-STS; 19009 Examinator: Joakim Widén Ämnesgranskare: Reza Fachrizal Handledare: Joakim Munkhammar
1
Table of content
Table of content _____________________________________________________________ 1 1. Introduction ___________________________________________________________ 4 1.1 Aim _______________________________________________________________ 4 1.2 Resarch Questions ___________________________________________________ 5 1.3 Limitations and delimitations ____________________________________________ 5 1.4 Disposition __________________________________________________________ 5 2. Background ___________________________________________________________ 6 2.1 Paris Agreement on Climate change _____________________________________ 6 2.1.1 Transport industry and Paris Agreement ______________________________ 6 2.2 Electric Vehicles _____________________________________________________ 6 2.2.1 History of EVs ___________________________________________________ 7 2.2.2 The future of EVs ________________________________________________ 7 2.3 The Power Grid ______________________________________________________ 8 2.3.1 Challenges _____________________________________________________ 8 2.4 PV system __________________________________________________________ 8 2.4.1 PV power today and future predictions _______________________________ 9 2.5 PV self-consumption __________________________________________________ 9 2.6 Smart Charging _____________________________________________________ 10 3. Methodology and data __________________________________________________ 11 3.1 Overview __________________________________________________________ 11 3.2 Data ______________________________________________________________ 11 3.2.1 Departure time and Arrival time ____________________________________ 11 3.2.2 Driving distance ________________________________________________ 11 3.2.3 House load ____________________________________________________ 11 3.2.4 Solar irradiance ________________________________________________ 12 3.3 Calculations ________________________________________________________ 12 3.3.1 Net-zero Engergy Building and PV power ____________________________ 12 3.3.2 Charging capacity and time _______________________________________ 13 3.3.3 Net resultant energy _____________________________________________ 13 3.3.4 Self-consumption and coverage of PV-power _________________________ 14 3.4 The Model Principle _________________________________________________ 15 4. Results ______________________________________________________________ 17
2 4.1 Simulations of charging scenarios for 1 EV _______________________________ 17
4.1.1 Self-consumption and Self-sufficiency _______________________________ 18 4.2 Sensitivity analysis on number of cars ___________________________________ 18 5. Discussion ___________________________________________________________ 21 6. Conclusions __________________________________________________________ 23 7. Reference list _________________________________________________________ 24 8. Appendix _____________________________________________________________ 27
3
List of terms
EV Electric Vehicle, a vehicle which is partly or fully driven by electricity. In this thesis EV corresponds to Battery Electric Vehicle (BEV) and Plug in Hybrid Electric Vehicle (PHEV).
PV system Photovoltaic system, a power system that uses irradiation of the sun to generate electricity
Self-consumption The use of locally produced power without using the power grid
Self-sufficiency The household load coverage from self-produced energy
Dumb charging Uncontrolled charging
Smart charging Controlled charging
4
1. Introduction
The transport industry accounts for almost 24 percent of the world's greenhouse gas (GHG) emissions, and it is also the primary cause of air pollution in the cities
(International Energy Association (IAE), 2018a). In comparison to other sectors, the transport sector has not seen the same gradual decline of emissions (European
Commission, n.d). To achieve the Paris agreements 2 degrees Celsius-target (2DS), the sectoral emissions must decline in the coming decade (United Nation Climate Change (UNCC), 2019). Calculations suggest the transport industry must achieve a reduction of welll to wheel GHG emission by a minimum of 20 percent for OECD economies until 2025 and 18 percent in non-OECD countries (IEA, 2017a). To reach the 2DS, one highlighted approach is electrification of the transport industry. The United Nation of Climate Change suggested a minimum of 20 percent of all global road transport
vehicles to be electrically driven by 2030 (UNCC, 2015). To target the presented goal, The International Energy Agency set a new Policy Scenario, the EV30@30, to speed up the development of electric vehicles (EV). A collective goal has been established that 30 percent of all global sales of cars must be EVs by 2030 (IEA, 2018b). During 2017, the worlds EV fleet reached over 3 million, a notable increase of 54 percent compared to 2016 (IEA. 2018). However, the EV30@30 scenario suggests 225 million EVs on the roads by 2030, a significant increase compared to today's number.
The penetration of EV to the car fleet has and is still encountering challenges. In comprehension to internal combustion engine vehicles (ICEVs), EVs has a limited driving range and requires a developed charging infrastructure. However, an expansion of the charging infrastructure will challenge the power infrastructure, especially the grid, (Turker et al. 2013) something that must be taken into consideration (IEA, 2017).
Most people have the same daily routines; they go to work in the morning and come home around the same time. After the time of arrival at the home, there is a major use of electricity, resulting in a high electricity demand needed from the grid. The electrical grid has a limited capacity of transported power, if the grid experiences maximum load, overloading of transformers and power lines can occur (Nordling, 2016, p.15). To minimize EVs impact on the grid, especially with the predicted future increase of EVs in the car fleet, research suggests increased use of renewable energy sources (Freire, 2010). Recently, a notable change has been seen in the increasing use of renewable energy sources, especially photovoltaic (PV) generation systems. Combining the technologies of EV and PV could not only reduce the carbon emissions but also reduce the EV load from the grid by smart charging. In order to meet the predicted future increase of an EV car fleet, smart charging is an option to reduce the household peak loads and to increase self-consumption from renewable energy sources.
1.1 Aim
This project aims to examine the potential of smart charging to reduce the household peak loads, increase the PV power self-consumption and self-sufficiency by developing a smart charging algorithm and conducting simulations of smart charging schemes.
5
1.2 Resarch Questions
Can the implementation of smart charging algorithm based on collected data for household electricity consumption and PV production reduce the peak load and increase PV self-consumption and self-sufficiency?
How does the load curve in smart charging scenario differ from the one in the uncontrolled charging scenario?
1.3 Limitations and delimitations
The report examines simulations of smart charging and dumb charging of EVs. The simulations are delimited to EVs at houses with a PV system. During the simulations, up to10 EVs are simulated and all EVs are assumed to be Teslas model S. Therefore, the power consumption characteristics is assumed to be equal for all cars.
The data of driving distance and time of arrival and departure used in the report is limited to the driving patterns provided in the Swedish travel survey (SIKA, 2007). The mobility pattern and parking period of each car is generated from those data using Monto Carlo simulation to estimate the charging demand. The collected data is only valid for Swedish conditions and does not represent a particular domain, such as urban, suburban, or agricultural. Therefore, the results are only valid for Sweden and may appear different in another country or a specific domain.
The data used regarding solar irradiation is obtained from SMHI 2018 for Stockholm region. When using the model in other regions, other results may occur. The size of the installed PV-system is limited to the assumption that the houses are Net-zero Energy Buildings where the buildings produce the same amount of energy as they consume.
For this project, the charging of the EV is delimited to the household level. The study assumes the EV charging only occur at homes. The schemes which includes workplaces or other places of charging are left for further studies.
1.4 Disposition
The report consist of six chapters, the following chapter (2) presents essential
background information of the study. In chapter 3 Methodology and Data are presented.
Furthermore, chapter 4, presents the result of the study, where a sensitivity analysis is included to investigate the reliability of the method. Chapter 5 discuss the results obtained, the following chapter (6) presents conclusions drawn from the study.
6
2. Background
This section presents essential background information of the study. This part is divided into six main sections; The Paris Agreement on Climate change, Electric Vehicles, The Power Grid, PV system, PV self-consumption, and Smart charging.
2.1 Paris Agreement on Climate change
The Paris agreement is an agreement dealing with the threat of climate change and the actions and investments needed for a sustainable low carbon future. The Paris
Agreements aim to act against the threat of climate change by keeping the rising global temperature below 2 degrees Celsius above pre-industrial levels. The agreement brings all nations together to lower the emissions and achieve the 2 degrees Celsius target (2DS) (UNCC, 2018b).
2.1.1 Transport industry and Paris Agreement
The transport industry accounts for almost a quarter of Europe's greenhouse gas (GHG) emissions (European Commission, n.d) and 24 percent in the world 2017 (IAE, 2018b).
The emissions are also the primary cause of air pollutions in the cities. Between 2010 and 2015, the transport industries GHG emissions increased by 2.5 percent annually.
However, to reach the Paris Agreements 2DS and reduce the impact of air pollutions, this trend must be reversed (IAE, 2017). To target the 2DS, the transport emissions must peak around 2020 and then decline by at least 9 percent by 2030 (IEA, 2018a). OECD countries must reduce well to wheel GHG emissions by 25 percent and non-OECD countries by 18 percent until 2025 (IAE, 2017). However, there are some positive trends. In 2017, the emissions increased by 0.6 percent, compared to 1.7 percent annually the past decade, due to electrifications, biofuels, and efficiency improvements.
One of the strategies to decrease the emissions of GHG and air pollutions is electrification of the transport industry (UNCC, 2015).
2.2 Electric Vehicles
An Electric Vehicle (EV), is defined by the U.S. Energy Information Administration (U.S. EIA), as a motor vehicle partly or fully driven by electricity. Unlike fossil fuel- powered vehicles, the EVs are powered by electricity from a rechargeable battery (UCS, 2018). There are different types of EVs, usually categorized into three main categories;
Battery Electric Vehicles (BEV), Plug-in Hybrid Electric Vehicles(PHEV) and Hybrid Electric Vehicle (HEV). The BEV is a purely electric vehicle powered by internal batteries. It must be connected to the electrical grid to charge the batteries, which in turn provides the vehicle´s propulsion. The PHEV and HEV both have internal batteries and internal combustion engines for propulsion. Their main power source for longer
distance driving is the combustion engine, with the electric engine kicking in at certain
7 phases, reducing fuel consumption mainly during start-stop, acceleration and when driving at lower speeds. Hence the means of operation is very similar between the PHEV and HEV. What differs them is the way the batteries are charged. The PHEV can be connected to an external electrical grid to charge the batteries, while the batteries of the HEV are charged solely by its internal combustion engine (Wi, 2013).
2.2.1 History of EVs
Electric vehicles were introduced more than 100 years ago, and at one point a third of the total car fleet consisted of electric vehicles. The electric vehicles were appreciated by users since it was silent and did not consume fuel. However, in 1935, the EVs
disappeared from the market partly due to improved infrastructure, an increased number of gas stations, and fuel price reduction. These improvements made it possible to travel longer distances, and at this point, EVs could no longer compete with the fuel driven vehicles. At 1970, when the fuel price increased, the EVs appeared on the market again.
Although, it was not until the beginning of 2000 when the EVs reached a turning point, and the development of EVs proceed. Until this point, the driving range for EVs was still limited. In 2006, the Silicon Valley startup, Tesla Motors, announced the
development of an EV with a driving range of 200 miles This, along with various other motives such as the environmental advantages contributed to the development of EVs in already established vehicle manufacturing companies (Matulka, 2014).
2.2.2 The future of EVs
To reach the Paris agreements 2 degrees Celsius target, electrification of the transport sector is needed. The EV30@30 initiative presents the opportunity to simplify the transition to a fully renewable energy system through transport electrification.
Furthermore, a collective goal has been established that 30 percent of all global sales of cars must be electric vehicles by 2030 (IEA, 2018). In Sweden, a member of the
EV30@30 initiative, the Cross-Party Committee on Environmental Objectives proposed a goal for reduction of emissions in the domestic transport sector by at least 70 percent by 2030 compared to the levels of 2010 (Government Office of Sweden, 2016). To target the presented goal, Fortum Charge & Drive together with Swedenergy, compiled The Almedalen Manifesto 2016. The Manifesto states how electric vehicles and
charging infrastructure should be promoted to reach the Cross-Party Committee on Environmental Objects proposed goal. The Manifesto states that electric vehicles and the charging infrastructures technical development have reached a point where an introduction to the broader market is possible. Therefore, the Manifesto proposes a goal of two million EVs in Sweden until 2030 to achieve the proposed reduction. At the first quarter of 2019, 20447 EV were registered and together with Plug-in Hybrid Electric Vehicles (PHEV), the number was roughly 74000, an increase of 50 percent from the year before (Elbilstatistik, 2019).
8
2.3 The Power Grid
The Swedish national power grid is one of the oldest electrical grids in the world and consists of 15000 km of power lines. The Swedish electricity market has been
dominated by nuclear and hydropower, contributing to around 80 percent of the national generated electricity. The power grid transports and distributes the generated electricity from where it is produced to where it is needed (Svenska Kraftnät, 2017a). The grid is divided into three levels; the national grid, regional grid, and local grid. In Sweden, the authority responsible for the national grid is Svenska kraftnät, and around 170
companies are operating the local- and regional grid (Nordling, 2016, p.7). Svenska Kraftnät is responsible for ensuring the balance of energy production in Sweden,
making sure that the energy produced is the same as the energy consumed. If there is an imbalance between these two in the electrical system, there is a risk of disruptions on the grid and power failure (Svenska Kraftnät, 2017b).
2.3.1 Challenges
The society is dependent on reliable energy production to maintain the functions of industries, households, infrastructures, and other vital services. One of the challenges today is the limited capacity of transported power from the power grid, something that could contribute to power distributions. Sweden is already experiencing problems with the capacity of the grid, especially during peak hours when there is major use of electricity in the households, resulting in a high electricity demand needed from the grid. The electrical grid has a limited capacity of transported power. If the maximum load is withdrawal from the grid, overloading of transformers and power lines can occur (Nordling, 2016, p.15). Future prognosis predicts further electrification of society and an increasing EV fleet, something that will increase the load on the grid even more. If charging of the EVs will occur during peak hours, the peak load will increase and the risk of disruptions on the grid and power failure will emerge. Therefore, if the charging demand is regulated, the charging load of the EVs could be limited during peak hours.
Regulated charging could also increase the use of intermittent renewable energy sources (Nordling, 2016, p.17). However, the electrical grid in Sweden must be developed to meet future demand (Svenska kraftnät, 2019).
2.4 PV system
Photovoltaics are electrical power generating systems converting light into electricity by the use of semiconductor materials. The sun rays consist of packets of energy (photons).
When the photons hit the PV cell, the electrons in the cell excite due to charge
separation in the absorbing material. The movement of the electrons generates a small voltage, and if the circuit is closed, a direct electrical current (DC) is generated. The DC electricity is converted into an altering current (AC) by a solar inverter (Schavemaker
9 and van der Sluies 2008:62). Since the produced electrical current from one cell is relatively small, more PV cells are usually linked together into a PV panel to produce a higher amount of energy.
PV systems are usually categorized as off-grid- or grid-connected systems. Where off- grid systems are initially set up to generate power for individual systems and devices similar to boats, cabins, and farm appliances. On-grid connected PV systems are dominating the market of installed PV power. The grid-connected PV power is mainly roof-mounted systems installed by companies and private persons (Swedish Energy Agency, 2016a). The most common types of PV panels are Polycrystalline solar cells, monocrystalline solar cells, and thin-film solar cells, whereas the first two dominates the market. The type of PV panels differs in price, material, efficiency, and flexibility (Swedish Energy Agency, 2016a).
2.4.1 PV power today and future predictions
One of the advantages of PV systems is the modular technology. The systems can be small, such as calculators and off-grid applications, and scaled up to extensive power generations facilities connected to the grid (IEA, n.d solar energy). The disadvantages of PV power is the reliability of its intermittent power generation. Therefore, to maximize the benefits of PV power, an energy storage system is required, and the development of smart charging systems.
During the last decade, the installed global PV power capacity has increased. At the beginning of 2000, the number of globally installed PV power capacity was around 1GW (Swedish Energy Agency, 2018). Since then, the number has increased, and at the beginning of 2017, the installed PV power capacity was around 398GW. Solar PV is dominating the growth of renewable energy, and a future prediction expects growth of 575GW to become operational until 2023 (IEA, 2018). In Sweden, the installed capacity of grid-connected PV power increased by 65 percent in 2017 compared to 2016
(Swedish Energy Agency, 2018).
2.5 PV self-consumption
PV self-consumption refers to the share of energy produced from an installed PV system which is directly used in the building of the installed PV system. There are several advantages with self-consumption, both technical and economic. When the PV power is used directly in the building, the amount of power needed from the grid decreases. Consequently, the load on the grid is reduced. Another benefit of increased self-consumption is the economic advantage when less energy is bought from the grid.
One technical advantage is the reduction of cable losses due to transportation. In the year of 2014, 6.7 percent of the produced energy in Sweden was lost as a result of cable loses whereas the losses from an installed PV system connected to the building of consumption was no more than 1 percent (Swedish Energy Agency, 2016b, p.14).
However, to reach the full potential of self-consumption further technological
development is needed. Development of regulatory smart battery-charging algorithms is
10 highlighted as a solution to reach the full potential of self-consumption (Dehler et al., 2017).
2.6 Smart Charging
During 2017, the global EV fleet reached over 3 million, a significant increase of 54 percent compared to 2016. The International Energy Agency predicts the number of EVs to reach 125 million by 2030, a number that might increase further (IEA, 2018b).
The penetration of EVs to the car fleet encounter challenges. EVs have a limited driving range and requires developed charging infrastructure, which will challenge the grid, especially during peak hours (Nordling, 2016). The ability to control the batteries charging allows for smart charging to maximize self-consumption of renewable energy sources and to reduce the load on the grid (Pecas Lopes et al, 2009). By charging when the power production from renewable sources is high, and by avoiding charging during peak hours, the negative impacts on the electric grid will be reduced. Other benefits are the increased use of self-consumption from renewable energy sources, reduced GHG emissions, and cost reductions. In the future, the power system requires not only flexible power generation but flexible power usage as well. In this way, excessive current loads on the local grid can be prevented (IEA, 2018b).
11
3. Methodology and data
In this section, the methodology, calculations, and the use of collected data to conduct the study are presented.
3.1 Overview
In this report, a rule based algorithm for smart charging is developed for simulations in MATLAB. The smart charging algorithm uses input parameters of driving distance, park period, solar power production and household energy consumption. With this information the model calculates and allocates the charging load for each hour,
depending on if smart charging and self-consumption from the installed PV-system are possible to see Figure 1, Flowchart. However, if charging from the PV system is not possible, the energy is supplied to the EV from the electrical grid. The algorithm allocated the charging load for each hour during the parking time. The smart charging scenario will be compared to dumb charging scenario to evaluate the potential of smart charging.
3.2 Data
3.2.1 Departure time and Arrival time
The arrival time represents the specific time when the EV arrives at the home, whereas the departure time is the time when the EV leaves the home. The data consist of driving sessions collected from the Swedish travel data survey during the year of 2005-2006 (SIKA, 2007). In each simulation of the model, for each day during the year, a departure and arrival time is generated from the data through the Monte Carlo method, then the parking period duration is determined from that.
3.2.2 Driving distance
The driving distance is generated from the Swedish travel data survey conducted the year of 2005-2006. It represents the driving distance for several cars every day for one year. A randomized driving distance is assigned to each EV in the simulation, each day of the year through the Monte Carlo method.
3.2.3 House load
The household load data is generated from the Widén-Markov model for typical Swedish detached houses.The Widen-Markov model is a stochastic generation for household electricity load patterns with probabilistic approach (Widén and Wäcklegård, 2009).
12 3.2.4 Solar irradiance
For the solar irradiance, data from the Swedish Meteorological and Hydrological Institute is used. The data is conducted from Bromma Airport and consists of hourly resolutional horizontal solar irradiance, [W/m2], throughout the year of 2018. The solar irradiance for each month during 2018 is illustrated in Figure 1.
Figure 1. Monthly solar irradiance for each month during 2018.
3.3 Calculations
3.3.1 Net-zero Engergy Building and PV power
To estimate the produced PV power, 𝑃𝑝𝑣, scaling of a Net-zero Energy buildings is used. A Net-zero Building produces the same amount of renewable energy to meet the buildings annual energy consumption (U.S Department of Energy, n.d). The Net-Zero Energy building, 𝐸𝑝𝑣, is calculated using equation 1.
𝐸𝑝𝑣 = 𝐸ℎ𝑜𝑢𝑠𝑒+ 𝐸𝐸𝑉 (1)
Where Epv is the Net-Zero Energy Buildings consumption annually [W], Ehouse the household load [W], and EEV the load from the EV [W].
Equation 1 is used to calculate the amount of energy needed to cover the consumption of a Net-zero Energy Building. To estimate the size of the PV power system, equation 2 is used where Ppv=EPV.
𝑃𝑝𝑣 (𝑡) = 𝑆 × 𝐼(𝑡) (2)
Where Ppv is the produced PV power for every hour, I the solar irradiance from SMHI [W/ m2], and S the size of the PV system.
13 3.3.2 Charging capacity and time
All EVs are assumed to be Tesla Model S with the battery capacity (𝐵𝐶) of 75 kwh. The average energy consumption, 𝐸𝑐𝑜𝑛, is set to 0.2 kWh/ km. The battery's state of charge (𝐵. 𝑆𝑂𝐶) at time of arrival is calculated using equation 3;
𝐵. 𝑆𝑂𝐶 = 𝐵𝐶 − 2 × 𝐷𝑑𝑖𝑠𝑡𝑎𝑛𝑐𝑒× 𝐸𝑐𝑜𝑛 (3)
Where 𝐵. 𝑆𝑂𝐶 is the battery state of charge when arriving at the home [kWh], 𝐵𝐶 the maximum battery capacity [kWh], 𝐷𝑑𝑖𝑠𝑡𝑎𝑛𝑐𝑒 the driving distance [km] and 𝐸𝑐𝑜𝑛 the energy consumption for the EV [kWh/km].
The energy required until fully charge is calculated using equations 4;
𝐸𝑟𝑒𝑞= 𝐵𝐶 − 𝐵. 𝑆𝑂𝐶 (4)
Where 𝐸𝑟𝑒𝑞 is the energy required [kWh], 𝐵𝐶 the maximum battery capacity [kWh] and 𝐵. 𝑆𝑂𝐶 the battery state of charge when arriving at home[kWh].
To estimate the charging capacity, 𝐶𝑥(𝑡), needed for each hour during the charging time, 𝑡𝑥, equations 5 is used:
𝐶𝑥(𝑡) =𝐸𝑟𝑒𝑞 𝑡𝑥
(5)
Where 𝐶𝑥(𝑡) is the charging capacity [ kW] and 𝑡𝑥 the charging time during the parking period [h]. 𝐶𝑥(𝑡) and 𝑡𝑥 varies depending on parameters illustrated in figure 3. An explanation of the parameters for x is presented in table 1.
3.3.3 Net resultant energy
To increase the PV self-consumption, the variable net resultant energy (𝑁𝑟𝑒𝑞) is introduced. The net resultant energy is the produced PV power minus the household load, calculated using equation 6;
𝑁𝑟𝑒𝑞= 𝑃𝑝𝑣 − 𝑃ℎ𝑜𝑢𝑠𝑒𝑙𝑜𝑎𝑑 (6)
Where 𝑁𝑟𝑒𝑞 is the net resultant energy [kWh], 𝑃𝑝𝑣 the produced PV power calculated in equation 2 and 𝑃ℎ𝑜𝑢𝑠𝑒𝑙𝑜𝑎𝑑 the load from the household [kW].
14 When the net resultant energy is greater than the charging demand for the specific hour, charging from the produced PV should occur to improve self-consumption from the PV system. Therefore, the charging during these hour/hours, 𝐶𝑝𝑣 are equal to the net resultant energy. Using equation (5) and (6):
𝐶𝑝𝑣 = 𝑁𝑟𝑒𝑞 (7)
Where 𝐶𝑝𝑣 is the charging capacity when 𝐶𝑥< 𝑁𝑟𝑒𝑞
3.3.4 Self-consumption and coverage of PV-power
PV Self- consumption (𝜑𝑠𝑐) is the share of energy produced from an installed PV system which is directly used in the building of the installed PV system. In this report, the energy is consumed by household and the EV, illustrated as the black line in figure 2.
Figure 2. The yellow line represents the produced PV power and the black line represents the electricity load from the household.
The level of self-consumption illustrated in figure 2, can be expressed as:
𝜑𝑠𝑐 = 𝐶 𝐵 + 𝐶
(8)
The self-sufficiency (𝜑𝑠𝑠) is how much of the total household load that is consumed from the produced PV power. The level of 𝜑𝑠𝑠 illustrated in figure 2, can be expressed as:
𝜑𝑠𝑠 = 𝐵 𝐴 + 𝐵
(9)
15
3.4 The Model Principle
The flowchart describes the model of the smart charging algorithm simulated in MATLAB. It presents the charging capacity of the EV each hour during the parking period. When the EV arrives at home, the charging demand is calculated using equations 2. In the smart charging scenario, the charging capacity for every hour is regulated after specified scenarios, illustrated in figure 2. During the dumb charging scenario, the charging capacity is not regulated. When the EV arrives, the charging starts with maximum charging capacity until the battery is full or when the EV departures.
Since the driving distance, time of arrival and departure is randomly generated each day, the EVs charging demand and parking time is different for each day of the simulation throughout the year. The idea of the smart charging algorithm is to reduce the peak load and increase the PV self-consumption. The peak hours are set to 17.00, 18.00 and 19.00, during these hours the EV should not charge if possible or be limited, illustrated in figure 3. To increase the PV self-consumption and self-sufficiency, charging from the produced PV power should occur when possible, calculated by equation 7.
16
[RF1]
Figure 3. The Flowchart describing the smart charging algorithm. Equation 2 is used to estimate the energy required for each hour, 𝐸𝑟𝑒𝑞. The parameters used for the charging time, 𝑡𝑥, and charging capacity, 𝐶𝑥, are presented in table 1.
Table 1. Explanation of 𝑡𝑥 and 𝐶𝑥 tpp Hours during parking period
top Hours during off-peak tp Hours during peak
tpv Parking hours when Nre> Cop
Cop Charging capacity during time off-peak using equation 3 Cp Charging capacity during peak time using equation 3 Cmax Maximum charging capacity
Cpv Charging capacity equals Nre
Crest Charging during off-peak, when charging of PV have occurred
17
4. Results
In this section the result of the simulations are presented. Section 4.1 presents the simulations of 1 EV during for dumb charging and smart charging. Section 4.2 presents a sensitivity analysis.
4.1 Simulations of charging scenarios for 1 EV
The result of the simulations with and without the smart charging algorithm is presented as a mean daily net-load profile. Two simulations are presented, the first is the dumb charging simulation, the second simulation is for when the smart charging algorithm is used. During both simulations, the maximum charging capacity was 3.7kW.
For the dumb charging simulation, 1 EV is simulated during all hours of the year and the mean value for the daily generated net-load profile is illustrated in figure 4. For the dumb charging simulation, charging of the EV increased the peak load, and most of the charging occurred during peak hours, illustrated in figure 4.
For the simulation of the smart charging algorithm presented in figure 3, 1 EV is simulated during all hours of the year and the mean value for the daily generated net- load profile is illustrated in figure 4. For the smart charging simulation, the peak load is reduced by 38.64 percent on average during peak hours compared to the dumb charging simulation. Table 2 presents the results of self-consumption and self-sufficiency for the smart charging and dumb charging scenario.
(a) (b)
18 Figure 4. Graph of the mean daily net-load profile of dumb charging (a) and smart
charging (b). The simulations are for 1 EV.
4.1.1 Self-consumption and Self-sufficiency
The PV self-consumption, the share of energy produced from the installed PV system directly used by the household, is presented in table 2 and was calculated using equation 8. From the result of the mean household load, the self-sufficiency is calculated using equation 9. Three different scenarios were simulated for the self-consumption and self- sufficiency, all using the mean daily net profile for a simulated year and 1 EV. The simulated scenarios were the household load without the EV and the household load with smart charging and dumb charging illustrated in figure 4. The results are presented in table 2.
Table 2. Self-consumption and self-sufficiency for the household with dumb charging and smart charging of 1 EV and without the EV.
Self-consumption [%] Self-sufficiency [%]
without EV 24.95 39.02
With EV using dumb charging
29.49 29.49
With EV using smart charging
34.18 34.46
Difference using smart charging versus dumb
charging
4.69 4.97
4.2 Sensitivity analysis on number of cars
In the sensitivity analyses, all simulations are made using the same variables as in the result section, apart from the number of cars. Simulations of 5 EVs and 10 EVs are made to see how the result of self-consumption, self-sufficiency and peak load reduction is affected, presented in table 3.
For the simulation illustrated in figure 5, the number of EVs are set to 5.
19
(a) (b)
Figure 5. Graph of the mean daily net-load profile of (a) dumb charging and (b) smart charging. The simulations are for 5 EVs.
For the smart charging simulation 5 EVs, the peak load was reduced by 31.91 percent compared to the dumb charging simulation. Table 3 presents the self-consumption and self-sufficiency calculations for the house load without the EVs, and with smart charging and dumb charging.
For the simulation illustrated in figure 6, 10 EV are used.
(a) (b)
Figure 6. Graph of the mean daily net-load profile of (a) dumb charging and (b) smart charging. The simulations are for 10 EVs.
For the smart charging simulation with 10 EVs, the peak load was reduced by 28.35 percent compared to the dumb charging simulation. Table 4 presents the self-
20 consumption and self-sufficiency calculations for the house load without the EVs, and with smart charging and dumb charging.
Table 4. Self-consumption (SC) and self-sufficiency (SS) for houses with dumb charging and smart charging for 1 EV and aggregation of 5 EVs and 10 EVs and peak load
reduction using smart charging for the three scenarios.
Number of EV-house couplings
smart charging SC SS
dumb charging SC SS
without EV SC SS
Peak load reduction with smart charging compared to dumb
charging 1 34.18 34.46 29.49 29.49 24.95 39.02 38.64 5 35.53 35.05 33.30 33.34 25.36 39.82 31.96 10 35.94 35.30 34.12 34.15 25.49 39.95 28.35
21
5. Discussion
In this section, the assumptions and results are analyzed and discussed.
For the result of 1 simulated EV the peak load was reduced by 38.64 percent compared to dumb charging. Today, the Swedish electrical grid is already experiencing a shortage of capacity since the grid has a limited capacity of transported power. A challenge not only seen in Sweden but also globally. If the predicted EV fleet in Sweden will reach the prognosis of 2 million cars before 2030, the charging infrastructure, if all EVs charge directly when arriving at the house, the peak load will increase significantly, illustrated in figure 4. The necessary charging infrastructure for the predicted EV fleet would most likely not be possible with today's grid capacity. The proposed smart charging algorithm could, therefore, be one solution since the peak load reduction was reduced by an average of 38.64 percent and all EVs reach their required battery capacity. Although, the electrification of the society and the transport sector will most likely require and combination of smart charging systems and development of the grid.
The PV self-consumption increased by 4.69 percent with the smart charging scenario compared to dumb charging scenario, when 1 EV was simulated. For a household without an EV, the PV self-consumption was an average of 24.95 percent. With dumb charging, the number increased to 29.49 percent, and for smart charging, the result was 34.18. Since the PV power is produced mostly during day time, a large share of the produced power cannot be used directly. To increase the PV self-consumption, an ideal future scenario is if the PV power could be stored and used later when the electric consumption is high. By using the smart charging algorithm, the PV self-consumption is increased, resulting in increased use of renewable energy for the studied houses. When the PV power is directly consumed, the cable losses are limited compared to if
withdrawn from the grid. Therefore, the smart charging algorithm is beneficial since it reduces power losses and increases the use of renewable energy sources. This aspect should be considered further in future studies.
The self-sufficiency increased when the smart charging algorithm is implemented compared to the self-sufficiency for the household without an EV. When dumb charging scheme is implemented, the self-sufficiency decreased compared to the household without the EV. All houses with an EV should, therefore, use smart charging since the self-sufficiency level would decrease otherwise. Furthermore, smart charging is one option to increase the houses self-sufficiency, resulting in more sustainable living, both economically and environmentally.
For the simulation of 5 and 10 EVs, the peak load reduction was 31.96 percent and 28.35 percent. The self-consumption and self-sufficiency were an average of 35.53 percent and 35.05 for 5 EVs. The level of self-consumption and self-sufficiency
22 increased to 35.94 percent and 35.30 percent for 10 EVs while implementing the smart charging scheme. The average of peak load reduction by smart charging decrease by 10.29 percent for 10 EV-house couplings compared to single EV-house coupling. This is most likely due to the aggregation effect that already reduce the system peak loads.
An improvement of the reliability and the results would be to increase the numbers of simulated cars. In further studies, the house load, type of EV, driving distance, parking time, and solar irradiance should be simulated over an expanded time period. Since the data obtained might vary over different years.
The smart charging algorithm is beneficial since it can reduce the peak loads and increase both self-sufficiency and self-consumption. A future predicted penetration of EVs into the car fleet would, therefore, require controlled charging to meet the
increased electricity demand from electrification of the transport sector. Smart charging would not only smoothen the transition for the future EV fleet but also accelerate the transition of renewable energy. This is crucial to achieve the United Nations climate change GHG emissions goal of reduction.
This study could be elaborated with more simulations of different scenarios for the smart charging algorithm and compared with dumb charging. An idea for future studies is a simulation of different car models with different battery capacity and consumption where the storage of the produced PV power is considered. Some of the houses may not only have one car. Therefore, the smart charging algorithm should be further developed to include a different number of cars. Importantly, the results from the smart charging scenarios highlight the benefit of smart charging. Therefore, this should be studied further.
23
6. Conclusions
The implementation of the developed smart charging algorithm can increase the houses PV self-consumption compared to the scenario without smart charging by 4.69 percent on average. Therefore, the smart charging scheme would increase the users share of renewable energy only by controlling the charging of the EV. Also, the self-sufficiency increased by 4.97 percent on average compared to dumb charging scenario. The
scenarios of 5 EV-house couplings and 10 EV-house coupling increased self-
consumption and self-sufficiency further. Therefore, the effect of aggregation by using smart charging could have a positive impact to increase the use of renewable energy sources.
The predicted future increase of EVs requires a developed charging system since it will increase the load on the grid. If only dumb charging schemes are used, the peak of the load curve would be significantly higher compared to the smart charging scenario. The smart charging scheme reduces the peak load by an average of 38.64 percent. A smart charging algorithm is therefore crucial to reduce the peak load.
24
7. Reference list
Dehler J. 2017 et al, Self-Consumption of Electricity from Renewable Sources.
Available online:
https://www.sciencedirect.com/topics/engineering/self-consumption (2019-05-02) Elibilstatistik, 2019, Elbilsstatistik. Available online:
https://www.elbilsstatistik.se/elbilsstatistik (2019-04-17).
Energiforetagen, 2017. Almedalsmanifestet för fler elfordon - Energiföretagen Sverige.
Available online: https://www.energiforetagen.se/sa-tycker-vi/almedalsmanifestet/
(2019-04-19).
European Commission, n.d, A European Strategy for low-emission mobility. Available online: https://ec.europa.eu/clima/policies/transport_en (2019-04-14)
Freire, R. Delgado, J. Santos, J. Almeida, A. “Integration of Renewable Energy Generation with EV Charging Strategies to Optimize Grid Load Balancing” IEEE, 2010
Goverment Office of Sweden, 2016. Broad consensus in Riksdag on proposal for Sweden’s future. Available online: https://www.government.se/press-
releases/2016/06/broad-consensus-in-riksdag-on-proposal-for-swedens-future/
(2019-04-14)
H. Turker, A. Hably, S. Bacha, “Smart Charging of Plug-in Hybrid Electric Vehicles (PHEVs) on the Residental Electric grid regarding the Voltage Plan.” IEEE, 2013 International Energy Association (IAE), 2007, Strong policy and falling battery costs
drive another record year for electric cars. Available online:
https://www.iea.org/newsroom/news/2018/may/strong-policy-and-falling-battery- costs-drive-another-record-year-for-electric-ca.html (2019-04-10)
International Energy Association (IAE), 2017a, Tracking Progress: Transport.
Available online: https://www.iea.org/etp/tracking2017/transport/ (2019-04-14) International Energy Association, (IAE), 2018a, Transport, Tracking Clean Energy
Progress. Available online: https://www.iea.org/tcep/transport/ (2019-05-09) International Energy Association, (IAE), 2018b, Global EV outlook 2018. Available
online: https://www.iea.org/gevo2018/ (2019-04-14)
Nordling, A. (2016). Sveriges framtida elnnät, IVA. Available online:
https://www.iva.se/globalassets/rapporter/vagval-energi/vagvalel-sveriges-framtida- elnat.pdf (2019-05-07)
Peças Lopes, A. Soares, F. J. Almeida, P. M. Moreira da Silva, M. (2009), ”Smart Charging Strategies for Electric Vehicles: Enhancing Grid Performance and
25 Maximizing the Use of Variable Renewable Energy Resources” Instituto de
Engenharia de Sistemas e Computadores do Porto (INESC Porto) and Faculdade de Engenharia da Universidade do Porto (FEUP), Porto, Portugal
Schavemaker, P. Van der Sluis, L (2008). Electrical power system essentials, Wiley, Chichester 2008.
Svenska Kraftnät, 2017a, Drift av stamnätet. Available online: https://www.svk.se/drift- av-stamnatet/ (2019-05-07)
Svenska Kraftnät, 2017b, Trygg elförsörjning. Available online:
https://www.svk.se/drift-av-stamnatet/trygg-elforsorjning/ (2019-05-07) Svenska Kraftnät, 2019, Drift och Marknad. Available online:
https://www.svk.se/drift-av-stamnatet/drift-och-marknad/ (2019-05-07)
Swedish Energy Agency, 2016a, National Survey Report of PV Power Applications in Sweden. Available online:
http://www.energimyndigheten.se/globalassets/fornybart/solenergi/national_survey_r eport_of_pv_power_applications_in_sweden_-_2016.pdf (2018-05-07)
Swedish Energy Agency, 2016b, Solceller i omvärlden. Available online:
http://www.energimyndigheten.se/globalassets/fornybart/solenergi/solen-i- samhallet/solceller-i-omvarlden.pdf (2019-07-08)
Swedish Energy Agency, 2018a, Kraftig ökning av nätanslutna solcellsanläggningar.
Available online:
http://www.energimyndigheten.se/nyhetsarkiv/2018/kraftig-okning-i-natanslutna- solcellsanlaggningar/ (2019-05-09)
Swedish Energy Agency, 2018b Systemperspektiv i Världen. Available online:
http://www.energimyndigheten.se/fornybart/solelportalen/lar-dig-mer-om- solceller/systemperspektiv-i-varlden/ (2019-05-07)
United nation Climate Change (UNCC), 2019, Nationally Determined Contributions (NDCs). Available online: https://unfccc.int/process/the-paris-agreement/nationally- determined-contributions/ndc-registry
Unit Nations Climate Change, 2018, The Paris Agreement. Available online:
https://unfccc.int/process-and-meetings/the-paris-agreement/the-paris-agreement (2019-05-09)
United nation Climate Change (UNCC), 2015, Paris Declaration on Electro-Mobility and Climate Change and Call to Action, United nation Climate Change, Paris.
26 U.S. Department of Energy, n.d, A Common Definition for a Zero Energy Building.
Available online: https://www.energy.gov/eere/buildings/downloads/common- definition-zero-energy-buildings (2019-05-12)
Y-M. Wi, J-U. Lee, S-K. Joo, (2013). “Electric Charging Method for Smart Homes/Buildings with a Photovoltaic System” IEEE,
Widén, J. Wäckelgård, E. (2009) “A high-resolution stochastic model of domestic activity patterns and electricity demand”, Applied energy 2010:87:1880-1882.
Uppsala
27
8. Appendix
clear close all rng(1)
%% Loading data and declare cases
load('household_load') load('solar_irr_dat')
load('distance_arrival_departure_sample_hour')
numcar = 10; %number of cars
days = 365; %number of the day in a year eff = 1; %charging efficiency
%% monte carlo distance and time arrival and departure s = RandStream('mlfg6331_64');
dist = zeros(days,numcar);
t_arr_day = zeros(days,numcar);
t_dep_day = zeros(days,numcar);
for rs = 1:numcar
dist(:,rs) = randsample(s,dist_sample,days);
t_arr_day(:,rs) = randsample(s,t_arr_sample_hour,days);
t_dep_day(:,rs) = randsample(s,t_dep_sample_hour,days);
end
%convert hour of the day into hour of the year slot in a year day_to_year_reso = (0:24:24*365-1)';
t_arr = t_arr_day + day_to_year_reso;
t_dep = t_dep_day + day_to_year_reso + 24;
%% EV type, and solar, ev, load calculation
%battery capacity of EV soc_max = 75e3; %wh?
charge_max = 3.7e3; %charging capacity
%km -> kwh
%Tesla 20 kWh/100 km
soc_arr = soc_max-2*dist*20000/100;
%hyundai IONIQ 15 kWh/100 km
%soc_arr = soc_max-2*dist*15000/100;
soc_arr(find(soc_arr<1000))=1000;
%Energy EV and household needed soc_target = soc_max;
energy_req_ev_day = (soc_target - soc_arr)/eff; %already in kWh, divided by charging efficiency energy_req_ev_year = sum(energy_req_ev_day); %already in kWh
energy_house_year = sum(load_60(:,1:numcar)); %kW15m -> kWh
%PV power production
energy_solar_irr_year = sum(solar_irr); %kW15m -> kWh
%solar-ev-load NZEB
solar_NZEB_plant = (energy_req_ev_year+energy_house_year)/energy_solar_irr_year; %area solar_plant = 1*solar_NZEB_plant;
solar_prod = solar_irr*solar_plant; %+1gmt
%solar_prod(89*24:299*24)= solar_prod(89*24-1:299*24-1);% with daylight saving
%solar_prod(2:8760)= solar_prod(1:8759);%gmt+2
h_load = load_60(:,1:numcar);%(idx_mod,:);
28
%solar_prod = solar_prod%(idx_mod,:);
h_load_net = h_load - solar_prod;
%copying the house load as a initial condition for dumb/smart charging h_load_net_ev_dumb = h_load_net;
h_load_net_ev_smart = h_load_net;
h_load_PV=solar_prod-h_load;
%charging load
smart_charge = zeros(8760,numcar);
dumb_charge = zeros(8760,numcar);
%% Charging scheduling
for i = 1:365
for id=1:numcar
if i<days %swithchin the last date of the year park_period = t_arr(i,id):t_dep(i,id);
else
end_day_time = days*24;
idt1 = t_arr(i,id):days*24;
idt2 = 1:(t_dep(i,id)-1)-days*24;
park_period = [idt1,idt2];
end
%% dumb charging b_soc = soc_arr(i,id);
j = t_arr(i,id);
while b_soc < soc_target && j < max(park_period)
if soc_target - b_soc >= eff*charge_max dumb_charge(j,id) = charge_max;
b_soc = b_soc + eff*dumb_charge(j,id);
else
dumb_charge(j,id) = (soc_target - b_soc)/eff; %divided by effi to make the battery full b_soc = soc_target;
end
h_load_net_ev_dumb(j,id) = h_load_net_ev_dumb(j,id) + dumb_charge(j,id);
j = j+1;
end
%% smart charging ip = 0; %index_peak inp = 0; %index_offpeak
%peak and off peak parking park_peak = [];
park_offpeak =[];
for m = 1:length(park_period)
if park_period(m) == 17 || park_period(m) == 18 || park_period(m) == 19 %or statement, peak time 17, 18, 19- ip = ip+1;
park_peak(ip) = park_period(m); %parking time during peakpeak else
inp = inp +1;
park_offpeak(inp) = park_period(m); %parking time during offpeak end
end
t = length(park_period); %number of hours of parking period
t_p = length(park_peak); %number of hours of parking period during peak load t_op = length(park_offpeak); %number of hours of parking period during offpeak load
%algorithm implementation b_soc = soc_arr(i,id);
j = t_arr(i,id);
energy_req=soc_max-b_soc; %kwh that the cars has consumed
if energy_req/t_op>=3.7e+03*eff
29
Enp=soc_max-(t_op*eff*charge_max)-b_soc; %+(eff*charge_max*t_op); %batterys state of charge with maximum charging power during no peak
charge_nopeak=charge_max*eff;
charge_peak=(soc_max-Enp)/t_p;
if t_p<=0 t_p=1;
end
if (soc_max-Enp/t_p)>=(charge_max*eff)
smart_charge(park_period,id)=charge_max*eff; %charge with charge_max*eff during all hours % b_soc = b_soc +smart_charge(j,id);
else
smart_charge(park_nopeak,id)=charge_nopeak;
smart_charge(park_peak,id)=charge_peak; %charge with charge_peak end
else %if the EV must not charge during peak hours
charge=energy_req/t_op; %charging amount for each hour ipv_x = 0; %index_solar_excess
ipv_s = 0; %index_s park_PVexcess = [];
park_PVshortage =[];
for n = 1:length(park_offpeak) %for all hours during parking period
if h_load_PV(park_offpeak(n))>charge %om houseload is greater than zero %pv_charge(n)=h_load_PV(pv_hour,id);
ipv_x = ipv_x+1;
park_PVexcess(ipv_x) = park_offpeak(n);
else
ipv_s = ipv_s+1;
park_PVshortage(ipv_s) = park_offpeak(n);
end end
t_pv_x=length(park_PVexcess); %numbers of hours when PV produced-houseload>0;
t_pv_s=length(park_PVshortage); %numbers of hours when PV produced-houseload<0;
if t_pv_x>0 %if there is more produced PV than the housload during the parking period
if sum(h_load_PV(park_PVexcess)) < energy_req
smart_charge(park_PVexcess,id) = min(3.7e3,h_load_PV(park_PVexcess));
smart_charge(park_PVshortage,id) = (energy_req-sum(smart_charge(park_PVexcess)))/t_pv_s;
elseif sum(h_load_PV(park_PVexcess)) > energy_req smart_charge(park_PVexcess,id) = energy_req/t_pv_x;
smart_charge(park_PVshortage,id) = 0;
end
else %if there is no hour where PV is higher than load smart_charge(park_period,id)=charge;
end end
h_load_net_ev_smart(park_period,id) = h_load_net_ev_smart(park_period,id) + smart_charge(park_period,id);
end end %%
%system net-load sys_load = sum(h_load,2);
sys_load_net = sum(h_load_net,2);
sys_dumb_charge = sum(dumb_charge,2);
sys_smart_charge = sum(smart_charge,2);
sys_load_net_ev_dumb = sum(h_load_net_ev_dumb,2);
30
sys_solar_prod = sum(solar_prod,2);
sys_load_net_ev_smart=sum(h_load_net_ev_smart,2);
%performance comparison in statistics
mean_load_net_no_ev = mean(mean(h_load_net));
mean_load_net_ev_dumb = mean(mean(h_load_net_ev_dumb));
std_load_net = std(sum(h_load_net,2));
std_load_net_ev_dumb = std(sum(h_load_net_ev_dumb,2));
std_sysload = std(sys_load);
std_sysload_net = std(sys_load_net);
std_sysload_net_ev_dumb = std(sys_load_net_ev_dumb);
%daily average, declaring variable to optimize the code h_load_davg = zeros(24,numcar);
h_load_net_davg = zeros(24,numcar);
solar_prod_davg = zeros(24,numcar);
h_load_net_ev_dumb_davg = zeros(24,numcar);
dumb_charge_davg = zeros(24,numcar);
smart_charge_davg = zeros(24,numcar);
sys_load_davg = zeros(24,1);
sys_load_net_davg = zeros(24,1);
sys_solar_prod_davg = zeros(24,1);
sys_load_net_ev_dumb_davg = zeros(24,1);
sys_dumb_charge_davg = zeros(24,1);
sys_smart_charge_davg = zeros(24,1);
for k = 1:24
idx_day = k:24:8760;
h_load_davg(k,1:numcar) = mean(h_load(idx_day,1:numcar));
h_load_net_davg(k,1:numcar) = mean(h_load_net(idx_day,1:numcar));
solar_prod_davg(k,1:numcar) = mean(solar_prod(idx_day,1:numcar));
h_load_net_ev_dumb_davg(k,1:numcar) = mean(h_load_net_ev_dumb(idx_day,1:numcar));
dumb_charge_davg(k,1:numcar) = mean(dumb_charge(idx_day,1:numcar));
smart_charge_davg(k,1:numcar) = mean(smart_charge(idx_day,1:numcar));
%system level
sys_load_davg(k,1) = mean(sys_load(idx_day,1));
sys_load_net_davg(k,1) = mean(sys_load_net(idx_day,1));
sys_solar_prod_davg(k,1) = mean(sys_solar_prod(idx_day,1));
sys_load_net_ev_dumb_davg(k,1) = mean(sys_load_net_ev_dumb(idx_day,1));
sys_dumb_charge_davg(k,1) = mean(sys_dumb_charge(idx_day,1));
%sys_load_net_ev_smart_davg(k,1) = mean(sys_load_net_ev_smart(idx_day,1));
sys_smart_charge_davg(k,1) = mean(sys_smart_charge(idx_day,1));
end
%self consumption
self_cons_no_ev = zeros(8760,1);
self_cons_ev_dumb = zeros(8760,1);
self_cons_ev_smart = zeros(8760,1);
for id = 1:8760
self_cons_no_ev(id) = min([sum(solar_prod(id,1:numcar)),sum(h_load(id,1:numcar))]);
self_cons_ev_dumb(id) = min([sum(solar_prod(id,1:numcar)),sum(h_load(id,1:numcar)+dumb_charge(id,1:numcar))]);
self_cons_ev_smart(id) = min([sum(solar_prod(id,1:numcar)),sum(h_load(id,1:numcar)+smart_charge(id,1:numcar))]);
end
self_cons_no_ev_rate = sum(self_cons_no_ev)/sum(sum(solar_prod(:,1:numcar)));
self_cons_ev_dumb_rate = sum(self_cons_ev_dumb)/sum(sum(solar_prod(:,1:numcar)));
self_cons_ev_smart_rate = sum(self_cons_ev_smart)/sum(sum(solar_prod(:,1:numcar)));
self_suf_no_ev= sum(self_cons_no_ev)/sum(sum(h_load(:,1:numcar)));
self_suf_ev_dumb= sum(self_cons_ev_dumb)/(sum(sum(h_load(:,1:numcar)))+ sum(sum(dumb_charge(:,1:numcar))));
self_suf_ev_smart= sum(self_cons_ev_smart)/(sum(sum(h_load(:,1:numcar)))+sum(sum(smart_charge(:,1:numcar))));
%% Peak load reduction
31
peak_load_red=zeros(days,1);
for m=1:days n = (m-1)*24;
peak_load_red(m) = (max(sys_load_net_ev_dumb(n+(1:24)))-
max(sys_load_net_ev_smart(n+(1:24))))/max(sys_load_net_ev_dumb(n+(1:24)));
end
peak_load_red_rate=mean(peak_load_red);
%%
hour = [0.5:1:23.5]';
figure
% you can use plot function instead of area function
plot_dumb=area([hour,hour],[sys_load_davg/1000,sys_dumb_charge_davg/1000]);
plot_dumb(1).FaceColor = 0.7*[1 1 1];
plot_dumb(2).FaceColor = [0 0.75 0.75];
hold on
plot_solar=area(hour,sum(solar_prod_davg,2)/1000);
plot_solar(1).FaceAlpha = 0.6;
plot_solar(1).FaceColor = [1 1 0.25];
plot_solar(1).EdgeColor = 'none';
legend('Household load','EV charging load','Solar power production');
xlabel('hour of the day') xlim([0.5 23.5]) ylabel('kW')
%title('mean daily net-load profile with dumb charging, daylight savings 5 EVs ')
disp('done')
%%
figure %smart charging
% you can use plot function instead of area function
plot_dumb=area([hour,hour],[sys_load_davg/1000,sys_smart_charge_davg/1000]);
plot_dumb(1).FaceColor = 0.7*[1 1 1];
plot_dumb(2).FaceColor = [0 0.75 0.75];
hold on
plot_solar=area(hour,sum(solar_prod_davg,2)/1000);
plot_solar(1).FaceAlpha = 0.6;
plot_solar(1).FaceColor = [1 1 0.25];
plot_solar(1).EdgeColor = 'none';
legend('Household load','EV charging load','Solar power production');
xlabel('hour of the day') xlim([0.5 23.5]) ylabel('kW')
%title('mean daily net-load profile with smart charging, daylight savings 5 EVs')
disp('done')
