Key microgrid features include its assets and architecture, the available or desirable resources it may have, and the specifications concerning Electric Mobility and Vehicle-to-Grid (V2G) technology. Based on the overview of these features and the assessment of the field research analysis results, three alternative energy schemes (Alternative Energy Scheme − AES) are developed:

• AES-1: Grid – Battery Installation (home & work),

• AES-2: Grid – Battery Installation (home & work) – PV Installation (Work),

• AES-3: Grid – Battery Installation (home & work) – PV Installation (Work) – Electric Vehicles (V2G).

These alternatives represent the current state (AES-1), a future state with the installation of RES systems (AES-2), and a future state based on the integration of both RES systems and EVs (AES-3).

The main aim of the analysis is to assess the energy scheme that results in resource exploitation in a sustainable and economically viable way. In this direction, four parameters (criteria) are considered to contribute toward these objectives: grid energy cost, utilization of renewable energy sources, use of energy storage elements, and balanced energy demand. The decision-making process is based on a hierarchy structure consisting of four levels, as presented in Figure 2, that depicts the interconnections between the different elements of each layer. The overall goal of the proposed approach lies at the top level. The decision criteria and sub-criteria are visible in the second and third levels. The level of sub-criteria can be recalibrated and include the corresponding data of each case study. The alternative energy management schemes lie on the fourth level of the hierarchy.

Figure 2.

Decision-making hierarchy structure

The former hierarchy provides AHP with input for assigning weights toward shaping the objective function, which will govern the Energy scheme selection in the following stages. The AHP is developed with the use of a pairwise comparison matrix. Table 1 presents the results from the criteria comparison based on a scale of importance and the values geometric means.

The weights resulting from the AHP calculation are presented in Table 2, along with the calculations related to the consistency ratio (CR), which is 0.065 < 0.1. One can safely conclude that the assigned weights are consistent and can be approved and proceed to the next stages of the process. Based on the assessment, the balanced energy demand and the utilization of renewable sources appear to be the most important criteria, which complies with the logic supporting the approach and increases the likelihood of achieving the overall goal.

A set of constraints should supplement the objective function to make the system economically viable and promote sustainability. Because of this problem's complexity and multilevel nature, the objective function is of a polynomial form. It consists of a set of variables considering the different EVs operational scenarios, indicating the state and rate of charge/discharge (−1 for a minimum to −10 for maximum discharging rate, while 1 for a minimum to 10 for maximum charging rate). Although the system is dynamic, it could work on a continuous operation, the algorithm configuration process charging/discharging time is set at a steady time frame. This time frame is selected to ensure the orderly operation of the bidirectional energy transfer process between the vehicle and the building installation network. Therefore, the optimization algorithm uses a 30 min time frame (period duration) of reference for configuring the power management program.

Figure 3.

Time of use, with on-peak cost, vs. conventional energy pricing (PPC pricing modelling)

The optimization procedure starts with setting up the decision-making parameters (design) and developing the objective function based on these parameters. The problem constraints are then formed to represent functional relationships among the design variables and other parameters describing physics and resource-based limitations [20]. The proposed optimization tool focuses primarily on sustainability performance. In this respect, it is not only the economic parameters, such as the energy cost per kWh, that solely determine the optimization output, but also the level of demand fluctuation during the day, expressed in periods (which show about the same demand profile) and the energy produced by RES. Although a common measurement unit is required, all parameters are converted into costs (expressed in Euro) based on information gathered from the current energy market.

As mentioned above, the criteria used for AHP calculations represent the objective function decision-making parameters. In each reporting period, calculations are made on energy requirements related to the cost per kWh provided by the Time-of-Use strategy. All parameters of the objective function are being expressed in cost units (based on energy market cost values) to facilitate the following stages of the process. Another reason for choosing a cost unit as an overall measurement unit is the ability to follow the dynamic changes in the energy market, expressed by the cost price of a kilowatt-hour. Based on the AHP calculations, the objective function of the system is described by the following equation:

$\text{Energy Scheme}=-\left(0.144\times C_{Grid}\right)+\left(0.276\times B_{Res}\right)+\left(0.06\times B_{Bat}\right)-\left(0.06\times C_{Bat}\right)+\left(0.462\times D{B}_{groups}\right)$ (1)

Where: C_Grid is the grid energy cost [€], B_Res is the price benefit from renewable sources utilization [€], B_Bat is the benefit from managing energy stored in batteries in and out at low and high energy pricing levels from the grid [€], C_Bat is the cost of the energy attributed by the batteries [€], and the last term is determined using the formula:

$D{B}_{groups}=D{N}_{groups}\times 0.1\times C_{Grid}$ (2)

where DN_groups denotes the difference in the number of energy demand groups (groups of periods with common energy demand profile) between the initial state and the state proposed by the energy management scheme. Each of the respected criteria and sub-criteria presented above is further analysed and correlated with energy unit cost.

Grid Energy cost is a straightforward matching of Energy cost based on ToU and the energy demand according to the following formula:

$C_{Grid}={\displaystyle \sum _{i=1}^{48}{E}_{t(i)}\times C_{t(i)}}$ (3)

where E_t(i) is the energy demand at period t(i) based on time of use [kWh] and C_t(i) is the energy price at period t(i) [€/kWh]. The utilization of renewable sources applies to the energy surplus created by the photovoltaic installation. This parameter is calculated according to the formula:

$B_{Res}={\displaystyle \sum }_{i=1}^{48}EP{V}_{t(i)}\times C_{Res}$ (4)

where EPV_t(i) is the energy production from PV installation at period t(i) [kWh], and C_Res is the price per kWh for energy production (due to lack of dynamic connection with the Energy Market, it is considered equal to the lowest energy cost price C_Res =0.08 [€/kWh]). In real life, this value will be linked with the corresponding value provided by the Energy Regulating Authority, based on the period of the day and the type of renewable sources.

The use of storage elements includes energy from either the storage units of an office building (Bat_ESS) or the batteries installed in houses (ΣHBat, corresponding to the house units included for this application) and the EV batteries (ΣEV_Bat). This criterion is subdivided into two equally weighted sub-criteria, so a weight of 0.500 is assigned to each of them. The equations describing the above are the following:

$B_{Bat}=0.5\times \sum _{i=1}^{48}c{d}_{state(i)}\times E_{bat(i)}\times C_{t(i)}-B_{Res}$ (5)

$E_{Bat(i)}=Ba{t}_{ESS}\times {\displaystyle \sum }_{i=1}^{N}HBat+{\displaystyle \sum }_{i=1}^{N}E{V}_{Bat}$ (6)

where cd_state(i) is the state of charge or discharge of each battery installed in the scheme, and C_t(i) is the energy price at period t(i)based on time of use [€/kWh].

The second sub-criterion is the cost of the energy stored in the storage elements at the end of the optimization (E_{Bat (t=48)} – end of the day as indicated based on the 48 periods of reference and EMS selection period). This energy is considered a cost in the objective function; it will be available for use beyond the examined period of reference. Therefore, the microgrid will not benefit from the unused resources. The said cost corresponds to the following equation:

$C_{Bat}=0.5\times E_{Bat(t=48)}-C_{t(t=48)}$ (7)

where C_t(i) is the energy cost at period t(i) based on time of use [€]

Balanced energy demand and minimized demand fluctuations are important aspects of a sustainable microgrid. Based on the standard deviation in each period of reference, the spread of the Grid Energy Cost is calculated and grouped into respective periods, with energy demand deviating up to 15% from the average value of the respective consumption. The goal is to have fewer groups than the original state and a stable demand profile related to the grid. This parameter is expected to be greatly valued in a fully integrated and dynamic energy market soon. The current research adopts a trade-off of 1/10 of the mean energy cost as the benefit for each reduction of the groups calculated by the proposed procedure.

After selecting the proposed energy management scheme, the next and final stage of user approval follows. Each user, in a real-life scenario, will have to approve the proposed scheme to be adopted by the microgrid. If there is no unanimous approval of the proposed scheme, the procedure must run again. Suppose the users do not approve the proposed scheme unanimously after the second effort. Then, the system continues with the charging/discharging schedules for the users who have approved it and rerun the procedure for those who have not, to meet their needs in the best possible way. If no approval is provided, these users are excluded from the procedure.

Following the methodological approach presented above, the first stage considers the citizens' opinions regarding the development of the energy management scheme. In 2020, the research team organised and conducted a questionnaire-based survey in the reference area and analysed the responses. The survey targeted 104 citizens of the City of Patras. An overview of the demographic characteristics of the respondents is given in Table 3. Figure 5 presents the most important reasons for users to buy an EV. Based on the responses, environmental concerns, energy price, cost, and the utilization of EV batteries lie among the most important motives. More specifically, there is an alignment between the survey responses and the architecture and operation of the proposed system, as the above-mentioned most popular choices have been considered for the design and development of the decision-making process.

Figure 5.

Electric Mobility selection parameters specified by future users

The user (EV owner) profile should be created within the system. This profile includes the average route distance (home to work commute) and other parameters, such as expected daily travelling time and EV systems' energy consumption needs (e.g., AC and media player). Based on this dataset, the system will estimate the energy required for each EV. One can then calculate the energy amount available to be supplied to the microgrid over any reference period by using the following equation:

$E_{req}^{trip}\left(AV\right)=DS\times n_i+E_{fact}^{trip}$ (8)

where $E_{req}^{trip}$ is the total energy required for a trip by an EV [kWh], DS is the round-trip distance [km], and $E_{fact}^{trip}$ is the energy required for a trip accounting for the desired comfort level at specific meteorological conditions [kWh]. A safety factor n_i = 1.3 is introduced to ensure that in all cases there will be sufficient energy available for the users to complete their everyday commute, regardless of unplanned needs that might require additional energy.

An important goal within the energy management system is to ensure that the EV owner's comfort level is not disrupted even in cases that the microgrid requires to absorb the maximum of the available energy provided by the EV. The available EV energy for exploitation by the microgrid is:

$E{V}_{av}=Ba{t}_{cap}^{av}-Ba{t}_{min}^{disc}-E_{req}^{trip}$ (9)

where $Ba{t}_{cap}^{av}$ is the available battery capacity at the start of the scheduling [kWh]. Nowadays, EVs have higher battery capacity than previous generations and faster or even supercharging capability. The increasing EV mass production and competition among the world's biggest car manufacturers have helped decrease purchase costs to levels close to conventional vehicles. These are some important factors that have led to increasing EV penetration in the market, with a further increase anticipated in the future. Based on current trends in the case study, a 40% EV penetration rate has been set.

The microgrid consists of the following components: the 10-office building, the photovoltaic plant, the energy storage system (ESS), the electric vehicle charging station, and the electric vehicles connected to the charging station. The microgrid is designed with four charging sites (one at each building), considering the suggested penetration rate. Apart from the number of available EVs connected to the system at any time, there are some constraints regarding the EV batteries. The battery is among the most expensive EV components that determine EV usability and viability as a product. The battery life cycle and performance are the main parameters considered in the proposed method.

EV batteries nowadays provide a capacity higher than what daily usage would demand. A battery of a 75 kWh capacity is a standard for vehicle manufacturers, while statistically, the average daily demand is not more than 10 kWh. It is possible to use the available excess energy for alleviating building energy demand peaks during the day.

EV batteries' life span and charging cycles are also important parameters. Tesla batteries present about 5,000 to 6,000 cycles before losing their capacity. Studies argue that the lifespan loss is minimised if the battery is used to 80% of its capacity [22]. Accordingly, eq. (10) is:

$\text{0 kWh}\le Ba{t}_{cap}^{sys}\le \text{60 kWh}$ (10)

where $Ba{t}_{cap}^{sys}$ is EV battery energy that may be available for use by the system. Based on the average EV life charging cycles, a soft constraint is set in the optimization process to allow no more than four cycles of EV charge/discharge daily. The limit is set according to the results of a survey reporting an 89% EV owner sample that would follow the constraint (see Table 4):

$\text{0}\le Ba{t}_{cycles}^{av}\le \text{4 }\left(\text{3 as an option}\right)$ (11)

where $Ba{t}_{cycles}^{av}$ is the acceptable daily number of EV battery charging and discharging cycles.

The user preferences and willingness to participate in the Energy Management Scheme are also studied by introducing a value to show the degree of desired participation on a scale from 1 to 10 (with 1 indicating the lowest and 10 the highest potential participation). This input, along with the number of available cycles of the EV battery, corresponds to two important indications of availability, which have an impact on the energy that the system can use for the energy management scheme. The energy of EV battery that can be available for use by the system is assumed and shown in Table 4.

The last constraint regarding the EV battery is the level of charging and discharging provided by the charging station. There are three levels of charging, starting from level 1, provided by the national grid simple outlet, to level 3, provided by superchargers. In this work, the intermediate level 2 chargers are considered, which are 5.6% more efficient than level 1 and way less expensive than superchargers. Level 2 chargers typically range from 16 A to 40 A and may be referred to as 3.3 kW and 7.2 kW, respectively. Superchargers installed in a specific microgrid solution would not affect the national grid much. However, having more systems (including more EVs and batteries that need to be charged simultaneously) would affect the national grid as they are 480 V and require installing special equipment. The possibility of their common use in homes would create risks of more frequent demand peaks.

The optimization process is designed and executed using the methodology of genetic algorithms and the support of the computer package Evolver of the Palisade company. This software provides the ability to work with Microsoft Excel, where the problem and all the parameters have been formed to evaluate the algorithm's effectiveness. The software enables changing the three basic parameters of the genetic algorithm methodology, namely the initial population size, the crossover rate, and the mutation rate.

The purpose of the presented application is to efficiently manage the microgrid energy resources and plan the charging and discharging cycles of electric vehicles. Based on the results, the scheme appears to have a positive performance by using the proposed Energy Management Scheme − including photovoltaic system installations and utilising the energy available from electric vehicle batteries.

For the purposes of this application, a total of 3 EV users have been specified, with the same car and specifications, although with different travel requirements and energy availability. The EV's preferences are presented in Table 5.

Figure 6 presents the proposed schedule for charging and discharging Electric Vehicles. All constraints set above have been taken into account. The system has responded well to the expectations, assisting the microgrid in achieving its environmental and economic goals. Of decisive importance is scheduling the charge of the EV when the energy price is low and discharge when it is high, as provided by the ToU model (Figure 3). Another indication of efficient functionality is the concentration of scheduled discharges between 12:00 to 14:00 when peak demand and a high energy cost may occur.

Figure 6.

EV charging and discharging schedule

Figure 7 shows an overview of energy consumption, the use of the EV batteries and PV installation, and the savings resulting from the proposed Energy Management scheme. More specifically, the demand during the reference day is visible, where the initial state (original state) is compared with the one proposed by the energy management scheme. In the demand profile proposed when EMS is enabled, 17 demand groups reduce to 15.

Figure 7.

Comparison of initial and final energy demand curves of the Microgrid of Reference

There is a distinct reduction in demand during the peak hours indicated by the ToU, which is also combined with the maximisation of energy storage utilization, towards reducing peak demand on the grid. The scheme follows the users' preferences regarding the energy availability and the number of charging/discharging cycles of the EV batteries scheduled for the day of reference. These observations indicate the scheme's proper operation and the adaptation capabilities to the energy market needs, based on the case-by-case ToU policy.

There are some metrics based on the adopted criteria set that supplementary confirm this conclusion. The energy grid cost for the initial state was 55.19 € (for the day the scheme was proposed) and has been reduced to 42.63 €, presenting economic efficiency equal to 22.75%. From this reduction, a part is attributed to the utilization of renewable sources, which provided a 9.22% cost reduction equal to 5.09 €. It becomes evident that the use of storage elements, such as the batteries installed in the office building and houses, along with the EV batteries, are responsible for the major part of the cost savings achieved by using the proposed Energy Management Scheme.