A review of electric vehicle hosting capacity quantification and improvement techniques for distribution networks

The electriﬁcation of the transport sector to control the carbon footprint has been gaining momentum over the last decade with electric vehicles (EVs) seen as the replacement for conventional internal combustion engines. Economic incentives, subsidies, and tax exemp-tions are also paving the way for rising EV penetration in the power distribution networks. However, the exponential EV adoption requires careful technical and regulatory analysis of traditional networks to satisfy the network reliability constraints. Therefore, it would be vital to ﬁnd EV hosting capacity (HC) limits of networks from a multifaceted approach involving various market players, mainly distribution system operators and EV owners. This review provides a systematic categorization of EV hosting capacity evaluation and improvement methods, thus enabling researchers and industry personnel to navigate the advancing landscape of EVs. This novel framework extends beyond the theoretical implication of diverse objective functions and HC improvement methods to the actual numerical values of EV HC across varying geographical settings. Therefore, this unique synthesis of varying aspects of EV HC facilitates the in-depth understanding of the integration of sustainable energy and transport sector.


INTRODUCTION
The climate impacts stemming from growing energy needs are a significant motive behind the devotion of policymakers to adopt electric vehicles (EVs) as a means to lower emissions.The positive environmental contribution of EVs due to zero emissions is a subjective matter due to the dependence of EV charging on the power grid.However, the ongoing incorporation of renewable energy sources (RES) can address this concern and ultimately favours EVs as a solution to environmental issues.In the future, the strategic development of power networks is expected to accommodate the rising EV charging demand that is reshaping the existing load profiles.This dynamic EV load undergoes spatial and temporal variation due to unique arrival or departure times, driving patterns, state of charge (SOC), charging preferences, EV location, and driving distances.Besides, network expansion requires careful consideration of loads, infrastructure capacity, and future generation or consumption patterns.Thus, the widespread demand for EVs requires operational studies and an inclusive techno-economic analysis for the social acceptance of EVs as a clean energy alternative.
The quantification of EV hosting capacity (HC) is vital in the efficient utilisation of grid resources and future decisions about expansion planning.This knowledge not only enables future developments for meeting RES targets but also provides a common discussion ground for regulatory, academic, and industrial entities to make informed decisions for network stability.Thus, the ongoing debate on rising EV uptake motivated many researchers to investigate distribution systems regarding EV HC [1][2][3].Figure 1 shows the rising trend of EV integration studies focused on HC.The uncertainties, such as user behaviour and EV location, render EVs a stochastic variable and require HC studies to be conducted on probabilistic grounds instead of deterministic approaches.For instance, EV charging location is an important parameter for analysing network limits due to the size-varying chargers installed at homes, workplaces, or public charging stations.In this context, the location dependency of EV HC was analysed in [4,5] with a Monte Carlo (MC) based stochastic power flow research subject to voltage limits.These studies concluded a higher EV HC at the node near the service transformer (TF) and increased voltage drop instances with EV placement away from the substation.Thus, the locational impacts of EV charging are highlighted in terms of the violation of network limits, ultimately requiring a careful selection of performance constraints.
The performance impacts of EVs can be broadly classified into technical, reliability-based, and economic.To begin with, EV integration can be restricted by network operational limits for voltage, overloading, losses, or power quality (PQ) [6,7].In that respect, EV impacts on network performance were investigated in [8] and the authors found increased voltage drop and power losses due to EV addition.Likewise, network reliability can be compromised by the addition of EV loads on top of the existing network demand.The overloading of grid assets due to EV uptake gives rise to increased network outages and thus increases the associated reliability risks.In addition, EV charger size substantially impacts the outage probability, and a higher charger rating might exhaust the network limits sooner than the lower charging power.Thus, the outage and reliability analyses are vital to performing corrective actions.The general objective of reliability-based EV impact studies is to maintain the initial reliability indices, such as loss of load probability (LOLP) or expected demand not supplied (EDNS), and perform the mitigation means if the index limits are violated [9].
Finally, the economic factors mainly dictate the economic viability of EVs and include EV pricing, charging costs, distribution system operator (DSO) incentives, and power losses.The first two aspects of pricing and charging cost influence the EV acceptability for EV owners.DSOs, while being inclined toward EVs for grid support as vehicle-to-grid (V2G), are concerned by the incremental losses due to peak loads.Moreover, incentivising the public can be an added expense for DSOs if the charging demands are not optimally coordinated to maintain the supply-demand balance.This is because of the fact that network balance is not merely a function of individual peaks, and the network capacity should accommodate the coincident peak loads.Accordingly, the customer incentives can be modelled based on the tariff designs focused on reducing the aggregated load of a customer group.The findings of [10] highlight a similar situation of higher demand and losses due to the overlap of the existing load and EV charging load of an IEEE-34 node test network.In essence, optimising DSO and EV owner benefits on an integrated level might prove more beneficial than focusing on individual objectives.This coordinated operation not only maintains the grid limits to raise HC but optimises economic objectives such as cost minimisation and profit maximisation [11].Thus, the grid assets can be efficiently utilised by designing appropriate tariff structures using the flexible nature of EV loads.
EV HC of networks can be enhanced by addressing the technical and economic concerns through the application of improvement methods.Various optimisation-and simulationbased approaches are utilised in the literature to meet the objective functions focused on HC evaluation or increasing the penetration level [12,13].The EV nodal capacity was analysed for industrial, commercial, and residential feeders for their ability to host different charger types, such as 350 kW, 7.2 kW, and 3.3 kW chargers [13].Similarly, a combination of linearised AC power flow and MC simulations was applied in [14] for HC maximisation and in [15] for determining HC using the undervoltage probability considering EV uncertainties as arrival time, SOC, number of EVs, and selection of EV charging station.On the other hand, HC enhancement can be indirectly realised by addressing network issues such as voltage profile improvement, congestion management, or loss reduction, which will be discussed in Section 4.
The role of EVs in shaping sustainable transportation and energy networks can not be underestimated due to the inclination towards a carbon-neutral future.However, EV integration into the energy paradigm requires a connection between theoretical insights and practical outcomes to facilitate informed decision-making.Thus, a comprehensive review of the quantification and improvement of EV hosting capacity serves as a strategic guideline for future researchers.
A literature review of EV integration reveals the inadequacy of a systematic review focused on EV HC quantification and the HC improvement techniques for low-voltage networks.The primary aim of this review article is to address this research gap and further identify the definitional vagueness of EV HC.This review article serves as a pioneering synthesis in the domain of EV HC estimation and future expansion possibilities.It provides a novel compilation of the numerical values of HC across varying geographical settings-thus facilitating the streamlined perspective of EV HC evaluation.In addition, a clear organization of the simulation/optimization and improvement methods of EV HC is presented.This is further enhanced by the tabular presentation to broadly map the objective functions and tools for EV HC estimation.Finally, our contribution extends to providing the percentage rise in EV HC, thus assisting the stakeholders with actionable insights into the impacts of the evolving EV industry.Thus, the above-mentioned contributions set this work apart, proving to be a significant resource for strategic planning and practical applicability for EV integration into energy systems.
The review is based on many scientific databases, specifically Scopus, ScienceDirect, and IEEE Xplore, using the EV hosting capacity, EV charging capacity, penetration limits, distribution network, and hosting capacity enhancements as the main keywords.Section 2 includes EV HC definitions and estimated HC for different networks, followed by Section 3, which focuses on the categorisation of evaluation methods for HC quantification.Section 4 takes the HC concept further by classifying the enhancement methods, and evaluation of network performance.Section 5 describes the current and future EV projections and the significance of EV charging characteristics.Section 6 provides the discussion, whereas concluding remarks and future research directions are presented in Section 7. Figure 2 provides a general framework for the review of EV hosting capacity.

ELECTRIC VEHICLE HOSTING CAPACITY ESTIMATION
The ongoing debate on transport electrification calls for network analyses to host the rising amount of EVs.However, HC results are not a hard limit on EV integration but how various performance indicators (PI) manifest themselves at varying penetration levels.Accordingly, the HC value can vary depending on the violation of different indicators with rising EVs.Two terms, namely penetration level and HC, are interchangeably used in the literature to refer to the highest EV integration in the network.However, these terms should be independently analysed as the penetration represents a distributed generation (DG) level concept and the HC refers to a system-level idea for EV integration [16].

EV HC definitions and limiting factors
The two major themes of EV studies in the literature are based on either the HC calculation or the enhancement of HC.HC reference, scaled with respect to some fixed network parameter, helps in analysing EV addition in the network, and five reference definitions are generally employed to represent the percentage value of EV penetration as shown in Figure 3.It has been defined as a ratio of EVs to the total number of customers in the network subject to network performance constraints [17][18][19][20][21][22][23][24][25][26][27].
Alternatively, HC is reported as the absolute value of the number of EVs or EV charging power that resulted in the violation of performance limits [28,29].The remaining three HC reference points in the literature are the vehicle fleet [30,31], available power [32] or network demand [33], and annual electricity use [34].The number of customers is used as a reference by 25% of research studies, followed by 13% of articles using the EV numbers or charging power as HC reference.Whereas, only 7% used either of the remaining references, such as vehicle fleet, network demand, or annual energy.A substantial share of the HC studies explicitly defined HC reference and the "number of customers" is found to be the most widely adopted definition among these studies.Nevertheless, some research studies concluded the HC as the percentage value without defining a reference, as shown in Table A1, Appendix.The network limits are additional criteria for identifying HC since their violation refers to the point of penetration after which the network cannot sustain further EV addition without any mitigation means.These limits can be divided into technical and non-technical where the former refers to voltage limits, thermal limits, voltage unbalance, and total harmonic distortion (THD), whereas the latter refers to the financial side, such as bid price limits while interacting as market players for EV integration.The literature survey reveals the widespread adoption of voltage as a performance indicator, followed by thermal limits of grid assets such as transformers or cables, and power quality as the third most deployed HC indicator.About 57% of the reviewed literature used voltage as a limiting factor, among which 30% of papers defined it as ±5% of nominal voltage (U n ) and 16% of papers as ±10% of U n .Similarly, about 41% of papers employed transformer overloading as a limiting factor, and 36% used cable ampacity rating for determining HC.Finally, voltage unbalances, THD, and pricing limits are used in about 15%, 4%, and 4% of research papers, respectively.The distribution of the limiting factors used in the literature is shown in Figure 4.
The impacts of weather conditions on EV HC are mostly viewed in terms of technical parameters such as charging efficiency or battery degradation.However, weather conditions might alter charging demand patterns and utilization of public charging stations.For example, extremely hot or cold weather leads to battery degradation and reduced efficiency, respectively.This discussion implies that weather conditions can be considered a limiting factor in future HC studies.
EV loads can be characterised by their unique spatial and temporal dimensions due to charging station placement and random charging schedules arising from EV owner behaviour.Thus, the EV characteristics help calculate HC apart from the performance constraints used in the HC studies.EV charging rate and charger size have a significant impact on EV HC with the fast chargers lowering the HC as the network limits are exhausted soon due to high power drawn in a shorter time duration [17][18][19].In this context, EVs are added in steps of 10% as a single-phase connection to a Flemish urban radial network to check the impact of various charging schemes [17].The network HC of this work is calculated without and with the inclusion of fast chargers as 80% and 70%, respectively.The authors of [18] found a 4.7% HC value, and the findings of [19] state that the network can host 90% EVs after which the network performance would deteriorate.Both studies further strengthen the concept of HC dependency on charging mode where the HC with slow chargers exceeds the HC using fast chargers.
The EV location, in addition to the charging mode, is another important factor that influences the HC of the network with pronounced network violations if EVs are located at the far end of the service transformer.The locational impacts of EVs appear as increased voltage drop while moving away from the substation, and thus, grid expansion should consider the scale and hotspots of future EV loads.Considering the placement aspects, EV addition was monitored for two charging cases: regular and worst-case charging [20].The findings of this research proved the point of pronounced voltage deviations due to EV connection at high-distance nodes.An analysis of the worst scenario can give an idea about the maximum stress on the network, and it was defined as the EV charging demand as 100% SOC [20].Moreover, the worst-case scenario can be realised as the charging simultaneity of EV loads and its definition varies with the problem formulation.In this regard, the research of [33] analysed EV charging simultaneity considering the spatial pat-terns of EV loads in terms of magnitude and location.The findings of this research are in line with the EV charging mode analysis as higher HC with slow charging as compared to fast charging.Thus, the charging mode, locational and simultaneity aspects of EV loads can be studied on an integrated basis for a better HC understanding.
EV HC is also sensitive to network dimensioning and asset utilisation in addition to the dependence on EV characteristics [21].The HC of two Malaysian low voltage (LV) networks: newly built and mature, was analysed in [22] with the results complementing the fact that the HC of an old network (10%) is lower than the new network (20%).Thus, it can be concluded that asset utilisation influences the HC and the traditional networks exhaust soon to support the modern loads.Similarly, EV impacts on distribution transformer ageing and loss of life were analysed for HC calculation (25%) in [23].The authors further deployed demand response (DR) using time-of-use (TOU) tariff for EV load shifting depending on TF ageing acceleration factor and loss of life.The findings of different research studies support the fact that the HC varies with EV parameters and network characteristics, thus leading to the idea of unique HC values under different conditions.

HC estimation without mitigation means
HC of the networks can vary widely, for example, 4%-80% for urban regions, as the EV penetration analyses are casedependent.This diversity of HC results cannot be explained without the knowledge of additional information regarding housing density, network structure, peak network load, and HCdefining parameters.Moreover, it requires adequate information on network parameters, that is, the main feeding circuit for meeting the EV load in addition to existing loads.Thus, network structure influences the HC results as network layout predominantly dictates the appearance of limiting factors.For instance, the different transformer sizes in North American and European networks and different voltage standards pose a significant impact on the HC value.This difference in HC can be attributed to the fact that the limiting factors for these networks manifest differently under different EV penetrations.This implies that even an EV penetration of 100% can be warranted with the use of dedicated transformers and/or cables.Finally, HC is sensitive to the region and country-specific requirements since the grid codes vary widely in different geographical areas.
The comparative analysis of HC results of the reviewed literature on a single scale was not possible due to the disparity in HC results.The numerical values of HC ranged from 4.7%-80%, 10%-40% and 10%-56% for urban, suburban and rural regions, respectively.Around 19% of studies reported HC ranging from 10%-60% without providing information on the geographical location of the test network.It is apparent that the networks, despite being positioned in the same type of regions, host varying EV penetrations.Therefore, HC results are depicted in Figure 5 to reveal the HC range for diverse regions.The x-axis shows an estimate of typical EV HC results of the reviewed papers and the y-axis displays the occurrence frequency of research papers with HC results corresponding to the x-axis.It is apparent that most HC values fall between 0%-40% and Table A1 in the Appendix provides the numerical HC values of the reviewed literature.
In short, HC research needs a common ground for their comparison to other studies to make future decisions about EV adoption considering network limits.This problem can be solved by scaling the networks to some reference system and providing all necessary information utilised in HC calculations.Location of EV charging, charging power and time, housing density, peak load, SOC, network conditions, region, and rating of components should be reported, among other parameters, for comparison purposes.

PROBLEM FORMULATION AND EVALUATION METHODS
The flow of HC analysis starts from the selection of reference definition and limiting factors leading toward the problem formulation and, ultimately, the selection of the method/tool for intended research.Therefore, this section reviews the approaches for HC assessment and the classification of the main objective functions of the research studies.

Problem formulation
The generalised framework for the investigation of EV integration revolves around certain objectives such as HC analysis, HC enhancement, economic gains, technical stability, reliability, and computational efficiency, as shown in Figure 6.The problem formulation is the mathematical form of optimising the objec- tive function subject to technical or economic constraints, and the following subsections explain commonly used approaches in the literature.

HC calculation/enhancement
The HC calculation or enhancement is the primary objective for most of the studies focused on EV uptake that can be subdivided into individual goals of various research articles.These goals are categorised into different areas of interest and will be explained in detail.

EV charging variability
The inherent difference in EV charging stems from numerous parameters such as battery capacity and SOC, number of EVS, location, charging duration and EV owner preferences.This uncertainty can be expressed by probability density functions and solved through designated numerical or analytical methods.Therefore, the response of a distribution system to dynamic EV loads can be substantially captured by undertaking EV stochasticity in the form of random charging schedules and EV owner behaviour.EV locational aspects and EV distribution among network nodes were analysed in the literature in an attempt to find the EV HC [29,32].Both studies supported the fact of a higher EV HC in the case of EV distribution among all available sites instead of clustering EVs into specified regions.An IEEE 123 node test system used in [32] could host only 125 EVs when EV distribution was limited to two residential regions as compared to 225 EVs with even EV distribution among all nodes.So, this study highlighted the importance of site utilisation for enhancing EV integration.In addition to EV location, the HC analysis can be performed in the context of modelling other stochastic variables of EVs [35][36][37].Accordingly, the authors of [37] performed a probabilistic MC simulation analysis by using EV location and starting time of charging as random variables.The authors of this article studied the maximum EV penetration in step addition of 1 EV until 100% under-voltage probability occurred at the service points.In addition, the concept of stochastic EVs is implemented as random EV distribution among network nodes and phases as performed in [38] for determining the HC of an urban LV residential grid in Macedonia.The problem formulations for HC maximisation can be extended towards the combined DG and EV HC for optimal utilisation of available capacity.Thus, the coordination of EV and DG was discussed in [39] by using a metaheuristic optimisation approach to maximise HC and minimise network losses.This study adopted a multi-period formulation for modelling random variables such as demand, DG production, driving patterns, and EV energy needs.

Optimal charging location
The increasing popularity of EVs might prove a fundamental challenge for DSOs to maintain network stability.At one end, EV owners' range anxiety ultimately gears toward fast charging as the viable solution for these issues.In addition, the problem of limited charging facilities is one of the challenging barriers to EV adoption and requires infrastructural reforms by DSOs to sustain the rising EV demand.In this context, the optimal location of charging facilities is advantageous from a multidimensional approach considering both technical and financial aspects.The optimum placement of charging stations can maintain grid limits and the HC of home charging facilities was maximised in [40] by optimising the placement of 4 fast charging stations among 6 locations in a UK distribution grid.However, the placement and sizing of a charging station could be considered a multi-objective optimization problem due to the interconnection of the power system and the coupled traffic system.Therefore, the authors of [41] studied this interconnection to optimize the annual investment cost and energy losses through the optimal placement of fast charging points.The communication aspects of modern power systems allow for coordination among multiple entities and [42] investigated the combined operation of on-load-tap-changer (OLTC) and charging point placement for HC enhancement.Nevertheless, the planning decisions regarding the location of the charging station partially depend on customer-oriented aspects such as the number and location of potential EV owners and the number of EVs per household.Thus, the EV user perspective plays an equally important role in HC maximisation in addition to individual EV parameters.

EV owner behaviour
The objective of HC maximisation can be formulated by satisfying the charging requirements of the EV owners within distribution system constraints.Nevertheless, EV owner behaviour is multi-layered depending on customer and utility economics, such as the EV market bidding, DSO incentives, and infrastructural availability.Thus, the potential for EV expansion follows not only the technical aspects but the EV owners' response toward future EV adoption.In this context, HC studies are formulated from the user perspective [43,44].The EV attractiveness, apart from market offerings, is a function of adequate energy delivery which is investigated in [45] for an LV suburban residential grid in Dublin with a total required energy of 819 kWh.This article verified two objective functions (OF) using linear programming for energy delivery maximisation as standard OF and weighted OF that differed in the way EVs are charged depending on their location on the feeder.The former OF was designed for EV charging at a higher rate in case of their proximity to the TF, whereas the latter OF proposed a fair distribution of EV charging regardless of their location.
Similarly, not only does the EV owners' behaviour impact EV adoption, but the diversity in consumer charging behaviour can be utilised in the HC estimation of a network.The research findings of [46] are in-line with this approach and adopted a conditional probability and clustering-based approach for determining EV owner diversity.Thus, the objective of HC maximisation is a function of technical limits, random EV variables, charging station planning, and the behaviours of market participants such as EV owners, charging station (CS) owners, and utility.

Multiplayer objectives
The problem formulation of these objectives entails three entities impacting the EV HC of the networks: EV owner, CS owner, and DSO.The objectives of the first two units are mostly economic while DSOs focus on maintaining network stability along with economic aspects.To begin with, EV owners try to minimise charging & travelling costs, waiting and charging time, whereas the CS owner's concern lies in the placement of stations to minimise the investment cost and maximise the charging cost.
On the other hand, the primary concern of DSOs is to maintain network stability, subject to technical concerns due to infrastructural limitations.In this context, several research studies investigated the individual or combined objectives and highlighted their importance when assessing the EV HC.The EV owner objective was analysed in [47] using the MC simulations and linear programming to minimise the cost of EV charging based on day-ahead electricity price, EV uncertainties, and the degradation cost of the battery.It is also possible to utilise the actual EV charging currents of discrete charging sessions instead of individual EV SOC signals, as highlighted in [48].
The authors of this study performed a DSO-based approach to optimally utilise the available charging capacity by an adaptive algorithm using actual EV charging currents.
In addition, the combined objective to increase the benefits for EV owners and utility was investigated by researchers working on multi-objective problems.A fuzzy logic (FL) controller was used with inputs as EV SOC and the price signal from the utility and output as the charging power in [24] for assessing charging coordination.This study strengthened the concept that smart charging can maximise the financial benefits of EV owners by scheduling off-peak charging and enabling utilities to design optimal tariff structures for charging within network limits.It is worth mentioning at this point that the concept of price-based EV charging is becoming popular with the benefits of either low-cost EV charging, government incentives, or both.Nevertheless, price-based approaches depend on communication from EV owners that might be a comfort concern for them.Thus, it is worthwhile to explore user comfort approaches for charging coordination.In this context, ref. [49] performed optimal coordination between owner travel commitments and dynamic energy tariffs by minimising the communication from the user end to address the concerns of user comfort.
And finally, the charging station owners' objective can be optimised along with the EV owner and utility [50,51].A mixed integer linear programming (MILP) approach was applied to an IEEE 123 node test system to maximise the extra load HC [51].This study fulfilled multiple objectives by satisfying the EV owner charging demand, satisfying network technical limits, and minimising CS investment costs.The V2G strategies can be beneficial for increasing EV HC and mitigating the network technical issues and ref. [52] used a metaheuristic optimisation method to consider both the charging and discharging of EVs.In this study, the charging cost and peak-to-average ratio (PAR) were minimised along with HC maximisation from 60% (G2V) to 80% (G2V+V2G), proving the benefits of discharging approaches.

Efficiency and reliability
The HC analysis of distribution networks involves numerous scenarios of stochastic variables and requires large calculations due to annual surveys, forecasting data and high time resolution requirements.Therefore, computationally intensive studies have underlying costs, such as long simulations spanning over many days, that must be considered for such analyses.As a solution, parameter reduction or parallelisation methods can be adopted for reducing the simulation time and ref. [53] used a multicore parallel computing (PC) method for HC analysis of an IEEE 123 node test system.The main objective of this study was to analyse the influence of network topology on HC with the additional goal of reducing computational time by 73% using multiple cores.Additionally, pre-processing of data was proposed as a solution for reducing the computational time by the analysis of the most critical period, known as the worst-case, using MC simulations and the two roulette-wheels-selection (RWS) approach [2].And finally, some research studies focused on network reliability aspects for HC calculation and an IEEE 33 bus distribution network was investigated in [54] by monitoring the network reliability indices as an HC limiting factor.The value of these indices, used as limiting factors, was calculated before EV addition to the network, and EVs were later added to the network until they reached the initial level of reliability calculated through LOLP, EDNS etc.The influx of EVs can potentially cause variability in energy demand and increase the demand beyond grid limits.Thus, stability metrics like voltage and frequency stability and load balancing require attention to increase the HC of existing networks.

Evaluation methods
The selection of an appropriate mathematical method for HC analysis is subject to the data requirements, desired accuracy and motivation of the intended research.The reviewed literature can be classified into two main types of HC evaluation methods: simulation and optimisation-based methods, as shown in Figure 7.

Simulation techniques
Stochastic methods consider the variability of load/generation profiles and other random parameters such as phase connection, location, or other EV factors.These methods are based on the execution of numerous scenarios for the realisation of random variables and perform HC analysis until the violation of performance limits.In addition, they provide a near-realistic realisation of the distribution network due to scenario formation based on the probabilistic occurrence of each outcome.
A similar kind of probabilistic study was performed in [55] to investigate the impacts of new photovoltaic (PV) or EV installations using EV location, PV production, and EV consumption patterns as stochastic variables.Nevertheless, MC simulation is the widely used approach for the consideration of stochastic variables for multiple scenario generation and analysing the HC for each of the possible scenarios.At least 16% of research studies employed MC for HC evaluation, as given in Table 1, which classifies different studies based on problem formulation and evaluation methods.
The appearance of random variables might vary over time and may not be captured, and thus time-series (TS) or quasidynamic simulations can be utilised for EV HC studies.A time series method is incorporated in [34] using a linear regression model (LRM) and graphical analysis for an annual detailed stochastic modelling.This study not only focused on the maximisation of EV HC but PV HC using smart charging and curtailment, respectively.Similarly, the impacts of rising EV penetration on TF loading were studied in [56] by using quasidynamic simulations for rural and urban regions with different initial TF loading as 50% and 80%, respectively.This study highlighted that EV penetration could be more critical for the areas with higher initial TF loading and thus needs a TF replacement.Besides, some authors utilised miscellaneous approaches for a better demonstration of EV HC, including MC, time-series, quasi-dynamic, fuzzy regression models (FRM) or quantitative evaluation (QE) methods.However, the superiority of any method is a subjective matter due to the different input data requirements, limitations of technology and the possibly conflicting objectives of varying market participants.Moreover, the time resolution, spanning period and scenario count depend on the accuracy of the required results.Thus, the need for an optimal solution has motivated researchers to find more accurate methods, such as optimisation approaches, despite the ubiquitous utilisation of stochastic methods, as depicted in Figure 8.

Optimisation-based methods
Optimisation methods can be viewed as a suitable solution for circumventing scenario-based computationally intensive simulation methods.They generally comprise an objective function, decision variables, and constraints with voltage, overloading and power balance as widely employed constraints.However, the objective function optimisation follows research scope,   required accuracy, dimensionality or computational time constraints, where losses, cost/profit and penetration level are commonly used OFs.Optimization methods generally adopt optimal power flow to determine the EV hosting capacity.Power flow equations monitor the changes in line flow between buses, nodal voltages, branch currents and angles.The active (PL mn ) and reactive line flow (QL mn ) can be calculated as follows in (1) and ( 2) respectively where m and n represent the buses connecting line mn.V m , θ m , V n , and θ n represent the voltages and angles at buses m and n respectively.
The conductance g mn and susceptance b mn are calculated from the admittance of line mn in (3) and ( 4) where r mn and x mn are the resistance and reactance of line section mn.
The line flows, bus voltages and branch currents must be within specified limits for a stable energy network in case of addition of any load or generation in the system as given by ( 5)- (8).
−I max mn ≤ I mn ≤ I max mn (8) Moreover, an extra set of constraints can be added that impose additional limits to EV HC determination depending on the specific problem formulation.
A variety of optimisation methods are used in literature for EV HC analysis, such as linear/non-linear programming (LP/NLP), metaheuristic (MH), robust, interior point method, game-theory, weighted-sum multi-objective, fuzzy logic, receding horizon (RH) and model predictive control (MPC).The proposed techniques are either focused on HC maximisation, mitigating network issues such as voltage drop etc., or both and modelled as either single or multi-objective studies.The concept of marginal HC is introduced in [57] for finding the HC of each bus using a robust optimisation approach.Moreover, EV charging is a complex problem that needs optimisation approaches for finding the optimal charging schedules based on pricing signals, user comfort, and technical parameters.Thus, a modern optimisation approach (interior point method) was used in [58] to maximise EV charging considering technical and customer preferences.This work investigated a test network of the IEEE-EU distribution system, and the voltage and ampacity violations were mitigated by aborting 13% of EV charging requests using an optimal decentralised charging scheme.
EV HC analysis involves complex variables that require computationally intelligent techniques for finding optimal solutions considering various EV uncertainties.The widely employed optimisation constraints are either technical, for example, voltage and thermal limits, or EVs-related, for example, SOC requirements, EV availability etc.The literature survey reveals that heuristic or metaheuristic approaches outnumber other methods for the analysis of EV HC using various alternatives such as particle swarm optimization (PSO), genetic algorithm (GA), tabu search (GRASP-TS), differential algorithm (DA) and evolutionary algorithm (EA) with different applications.For instance, PSO and GA methods reach global optima solutions using different approaches.The former performs this task based on the movement of potential solutions, known as particles, in the problem space to avoid the problem of local minima and later by an iterative method inspired by an evolutionary process.Thus, the authors of [59,60] used PSO for evaluating the EV integration impacts such as network security, voltage and energy losses.Whereas, a multi-objective differential algorithm is used in [61] for EV charging control to improve the power quality of the network.
Moreover, the linear programming approaches were adopted by some authors, and the literature review ranks them as 2nd widely used optimisation method after MH.Thus, the linear optimisation problem in [62] was solved by considering an EV control technique using a TOU pricing environment to minimise network congestion and customer payment cost.A similar objective of charging cost minimisation and HC maximisation was optimised using a receding horizon approach in [63].In addition, multi-objective problems can be solved from a game-theoretic optimisation approach, as in [64] where the EVs represent the agents trying to optimise their charging strategy.This study aimed at meeting the local and global targets of EV charging and management of voltage profiles, respectively, through a multi-agent decentralised system.

Comparative analysis
Simulation techniques, while being computationally intensive, can provide comprehensive details about hosting capacity.Stochastic methods can be used for forecasting and sensitivity analysis of large data sets involving EV load and charging patterns.For instance, MC simulations are the commonly used approach for stochastic analysis by creating multiple scenarios to account for uncertain variables such as size, location, and time.However, this method face challenges for real-time network monitoring and data analysis due to the computational burden of multiple scenarios [65].Similarly, time series methods work well in situations demanding analysis of seasonal and diurnal variations in EV charging profiles.However, accuracy of this method varies with data availability and large data sets are required for more accurate results [66].Nevertheless, simulation methods undergo extensive computation, thus requiring substantial time and processing resources.Therefore, optimization methods can be seen as the future of hosting capacity studies due to their advantages in systematically capturing optimal solutions to complex situations.For instance, metaheuristic methods provide optimal or near-optimal solutions, and robust methods provide resilient solutions by addressing data uncertainties.Moreover, game theory-based approaches can be adopted to deal with the problems of multiple players with conflicting objectives and thus optimally find an equilibrium point.Linear programming is increasingly employed to reduce the computational burden of optimizationbased methods.A set of linear power flow equations are solved in these methods to circumvent the iterative process [67].Thus, the relevance of optimization methods lies in the ability to navigate complex solutions and strategically reach optimal solutions.On the whole, the selection of an appropriate method for EV HC evaluation, lies in the specific objectives, available resources, accuracy and time constraints.
An overview of the problem formulation and evaluation methods for EV HC is given in Table 1.

ELECTRIC VEHICLE HOSTING CAPACITY IMPROVEMENT
The growing EV penetration substantially impacts the network performance and demands adequate measures to accommodate EVs.Thus, hosting capacity improvement methods are of particular interest in further literature as some networks already reached the point of maximum EV penetration.This section reviews the improvement methods and associated benefits in detail as follows.

HC improvement methods
The HC enhancement strategies address the network issues specifically resulting from EV integration.This improvement analysis broadly classifies the HC enhancement into two categories: a percentage rise in EV HC and an improvement in network performance by mitigating network issues.

Smart charging
The EV charging during peak demand periods might overload the network due to an additional EV load over the existing load.
This additional demand due to uncontrolled EV charging can limit the network's EV HC due to the associated violation of network limits.However, such undesirable charging impacts can be mitigated by adopting smart charging schemes that include but are not limited to delayed or off-peak charging, peak shaving or curtailment, and valley filling [74].In this context, a regulated EV charging approach was found to improve EV HC of a real suburban distribution network in [75].Similarly, the research results of [76] favoured smart EV charging for raising HC and mitigating network issues.A network upgrade is a conventional approach for hosting rising distributed generation/loads in the network by replacing the network components.However, the cost associated with these upgrade solutions is the main deterrent for DSOs that are trying to accommodate EVs using the existing network capacity.Thus, the traditional upgrade solutions were avoided by more than a 50% reduction in network reinforcement instances using smart charging in a Liverpool network [77].It follows that smart charging can maintain the technical limits of the network along with circumventing the cost of other expensive alternatives.Correspondingly, charging schemes can increase EV HC if EV owners are willing to adjust their charging habits and the research of [78] concluded that slow charging raised HC (301 EVs) as compared to fast charging (42 EVs).Similarly, the impact of charger size on HC and network performance was observed as reduced outage probability with slow charging in [79] and a low-rating charger allowed simultaneous charging of more EVs than a high-rating charger in [80].

4.1.2
Tariff reforms and DSM Demand side management (DSM) is a widely used approach in maintaining the demand and supply balance by implementing various schemes depending on user involvement.For instance, the application of incentive-based DR and price-based DR depends on user comfort and participation.These schemes impact user comfort by controlling EV charging where the former is a direct load control approach, and the latter motivates the users to change charging patterns depending on price.TOU price-based charging [81] is a common variant of pricebased DR by shifting the EV load to off-peak hours depending on pricing signals.It can be implemented by different charging prices at peak and off-peak demand hours.Thus, the customers are motivated to charge at off-peak hours due to lower prices, subsequently reducing the peak demand of the network.However, the proper implementation of market-based solutions depends on the customer's responses to the EV tariff structures and EV load uncertainties.Therefore, it is important to consider the EV load on top of the peak network load while analysing the prices of peak and off-peak load hours in the TOU price schedule.Otherwise, the simultaneous charging shift of EVs during the lower electricity rate timings might result in other peaks that are even higher than the uncoordinated EV charging resulting in a state known as the rebound effect.The findings of [82] are aligned with this phenomenon, where the authors analysed the charging impacts of peak network load on a UK distribution system with uncoordinated and off-peak charging.
The EV clustering during the off-peak charging period created an even higher peak load of 470 kVA than the uncontrolled charging load of 400 kVA.Altogether, the charging coordination should consider the EV charging simultaneity in addition to the existing loads while developing any market structures.
Nevertheless, the TOU methods are unable to account for the transient variations in demand due to significantly early pricing signals using the historical supply-demand conditions instead of current load curves.Therefore, a dynamic pricing approach is relevant to account for the actual power system balance.Accordingly, the findings of [50] are aligned with using dynamic pricing for DR to increase the HC using a bi-level optimisation approach with price as the upper level and energy usage as the lower level of the problem.

Power quality improvement
The power quality or voltage regulation problems arising due to massive EV integration are a challenge for the power grid.However, these issues can be handled by reactive power support (RPS), OLTC or power factor (PF) control approaches.The potential of RPS for increasing EV HC was investigated in various studies.The fast charging in a Norwegian distribution grid resulted in an additional voltage deviation of 0.03 p.u. as compared to slow EV charging [83].However, reactive power injection of 0.74 (leading) permitted fast charging even in the weakest part of the network while complying with the voltage drop limits.Likewise, the RPS was complemented by coordinating with other techniques for increasing the EV hosting capability of the network.This concept has been analysed in [84] as a coordinated approach of reactive power control, load curtailment and load shifting.Thus, the HC was raised from 160 EVs to 5450 EVs by simultaneously synchronising the potential of multiple schemes with charging cost reduction as an added benefit.Finally, harmonic distortion is an important phenomenon regarding power quality and a real parking area was studied regarding EV integration and total harmonic distortion limits in [85].The authors of this study proposed to maintain the THD limits, at the point of common coupling, by detaching 24 EVs which reduced THD from 10.7% to 7.5%.These studies complement the notion of increased EV penetration by addressing power quality concerns.

ESS or PV support
Energy storage systems (ESS), while being expensive, overcome technical issues for integrating modern loads such as EVs.The potential of ESS can be utilised to reduce the grid stress resulting from peak load demand and ref. [86] studied the increase in charging capability by optimal charging and battery energy storage (BES) to reduce the charging demand from the grid.In addition, ESS provides financial benefits to DSOs by postponing traditional grid upgrades with the proper sizing and location of storage.In this regard, the optimal location of ESS was investigated in [87] for increasing the HC using a model predictive control system.The placement of ESS at the end of the feeder provided a twofold HC improvement that further highlighted the importance of the optimal placement of the storage system.EV-related issues can be addressed by PVs and the benefits of reactive power support from the PV inverter were investigated in [88,89].Ref. [88] used an under-voltage index that defined the voltage support needed from the PV inverter by tracking the severity and extent of voltage drop below the defined limit (0.94 p.u.).Likewise, the PV potential can also be realised by mitigating network issues such as overloading, power losses and voltage profile deterioration due to EVs [90].However, the financial suitability of ESS and PV support cannot be decided without a comparison with the competing solutions, and thus a complete cost and benefit analysis is required in this regard.

NR, DSO and end-user contracts
Network reconfiguration (NR) can be broadly classified into feeder and phase reconfiguration.The former approach deals with the topological changes by opening and closing sectionalising and tie switches, respectively and the latter by reducing the unbalance between the three phases.Thus, a rerouting of network loads maintains security and reliability.This indirect increase in EV HC by feeder reconfiguration was observed in [59] and [71] where the network reliability issue was addressed by improving the power supply capacity index and by maximising the energy index of reliability value, respectively.Moreover, rephasing was adopted to reduce voltage unbalance for reaching the optimal network configuration in [91] by solving a nonlinear HC improvement problem for an IEEE 37-bus network.Thus, the reconfiguration approaches can improve the hosting capability of networks by addressing the underlying technical concerns.
In addition, DSO contracts can be utilised for HC enhancement and network congestion was addressed by using the DSO's owned flexible capacity that could be activated through contracts upon network violations [92].However, the network congestion problem cannot be solely attributed to the individual peak loads, but the coincident peak loads should be considered for mitigating overloading.In this context, a customer capacity trading approach was investigated in [70] to reduce the coincident peaks by utilising the flexibility potential of EV loads.This study highlighted the benefits for the DSO and EV owners, where the former gets the benefit of system stability and reduction of total system cost and later as customer cost reduction.Therefore, the perspective of DSOs and EV owners can be coordinated for the optimal solutions to raise HC by reducing the rebound effect and charging price.Table 2 gives an overview of the relative improvement in EV HC using different techniques.
Each HC improvement method has its limitations and advantages depending on available technology, regulatory systems and resources.For instance, smart charging or DSM controls EV charging for peak load management or grid support.However, these methods rely on customer consent and behavioural patterns for successful implementation.Tariff reforms facilitate load distribution through incentive-based charging at off-peak times but require proper economic and regulatory setup to execute this interaction.Similarly, ESS or PVs offer flexibility by optimal utilization of available resources while accompanied by the limitations of investment and resource management complexities.Finally, NR and DSO/customer contracts can potentially improve the EV HC through temporary service disruptions and transparent and secure communication, respectively.Thus, Table 2, encompassing the HC improvement methods, provides a panoramic overview of the effectiveness of proposed methods by various researchers.This could benefit the respective stakeholders to make informed choices depending on the specific objectives of integrating EVs into the modern energy landscape.

Improvement in network performance
The HC improvement is not always materialised in terms of percentage values, but it can be expressed as improved network performance compared to the base case without any mitigation means, as shown in Figure 9. Therefore, this section covers literature that expounds on improving network performance without providing numerical values of HC improvement.The enhancement in EV integration capacity is observed by the improvement of technical indices such as voltage profile, overloading, and power quality that undergo deterioration due to EVs as observed in [93][94][95].Additionally, charging cost reduction is another benefit of mitigation means.The research of [96] showed a cost reduction from 3912 (€) with uncontrolled charging to 3750 (€) and 2769 (€) with smart charging and BES, respectively.However, the initial investment cost of battery storage must not be understated despite a higher cost reduction in the above-mentioned research analysis.Thus, the superiority of any method considering economics is complex due to the initial investment cost and needs a cumulative cost analysis for evaluating the competitiveness of the proposed method.In addition to the cost benefits, the reduction of greenhouse gases and energy loss was observed as an added advantage along with cost reduction in [97,98] where coordinated charging was employed as the HC improvement method.Certain mitigation methods, such as TOU pricing schemes, might give rise to the rebound effect due to off-peak clustered EV charging.In this respect, the authors of [99] adopted a novel pricing scheme using a customer division approach to avoid the rebound effect due to TOU pricing.This approach was based on multiple TOU signals to different customer groups with a time delay to avoid the simultaneous charging shift to low price periods and, thus, provide an improved valley filling.
Finally, computational time reduction, enhanced RES integration and reliability improvement are some additional benefits of using HC enhancement means observed in the literature.Figure 10 displays a histogram plot using values from Table 2 where the x-axis shows relative HC rise and the y-axis shows the number of research papers corresponding to a particular HC rise.It shows that the initial HC can be raised by more than three times through mitigation methods which proves the point that HC is not a hard limit on EV penetration.The relative HC rise is defined here as (final HC−initial HC) divided by the initial value of HC.For instance, a relative rise of 1.81 corresponding to the initial and final HC values of 11% and 31% depict that the HC is raised by more than twice as compared to the initial value [53].Nevertheless, the efficiency of mitigation solutions for addressing network issues depends on a variety of factors, including compliance with existing grid codes, social acceptance, associated costs and availability of technological infrastructure.

Current and future EV projections
As a showcase, we consider Norway, which is often regarded as the pioneer in EV adoption due to its progressive policies towards sustainable mobility, with the Parliament having a target of zero-emission cars by 2025 [100].Norway's transport sector is expected to be electrified within 2030 [70].It is estimated by the Water Resource and Energy Directorate of Norway that the status of EV adoption in Norway would substantially increase the load levels by 2030.This EV adoption might count to 1.5 million EVs with an estimated annual energy consumption of up to 4 TWh [83].It has been reported that EV registration surged 41% globally in 2020 regardless of the Covid-19 pandemic.In Norway, EV sales experienced a percentage increase of 120% by a comparison of 80 (thousand) EVs in 2018 and 176.3 (thousand) in 2021 [101].However, the energy required for charging this EV fleet can strain the existing networks and the simultaneous EV charging can further cause the problem of system stability.Therefore, it is imperative to understand the charging characteristics of EVs due to the unprecedented surge in EV adoption across the world.EV charging criteria is mainly dependent on the charging time, location and charger size.

EV charging characteristics
EV charging mode impacts network performance depending on the charger type and placement of the charging points [102][103][104][105][106]. It is majorly classified into three levels that are distinct based on their location, charging duration and driving range.Level 1 charging is generally the slowest among all with residential locations whereas level 2 charging provides a faster charging time and rate, as compared to level 1, with residential and commercial locations.Finally, level 3 charging, also known as DC fast charging, provides the highest charging rate.The charging powers of level 1 and level 2 are up to 4.6 kW and 44 kW, respectively, whereas the maximum power corresponds to level 3 as 50-150 kW [107].The state of charge is an important metric for EV HC calculation as it is used in EV modelling and calculating the charging energy of an EV.Equations (9a) and (9b) give SOC calculation where d is the distance driven by an EV and AER is "all electric range".
EV charging energy can be expressed as (10) where C represents the EV battery capacity in kWh.Moreover, depth of discharge is defined as the amount of discharged EV capacity from fully charged battery and it is directly linked with SOC as given in (11).
Charging time and driving range is a major factor that varies with the expectations and requirements of EV owners.For instance, slow/home charging is not suitable for long travel distances and shorter charging time requirements.Fast charging, on the other hand, provides benefits in terms of high-power rating, but the high-cost constraints can prove a deterrent to such installations.Most studies comprising 35% of articles, investigated the maximum network capacity for EV integration using slow charging with a mix of only home or home/workplace charging.Whereas normal and fast charging is investigated by only a few research papers.This points toward a scarcity in the HC studies considering different charging powers and needs research on higher power ratings considering a future shift in the transport sector and EV adoption.Figure 11 provides a general overview of the important charging criteria of EVs.
Additionally, the EV-related behavioural aspects are of particular importance to improve the HC in addition to mitigating the technical violations of the network.In this context, charging time scheduling dictates the range anxiety of EV owners due to EV mobility constraints and EV owners' comfort [108][109][110].Thus, the implementation of DR programs is a function of consumer acceptance and substantially varies with EV owner's choice.It follows that the charging time and consumer responses are interlinked, and this aspect needs consideration while calculating the EV HC of networks.

DISCUSSION
The evolution of EVs and methodological advancements over time necessitated studying the grid dynamics and future innovative technologies.Over the past years, the significance of EVs has risen in the pursuit of sustainable mobility and reduced reliance on fossil fuels to transform the transportation landscape.However, this evolutionary journey of EVs leads to unique challenges despite providing environmental benefits and opportunities.A rise in EV adoption due to competitive technology and low carbon footprint is accompanied by their impacts on energy networks.This continuous surge in EV charging unfolds the complexities involved with the traditional energy networks regarding grid capacity and dynamics of energy demand.Thus, the question of accommodating everincreasing EVs into the current and future energy landscape is of paramount importance to bridge the gap between the mobility and energy systems.However, increasing EV penetration demands a shift in the technological paradigm towards an emerging end to sustain grid stability.No doubt, the initial interplay of EV and the grid could be studied by simple load forecasting and basic modelling due to limited EV adoption.However, the emergence of technology and EVs paved the way for adopting sophisticated simulation techniques to understand EV-grid interaction on real data sets.As time progresses, time and grid resources are valued increasingly, and robust and accurate solutions are required to ensure grid stability.Thus, optimization-based methods offer optimal solutions by strategically managing the EV and grid interactions.Data sorting and analysis have seen revolutionary shifts due to the use of machine-learning and deep-learning-based modern analysis tools.Therefore, future research is continuously moving forward to unfold novel methods to empower EV integration for sustainable transportation.
It follows from the above discussion that the EV integration problem is not just a transportation problem but a power system problem to handle multi-objective functions.Generally speaking, EV penetration studies are performed by optimising various objectives such as losses, cost and technical limits while maximising HC.It is noteworthy to realise that the cost, as an optimisation objective, is a complex variable while evaluating some mitigation methods for raising the HC.For instance, the storage systems and switch reconfiguration approaches involve underlying expenses incurred on their implementation.The storage systems incur an additional investment cost, while the NR approaches must be monitored carefully, considering the influence of the number of switches on system cost and complexity.Therefore, the metaheuristic approaches are widely employed in the literature, which can be attributed to their applicability to broad problems and their usefulness in solving multi-objective formulations.
Alternatively, the EV load shifting aspect can be exploited to avoid the expensive methods for improving HC.However, certain charging coordination schemes based on tariff reforms are even more problematic than uncontrolled charging due to the rebound effect.In addition, EV owner behaviour is a determining factor toward EV adoption and HC analysis because of the customer preferences for charging modes.Most of the existing EV HC research is based on network capacity analysis for slow charging, whereas EV adoption is a function of charging time.This signifies that the customer range anxiety and social acceptance can be improved by fast charging.It not only addresses the range and time concerns of EV owners but reduces the need for charging points to minimise the charging station planning costs for CS owners.However, DSOs must take corrective actions to prevent imminent problems due to future EV uptake and potential installations of fast charging points despite their potential benefits.It follows that the social barriers, in addition to technical and economic concerns, hold significance for HC analysis and must be accounted for in optimising multiple objectives and mitigation methods.

CONCLUSIONS AND FUTURE WORK
The key takeaways of this article are methodological analysis to outline the holistic overview of EV HC.It is a unique combination of concrete data paired with the practical evaluation framework of actual networks.It not only encompasses the theoretical background of the state-of-the-art EV HC studies but offers a unique compilation of the actual EV HC and percentage improvement values.Thus, the implications of this review go beyond academic researchers to the industry stakeholders to assist them as a reference point for aiding research and collaboration among different entities.EV adoption is in the early development stages and the EV data is very limited and specific to a country.Moreover, the communication between utility and EV owners involves underlying security risks and owners' range anxiety is a real concern considering the location of the charging station.Broadly speaking, the existing EV fleet might not be an immediate stability problem for the distribution networks in developed countries due to huge RES and charging infrastructure availability.However, the future fast charging facilities can soon exhaust the networks if the proper measures are not undertaken.The EV HC results for most studies are based on benchmark test networks instead of real networks due to a lack of input data.Despite this, the research findings can be used as an estimate for planning real networks considering the future influx of EVs.Furthermore, the scalability of EV HC research can be extended to capture the real case scenarios for the application of theoretical approaches in a broader context.
It is important to foster collaborative research between the energy industry, transportation sector, urban planning and data scientists for a holistic overview of future EV opportunities.Data science and machine learning algorithms can provide a promising future for EV HC enhancement through the application of predictive analytics.A potential research direction could be the integration of state estimation and predictive analytics concepts to enhance the accuracy of HC predictions by observing the current grid state and future EV events.Generally, state estimation provides valuable insights into the network condition, including voltage, current and power flows.Whereas, predictive analysis utilizes historical data and load patterns to forecast future EV events.Thus, a synergy between these approaches holistically converges the real-time grid data and future events for the proactive planning of future EVs.
In addition, the data-driven integration of EV HC analysis into the urban planning framework can enhance the optimized EV charging and urban development.The physical location of charging station planning and EV HC analysis have been significantly explored by researchers as separate entities.However, their holistic integration remains largely unexplored and requires an inclusive approach to city planning that extends beyond the traditional physical location of EV charging stations.This would enable an optimized utilization of public spaces, improved charging infrastructure, and environmental sustainability along with technical insights into the potential impacts on the grid dynamics.
A large part of the literature covers centralized EV charging primarily managed by utility companies.While peer-to-peer transactions and monetization of assets are important future avenues, a blockchain-enabled peer-to-peer charging network can revolutionize this idea.It has the tendency to create a decentralized, transparent and sustainable environment through direct interactions, verified transactions and the use of renewable energy sources at host stations.Thus, blockchain technology can be investigated in terms of its technical suitability for hosting more EVs and its scalability to handle larger networks.

FIGURE 1
FIGURE 1The research trends of electric vehicle hosting capacity over the course of the last decade in the studied literature.

FIGURE 2
FIGURE 2 General framework for electric vehicle hosting capacity review.

FIGURE 3
FIGURE 3 Reference definitions of electric vehicle hosting capacity and their distribution in the literature.

FIGURE 4
FIGURE 4 Distribution of limiting factors used in the literature.

FIGURE 5
FIGURE 5 Estimated electric vehicle hosting capacity of different networks without the application of any mitigation means in the studied literature.

FIGURE 6
FIGURE 6The estimation criteria of electric vehicle hosting capacity.

FIGURE 7
FIGURE 7 Process model for electric vehicle hosting capacity calculation.

FIGURE 8 A
FIGURE 8 A share of different evaluation methods for the analysis of electric vehicle hosting capacity in the reviewed literature.

FIGURE 9
FIGURE 9Indirect network improvement using mitigation means.

FIGURE 10
FIGURE 10 Relative rise in hosting capacity values using different mitigation means observed in the studied literature.

FIGURE 11
FIGURE 11 Important features of electric vehicle charging and load modelling.