Enhancing smart charging in electric vehicles by addressing paused and delayed charging problems

Smart charging of electric vehicles can alleviate grid congestion and reduce charging costs. However, various electric vehicle models currently lack the technical capabilities to effectively implement smart charging since they cannot handle charging pauses or delays. These models enter sleep mode when charging is interrupted, preventing resumption afterwards. To avoid this, they should be continuously charged with their minimum charging power, even when a charging pause would be desirable, for instance with high electricity prices. This research examines this problem to inform various stakeholders, including policymakers and manufacturers, and stimulates the adoption of proactive measures that address this problem. Here, we demonstrate through technical charging tests that around one-third of tested car models suffer from this issue. Through model simulations we indicate that eliminating paused and delayed charging problems would double the smart charging potential for all applications. Lastly, we propose concrete legal and practical solutions to eliminate these problems.


REVIEWER COMMENTS
Reviewer #1 (Remarks to the Author): -justify the selection of the currents given in Fig. 1. why these current ranges are selected.
-Is the aim of this work only raising awareness?as it is written in the last paragraph of intro.What about solutions regarding these challanges?Some are already given.Hence, eloborate on the contribution more in this section.
-EV can handle 6 and 32 amperes.Is it for all type of EVs?What about the size of EVs?There is a need to conduct sensitivity analysis.
-When there is minimum charging current; what about charging time optimization?Charging speed is critical for some custometers, not having time.Fast charging etc.Hence, discussion on the charging times and its effect on the customer satisfaction should be provided.
-Bsed on the conducted simulation, i suggest to add few bullet points for policy making in charging.
-Is it possble ti validate the model in real environment?-Add discussion on renewable based EV charging since the timings will be dependent on availability of solar, wind etc.
Reviewer #3 (Remarks to the Author): This paper investigates the minimum charging power issue that may lock the potential of smart charging in reducing cost, mitigating grid congestion, and offering flexibility services.To prevent EVs from shifting to sleep mode when charging is interrupted, a minimum charging power is necessary, even at moments when a charging pause is desirable.This paper uses extensive model simulations to show the impact of this technical issue on the benefits of smart charging in the above scenarios.A quantitative analysis is conducted from the real-world data, showing that almost half of the potential benefits are lost.This paper aims to raise awareness of this technical issue and also proposes several legal and practical solutions.
In general, this paper focuses on a small problem yet shows that its impact could be surprisingly huge.The three main plots from running different charging models on the real-world data are informative and offer nice insights.However, I have major concerns about the technical contributions of this work.First, the minimum charging power is easy to model and incorporate into most of the EV charging literature.Therefore, technically it is not a novel problem.Second, the three charging models are very standard offline formulations to schedule EV charging.However, they are not directly implementable in an online fashion in practice.Thus, the computed reduction in potential benefits is, at most, a lower bound.It may not reflect the actual impact of the minimum charging power.Last, the discussion of the options for eliminating this charging problem is a bit abstract.The enforcement methods from governments/legislation, charging operators, and consumers are briefly mentioned, but details are missing.
I also have a few other comments below.
1.The authors highlighted delayed charging problems in the paper.However, given Figure 1, it is a bit confusing since the "delay charging" test seems the least relevant (no pause or intermittency).I suggest using a more clear terminology.
2. From the smart charging test results (Figure 2), what is the percentage of unsuccessful tests that were caused by charging currents exceeding charging signals?This unsuccessful label does not seem relevant to the necessity of minimum charging power.
3. The three charging models are also stylized.If the authors consider discrete charging power instead of continuous one, I suspect the results would be much different.
Reviewer #2 (Remarks to the Author): Please see the attached comments / suggested edits / queries on the PDF.
I think the Abstract is fairly clear -but not fully so.
Whose awareness should be raised, and with hat 'theory f change' for impact?
In the Introduction: Really needs to say more about peaks for which type of EV charging context ['on-street', home, workplace'].
Is it just EV owners, or (also) users -and how do you see the distinction between the two in your article / argument?You really need to give the reader a definition of 'smart charging'.
Is there a pattern of EV models to which you concerns apply: Hybrid EVs? Full EVs? Depending on EV battery capacity?How do you mean by having paid little attention?Researchers?Regulators?Governments?Or perhaps [less likely, and if so more deliberate?] EV manufacturers?Current legislation?Do you mean standards / regulation / legislation?At global/international level [IEC], EU?, national, regional/local?Results: see Figure 2: Is there a pattern of EV models to which you concerns apply: Hybrid EVs? Full EVs? Depending on EV battery capacity? 2.2.1.You should have mentioned mentioned [both] Time-of-Use and Dynamic Electricity [EV charging] Prices much earlier before -even in the Abstract.
2.2.2.I don't follow with the logical connections that "grid congestion solutions will fail with lower EV adoption rates".Why / how so? 2.3 Needs a bit re-writing in more precise language -see my queries on the PDF.
Figure 5: which types of charging stations, and where?Will most EV models "soon" be able to perform V2G / bi-directional charging?Think a bit more about at which level [EU / national] these potential voluntary agreements with manufacturers?
Now you are talking about EV "buyers", previously "owners and at times "users".Work this through and clarify this right through.And a significantly share of EVs is now leased ...

Reviewer #1 Response
Dear reviewer, Thank you for your constructive review comments.We believe that due to your comments, the quality of the manuscript has improved considerably.
-justify the selection of the currents given in Fig. 1. why these current ranges are selected.
Figure 1 presents the considered charging currents for the different conducted smart charging tests.As Figure 1 shows, charging currents of either 6 or 32 amperes were regularly considered in the smart charging tests.These currents refer to the minimum and maximum charging current for AC EV charging, as defined in the International Electrotechnical Commission's communication standard for EV charging (IEC 61851).
Inspired by this comment, we have thoroughly revised the text in the 'Results -Technical smart charging tests' section and included more details about the technical smart charging tests that were conducted and their relevance.The paragraph discussing these aspects now reads as follows: The smart charging tests conducted at the Testlab aim to evaluate the EV's response when exposed to various charging patterns that could occur with different applications of smart charging.-Is the aim of this work only raising awareness?as it is written in the last paragraph of intro.What about solutions regarding these challanges?Some are already given.Hence, eloborate on the contribution more in this section.
Thank you for this suggestion.Inspired by this comment, we have elaborated on the contributions of this work in the introduction, highlighting that we quantify the impact of EVs' inability to handle charging pauses on the effectiveness of EV smart charging and that we present different solutions for these challenges (last sentence).The contributions paragraph now reads as follows: In this work, we shed light on these problems to raise awareness among relevant stakeholders (e.g., grid operators, EV manufacturers and policymakers) about the prevalence of technical smart charging problems and their impact the effectiveness of smart charging.We present the results of large-scale technical charging tests, which indicate that around one-third of the EV models in the market cannot handle charging pauses or delays.Moreover, this study presents the results of model simulations that quantify the impact of EVs' inability to perform paused or delayed charging on three different applications for which smart charging can be used, namely: i) charging cost reduction, ii) mitigation of grid congestion, and iii) offering flexibility products to grid operators.The outcomes of these model simulations show that the potential impact of smart charging is halved for all applications if paused or delayed charging cannot be considered.Lastly, the current international regulations and standards on this topic are discussed, and options to eliminate paused and delayed charging problems are analysed.
Is it for all type of EVs?What about the size of EVs?There is a need to conduct sensitivity analysis.
Most EV models and EV charging stations accept charging currents between 6 and 32 amperes.This minimum current is defined in the IEC 61851-1 communication standard between the EV and the charging station.It states that a minimum charging current of 6 amperes is required before an EV is allowed to start charging.This is why we considered the minimum charging current of 6 amperes in this work.
The maximum charging current of 32 amperes is also based on this standard.This standard defines that the maximum charging current for all AC charging modes (i.e., non-fast charging modes) equals 32A.The on-board chargers of some EV models limit the charging power with higher charging currents.For instance, the Tesla Model 3 limits itself to 16A when using 3-phase AC charging (e.g., see this link).This did not affect our analysis.As described in the section named 'Methods -Model simulationsdata inputs & preparation', this study considered historical EV charging session data and the maximum charging power of a charging session was obtained from the highest measured charging power in a session, and not based on the maximum allowed charging current of 32A.Hence, if a vehicle limits the charging power when exposed to a charging current of 32A, this is reflected in the observed maximum charging power in the used data.
The minimum charging current of 6A was used to compute the minimum charging power of a vehicle (see the section named 'Methods -Model simulationsdata inputs & preparation').As discussed above, all vehicles should be able to charge with at least 6A according to the IEC 61851-1 charging standard, which means that the minimum charging power of an EV can be derived with high accuracy from the minimum charging current of 6A.For these reasons, we believe that there is no direct need to conduct an additional sensitivity analysis.This reviewer's comment has motivated us to provide additional explanation about the charging limits for EV models in the manuscript.We have included the following short explanation in the section named 'Results -Technical smart charging tests' of the revised manuscript: As shown in Figure 1, this test considers a fluctuating charging signal between 6 and 32 amperes at a 60-second interval.These current values correspond to the prescribed minimum and maximum charging currents for EV charging, as defined in the International Electrotechnical Commission's communication standard for EV charging [23].
-When there is minimum charging current; what about charging time This is an important point that is being addressed.Our research demonstrates that incorporating charging pauses significantly enhances the efficiency of various smart charging applications.However, the introduction of charging pause may potentially trigger range anxiety concerns optimization?Charging speed is critical for some custometers, not having time.Fast charging etc.Hence, discussion on the charging times and its effect on the customer satisfaction should be provided.
among EV users.In scenarios where a minimum charging current of 6A is consistently considered after connecting an EV to the charging station, and an EV owner departs from the station prior to the estimated departure time, it is ensured that the EV has at least partly been charged, as the vehicle was continuously supplied with a charging current of at least 6A.On the other hand, if charging pauses are implemented and the connection duration of an EV to the charging is not accurately predicted, there is a risk the EV has hardly been charged at departure, resulting in reduced customer satisfaction.
To further address this point, we have renamed the 'Conclusions' section into 'Discussion'.This section has been fully revised and different discussion points have been added to this section, including the discussion point brought up in this reviewer comment: Although this research indicated that it is important that EVs are able to deal with paused and delayed charging, it should be acknowledged that actual implementation of paused and delayed charging could trigger range anxiety issues among EV users.When scheduling the charging of an EV, its departure time from the charging station has to be estimated through user input and/or by applying forecasting methods.If an EV departs from the charging station before the anticipated departure time and the vehicle has continuously been charged with a charging current of at least 6 amperes, it is ensured that the EV has at least partly been charged.However, with the application of paused and delayed charging, there is a risk that the EV will receive a minimal charge if it departs before the expected departure time.Therefore, smart charging operators must exercise caution regarding the uncertainties in their models when employing paused and delayed charging scheduling.The reader should bear in mind that this additional uncertainty was not considered in this work's model simulations for paused and delayed charging.Nevertheless, it should be realized that this real-world challenge may be largely mitigated by actively requesting user information about their charging sessions (e.g., expected departure time & charging demand), for instance through a mobile application (e.g., [37,38]), either as user-defined defaults with optouts or as per-session preferences.
-Bsed on the conducted simulation, i suggest to add few bullet points for policy making in charging.
The section named 'Results -Options for eliminating charging problems' discusses a range of enforcement or stimulation options that governments and policymakers could use to eliminate the delayed and paused charging problems of EVs.These enforcement and stimulation options were not presented as bullet points, as we discuss each option in a separate paragraph, and the text following the bullets would become very long.We believe that this would affect the flow and readability of the manuscript, which is why we decided to not use bullet points when discussing these options.
We would also like to highlight that the discussion of policy options has been considerably extended compared to the previous version of the manuscript.This discussion now reads as follows: If the public sector, including governments and regulatory agencies, deem that the paused and delayed charging problems are too severe to be resolved without intervention, there are multiple enforcement or stimulation methods available.First, these organizations could stimulate manufacturers to comply with the existing standards that address this issue.This can be done by taking an active role in informing manufacturers about the importance of EVs being able to handle paused and delayed EV charging.Alternatively, the public sector has the option to establish an EV model certification program, where EV models that successfully completed a set of EV smart charging tests are granted a certificate, which could make the specific EV model more appealing to consumers.
The public sector can also enforce the elimination of paused and delayed charging problems by implementing regulations on this topic.This could be modelled after the type-approval system that is currently in place.Before getting access to public roads in the EU, all car models need to acquire type-approval [36] This work describes the results of a real-world smart charging experiment.As described in this work, no charging pauses could be considered when scheduling the EVs, and a similar model formulation was adopted when scheduling EV charging in this experiment.
This has been further highlighted in the manuscript, in the section named 'Methods -Model simulationscharging models': The validity of this model has been confirmed through real-world application [42] and the model is formulated as follows: -Add discussion on renewable based EV charging since the timings will be dependent on availability of solar, wind etc.
We agree that this requires further discussion.As highlighted in Figure 2, many EV models cannot deal with smart charging applications that consider fluctuating or intermittent charging profiles.This is problematic when applying solar charging, in which the charging power of an EV is directly linked to the output of a solar PV system.However, most of the paper is focused on the impact of EVs not being able to perform delayed or paused charging.For this reason, we have added the following text to the renamed 'Discussion' section: In addition, it should be recognized that the model simulations in this research exclusively focused on quantifying the impact of the EV's inability to perform paused and delayed charging.The results of the technical smart charging tests indicated that fluctuating or intermittent charging problems also occur frequently.This could harm the roll-out of different smart charging applications, including solar charging, in which the EV charging power depends on the output of a PV system.Solar charging could contribute to increasing PV self-consumption and facilitate the grid integration of PV technology [40,41].For this reason, policy addressing technical charging problems for EV smart charging should also encompass the resolution of technical charging problems related to fluctuating or intermittent charging signals.

Reviewer #2 Response
Please see the attached comments / suggested edits / queries on the PDF.
Dear reviewer, Thank you for your constructive review comments and thank you for providing a PDF with detailed comments.We believe that due to your comments, the quality of the manuscript has improved considerably.
I think the Abstract is fairly clear -but not fully so.Whose awareness should be raised, and with hat 'theory f change' for impact?
We have extended the abstract and further specified whose awareness should be changed, and what is the desired impact of this work.This is formulated as follows in the revised abstract: This research examines this problem to inform various stakeholders, including policymakers and manufacturers, and stimulates the adoption of proactive measures that address this problem.
In the Introduction: Really needs to say more about peaks for which type of EV charging context ['onstreet', home, work-place'].
In this work, we mostly focus on charging in residential low-voltage grids, which happens at 'home' or 'on-street' charging stations.
Inspired by this comment, we have further specified this in the introduction: Most EV charging occurs in Low-Voltage (LV) grids at home or on-street charging stations [3], and the majority of these grids were designed decades ago without the concept of EV charging in mind.The charging power of an EV is significantly higher than the typical peak-time power consumption of a household, and since most EV users tend to arrive at their charging station at a similar time, concentrated charging moments are expected in residential LV grids [4,5].
This has also been further outlined in the 'Methods' section.
Is it just EV owners, or (also) users -and how do you see the distinction between the two in your article / argument?
Thank you for addressing the inconsistencies in terminology here.In the original version of the manuscript, both references to "EV owners" and "EV users refer to the person driving the EV and charging it at the charging station.To improve consistency, we have replaced 'EV owners' with 'EV users' throughout the manuscript.
You really need to give the reader a definition of 'smart charging'.
The introduction has been extended by adding the following definition of 'smart charging': This provides ample opportunities for EV smart charging.With smart charging, EV charging sessions are optimized for different objectives by aligning the charging moments and charging speed over time with user preferences and current market or grid conditions [3,13].
How do you mean by having paid little attention?Researchers?Regulators?Governments?Or perhaps [less likely, and if so more deliberate?] EV manufacturers?
Thank you for pointing out that our initial statement 'Remarkably, the technical problems associated with smart charging have received little attention' could lead to confusion.The intention behind this statement is to convey that technical charging problems are not widely discussed in both scientific literature and the media, resulting in limited awareness among policymakers and EV manufacturers.To avoid confusion, this statement has been rephrased as follows: Remarkably, the technical problems associated with EV smart charging have hardly been addressed in scientific literature and the media, leading to low awareness about these issues among different stakeholders, including policymakers, EV manufacturers and grid operators.
Current legislation?Do you mean standards / regulation / legislation?At global/international level [IEC], EU?, national, regional/local?
In our initial statement in the introduction, we used the term 'current legislation' to refer to all regulations and standards that are currently in place on this topic.As highlighted in the section named 'Resultsoptions for eliminating charging problems', the current regulations and standards on this topic have an international scope.Consequently, we have substituted the term 'current legislation' in the introduction with 'current international regulations and standards'. Results: see Figure 2: Is there a pattern of EV models to which you concerns apply: Hybrid EVs? Full EVs? Depending on EV battery capacity?
No clear pattern is observed regarding the specific types of EV models experiencing technical charging issues; these problems manifested with comparable frequency in both hybrid and battery electric vehicles.In addition, the occurrence of the technical charging problems seem to be independent of the battery capacity of the EV models; the technical charging problems occurred with similar frequency between EV models with a small and large battery capacity.
It has been agreed with the EV manufacturers when performing the charging tests that the results could only be presented in an aggregated manner, to ensure that no individual EV model can be identified from the analysis.For this reason, we are not allowed to present more detailed results of the technical charging tests.

You should have mentioned mentioned [both] Time-of-Use and Dynamic Electricity [EV charging]
Prices much earlier before -even in the Abstract.
We agree that our references to 'dynamic electricity prices' in the initial version were incomplete.This has been replaced throughout the manuscript by 'static and dynamic time-of-use tariffs'.We believe that the term 'static and dynamic time-of-use tariffs' covers all applications for which EV smart charging can be used to reduce costs: -Static time-of-use tariffs: Prices vary throughout the day, but in a repeating manner, e.g.fixed night and day tariffs.
-Dynamic time-of-use tariffs: Prices vary throughout the day, but in a non-repetitive manner, e.g.day-ahead market prices Next to replacing 'dynamic electricity prices' by 'static and dynamic time-of-use tariffs', we have introduced 'static and dynamic time-ofuse tariffs' earlier in the paper, in the introduction.This is the updated text in the introduction: Additionally, it can help EV users reduce their charging costs by taking advantage of moments with low electricity market prices when considering static or dynamic Time-of-Use (ToU) pricing schemes [19,20].
Given the strict word limit for the abstract of 150 words, we have not introduced static and dynamic time-of-use tariffs in the abstract.
2.2.2.I don't follow with the logical connections that "grid congestion solutions will fail with lower EV adoption rates".Why / how so?
The initial version of the manuscript contained the following statement: "When a minimum charging current needs to be applied, EV charging cannot be completely shifted away from peak hours.Hence, grid congestion solutions will fail with lower EV adoption rates.".
We acknowledge the reviewer's concern about the clarity of the final sentence in this statement.Our intent in that sentence was to convey that the application of smart charging to mitigate grid congestion problems which are induced by EV charging is less effective if no paused or delayed charging can be applied, as the EV charging cannot be fully shifted away from peak hours.As a result, grid congestion problems are more likely to manifest at lower EV adoption levels when smart charging does not involve paused or delayed charging, in contrast to a scenario where such features are incorporated.
To improve the clarity of the manuscript, this statement has been rewritten as follows: Smart charging can also be used to address grid congestion problems induced by EV charging by shifting the charging from moments with high local grid load to moments with low local grid load [28,29].When charging pauses cannot be considered, EV charging cannot be completely shifted away from peak hours.Consequently, grid congestion problems will manifest at lower EV adoption levels when deploying smart charging without paused or delayed charging compared to the deployment of smart charging with these features.The analysis used to create Figure 4 considered an LV grid in a residential area in the city of Utrecht, the Netherlands.This has been further specified in the caption of the figure in the revised manuscript.
2.2.3.Look at the grammar in the first sentence.
The first sentence of this section.has been reformulated as follows: "Moreover, smart charging can be applied to offer flexibility services to grid operators, for instance by supplying balancing reserves (e.g., frequency restoration reserves) to TSOs for restoring the balance between supply and demand [30][31][32]." 2.3 Needs a bit re-writing in more precise language -see my queries on the PDF.
Thank you for providing the very helpful and detailed comments on the PDF.We have thoroughly checked your comments and have implemented changes in all sections.The changes are highlighted in the 'revised manuscript with changes highlighted' document.This analysis used data from on-street charging stations located in the city of Utrecht, the Netherlands.This has been further specified in the caption of this figure.
Will most EV models "soon" be able to perform V2G / bi-directional charging?
This is an interesting point.There are different aspects that will affect the future roll-out of V2G/bi-directional charging: 1.The first production-scale EV models that support AC V2G were recently introduced to the market (e.g., Hyundai IONIQ5), but the majority of car models that are currently introduced to the market are not V2G-compliant.2.Not all AC charging stations are V2G compliant, although this is now a requirement in the tenders for public charging stations in different European cities, including Utrecht, the Netherlands.3.Only a small share of the car models are currently compliant with the ISO15118-20 standard that is required to perform bidirectional charging.4.There are some regulatory issues (e.g., certification of EVs by grid operators before the EV can supply electricity to the grid) which need to be tackled before large-scale implementation of V2G.
This shows that a separate study could be dedicated to the barriers to and current status of the roll-out of V2G technology.As we aim to keep this work concise, we have decided to not include this discussion into this work.
Think a bit more about at which level [EU / national] these potential voluntary agreements with manufacturers?
Thank you for addressing this interesting point.We believe that all measures described in the section named 'Resultsoptions for eliminating charging problems' are ideally taken on an international level (i.e., EU level), to enhance efficiency and avoid discrepancies in the policies between different nations.This has been further specified in the manuscript: Ideally, all discussed enforcement or stimulation methods should be implemented at an international level, within entities like the EU, to enhance efficiency and maintain consistency in policies across different nations.Thank you for addressing this point.We have expanded the introduction to cost-optimized charging in the section named 'Results -Impact on smart charging's charging cost reduction potential', in order to provide the reader with better understanding about the mechanisms behind this.The description is now as follows: The inability of EVs to perform paused or delayed charging can diminish the effectiveness of different smart charging applications, including smart charging for participating in an electricity market that considers static or dynamic ToU tariffs (e.g., day-ahead electricity market) [24][25][26].Charging costs can be reduced by shifting the charging demand from moments with high prices to moments with low prices.
The provided paper was relevant and a reference to this work has been included in the manuscript.

Reviewer #3 Response
This paper investigates the minimum charging power issue that may lock the potential of smart charging in reducing cost, mitigating grid congestion, and offering flexibility services.To prevent EVs from shifting to sleep mode when charging is interrupted, a minimum charging power is necessary, even at moments when a charging pause is desirable.This paper uses extensive model simulations to show the impact of this technical issue on the benefits of smart charging in the above scenarios.A quantitative analysis is conducted from the real-world data, showing that almost half of the potential benefits are Dear reviewer, Thank you for your constructive review comments.We believe that due to your comments, the quality of the manuscript has improved considerably.
lost.This paper aims to raise awareness of this technical issue and also proposes several legal and practical solutions.In general, this paper focuses on a small problem yet shows that its impact could be surprisingly huge.The three main plots from running different charging models on the real-world data are informative and offer nice insights.However, I have major concerns about the technical contributions of this work.First, the minimum charging power is easy to model and incorporate into most of the EV charging literature.Therefore, technically it is not a novel problem.
We agree that this paper focuses on a very specific problem, but as the results of this work report a tremendous impact of this problem on the effectiveness of EV smart charging applications, we believe that it is important to dedicate a focused study on it.
The authors agree that introducing a minimum charging power constraint to an optimization model to avoid charging pauses will generally be regarded as a relatively minor challenge: One needs to include binary variables and a minimum charging power constraint to the optimization problem (see the section named 'Methods -Model simulations -charging models') to assure that i) EVs charge continuously with at least the minimum charging power after their arrival to the charging station until their charging demand is satisfied, and ii) the minimum charging power constraint no longer applies when the charging demand of a charging session is satisfied.This is exactly the main point of this work: even though it is not extremely challenging to incorporate the continuously-required minimum charging power into charging optimization models, it is neglected in all scientific literature on the topic of EV smart charging.In this study, we aim to show, through extensive model simulations, that most literature in this field overestimates the potential of smart charging and that more realistic insight into the current potential of smart charging can be obtained by making a relatively simple alteration to charging models (i.e., introducing a continuously-required minimum charging current to prevent charging pauses).This in turn demonstrates the impact of these constraints, resulting in a call for action (in the section named 'Results -Options for eliminating charging problems' and the updated discussion section).
Second, the three charging models are very standard offline formulations to schedule EV charging.However, they are not directly implementable in an online fashion in practice.Thus, the computed reduction in potential benefits is, at most, a lower bound.It may not reflect the actual impact of the minimum charging power.
Thank you for this remark.We indeed considered an offline (i.e., not real-time) model formulation for scheduling EV charging in this work.
Delayed and paused charging problems apply to both offline and online charging.Even when scheduling EV charging realtime (i.e., online formulation), charging pauses still cannot be considered, to avoid that EVs switch to sleep mode and become unresponsive to charging signals.Hence, this means that EVs with an online model formulation still need to charge continuously with at least the minimum charging current at moments at which they ideally would not charge.
The main difference between an online and offline model formulation will be the uncertainty associated with the charging session characteristics (e.g., departure time and charging demand) of each session.These charging session characteristics are generally assumed to be known in this study prior the optimisation with an offline model formulation.This was also the case in this study.When optimising a charging session in practice, this often happens real-time (i.e., online formulation) and there will be more uncertainty regarding these charging session characteristics.To avoid that EV users face an empty battery if they (unexpectedly) depart from the charging station before the anticipated departure time, the applied charging schedules with real-time charging scheduling might be more conservative to reduce the risk of user dissatisfaction.
However, for two reasons, we believe that the impact of using an offline model formulation to evaluate the consequences of EVs' inability to perform paused or delayed charging is relatively minor.Firstly, in many cases, departure information may be known at the time a car connects through explicit user information (e.g.mobile applications in which users communicate their expected departure time and charging demand).In these cases, the difference in uncertainty between the offline and online model formulation is considerably reduced, and EVs will not necessarily be charged more conservatively in the online formulation.Secondly, the increased uncertainty associated with an online model formulation also holds true in scenarios where paused and delayed charging cannot be applied.Hence, EVs will be charged more conservatively with the online model formulation compared to the offline model formulation, regardless of whether delayed and paused charging can or cannot be considered.
To further address the uncertainty aspect of uncertainty in the manuscript, we have added a discussion to the discussion section: The reader should bear in mind that this additional uncertainty was not considered in this work's model simulations for paused and delayed charging.Nevertheless, it should be realized that this real-world challenge may be largely mitigated by actively requesting user information about their charging sessions (e.g., expected departure time & charging demand), for instance through a mobile application (e.g., [37,38]), either as user-defined defaults with opt-outs or as per-session preferences.
Last, the discussion of the options for eliminating this charging problem is a bit abstract.The enforcement methods from governments/legislation, charging operators, and consumers are briefly mentioned, but details are missing.
Inspired by this comment, we have considerably expanded the discussion of the different enforcement methods that governments could use to eliminate paused and delayed charging problems.This is highlighted in the 'revised manuscript with changes highlighted'.We have added a concrete example of a policy initiative to stimulate the elimination of paused and delayed charging problems, by explaining how governments could introduce a certification scheme for this.In addition, we have further explained how the smart charging performance of EV models could be linked to the type approval system.Moreover, we have included a discussion about the preferred level (national/international) of implementation of the different proposed policies.This section now reads as follows: If the public sector, including governments and regulatory agencies, deem that the paused and delayed charging problems are too severe to be resolved without intervention, there are multiple enforcement or stimulation methods available.First, these organizations could stimulate manufacturers to comply with the existing standards that address this issue.This can be done by taking an active role in informing manufacturers about the importance of EVs being able to handle paused and delayed EV charging.Alternatively, the public sector has the option to establish an EV model certification program, where EV models that successfully completed a set of EV smart charging tests are granted a certificate, which could make the specific EV model more appealing to consumers.
The public sector can also enforce the elimination of paused and delayed charging problems by implementing regulations on this topic.This could be modelled after the type-approval system that is currently in place.Before getting access to public roads in the EU, all car models need to acquire type-approval [36], issued by a national approval authority.The typeapproval tests assess whether the car models meet EU safety rules (e.g., crash tests) and noise and emission limits.Expanding these tests to include an evaluation of the technical charging capabilities of EVs would ensure that only EV models meeting technical charging standards are permitted on the road.If the public sector prefers not to incorporate technical charging tests into the type-approval process, they could introduce legislation that makes compliance with standards that address the paused and delayed charging problems, such as ISO 15118-20, compulsory, either through self-assessment or by requiring testing and approval by a national approval authority.Ideally, all discussed enforcement or stimulation methods should be implemented at an international level, within entities like the EU, to enhance efficiency and maintain consistency in policies across different nations.
I also have a few other comments below.
1.The authors highlighted delayed charging problems in the paper.However, given Figure 1, it is a bit confusing since the "delay charging" test seems the least relevant (no pause or intermittency).I suggest using a more clear terminology.
Thank you for raising this point.The delayed charging test is very similar to the paused charging test.In both tests, a charging pause is considered.The difference lies in the moment of the charging pause.With the paused charging test, the charging is paused after a vehicle has briefly charged.With the delayed charging test, the charging is paused directly after the arrival of the EV to the charging station.This could be relevant for cases in which the EV arrives at the charging station at moments with high electricity prices or a high grid load.The difference between both tests has been further specified in the manuscript: The last set of tests analyses the EV's response to paused and delayed charging.These tests are relevant for smart charging applications that require longer periods without charging, such as smart charging to reduce charging costs with static or dynamic ToU tariffs or smart charging to mitigate grid congestion.The paused and delayed charging tests have a similar setup.The paused charging tests assess the EV's ability to properly react to the charging signal after a charging pause, which is implemented after the vehicle has been charged for a brief period.The delayed charging tests also consider a charging pause, which starts directly after the EV arrives at the charging station.Both tests are conducted with pauses of 20 minutes and 6 hours.
In addition, we would like to point out that, as mentioned in the opening of this response letter, we have replaced 'delayed charging' by 'paused and delayed charging' throughout the manuscript.
2. From the smart charging test results (Figure 2), what is the percentage of unsuccessful tests that were caused by charging currents exceeding charging signals?This unsuccessful label does not seem relevant to the necessity of minimum charging power.
The reason why the testing procedure for the smart charging tests identifies a charging test as unsuccessful if the charging current of the EV exceeds the charging signal by 0.5 amperes, is that this is a violation of the EV communication standards (IEC61851-1).These state that the charging current may never exceed the charging signal.If the charging station recognises this, it could terminate the charging session.This is because if this current overshoot is too severe or occurs too frequently, main circuit breakers might break down because of these current violations.
This problem only occurred with the intermittent and fluctuating charging tests, and did not occur with the paused and delayed charging tests.For the intermittent and fluctuating charging tests, 80% of the charging tests that failed were caused by current violations.
Inspired by this comment, we have made two alterations to the manuscript.First, we have further explained in the section named 'Results -Technical smart charging tests'.why a charging test failed if the charging current was violated by at least 0.5 amperes, and that this only applies to the fluctuating and intermittent charging tests: Dear Reviewers, Thank you for your constructive review of the first revision of our submission of our manuscript "Unlocking the full potential of smart charging: Addressing paused and delayed charging problems in electric vehicles" to "Nature Communications".We have addressed all comments provided by the reviewers while revising the manuscript.In the table below, we detail how we have dealt with the reviewers' comments in the second revision of the manuscript.
Yours sincerely, on behalf of all authors, Nico Brinkel

Reviewer #1 Response
Most comments were addressed successfully.The success rate of the fluctuating charging test equals 71% for the tested EV models.The share of tested models that can follow the intermittent charging profile is lower and equals 63%.In both cases, charging problems were observed in both PHEVs and BEVs, with the majority of failed tests attributed to violations of current limits.
When considering a charging pause of 20 minutes, the success rate for the tested EV models equals 86% and 83% for paused and delayed charging, respectively.When the pause duration is extended to 6 hours, the share of tested EV models that successfully pass the charging test reduces to 71% and 67% for paused and delayed charging, respectively.These problems manifested with both PHEVs and BEVs.4. Lastly, the methods on the technical smart charging tests have been slightly extended, by outlining that both PHEV and models were tested: A large majority of the sold EV models PHEV and BEV models) in the Netherlands have undergone the technical charging test procedure at the Testlab.This study reported the results of the technical smart charging tests that were conducted at the Testlab between 1 June 2020 and 1 January 2023.
Detailed charging test results cannot be disclosed individually due to non-disclosure agreements with EV manufacturers.
The model simulations in this work were conducted using a dataset of charging sessions occurring at public, on-street charging stations in the city of Utrecht, the Netherlands (as outlined in the 'Methods' section).These charging stations are used by both PHEVs and BEVs and hence, both PHEV and BEV charging sessions were included in the data set.Since the EV model and the type of EV was not identified in the charging session data, it was not possible to make a distinction between PHEVs and BEVs in the model simulations.Inspired by this comment, the following adjustment has been made to the section titled 'Model simulationsdata inputs & preparation': 1.All model simulations only considered charging session data from public, on-street charging stations located in the city of Utrecht, the Netherlands.These stations were accessible to both PHEVs and BEVs. 2

Reviewer #2 Response
I find the Response Letter to the Reviewers / Editor both clear and well motivated and substantiated.
I also agree with all the changes made, and likewise with what has been retained.
Accordingly, I can now endorse this manuscript for publication.

Dear reviewer,
We highly appreciate your efforts in reviewing our manuscript once more.We are also grateful for your review of our response to the comments from reviewer 3.Your positive evaluation is greatly appreciated and we acknowledge that your comments from the first revision round enhanced the quality of our work.
(Regarding Reviewer #3's remaining concerns:) Having read the latest version, I think the revised manuscript deals satisfactorily with the review requests by Reviewer #3, and I think the authors' point-by-point response can be accepted.

Figure 5 :
Figure 5: which types of charging stations, and where?
, issued by a national approval authority.The type-approval tests assess whether the car models meet EU safety rules (e.g., crash tests) and noise and emission limits.Expanding these tests to include an evaluation of the technical charging capabilities of EVs would ensure that only EV models meeting technical charging standards are permitted on the road.If the public sector prefers not to incorporate technical charging tests into the type-approval process, they could introduce legislation that makes compliance with standards that address the paused and delayed charging problems, such as ISO 15118-20, compulsory, either through self-assessment or by requiring testing and approval by a national approval authority.Ideally, all discussed enforcement or stimulation methods should be implemented at an international level, within entities like the EU, to enhance efficiency and maintain consistency in policies across different nations.The section named 'Methods -Model simulationscharging models' presents charging models in which a charging session cannot be paused and EVs should continuously be charged with the required minimum charging current of 6A.Such models have already been validated in practice: different real-world smart charging experiments do not consider charging pauses to avoid the charging problems discussed in this work.Hence, these real-world experiments considered a similar model formulation as the model formulation in this work, validating the models in a real-world environment.
It has been highlighted twice in the same section that the charging problems occurred both at PHEVs and BEVs: ) The discussion on renewable-based EV smart charging is still not deep enough.Fluctuating nature Thank you for addressing this point.We have implemented different changes to the manuscript based on this comment: and energy storage aspects should be also discussed.1.First, we have extended the introduction by introducing renewable-based EV smart charging when discussing the different applications for smart charging: Lastly, the roll-out of smart charging could accelerate the energy transition by shifting the charging demand of EVs to moments with excess renewable generation [21-23], thereby reducing the dependency on fossil-based energy resources and mitigating the intermittency challenges associated with renewable energy sources.If vehicle-togrid (V2G) functions are considered, EVs could even act as a storage medium for excess renewable energy, which can be utilized to meet the electricity demand during periods of renewable energy shortage.2. The Discussion section has been extended by further discussing how renewable-based EV smart charging could mitigate the problems associated with the intermittency of renewables: This could harm the roll-out of different smart charging applications, including renewable-based charging, in which the EV charging power depends on the output of a PV system or wind turbine.Implementing renewable-based charging systems has the potential to boost the selfconsumption of renewable energy and enhance the integration of renewable energy technologies into the grid [21-23].This approach helps mitigate the intermittency of renewable energy generation and reduces dependence on fossil fuels to fulfil electricity demand.For this reason, policy addressing technical charging problems for EV smart charging should also encompass the resolution of technical charging problems related to fluctuating or intermittent charging signals.