Coordination of Multiple BESS Units in a Low-Voltage Distribution Network Using Leader–Follower and Leaderless Control

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

In this article, the authors propose two distributed control methods for the coordination of several Battery Energy-Storage Systems (BESS) in order to keep voltage within statutory limits on a low-voltage (LV) distribution feeder (the 18-bus CIGRE benchmark). For the leader-follower approach, the first BESS that recognizes a voltage violation is the leader; its utilization factor is iteratively updated and broadcast so that all other units (the followers) follow. Meanwhile, for the leader-less approach, each BESS calculates its own utilization factor from its local voltage measurement, shares that value only with immediate electrical neighbors, and comes to consensus via a gossip-style update. In addition, time-domain simulations for scenarios including over-/under-voltage, unequal states of charge (SoC) and unit outages demonstrate both approaches restore voltages rapidly, but the leader-less method has more uniform SoC usage and eliminates the transient non-compliance seen when using a single leader.

In general, this paper present some promising approaches to address the mismatch issue of generation and consumption in low-voltage distribution network. On the other hand, the manuscript requires further revision and editing, especially in terms of technical clarity, presentation quality, and applications to real-world scenarios. Therefore, I have the following questions and suggestions for the authors:

1. In equation (3), the authors stated that at t → ∞, the value of xi(t) = x* = c. However, since the definition and expression of x(t) are not given, it is not clear to understand such statement. Please consider adding some additional explanation of this equation.
2. What are the considerations for the placing/layout of PV units and BESS in the CIGRE LV Test Distribution Network shown in Figure 1?
3. In Figure 3, if the leader BESS or its communication link fails, then the coordination strategy can also fail, as discussed later in the results. Thus, it would be better to add a failure-mode scenario (leader loss, delayed replacement) and show recovery dynamics; consider a rotating-leader or backup mechanism.
4. Can the authors provide further explanation of the voltage magnitude of the buses with no control implemented shown in Figure 4? How were the results obtained and what test conditions were defined?
5. From the results of leader-follower coordination strategy shown in Figure 5 and Figure 6, no utilization factors are shown during the hours from 0 to 13 when the bus voltages are below 1 p.u. However, the utilization factors do become negative for a short duration around 22 hours when the bus voltages go below 1 p.u. Can the authors explain the underlying cause of this behaviour?
6. How were the SoC profiles in Figure 7–8 obtained? Please state whether SoC is model-based (Coulomb counting, Kalman filter) or ideal. If it is model-based, then it is better to provide details such as sampling rate and error bounds.
7. Please correct the wrong figure label (order) for Figure 10 and Figure 15. Also, please improve the image quality of the results by using Figure 17 as the good example.
8. Since the presented results are limited to one feeder and a single daytime PV profile, it is unclear how good the proposed coordination control strategies applied to different scenarios. Thus, if possible, it would be beneficial to extend the study to include, for example, Monte-Carlo runs covering seasonal PV variation, load uncertainty, and feeders of different sizes and impedances.

I believe addressing these points will greatly strengthen the manuscript’s technical rigour and practical relevance.

Author Response

Response to the Reviewers Comments

Reviewer 1:

Thank you for providing very valuable feedback. Please find next the details of how we considered it in our paper to enhance its quality and value:

 

Point 1: In equation (3), the authors stated that at t → ∞, the value of xi(t) = x* = c. However, since the definition and expression of x(t) are not given, it is not clear to understand such statement. Please consider adding some additional explanation of this equation.

Response 1: Thanks for your suggestion. We have added an explanation to the equations.

 

Point 2: What are the considerations for the placing/layout of PV units and BESS in the CIGRE LV Test Distribution Network shown in Figure 1?

Response 2: Thanks for your comment. The placement of PV and BESS follows the placement of loads of the 18-node low-voltage CIGRE test layout. We do so to mimic PC and BESS placement for a residential building; however, the placement of units can be arbitrary.

 

Point 3: In Figure 3, if the leader BESS or its communication link fails, then the coordination strategy can also fail, as discussed later in the results. Thus, it would be better to add a failure-mode scenario (leader loss, delayed replacement) and show recovery dynamics; consider a rotating-leader or backup mechanism.

Response 3: Thanks for your suggestion. Indeed, a study on possible failures such as communication loss, disconnection for system failure or maintenance, to mention some, would better demonstrate how robust the control strategies are. From a theoretical implementation perspective, the adjacency matrix would be updated based on the available communication with the agents. Therefore, if one suddenly becomes unavailable, the adjacency matrix would be updated and the system would see such node as a load and not an agent, and react accordingly. Nevertheless, to demonstrate such behaviour, a reliability analysis must be performed, which exceeds the scope of this manuscript, but will be considered for our future works.

 

Point 4: Can the authors provide further explanation of the voltage magnitude of the buses with no control implemented shown in Figure 4? How were the results obtained and what test conditions were defined?

Response 4: Thanks for your observation. We have added a net load curve to demonstrate base case. The voltage profile is obtained from the load flow calculation of the CIGRE LV network, based on the load shown.

 

Point 5: From the results of leader-follower coordination strategy shown in Figure 5 and Figure 6, no utilization factors are shown during the hours from 0 to 13 when the bus voltages are below 1 p.u. However, the utilization factors do become negative for a short duration around 22 hours when the bus voltages go below 1 p.u. Can the authors explain the underlying cause of this behaviour?

Response 5: Thanks for your question. Our proposed voltage control only activates when voltage exceeds the 1 ±0.5 p.u limit (following the standard EN 50160). Although bus voltages are below 1 p.u. from 0 to 13 hours, they are not below 0.95. Hence there is no need for control during this period.

 

Point 6: How were the SoC profiles in Figure 7–8 obtained? Please state whether SoC is model-based (Coulomb counting, Kalman filter) or ideal. If it is model-based, then it is better to provide details such as sampling rate and error bounds.

Response 6: Thanks for your suggestion. We used the battery model block available in Simulink for our simulation, using a discrete simulation with a sample rate of 1 s per simulation step. According to the documentation, SOC is estimated using current integration with a maximum error of 5% compared to experimental validations between 10% and 100% SOC, 0-2C charging, and 0-5C discharging (our operating region). Details regarding this model can be found in (https://nl.mathworks.com/help/sps/powersys/ref/battery.html). We have indicated these details in the manuscript.

 

Point 7: Please correct the wrong figure label (order) for Figure 10 and Figure 15. Also, please improve the image quality of the results by using Figure 17 as the good example.

Response 7: Thanks for the observation. We have revised the figures of our manuscript to improve their clarity, ensuring their labels are consistent.

 

Point 8: Since the presented results are limited to one feeder and a single daytime PV profile, it is unclear how good the proposed coordination control strategies applied to different scenarios. Thus, if possible, it would be beneficial to extend the study to include, for example, Monte-Carlo runs covering seasonal PV variation, load uncertainty, and feeders of different sizes and impedances.

Response 8: Thanks for the suggestion. Our scope for this work was limited to demonstrate the operation of the two proposed methods, providing a comparative analysis between them. In future works, we will extend on the stochasticity of the external variables and adoption scenarios, as such comprehensive study would require a complete separate analysis.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The work provides a framework for studying the contribution towards voltage balance of BESS in LV distribution networks. The chosen approach is based on consensus algorithm using leader and leaderless variants.

Introduction objectives and contributions are clear, however the motivation could be improved. Authors state that most literature is focused on medium to high-voltage networks but I beg to differ, as there are many works related to maintaining voltage balance in LV distribution networks, such as those pertaining to building microgrids http://dx.doi.org/10.1109/ACCESS.2025.3553350. 

I suggest that motivation be improved and clarified towards showing the importance of using BESS towards voltage regulation in LV grids, such as buildings, energy communities and microgrid systems.

There's a typo in line 63.

The used methods are well introduced and mathematically sound.
It would help to add the mathematical formula defining the utilization factor in section 2, since it is used there but not previously defined, only mentioned.

Results are sound but they are a bit extensive and the images are too detached and far away from when they are mentioned from the text which difficults interpretation. The paper would benefit if authors could concise some of the text and aproximate the figures to the paragraphs where they are mentioned.

Figure 15 is out of place or wrongly attributed.

Conclusions are also too short and could be extended while result discussion reduced.

Q1: Would this system remain effective in an off-grid scenario where voltage imbalance can be more drastic? It would be interesting to explore this scenario in microgrids installed in LV-networks.

Q2: This system only considers the BESS using power towards voltage regulation (if voltage is within limits, the utilization factor is zero). Altought this is understandable under the guise of scientific proof of the concept, it is not a realistic situation. How would results here be impacted if there was a normal utilization of a BESS system where it is used for, say, peak shaving/battery arbitrage and not all capacity can be used for voltage regulation?

Author Response

Response to the Reviewers Comments

Reviewer 2:

Thank you for providing very valuable feedback. Please find next the details of how we considered it in our paper to enhance its quality and value:

 

Point 1: Introduction objectives and contributions are clear, however the motivation could be improved. Authors state that most literature is focused on medium to high-voltage networks but I beg to differ, as there are many works related to maintaining voltage balance in LV distribution networks, such as those pertaining to building microgrids http://dx.doi.org/10.1109/ACCESS.2025.3553350. I suggest that motivation be improved and clarified towards showing the importance of using BESS towards voltage regulation in LV grids, such as buildings, energy communities and microgrid systems.

Response 1: Thanks for your comment and valuable reference. Indeed, there are some studies that focus in voltage support for LV networks as well, as we have mentioned in our literature review. However, congestion problems in LV networks are considerable more recent than in MV or HV network (due to the energy transition in the last decade); thus, the literature on how to address voltage challenges in distribution and transmission networks in higher voltages is more abundant and mature than on LV voltage distribution networks, particularly through aggregation strategies, which is the focus of our work. Nevertheless, we have adapted our motivation and provided a correlation between the research gaps and the works surveyed in the literature review.

 

Point 2: There's a typo in line 63.

Response 2: Thanks for bringing this to our attention. We have corrected it.

 

Point 3: The used methods are well introduced and mathematically sound. It would help to add the mathematical formula defining the utilization factor in section 2, since it is used there but not previously defined, only mentioned.

Response 3: Thanks for your observation. The utilization factor depends on the control strategy followed, therefore, it cannot be described as a general term. However, we provided a reference to the sections where the utilization factor is calculated.

 

Point 4: Results are sound but they are a bit extensive and the images are too detached and far away from when they are mentioned from the text which difficults interpretation. The paper would benefit if authors could concise some of the text and aproximate the figures to the paragraphs where they are mentioned.

Response 4: Thanks for your suggestion. We have rearranged the manuscript so that the figures are as close as possible to their first mention. Please consider that, if accepted for publication, the editorial team from MDPI does a format revision, optimizing the readability in terms of figures and tables location.

 

Point 5: Figure 15 is out of place or wrongly attributed.

Response 5: Thanks for bringing this to our attention. We have corrected the figure’s placement.

 

Point 6: Conclusions are also too short and could be extended while result discussion reduced.

Response 6: Thanks for your observation. We have adapted our conclusions, so the similitudes and differences between the strategies are clear.

 

Point 7: Would this system remain effective in an off-grid scenario where voltage imbalance can be more drastic? It would be interesting to explore this scenario in microgrids installed in LV-networks.

Response 7: Thanks for your question. From a theoretical point of view, the method would work in an off-grid system as well, but it would control only the storage assets, not the generation. However, since off-grid systems tend to be less robust than grid-connected networks in general, a separate analysis must be done to confirm the hypothesis.

 

Point 8: This system only considers the BESS using power towards voltage regulation (if voltage is within limits, the utilization factor is zero). Altought this is understandable under the guise of scientific proof of the concept, it is not a realistic situation. How would results here be impacted if there was a normal utilization of a BESS system where it is used for, say, peak shaving/battery arbitrage and not all capacity can be used for voltage regulation?

Response 8: Thanks for your question. In principle, additional operation of the battery would not affect neither of the control strategies, as they adjust to the current state-of-charge of each agent. If at some point the SoC falls outside the requirements of the system (either too low to discharge, or to high to charge), the agent would become unavailable and the control would ignore it, and continue only with the available agents. Nevertheless, this brings two possibilities, considering that there would be two separate control mechanisms: the local controller of the asset (for peak-shaving or energy arbitrage), and the aggregated controller (for grid support). Either the storage assets must obey the signals from the aggregator for grid support, or they could decide whether to follow or not the aggregator. In such cases, a full separate analysis must be conducted, which follows outside of the scope of this work; however, in another paper, we have discussed about this possibility (https://doi.org/10.1016/j.est.2025.117507).

 

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

COMMENTS TO AUTHORS:

In this work, the authors presented two approaches for coordinated voltage control using the consensus algorithm in the 18-node CIGRE low-voltage distribution network and demonstrated the voltage control capable of both strategies. This work is interesting and carries substantial importance in mitigating global energy and carbon-neutral issues. I recommend the acceptance of this manuscript after with some revisions. The detailed suggestions can be checked as follows:

1. On page 2, the authors did a very good literature review, and many works have been compared. I think if the author could summarize the features, difference or drawbacks in a table or use this information to provide a figure could improve interpretability for readers.
2. On page 3, the “1.2. Contributions” part, the authors identified some research gaps according to literature review. It might be better to list some literature at the end of related point.
3. The authors mentioned “One highlight is that the power profiles for all batteries are different when the utilization factors are uniform. In contrast, the power profiles are uniform when the batteries have different utilization factors. This behaviour is due to the variation in the batteries’ nominal power. As indicated earlier in (10), the reference charging/discharging power is determined by the utilization factor and nominal power; therefore, uniform utilization factors do not guarantee similar reference powers.” It might be better if the author could provide some detailed discussion on relevant considerations or proposed solutions.
4. In the conclusion section, the author should offer a comprehensive summary of two discussed approaches and a big picture of them.
5. It might be better if the author could to improve the clarity of figures.

Comments for author File: Comments.pdf

Author Response

Response to the Reviewers Comments

Reviewer 3:

Thank you for providing very valuable feedback. Please find next the details of how we considered it in our paper to enhance its quality and value:

 

Point 1: On page 2, the authors did a very good literature review, and many works have been compared. I think if the author could summarize the features, difference or drawbacks in a table or use this information to provide a figure could improve interpretability for readers.

Response 1: Thanks for your suggestion. We have added a table to summarize the main findings of our literature review, associating the sources to the identified research gaps.

 

Point 2: On page 3, the “1.2. Contributions” part, the authors identified some research gaps according to literature review. It might be better to list some literature at the end of related point.

Response 2: Thanks for the recommendation. We have provided explicitly the sources associated with the identified research gaps.

 

Point 3: The authors mentioned “One highlight is that the power profiles for all batteries are different when the utilization factors are uniform. In contrast, the power profiles are uniform when the batteries have different utilization factors. This behaviour is due to the variation in the batteries’ nominal power. As indicated earlier in (10), the reference charging/discharging power is determined by the utilization factor and nominal power; therefore, uniform utilization factors do not guarantee similar reference powers.” It might be better if the author could provide some detailed discussion on relevant considerations or proposed solutions.

Response 3: Thanks for your observation. The purpose of having a utilization factor defined by the ratio of a battery's instantaneous power to its nominal power is to ensure all batteries within the control coverage contribute according to their size. By doing so, we aim for all the batteries to have a participation proportional to their capacity, balancing their degradations rate. This is because the main degradation mechanisms found in the literature are associated with cycling (related with the depth of discharge) and calendar (elapsed time), hence, similar behaviours would result in similar degradation. We have indicated these details in the manuscript.

 

Point 4: In the conclusion section, the author should offer a comprehensive summary of two discussed approaches and a big picture of them.

Response 4: Thanks for your suggestion. We have adapted our conclusions, so the similitudes and differences between the strategies are clear.

 

Point 5: It might be better if the author could to improve the clarity of figures.

Response 5: Thanks for the observation. We have revised the figures of our manuscript to improve their clarity.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

I appreciate the authors' time and efforts on revising the initial manuscript based on my suggestions and recommendations. Although some further improvements can be made to enhance the overall technical strength and applications in real-world scenarios, the revised paper is generally well-organized with detailed discussions of the methodologies and results. One suggestion I have is to make the traces in the results thicker so that they are clearly visible with normal view in pdf reader, thank you!

Comments on the Quality of English Language

The quality of English language is generally fine with minor editing and grammatical checks recommended.

Author Response

Response to the Reviewers Comments

Reviewer 1:

Thank you for providing very valuable feedback. Please find next the details of how we considered it in our paper to enhance its quality and value:

 

Point 1: I appreciate the authors' time and efforts on revising the initial manuscript based on my suggestions and recommendations. Although some further improvements can be made to enhance the overall technical strength and applications in real-world scenarios, the revised paper is generally well-organized with detailed discussions of the methodologies and results. One suggestion I have is to make the traces in the results thicker so that they are clearly visible with normal view in pdf reader, thank you!

Response 1: Thanks for your observation. We have improved the readability of some of the figures.

Author Response File: Author Response.pdf