Optimal Coordination of Local Flexibility from Electric Vehicles with Social Impact Consideration
Open accessAbstract
The integration of renewable energy sources (RES) and the convergence of transport electrification, creates a significant challenge for distribution network management e.g. voltage and frequency violations, particularly in rural and remote areas. This paper investigates how smart charging of electric vehicles (EVs) can help reduce renewable energy curtailment and alleviate stress on local distribution networks. We implement a customised AC Optimal Power Flow (AC OPF) formulation which integrates into the optimisation an indicator reflecting the social impact of flexibility from EV users, based on the analysis of historical EV charging behaviours. The contribution of EV owners to reducing wind curtailment is optimised to enhance the acceptability of flexibility procurement, as the method targets EV users whose charging habits are most likely to align with flexibility requirements. Our method integrates social, technological, and economic perspectives with optimal flexibility coordination, and utilises clustering of EVs through a k-means algorithm. To ensure scalability, we introduce a polar coordinate-based dimension reduction technique. The flexibility optimisation approach is demonstrated on the Orkney grid model, incorporating demand and wind farm generation data, as well as multi-year charging data from 106 EVs. Results indicate that, by building upon the existing habits of EV users, curtailment can be reduced by 96.7% during a typical summer week, when curtailment can be most prevalent. This research demonstrates a foundational and transferable approach which is cognisant of social-techno-economic factors towards accelerating decarbonisation and tackling the stochastic challenges of new demand and generation patterns on local distribution networks.
AI snapshot
Written from the paper’s full text.
TL;DR
The paper develops an AC OPF-based coordination framework that targets EV smart-charging flexibility at users whose historical charging habits align with periods of wind curtailment, using k-means clustering with a polar-coordinate dimension reduction for scalability. Applied to the Orkney distribution network with 80% EV adoption, it reports a 96.7% curtailment reduction over a typical summer week.
Key findings
- Curtailment was reduced by 96.7% during a typical summer week (June, low demand/high generation) when EV demand turn-up flexibility was optimised, versus a baseline of 130 MWh of curtailment for that week.
- Clustering of 106 EV users' multi-year charging data (2020–2024, over 26,000 charging sessions) yielded k = 6 optimal clusters via elbow and silhouette methods, with total intra-cluster Euclidean distance of 5.7 kW for standard k-means and 6.5 kW for the proposed polar-coordinate method.
- The polar-coordinate approach trades a small loss in clustering optimality (6.5 vs 5.7 kW) for a reduction from 1,440 dimensions to 2, and produced a more uniform distribution of EVs across clusters (the standard method left 2 clusters with only 3 EVs each).
- Flexibility participation followed user habits: Cluster 1 (night chargers) participated mainly in nighttime curtailment events, Cluster 4 (day chargers) only in daytime events; in a major ~7 p.m. event all clusters except Cluster 4 participated.
- Required flexible EV demand power was usually higher than the optimal curtailment need, attributed to network losses and voltage drops since demand is not always located near generation.
Methods
- Unsupervised k-means clustering of average daily EV charging profiles (1,440 minute-level points per user), with a proposed polar-coordinate transformation reducing each profile to a 2-dimensional vector; clusters described in Table I (e.g., night home charging, daytime charging, rarely charged).
- A customised AC OPF formulation with semidefinite programming convex relaxation and branch flow model, building on [11], with a multi-objective function minimising losses and a social impact coefficient π_Ci,t = 1/(P_avg_Ci(t) + ε), where social impact is inversely proportional to a cluster's tendency to charge at time t.
- Case study on the Orkney distribution network (43 33 kV buses and 30 11 kV buses; 73-bus network), using DSO portal demand and wind generation data, with an assumed 80% EV adoption rate (over 13,000 EVs) distributed per current electric demand; simulation of the first day required one hour on an i7 @2.5GHz with 32GB RAM.
Limitations
- The paper states future work will incorporate market mechanisms to financially compensate EV users for flexibility, and will extend the framework to V2G technologies and broader grid constraints using connection-time information beyond active charging periods.
- Evident from the text: the polar-coordinate method is explicitly "not intended to improve clustering accuracy" and yields a higher distance sum (6.5 vs 5.7 kW) than standard k-means; results are demonstrated on a single typical summer week in Orkney, and the 80% EV adoption scenario is an assumption rather than an observed fleet; the optimisation focuses only on demand turn-up (ΔP+), not downward flexibility.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
3 of 7 cited sources checked · 2 supported · 0 flagged · accuracy not yet measured
- Premisecitation§ II. EV CHARGING BEHAVIOUR CLUSTERING · p. 2
Residential users are less inclined to provide flexibility through their EVs compared to other appliances, due to concerns that their vehicles may not be sufficiently charged when needed.
“However, studies have shown that residential users are less inclined to provide flexibility through their EVs compared to other appliances, due to concerns that their vehicles may not be sufficiently charged when needed [9].”
Quote in the paperSubstantial claim1 of 1 sources checked- cites [9] source
The Discussion (p. 12) reports lower willingness to enroll EVs than appliances and heating. It does not attribute this to fears of an insufficient charge when the vehicle is needed; it instead suggests low EV ownership and limited knowledge of EV charging as possible explanations (Discussion, p. 12; Limitations, p. 13). (LLM, conf=0.88)
Experimental — accuracy not yet measured
- Findingown data§ A. Standard k-means approach · p. 2
The clustering identified several distinct EV charging patterns, summarised in Table I.
“Based on the clustering results, we identified several distinct charging patterns, summarised in Table I.”
Quote in the paperNot a substantial claim - Findingown data§ III. OPTIMAL COORDINATION OF EV FLEXIBILITY FOR GRID MANAGEMENT · p. 4
For a typical week in June, the overall curtailment in the Orkney network reaches 130 MWh, as simulated with data from the DSO portal.
“For the considered week, the overall curtailment reaches 130 MWh, as simulated with data from the DSO portal”
Quote in the paperSubstantial claim - Findingown data§ III. OPTIMAL COORDINATION OF EV FLEXIBILITY FOR GRID MANAGEMENT · p. 5
In larger wind curtailment events requiring extensive flexibility, all clusters except Cluster 4 participated in flexibility provision.
“However, in cases of larger wind curtailment events, which require extensive flexibility, all clusters participate in flexibility provision, as shown in Fig. 5 for a major event occurring around 7 p.m., where all clusters except Cluster 4 (whose users never charge in the evening) were engaged.”
Quote in the paperNot a substantial claim - Findingown data§ III. OPTIMAL COORDINATION OF EV FLEXIBILITY FOR GRID MANAGEMENT · p. 4
For the considered week, the voltage throughout the network is regulated to 1.1 p.u., as shown for Bus 58.
“and leads to a regulation of the voltage throughout the network to 1.1 p.u., as shown in Fig. 4 for Bus 58”
Quote in the paperNot a substantial claim
- Findingown data§ Optimal Coordination of Local Flexibility from Electric Vehicles with Social Impact Consideration · p. 1
With EV flexibility and an assumed 80% EV adoption rate, curtailment was reduced by 96.7% on the considered typical summer week.
“Results indicate that, by building upon the existing habits of EV users, curtailment can be reduced by 96.7% during a typical summer week, when curtailment can be most prevalent.”
- Also stated in: § III. OPTIMAL COORDINATION OF EV FLEXIBILITY FOR GRID MANAGEMENT · p. 5 · boosted
“The results demonstrate that curtailment was reduced by 96.7% on the considered week thanks to demand turn-up flexibility provided by EVs that were not already charging”
- Also stated in: § IV. CONCLUSION & FUTURE WORK · p. 5 · boosted
“During a typical summer week with high curtailment (high generation and low demand), the proposed approach with 80% electrification of transport demonstrated a curtailment reduction of 96.7%”
Quote in the paperSubstantial claim- The authors state this as an assumption: EVs that are not already charging are assumed to be available for flexibility due to the lower effort required.
- Also stated in: § III. OPTIMAL COORDINATION OF EV FLEXIBILITY FOR GRID MANAGEMENT · p. 5 · boosted
- Premisecitation§ I. INTRODUCTION · p. 1
The ANM system operated by the distribution system operator in Orkney enforces curtailment of the most recently connected wind turbines.
“As a result, the ANM system operated by the distribution system operator (DSO) enforces curtailment of the most recently connected wind turbines [6].”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites [6] source
- Interpretationreasoning§ IV. CONCLUSION & FUTURE WORK · p. 5
By incorporating information on EV users' historical charging habits into the optimisation objective, the approach prioritises users who are more likely to have their EVs connected when support is needed.
“Moreover, by incorporating information on EV users' historical charging habits into the optimisation objective, the approach prioritises users who are more likely to have their EVs connected when support is needed.”
Quote in the paperSubstantial claim- The authors state this as an assumption: The social impact of requesting an EV user in cluster C_i to increase consumption at time t is inversely proportional to that cluster's tendency to charge at time t, as expressed by its centroid.
- Interpretationown data§ IV. CONCLUSION & FUTURE WORK · p. 5
The proposed polar-coordinate dimension reduction technique allows the method to scale effectively for large datasets containing thousands of EVs.
“A novel and improved dimensionality reduction technique based on polar coordinates was developed allowing the method to scale effectively for large datasets containing thousands of EVs.”
Quote in the paperSubstantial claim - Interpretationnone given§ IV. CONCLUSION & FUTURE WORK · p. 5
The proposed formulation ensures that flexibility is only requested from EVs located in grid areas where their contribution is technically effective.
“The proposed formulation ensures that flexibility is only requested from EVs located in grid areas where their contribution is technically effective.”
Quote in the paperSubstantial claim
- May combine two claims. 9
- Classified as probably not a substantial claim. 2
May combine two claims.
Premisecitation§ I. INTRODUCTION · p. 1EVs offer potential as residential flexible loads that could provide demand response and mitigate renewable generation curtailment.
“However, EVs also offer potential as residential flexible loads that could provide demand response and mitigate renewable generation curtailment [8].”
Quote in the paperSubstantial claim1 of 1 sources checked- cites [8] source
Classified as probably not a substantial claim. May combine two claims.
Premisecitation§ I. INTRODUCTION · p. 1Electrification of transport is driving an increase in the number of electric vehicles, which significantly raises electricity demand and introduces technical challenges including voltage imbalances, harmonic distortion, transformer and line overloading, and other operational and power quality issues.
“Simultaneously, the electrification of transport is driving an increase in the number of electric vehicles (EVs), which significantly raises electricity demand and introduces further technical challenges, including voltage imbalances, harmonic distortion, transformer and line overloading, and other operational and power quality issues [7].”
Quote in the paperSubstantial claim1 of 1 sources checked- cites [7] source
May combine two claims.
Findingown data§ III. OPTIMAL COORDINATION OF EV FLEXIBILITY FOR GRID MANAGEMENT · p. 5The optimisation requested flexibility from EV clusters in alignment with their typical charging times, e.g., Cluster 1 participated predominantly during nighttime curtailment events and Cluster 4 was only active during daytime curtailment events.
“For instance, Cluster 1, i.e. users who primarily charge their EVs at night, participate predominantly during nighttime curtailment events. In contrast, Cluster 4, which includes users who typically charge during the day, was only active during daytime curtailment events.”
Quote in the paperSubstantial claim- The authors state this as an assumption: The social impact of requesting an EV user in cluster C_i to increase consumption at time t is inversely proportional to that cluster's tendency to charge at time t, as expressed by its centroid.
May combine two claims.
Findingown data§ B. Polar coordinates based approach · p. 3The polar-coordinate k-means clustering achieved a sum of Euclidean distances J = 6.5 kW, which is not as optimal as the standard method but is achieved with a significant reduction in dimensionality, thereby improving scalability.
“The sum of Euclidean distances is J = 6 . 5 kW, which, although not as optimal as the standard method, is achieved with a significant reduction in dimensionality, thereby improving scalability.”
Quote in the paperSubstantial claimMay combine two claims.
Findingown data§ A. Standard k-means approach · p. 2Applying standard k-means to 106 EV users identified k = 6 as the optimal number of clusters, with a total intra-cluster Euclidean distance of 5.7 kW across all users.
“Using both the elbow and silhouette methods, we identify k = 6 optimal number of clusters with a total intra-cluster Euclidean distance of 5.7 kW across all users.”
Quote in the paperNot a substantial claimMay combine two claims.
Findingown data§ B. Polar coordinates based approach · p. 3The clusters obtained using the polar-coordinate approach are comparable to those from the traditional method, except for cluster 6, with a more uniform distribution of EVs across clusters than the standard method, which led to 2 clusters with 3 EVs only.
“The clusters obtained using this approach are comparable to those from the traditional method, except for cluster 6, with the added benefit of a more uniform distribution of EVs across the clusters than what was obtained with the first method, which led to 2 clusters with 3 EVs only.”
Quote in the paperSubstantial claimMay combine two claims.
Interpretationreasoning§ III. OPTIMAL COORDINATION OF EV FLEXIBILITY FOR GRID MANAGEMENT · p. 5The required flexible demand power from EVs is usually higher than the optimal curtailment need, because the approach integrates the grid in the optimisation, leading to larger need of flexible power to overcome network losses and voltage drops.
“Another important observation is that the required flexible demand power from EVs is usually higher than the optimal curtailment need displayed in Fig. 4. This is due to the fact that our approach is comprehensive and integrates the grid in the optimisation, leading to larger need of flexible power to overcome network losses and voltage drops”
Quote in the paperNot a substantial claim- The authors state this as an assumption: The social impact of requesting an EV user in cluster C_i to increase consumption at time t is inversely proportional to that cluster's tendency to charge at time t, as expressed by its centroid.
May combine two claims.
Study design and conductnone given§ IV. CONCLUSION & FUTURE WORK · p. 5The method was applied to a real 73-bus distribution network in Orkney, using multi-year charging data from 106 EV users.
“The method was applied to a real 73-bus distribution network in Orkney, using multi-year charging data from 106 EV users.”
Quote in the paperSubstantial claimMay combine two claims.
Premisecitation§ I. INTRODUCTION · p. 1Most previous studies using clustering techniques such as k-means, decision tree-based clustering, self-organising maps, and hierarchical clustering do not explore how such clusters can be directly used by DSOs, nor address the challenge of scalability for future large-scale EV deployment.
“While techniques such as k-means, decision tree-based clustering, self-organising maps, and hierarchical clustering have been used in previous studies [2], [3], most do not explore how such clusters can be directly used by DSOs, nor address the challenge of scalability for future large-scale EV deployment.”
Quote in the paperSubstantial claim0 of 2 sources checkedClassified as probably not a substantial claim.
Premisecitation§ I. INTRODUCTION · p. 1Integration of renewable energy sources and transport electrification introduces new challenges for distribution networks, which must accommodate increasingly large and dynamic patterns of energy production and consumption.
“However, these transformations introduce new challenges for distribution networks, which must accommodate increasingly large and dynamic patterns of energy production and consumption [2].”
Quote in the paperSubstantial claim0 of 1 sources checked- cites [2] source
The paper relies on this without support.
Many of the newer wind turbines in Orkney are community-owned, which amplifies the social impact of curtailment.
“Notably, many of these newer wind turbines are community-owned, amplifying the social impact of curtailment.”
The authors state this as an assumption.
The social impact of requesting an EV user in cluster C_i to increase consumption at time t is inversely proportional to that cluster's tendency to charge at time t, as expressed by its centroid.
“We consider that the social impact for requesting an EV user in cluster C i to increase its consumption at time t , is inversely proportional to its tendency to charge at time t , as expressed by its cluster's centroid.”
The authors state this as an assumption.
EVs that are not already charging are assumed to be available for flexibility due to the lower effort required.
“thanks to demand turn-up flexibility provided by EVs that were not already charging, but were assumed to be available for flexibility due to the lower effort required.”
The paper relies on this without support.
Wind turbines in Orkney are curtailed without compensation under their current non-firm connection agreement.
“Although wind turbines in Orkney are curtailed without compensation under their current non-firm connection agreement, their curtailment cost π g is modelled as virtually higher than social impacts to prioritise EV flexibility”
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Major event described as involving both all clusters and all but one
MinorThe paper says all clusters participate in larger curtailment events and presents the 7 p.m. event as an example. It then says Cluster 4 did not participate in that event.
“in cases of larger wind curtailment events, which require extensive flexibility, all clusters participate in flexibility provision, as shown in Fig. 5 for a major event occurring around 7 p.m.”
“where all clusters except Cluster 4 (whose users never charge in the evening) were engaged.”
Cites
Ray, S., Kasturi, K., Patnaik, S., & Nayak, M. R. (2023) · Journal of Energy Storage · doi:10.1016/j.est.2023.108672 · Cited by 118 · cited 2× · see citing works
“…techniques such as k-means, decision tree-based clustering, self-organising maps, and hierarchical clustering have been used in previous studies [2], [3], most do not explore how such clusters can be directly used by grid operators, nor address the challenge of scalability for future…”
Srivastava, A., Manas, M., & Dubey, R. K. (2023) · Journal of Engineering and Applied Science · doi:10.1186/s44147-023-00193-w · Cited by 106 · cited 1× · see citing works
“…is driving an increase in the number of electric vehicles (EVs), which significantly raises electricity demand and introduces further technical challenges, including voltage imbalances, harmonic distortion, transformer and line overloading, and other operational and power quality issues [7].”
Toward residential flexibility—Consumer willingness to enroll household loads in demand response
Sridhar, A., Honkapuro, S., Ruiz, F., Stoklasa, J., Annala, S., Wolff, A., & Rautiainen, A. (2023) · Applied Energy · doi:10.1016/j.apenergy.2023.121204 · Cited by 59 · cited 1× · see citing works
“However, studies have shown that residential users are less inclined to provide flexibility through their EVs compared to other appliances, due to concerns that their vehicles may not be sufficiently charged when needed [9].”
Responsive FLEXibility: A smart local energy system
Couraud, B., Andoni, M., Robu, V., Norbu, S., Chen, S., & Flynn, D. (2023) · Renewable and Sustainable Energy Reviews · doi:10.1016/j.rser.2023.113343 · Cited by 57 · see citing works
Self-citationIntegration of tidal energy into an island energy system – A case study of Orkney islands
Almoghayer, M. A., Woolf, D. K., Kerr, S., & Davies, G. (2022) · Energy · doi:10.1016/j.energy.2021.122547 · Cited by 56 · cited 1× · see citing works
“As a result, the ANM system operated by the distribution system operator (DSO) enforces curtailment of the most recently connected wind turbines [6].”
Smart EV Charging Strategies Based on Charging Behavior
Singh, S. K., Vaidya, B., & Mouftah, H. T. (2022) · Frontiers in Energy Research · doi:10.3389/fenrg.2022.773440 · Cited by 36 · cited 1× · see citing works
“Some studies attempt to reduce the dimensionality of x j further by considering only a few features—such as the typical start time of charging sessions and the total connection duration [10].”
Prediction of voltage distribution using deep learning and identified key smart meter locations
Mokhtar, M., Robu, V., Flynn, D., et al. (2021) · Energy and AI · doi:10.1016/j.egyai.2021.100103 · Cited by 28 · cited 1× · see citing works
“In response, smart grid solutions have been adopted [3], [4], along with zonal segregation of distribution networks to mitigate constraints through Active Network Management (ANM) schemes [5].”
Self-citationHasan, K. N., Rahman, M. T., Terrill, C., McLean, R., Rodricks, R., Sharma, A., & Islam, A. (2025) · Sustainable Energy Technologies and Assessments · doi:10.1016/j.seta.2024.104141 · Cited by 10 · cited 3× · see citing works
“…such as k-means, decision tree-based clustering, self-organising maps, and hierarchical clustering have been used in previous studies [2], [3], most do not explore how such clusters can be directly used by grid operators, nor address the challenge of scalability for future large-scale EV…”
Active network management in LV networks: a case study in the UK
Perez-Olvera, J., Green, T. C., & Junyent‐Ferré, A. (2020) · IEEE Power Engineering Society General Meeting · doi:10.1109/pesgm41954.2020.9281406 · Cited by 4 · cited 1× · see citing works
“In response, smart grid solutions have been adopted [3], [4], along with zonal segregation of distribution networks to mitigate constraints through Active Network Management (ANM) schemes [5].”
An AC OPF based Clearing Mechanism for Local Flexibility Markets
Couraud, B., Norbu, S., Andoni, M., et al. (2024) · IEEE PES Innovative Smart Grid Technologies Europe · doi:10.1109/isgteurope62998.2024.10863672 · Cited by 2 · see citing works
Self-citation
Flagged first, then by citations
Research works printed but not linked: 1
- 2024, 2022 uk greenhouse gas emissions
- Authors
- Chen, S., Couraud, B., Norbu, S., Andoni, M., Iqbal, Z., Djokic, S., Kirli, D., Kanugrahan, S. P., Cherubini, P., Krumdieck, S., Robu, V., & Flynn, D.
- Affiliations
- University of Glasgow (GB)
- University of Edinburgh (GB)
- University of Edinburgh,Edinburgh,UK
- Heriot-Watt University (GB)
- Centrum Wiskunde & Informatica (NL)
- ORCID
- Venue
- 2025 IEEE PES Innovative Smart Grid Technologies Conference Europe (ISGT Europe)
- Publisher
- IEEE
- Published
- 2025-10-20
- Pages
- 1-5
- OpenAlex
- W7117530220
- Semantic Scholar
- 006cefc96df300b5b8b97ef693fc4e3fd2303556
- Cited by
- 0
- References
- 10
- FWCI
- 0.00
- Type
- conference paper
- Language
- en
- Open access
- green
- doi.org
- Locations
- IEEE PES Innovative Smart Grid Technologies Conference Europe (ISGT Europe) (publishedVersion)
- ENLIGHTEN (Jurnal Bimbingan dan Konseling Islam) (acceptedVersion, cc-by)
- Funders
- Engineering and Physical Sciences Research Council (EP/V042955/1, EP/X038823/2, EP/Y024257/1, EP/X038823)
- Topics
- Physical Sciences › Engineering › Electrical and Electronic Engineering › Electric Vehicles and Infrastructure
- Physical Sciences › Engineering › Electrical and Electronic Engineering › Smart Grid Energy Management
- Physical Sciences › Engineering › Electrical and Electronic Engineering › Optimal Power Flow Distribution
- Keywords
- electric vehicle charging, renewable energy curtailment, wind power curtailment, distribution network management, AC optimal power flow, K-means clustering, electric mobility, decarbonization
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-19.
Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-19.
Cited sources are checked by gpt-6-sol.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
