The research, conducted by a team at the Electric Power Research Institute of Guangxi Power Grid in Nanning, China, tackles a problem that has grown acute as transport electrification accelerates. Large-scale electric vehicle integration can worsen nodal voltage deviations, overload lines and create local congestion in distribution networks. Yet the same flexible charging loads, if managed well, can support peak-valley regulation and help accommodate renewable energy. The challenge, the authors argue, is that most existing studies focus on individual operational goals such as voltage security, charging cost or station planning, without a unified framework for assessing how the various participants actually coordinate with one another.

To fill this gap, the team proposed representing the entire vehicle-charger-grid system as a service-oriented collaborative network. Rather than treating electric vehicles merely as adjustable loads, the method maps distribution network nodes, aggregators, charging stations and electric vehicle user groups into a multi-agent network in which participants are characterized by the coordination services they provide or receive. Each interaction between two participants is modeled as a directed edge carrying three distinct types of coupling: information, energy and value-price. The information dimension captures state sharing rates, data update frequencies, control-command arrival rates and prediction consistency. The energy dimension measures interactive power intensity, local energy matching, peak-valley complementarity and aggregation execution rates. The value-price dimension tracks how consistently participants respond to price incentives, how well benefits are shared and how reliably contracts are fulfilled.

A key technical innovation lies in how these three dimensions are weighted. The researchers use a state-driven softmax formulation in which a normalized state-feature vector reflects system security pressure, economic operation pressure and service quality pressure. Under heavy-load or abnormal conditions, security-related weights increase, emphasizing the coupling dimensions most relevant to keeping the grid stable. During normal economic operation, energy and value-price coupling receive greater emphasis. This dynamic weighting allows the assessment to shift its focus as operating conditions change, rather than applying a fixed set of priorities regardless of circumstances.

The framework also introduces a coordination-consistency correction mechanism. Even when two nodes exchange large amounts of information and power, their coordination can fail if control actions suffer from time mismatch, delays or execution deviations. The consistency coefficient penalizes response-delay deviations and execution deviations exponentially while rewarding a high action-direction consistency rate. The corrected edge-level coordination strengths are then assembled into a weighted adjacency matrix, which serves as the basis for computing node importance through weighted strength, betweenness and closeness centrality, and for deriving network-level indicators including network density, weighted clustering coefficient, network efficiency, average coordination response delay and operational risk.

Because the vehicle-charger-grid system is highly time-varying, the method applies exponential smoothing to filter short-term noise and uses exponential decay time weighting so that coordination behaviors closer to the end of an evaluation period carry more weight. The edge-level, node-level and network-layer results are aggregated into a periodic comprehensive coordination score, which is then classified into weak, medium and strong coordination levels using an adaptive quantile-driven threshold method rather than manually set empirical cutoffs.

To validate the approach, the team tested it on the IEEE 33-bus distribution system, a standard benchmark in power systems research, under three scenarios. In the non-collaborative scenario, charging stations operate independently based on local demand. The basic collaborative scenario introduces a coordination mechanism between aggregators and stations. The enhanced collaborative scenario adds the joint effects of information, energy and value-price coupling across all participants. The results were striking. The non-collaborative scenario scored 0.4355 on the comprehensive coordination index, with voltage deviation of 0.0405 and operational risk of 0.0045. The basic collaborative scenario improved the score to 0.4955 and reduced voltage deviation to 0.0280. The enhanced collaborative scenario achieved the highest score of 0.5031 while cutting voltage deviation to 0.0203 and operational risk to 5.55 times ten to the negative fifth, roughly eighty times lower than the non-collaborative case.

The physical benefits were equally clear. The evening peak-valley difference in main-grid power supply dropped from 0.2132 in the non-collaborative case to 0.1862 under enhanced coordination. Minimum voltages at weak buses rose from 0.9070 per unit to 0.9555 per unit at Bus33, providing stronger voltage support in terminal and heavily loaded areas. Maximum line loading fell from 0.4413 to 0.4090 per unit, well below the 1.0 per unit limit, indicating a better security margin. The time-weighted coordination strength matrices revealed that enhanced collaboration did not merely reinforce local links between charging stations and nearby users but significantly strengthened cross-layer interactions, with the aggregator-to-charging-station link increasing from 0.3146 to 0.3770 and the distribution-network-to-user-group link rising from 0.1731 to 0.2037.

What makes this framework particularly valuable is that it links coordination structure to physical operating performance rather than simply ranking scenarios in the abstract. The agreement between the assessment results and the observed changes in nodal voltage and line loading demonstrates that the method can explain why a coordination mode performs better and where the critical links lie. This addresses a persistent gap in the literature, where studies on coordinated charging and infrastructure planning have largely focused on peak-valley regulation, cost reduction or service capability without explicitly characterizing the multi-agent coordination structures that drive those outcomes.

The researchers suggest that their interpretable, multi-level tool can support the comparison of alternative coordination modes and provide a quantitative basis for coordination optimization, active distribution network regulation and multi-agent interaction mechanism design. As electric vehicle fleets grow and renewable generation deepens its penetration into distribution systems, the ability to measure, diagnose and improve the strength of collaboration among vehicles, chargers and the grid may prove essential to keeping power flowing reliably while the transportation sector completes its electrification transition.

Subject of Research: Coordination strength assessment for vehicle-charger-grid interaction in electric vehicle-integrated distribution networks

Article Title: A collaboration-strength-aware assessment method for vehicle-charger-grid coordination in distribution systems via multi-dimensional business coupling

Article References: Wu, N., Yu, X., Luo, Y., & Huang, S. (2026). A collaboration-strength-aware assessment method for vehicle-charger-grid coordination in distribution systems via multi-dimensional business coupling. Frontiers in Energy Research, 14, Article 1947122. https://doi.org/10.3389/fenrg.2026.1947122

Image Credits: AI Generated

DOI: 10.3389/fenrg.2026.1947122

Keywords: electric vehicles, smart charging, distribution networks, coordination assessment, vehicle-to-grid, aggregators, charging stations, power grid, voltage deviation, collaborative network, IEEE 33-bus, energy coupling