What question did this study set out to answer?

This research aims to develop an integrated approach for optimizing electric vehicle charging infrastructures under varying demand conditions.

June 13, 2026Open Access

An Integrated Framework for Adaptive Operational Control and Hierarchical Spatial Optimization in Multi-Tiered Electric Vehicle Charging Infrastructures

Key Points

This research aims to develop an integrated approach for optimizing electric vehicle charging infrastructures under varying demand conditions.
Utilized Analytic Hierarchy Process, K-means clustering, and GIS for macro-scale spatial optimization.
Employed Genetic Algorithms for scenario-adaptive allocation of charging stations.
Implemented agent-based simulations to model charging behaviors and operational controls.
Reduced average user waiting times from 1.54 hours to 0.65 hours.
Mitigated peak-to-valley charging load differences by 50%.
Suppressed maximum grid voltage deviations by 62%.

Abstract

Current electric vehicle (EV) public charging infrastructures encounter high spatiotemporal demand fluctuations, which are caused by users' different charging behaviours, density gradients between different regions, and grid structural constraints 1, 2. The paper introduces a multi-level integrated approach which combines the macro-scale spatial suitability optimization and micro-scale dynamic operation and power electronics control. At the macro planning layer, we integrate the Analytic Hierarchy Process (AHP), K-means clustering, and Geographic Information System (GIS) methods with scenario-adaptive Genetic Algorithms (GA) for the optimization of station allocation along the urban-rural gradient across the different policy pathways of Progressive and Thriving. To consider the reuse of infrastructure, a local model based on Mixed-Integer Linear Programming (MILP) is built, which considers the spatial visibility of the lampposts in the GIS and the capacity of the cabinets of public lighting 3. The micro-operational layer involves an agent-based simulation that realistically simulates queueing behaviours with directly enforced partial-charging rules (e.g., 70% and 80% State of Charge thresholds), while considering queue abandonment and different human patience levels. In the last, inner digital Proportional-Integral (PI) current loop with outer Model-Based Predictive Control (MBPC) voltage regulator for PV-integrated bidirectional DC-DC converters are used to guarantee dynamic stability 4 together with geometric algebra (GA)-based harmonic analysis and Reactive-Active power coordination control 5, 7 at the grid interface. The integrated architecture makes a significant contribution to the reduction of average user waiting times from 1.54 h to 0.65 h 6 and to the mitigation of peak-to-valley charging load differences by 50% 6 and the suppression of maximum grid voltage deviations by 62% 7.

Ask AI

Mark Helpful

Bookmark

Relay

View Full Paper