Supporting power balance in Microgrids with Uncertain Production using Electric Vehicles and Indirect Control

Problem and Research Question

Islanded microgrids with high renewable penetration must continuously balance supply and demand despite uncertain generation from wind and solar. When production fluctuates faster than conventional generation can respond, the microgrid relies on flexibility from other resources. Electric vehicles, if their charging can be coordinated, represent a large and growing pool of controllable load. The central question: how can electric vehicles be used as an indirect control resource to support power balance in microgrids with uncertain renewable production, without requiring direct, real-time command of every vehicle?

Direct control — the aggregator sending power setpoints to each vehicle — is intrusive and scales poorly. Indirect control instead uses a signal (price, incentive, or schedule) to steer aggregate charging behavior, which is more acceptable to users but must be designed so that the aggregate response reliably supports the microgrid’s balance under uncertainty.

Method

This paper studied direct and indirect control of EV charging for partially compensating an uncertain production disturbance:

• Indirect control formulation — rather than dispatching each vehicle directly, the aggregator broadcasts a control signal (e.g., a price or incentive) that shapes the aggregate charging behavior of the EV fleet. The fleet’s response is modeled and used as a flexibility resource.
• Response estimation — each EV followed a simple saturated linear price-response model whose parameters were estimated online.
• Biased forgetting — the identification scheme discarded parameter samples farthest from the current estimate to adapt the response model.
• Combined control — directly controlled EVs supplied predictable action, while a price signal steered the indirectly controlled group.

The formulation was evaluated in simulation with 30 synthetic EVs. The paper does not model rebound behavior or establish that the inferred parameter uncertainty is statistically calibrated.

Personal Contribution

Frederik Banis was the first author. He developed the indirect control framework and the EV integration formulation, designed the experiments, and wrote the manuscript. Co-authors Daniela Guericke, Henrik Madsen, and Niels Kjølstad Poulsen contributed supervision, methodological guidance, and review.

Validation and Key Results

• In the simulated example, direct control compensated part of the production disturbance and adding indirect price control supplied additional aggregate response.
• The study illustrated the interaction between online response estimation, direct dispatch, and price-mediated control.
• The evidence is qualitative and specific to the synthetic fleet and simple response model; it does not quantify reserve reduction or validate user acceptance in practice.

The paper was published in IFAC-PapersOnLine (2019), a peer-reviewed conference proceedings series (volume 52, issue 4, pages 371–376).

Related Outputs

• Conference paper: IFAC-PapersOnLine, 2019, vol. 52, no. 4, pp. 371–376. DOI: 10.1016/j.ifacol.2019.08.238
• Publication catalogue entry: data/publications.yaml, id 2.
• Related project: Target-Adjusted MPC for Microgrid Frequency Control — the broader microgrid MPC project this paper contributes to.
• Related journal paper: Load Frequency Control in Microgrids using target adjusted Model Predictive Control, IET Renewable Power Generation, 2019. DOI: 10.1049/iet-rpg.2019.0487
• PhD thesis: Efficient Operation of Energy Grids, DTU, 2020. DOI: 10.11581/DTU.00000334

Collaborators and Institutions

• DTU Compute, Technical University of Denmark — Henrik Madsen, Niels Kjølstad Poulsen (supervisors), Daniela Guericke (co-author)

Status and Next Steps

Status: Published (2019). The indirect control framework is a constituent strand of the broader microgrid MPC research program documented in the PhD thesis.

Transfer to current research: The indirect control idea — steering aggregate behavior with a signal rather than commanding each unit directly — is conceptually related to the closed-loop experimentation architecture used in current work, where a policy shapes a sequence of experimental actions rather than each action being commanded in isolation. The shift is from EV fleets to bioassay platforms, but the principle of policy-driven aggregate control under uncertainty is shared.

See the related project page for the broader research context, the Methods page for the methodological framework, and the Research page for the research program overview.