Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Experimental Validation of Distributed Dispatching of Multiple Active Distribution Networks Using the ADMM

Matthieu Jacobs, Hanmin Cai, Mario Paolone

TL;DR

This work experimentally validates an ADMM-based framework for the distributed dispatch of multiple Active Distribution Networks hosting DERs. It introduces a two-stage process—day-ahead stochastic scheduling and intra-day tracking via a bilevel MPC—to coordinate independent ADNs without exposing asset details. The approach enables flexible sharing across networks, demonstrated on two LV microgrids (EPFL-DESL and Empa-NEST) under PV uncertainty, achieving accurate dispatch tracking while revealing the economic weighting of flexibility. Limitations include algorithmic convergence speed and measurement delays, with future directions toward faster data exchange and alternative aggregation schemes such as virtual power plants to provide grid services.

Abstract

This paper presents the experimental validation of a framework for the coordinated dispatch and control of multiple active distribution networks (ADNs) hosting distributed energy resource (DER). We show that the presented method, which builds further on work done in [1], effectively allows to control multiple ADNs in a distributed way to ensure they achieve a common objective without revealing information on their DERs capabilities or grid model. This experimental validation is carried out using demonstrators at the DESL of EPFL and the NEST site at Empa, both in Switzerland. The coordination of the systems to share the flexibility of their controllable assets is demonstrated through a set of 24h experiments. Finally, the limitations of the method are discussed and future extensions proposed.

Experimental Validation of Distributed Dispatching of Multiple Active Distribution Networks Using the ADMM

TL;DR

This work experimentally validates an ADMM-based framework for the distributed dispatch of multiple Active Distribution Networks hosting DERs. It introduces a two-stage process—day-ahead stochastic scheduling and intra-day tracking via a bilevel MPC—to coordinate independent ADNs without exposing asset details. The approach enables flexible sharing across networks, demonstrated on two LV microgrids (EPFL-DESL and Empa-NEST) under PV uncertainty, achieving accurate dispatch tracking while revealing the economic weighting of flexibility. Limitations include algorithmic convergence speed and measurement delays, with future directions toward faster data exchange and alternative aggregation schemes such as virtual power plants to provide grid services.

Abstract

This paper presents the experimental validation of a framework for the coordinated dispatch and control of multiple active distribution networks (ADNs) hosting distributed energy resource (DER). We show that the presented method, which builds further on work done in [1], effectively allows to control multiple ADNs in a distributed way to ensure they achieve a common objective without revealing information on their DERs capabilities or grid model. This experimental validation is carried out using demonstrators at the DESL of EPFL and the NEST site at Empa, both in Switzerland. The coordination of the systems to share the flexibility of their controllable assets is demonstrated through a set of 24h experiments. Finally, the limitations of the method are discussed and future extensions proposed.

Paper Structure

This paper contains 14 sections, 4 equations, 9 figures, 1 algorithm.

Table of Contents

  1. Introduction
  2. Problem Statement
  3. Use Cases
  4. Scheduling and Control Framework
  5. Scheduling Stage
  6. Tracking Stage
  7. Experimental Setup
  8. Information Flow
  9. Experiment
  10. Results and Discussion
  11. Dispatch Tracking
  12. Tracking error reduction and the impact of delays
  13. Flexibility Sharing
  14. Conclusions and Future Work

Figures (9)

  • Figure 1: Configuration of the EPFL-DESL LV microgrid.
  • Figure 2: Configuration of the Empa-NEST LV microgrid.
  • Figure 3: Communication architecture with information flows between the main problem and the sub problem.
  • Figure 4: Prosumption profiles of stochastic DER for the coordination of similar ADNs, shown at one minute resolution. The top plot shows the load of the EPFL-DESL microgrid, defined as a negative injection. The next two show the total PV injection at the EPFL-DESL and Empa-NEST microgrids respectively.
  • Figure 5: Dispatch tracking performance over 24 hours for the coordination case. The power exchanged by both the EPFL-DESL and the Empa-NEST systems are also shown.
  • ...and 4 more figures