Research & Papers

A Convexified Eulerian Framework for Scalable Coordination of Massive DER Populations

Scalable coordination of millions of storage devices via PDE and convexified LP.

Deep Dive

A team of researchers has developed a novel framework for coordinating massive numbers of storage-like distributed energy resources (DERs), such as batteries and electric vehicles, at scale while preserving individual privacy. The framework uses a two-layer architecture: a macroscopic layer that models the DER population as a continuum using an Eulerian perspective, governed by a partial differential equation (PDE). This approach makes computational complexity independent of the population size, a key advantage for managing millions of devices.

The researchers address the bilinear non-convexity in the PDE-constrained optimization problem through a convexification method that combines finite-volume discretization with a flux-lifting technique, reformulating it into a sparse linear program (LP). The LP solution yields a unified, state-dependent broadcast signal for coordination. Additionally, a Wasserstein-based relaxation replaces rigid cyclic constraints for better economic performance. At the microscopic layer, individual resources autonomously recover local setpoints from the broadcast signal, and an upstream data-mixing protocol aggregates states into a density histogram without exposing raw individual data to the aggregator.

Key Points
  • Computational complexity is independent of DER population size, enabling scaling to millions of devices.
  • Convexification method uses finite-volume discretization and flux-lifting to reformulate the problem into a sparse linear program.
  • Privacy is protected via a data-mixing protocol that aggregates states without revealing individual data to the aggregator.

Why It Matters

Enables scalable, privacy-preserving coordination of millions of DERs, crucial for grid stability and renewable integration.