Research & Papers

New DLMP Bilevel Framework Optimizes Data Center-Grid Coordination

Researchers tackle phase imbalance and voltage regulation using DLMP-based optimization.

Deep Dive

A new research paper from arXiv introduces a bilevel optimization framework designed to improve coordination between data centers and distribution grids. The model uses Distribution Locational Marginal Prices (DLMPs) to align the data center aggregator's power demand decisions with the distribution system operator's network constraints. Crucially, it incorporates both active and reactive power consumption of data centers, allowing the framework to evaluate impacts on voltage regulation and phase imbalance – a key challenge in unbalanced three-phase systems. The approach is validated on the IEEE 37-bus unbalanced test system, demonstrating economically efficient operation without ignoring location- and phase-specific network conditions.

Two operating cases are analyzed: one without reactive power compensation and one with static var generator (SVG)-based compensation. Simulation results reveal that SVG-based compensation significantly improves voltage profiles and reduces phase unbalance, mitigating adverse network effects while maintaining cost efficiency. This work provides a practical path for grid operators to integrate growing data center loads without compromising power quality or reliability. The study was conducted by Arash Baharvandi and Duong Tung Nguyen, and is available on arXiv under ID 2606.26328.

Key Points
  • Bilevel model uses DLMPs to coordinate data center aggregator and distribution system operator.
  • Incorporates both active and reactive power to evaluate impacts on voltage and phase imbalance.
  • Static var generator compensation enhances voltage profiles and reduces phase unbalance in IEEE 37-bus test.

Why It Matters

Enables more efficient grid integration of data centers, reducing energy costs and improving power quality.

📬 Get the top 10 AI stories daily