Backend
Tie-Line Scheduling Using Real-Time Contraction Metrics: Hierarchical Reachable Sets for Multi-Area Inverter-Dominated Grids
arXiv cs.SY0 views
arXiv:2611.21858v1 Announce Type: new Abstract: Managing transmission congestion in multi-area power grids dominated by inverter-based resources (IBRs) requires bridging a severe timescale and modeling divide. Supervisory congestion optimizers -- such as Online Feedback Optimization (OFO) and two-tiered differentiable Model Predictive Control (MPC) -- rely on linearized, convex algebraic grid approximations that cannot guarantee dynamic stability during sub-second converter transients. Conversely, fast distributed consensus schemes like Alternating Direction Method of Multipliers (ADMM) depend on convex plant assumptions and iterative message exchanges that lag behind rapid tie-line surges. This paper replaces convex ADMM emergency consensus with a Distributed Nonlinear Separable Control Contraction Metric (SCCM) architecture spanning three nested tiers: the Device layer, the Area layer, and the Grid tie-line layer. By pairing online Prediction Error Method and Moving Horizon Estimation (PEM-MHE) parameter tracking with offline convex Linear Matrix Inequality (LMI) synthesis, the controller executes a closed-form nonlinear feedback law online without solving real-time optimization problems. We derive the Pontryagin attainable (reachable) ellipsoids for each layer, prove hierarchical tube containment via Schur-complement matrix inequalities, and project the certified Grid reachable tube through Power Transfer Distribution Factors (PTDFs) to supply guaranteed safe thermal limits to the supervisory coordinator. Validation on modified IEEE 39-bus (100% IBR, zero synchronous generators) and 3-area IEEE 118-bus systems demonstrates an 82.8% reduction in cumulative tie-line thermal overload while maintaining strict exponential contraction ($\lambda_{\max} \le -2.43$) even under a 3 s communication blackout.
Source:https://arxiv.org/abs/2611.21858PreviousMechanism Design and Active Systems Theory: 60 Years of Parallel DevelopmentNextWho Gets the Grid Capacity? Allocating Dynamic Access Among Multiple Virtual Power Plants
Related posts
Right Goal, Wrong Body: Re-grounding Frozen VLAs under Actuation Shift
0arXiv cs.RO
DAPPT: Domain-Aware Policies from Procedural Trajectories for Sim-to-Real Dexterous Tool Use
0arXiv cs.RO
Prompt Re-Specification and Target-Mask Availability in Mask-Conditioned Visual Reinforcement Learning for Moving-Platform Landing
0arXiv cs.RO
Drone Navigation for Fast, Safe Emergency Response
0arXiv cs.RO
Why Your Enterprise Needs Digital Transformation Strategy Consulting Now!
0dev.to AI
How do I get a receipt when one AI researches and another writes the final work?
0dev.to AI