Chordal SDP#
API path: power_flow_relaxations.models.chordal_shor
- class ChordalShor(scenario, configuration, clique_reduction=0.2, **kwargs)[source]#
Bases:
NodalBaseModelChordal-decomposed Shor SDP relaxation for ACOPF.
Uses clique-wise PSD matrices from a chordal extension of the network graph, plus overlap-consistency constraints, to approximate the dense Shor SDP with improved scalability on sparse grids.
Build the chordal Shor model and clique-wise PSD variables.
Forces relaxation mode, computes a chordal extension, extracts reduced cliques, creates one PSD matrix per clique and period, and precomputes edge-to-clique mappings used by the flow constraints.
- Parameters:
clique_reduction (float)
- chordal_consistency_constraints()[source]#
Enforce agreement between overlapping clique PSD matrices.
A maximum spanning tree of the clique-intersection graph is used as the stitching structure. For each tree edge (C_i, C_j), all shared real/imag lifted entries corresponding to node pairs in C_i ∩ C_j are constrained equal for every period.
This ensures local clique matrices define a globally consistent lifted representation on overlaps.
- get_V_vt_values()[source]#
Recover approximate bus voltages from clique-based SDP matrices.
For each period, aggregates clique matrix entries into a global lifted matrix, symmetrizes and PSD-completes it, extracts a rank-1 approximation from the dominant eigenpair, rotates the result to the reference angle, and returns (V_d, V_q) per node and period.
- Return type:
dict