Solver Overview

GDM-Flow provides four power flow solvers, each suited to different analysis needs. All operate on DistributionSystem objects from grid-data-models and share the same Y-bus construction infrastructure.

Solver Comparison¶

When to Use Each Solver¶

AC OPF¶

Use when you need accurate voltages and losses. The AC solver finds complex voltages that satisfy power balance at every node, including reactive power and $I^2R$ line losses. Best for:

• Voltage regulation studies

• Loss analysis

• Detailed power quality assessment

AC PF (Fixed-Point Iteration)¶

Use when you need classical power flow with fixed P/Q injections and a slack bus. Unlike the AC OPF which optimises voltage magnitudes within bounds, the AC PF solves the standard power-flow equations directly. Best for:

• Baseline power flow studies

• Validating AC OPF results against a traditional solver

• Steady-state analysis with known load/generation profiles

• Cases where you want exact bus voltages without optimization bounds

Features SI-unit formulation (avoids per-unit ill-conditioning across voltage levels), sparse LU factorisation of the Y-bus, and a direct initial solve ($V = Y^{-1} \cdot I$) for a physically correct warm start.

DC OPF¶

Use when you need economic dispatch with generation costs. The DC solver minimizes total generation cost subject to linearized power balance constraints. Best for:

• DER dispatch optimization (solar, battery, grid import)

• Market-clearing simulations

• Generation scheduling

Note: DC OPF uses the small-angle approximation ($\sin\Delta\theta \approx \Delta\theta$), which breaks down across center-tapped transformers where the S2 winding operates at 180° from the primary. On systems with significant split-phase residential load, DC OPF will underestimate total source power. Use AC OPF or LinDistFlow for accurate results on such systems.

LinDistFlow¶

Use when you need fast voltage drop estimates on radial feeders. LinDistFlow performs a single backward/forward sweep without iteration. Best for:

• Screening studies and quick assessments

• Large-scale parametric sweeps

• Hosting capacity analysis

Common Workflow¶

All solvers follow the same pattern:

from gdm.distribution import DistributionSystem

# 1. Load the system
system = DistributionSystem.from_json("model.json")

# 2. Run a solver (each has a *_from_components convenience wrapper)
result = solver_from_components(system, ...)

# 3. Inspect results
print(result.success)

The *_from_components wrapper functions automatically extract loads, solar, batteries, and other components from the system. For fine-grained control, use the lower-level functions that accept explicit parameter dictionaries.

Time Series Simulation¶

For models with time-varying load and solar profiles, GDM-Flow offers two simulation modes:

• QSTS is ideal for impact studies and monitoring — run any solver across thousands of timesteps

• Multi-Period OPF is ideal for dispatch scheduling and storage optimization — jointly optimizes battery SOC and generator dispatch

See the Time Series guide for details.