Crate gridoxide_core 

gridoxide’s solvers, without any file format.

• The network model: types, network (the Y-bus and island classification), topology, branch_flow, jacobian, ratings and json, gridoxide’s own network document.
• Power flow: run_power_flow, Newton-Raphson (solver) with its outer loops (outerloop) or one of the direct methods in linear, and what is built directly on it — batches (batch, bde, scenario), sensitivities (ac_sensitivity), continuation and Q-V curves (continuation, qv), HVDC (dc), node-breaker switching (switches, constrained), reactive dispatch (dispatch) and the injection Hessian OPF needs (injection_hessian).
• State estimation (se) and its measurement model.
• Short circuit (shortcircuit).

Every reader and writer of a file format lives in gridoxide, on top, as do OPF, remedial actions, dynamics and the C and Python bindings. The sparse LU is gridoxide-sparse’s.

Modules

ac_sensitivity

AC sensitivity analysis: how a converged operating point responds to a small change.

batch

Batched power flow: one topology, many scenarios, solved across all cores.

bde

Block-diagonal embedding: solve B scenarios as one sparse system.

branch_flow

Branch terminal power flows and their partial derivatives.

constrained

A switch as an exact equality constraint, rather than as a stiff branch.

continuation

Continuation power flow: how much further can this network be loaded?

dc

Real-valued DC network model and Newton-Raphson solver.

dispatch

Which machine produced what reactive power.

injection_hessian

Second derivatives of the bus power injections.

jacobian

Topology-derived Jacobian assembly: analyze the sparsity pattern once, then refill values every iteration at fixed offsets.

json

linear

Linearized power-flow methods: the two different things the literature and the reference tools both call “linear power flow”.

measurement

The measurement model state estimation is fitted against: what was observed, where, and how much it is trusted.

network

outerloop

Outer loops: the controls that sit around a Newton-Raphson solve.

qv

Q-V curves: how much reactive headroom does one bus have?

ratings

Branch operating limits — the permanent rating and its temporary overload allowances.

roots

Bracketed root finding, shared by the event locators in continuation and gridoxide’s dynamics.

scenario

What a batch varies: per-bus overrides and branch outages against a shared template, and the errors a batch reports for them.

se

State estimation: recovering the most likely grid state from noisy, redundant, partial measurements.

shortcircuit

Short-circuit calculation, IEC 60909 style.

solver

switches

Giving a retained switch a formulation.

topology

Network topology: the node/switch graph, and the reductions over it.

types

Structs

OuterLoopOutcome

The outer loops’ side of a solve, gathered so a caller need not downcast outerloop::OuterLoop trait objects to read their reports.

PowerFlowOptions

Everything crate::run_power_flow needs to know.

PowerFlowReport

Result of a full power-flow analysis: the solved (or placeholder, for unreferenced islands) buses, plus a per-connected-component breakdown of how each one was resolved — see solver::IslandReport/ solver::IslandStatus.

RemoteControlData

The regulating machines a solve may act on.

TapData

The tap tables and regulating controls a solve may act on.

Enums

PowerFlowInit

Where PowerFlowMethod::NewtonRaphson starts from. Ignored by the direct methods, which have no starting point.

PowerFlowMethod

Which method crate::run_power_flow uses.

Functions

initialize_newton

Sets Newton’s starting point in buses, as run_power_flow does before its first iteration.

run_power_flow

Runs a power flow by the method named in opts.

run_power_flow_analysis

run_power_flow_analysis_from_ybus

Runs a power flow analysis given a pre-built Y-bus matrix. Intended for the 3-phase case too, where buses is a 3N-element vector and ybus is the 3N×3N phase-domain admittance matrix from build_ybus_3ph — a physical bus’s 3 phase-rows always share the same BusType, so the island partitioning PersistentSolver::solve does internally is safe for that case too (worst case, in a network with perfectly balanced zero/positive-sequence impedances, phases may partition into separate components rather than staying grouped by physical bus; each still carries its own valid reference, so this doesn’t change correctness, only granularity).

run_power_flow_with_remote

run_power_flow with the regulating machines a remote-voltage control needs. The two-argument form is the shorter path for every caller that has none, which is every importer but CGMES.