Jd/pnm pf update by jd-lara · Pull Request #1619 · Sienna-Platform/PowerSimulations.jl

jd-lara · 2026-05-30T00:26:38Z

This PR updates the logic and the versions for PNM and PFs

Copilot

Pull request overview

This PR updates PowerNetworkMatrices/PowerFlows compatibility and adapts network reduction construction to the PNM 0.22 API, where irreducible buses are supplied to matrix constructors rather than reduction specs.

Changes:

Updates dependency bounds for PowerFlows and PowerNetworkMatrices.
Refactors PTDF/MODF construction and reconciliation into helper functions.
Threads irreducible_buses through Ybus, VirtualPTDF, and VirtualMODF constructors.

Reviewed changes

Copilot reviewed 1 out of 2 changed files in this pull request and generated no comments.

File	Description
`Project.toml`	Bumps `PowerFlows` and `PowerNetworkMatrices` compatibility versions.
`src/core/network_model.jl`	Updates matrix reduction construction and PTDF/MODF consistency handling for PNM 0.22.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

github-actions · 2026-05-30T01:25:56Z

Performance Results

Version	Precompile Time
Main	4.015227968
This Branch	3.941622487

Version	Build Time
Main-Build Time Precompile	128.344117189
Main-Build Time Postcompile	1.842040759
This Branch-Build Time Precompile	133.035351051
This Branch-Build Time Postcompile	1.832313978

Version	Build Time
Main-Solve Time Precompile	349.382837946
Main-Solve Time Postcompile	305.900130706
This Branch-Solve Time Precompile	3623.732303072
This Branch-Solve Time Postcompile	3566.697487472

…n model construction DecisionModel deepcopied the caller's template, which bit-copied the raw libklu factorization Ptr held by a user-provided VirtualPTDF/VirtualMODF. The copy aliased the original's handle while the cleanup finalizer stayed attached only to the original; when GC reclaimed the original template during build!, the finalizer freed the handle out from under the copy. The next klu_solve then read freed memory (Numeric.n == 0) -> KLU_INVALID -> SIGSEGV. Linux crashed deterministically; macOS survived on heap-layout luck. Exposed by PNM 0.22, whose Virtual matrices carry the KLU factorization through a finalizer-guarded Ptr. Add _copy_template_for_build: a deepcopy that shares PTDF_matrix/MODF_matrix by reference (by seeding the deepcopy memo) so the factorization handle is never copied, while still deep-copying the device/branch/service models and network reduction so the model can mutate them independently. Use it in DecisionModel (replacing the unsafe deepcopy) and EmulationModel (which previously mutated the caller's template in place — the same contract violation). https://claude.ai/code/session_01KNFhS4Z85haH3HZFT8VJC4

A failure on one platform (currently macOS, which runs Julia x64 under Rosetta on the arm64 runner) was cancelling the Linux and Windows jobs before they could report. Disabling fail-fast gives a clean per-platform verdict so the Linux/KLU result is visible while the macOS failure is investigated separately. https://claude.ai/code/session_01KNFhS4Z85haH3HZFT8VJC4

…Simulations.jl into jd/pnm_pf_update

Copilot

Pull request overview

Copilot reviewed 7 out of 8 changed files in this pull request and generated 1 comment.

Copilot

Copilot encountered an error and was unable to review this pull request. You can try again by re-requesting a review.

Copilot

Pull request overview

Copilot reviewed 14 out of 15 changed files in this pull request and generated 3 comments.

m-bossart · 2026-06-02T16:14:54Z

+        # Both matrices must share one reduction so the nodal-balance rows and the
+        # post-contingency MODF columns line up; reconcile (and fail fast) before
+        # outage consolidation populates the branch maps.
+        _ensure_ptdf_modf_consistent!(model, sys)
        _consolidate_device_model_outages_with_modf!(branch_models, get_MODF_matrix(model))


This is correct. I will push a different fix for this.

    for (device_type, devices) in contributing_devices_map
-        device_model = get(devices_template, Symbol(device_type), nothing)
+        device_model = get(devices_template, nameof(device_type), nothing)
        isnothing(device_model) && continue


+                "An interface is specified with only part of a double-circuit that has been reduced.
                Branches: $(branch_names[in_interface]) are in the interface and branches: $(branch_names[.!in_interface]) are not.
                Modify the data to include all of or none of the parallel segements.",
            ),


m-bossart · 2026-06-02T15:49:58Z

    name: Julia ${{ matrix.julia-version }} - ${{ matrix.os }}
    runs-on: ${{ matrix.os }}
    strategy:
+      fail-fast: false


Just confirming, is this meant to be committed or just for debugging?

Yeah, I want to keep it because if it fails in one OS throws all the CI out

m-bossart · 2026-06-02T16:14:54Z

+        # Both matrices must share one reduction so the nodal-balance rows and the
+        # post-contingency MODF columns line up; reconcile (and fail fast) before
+        # outage consolidation populates the branch maps.
+        _ensure_ptdf_modf_consistent!(model, sys)
        _consolidate_device_model_outages_with_modf!(branch_models, get_MODF_matrix(model))


This is correct. I will push a different fix for this.

Copilot

Pull request overview

Copilot reviewed 14 out of 15 changed files in this pull request and generated 2 comments.

m-bossart · 2026-06-04T15:38:47Z

+function _set_subnetworks_from_ptdf!(
+    model::NetworkModel{<:AbstractPTDFModel},
+    sys::PSY.System,
+)
    model.subnetworks = _make_subnetworks_from_subnetwork_axes(model.PTDF_matrix)
    if length(model.subnetworks) > 1
        @debug "System Contains Multiple Subnetworks. Assigning buses to subnetworks."
        _assign_subnetworks_to_buses(model, sys)
    end


I think this is right, we should only overwrite the subnetworks if the user input is empty.

users should never provide subnetworks here. It always needs to be derived from the data

fixed via 6995326

jd-lara added 2 commits May 29, 2026 18:24

update dependency

5cd2596

update the reduction management logic

34479fa

jd-lara requested review from Copilot and m-bossart May 30, 2026 00:26

Copilot started reviewing on behalf of jd-lara May 30, 2026 00:26 View session

Copilot AI reviewed May 30, 2026

View reviewed changes

jd-lara and others added 4 commits May 31, 2026 10:08

debug CI/CD failures

e559fe0

add more instrumentation to debug

eeb1be9

add more debug

f642037

jd-lara force-pushed the jd/pnm_pf_update branch from f642037 to 88b2806 Compare May 31, 2026 10:47

claude and others added 5 commits May 31, 2026 11:21

Merge branch 'jd/pnm_pf_update' of https://github.com/NREL-SIIP/Power…

ad7e1ad

…Simulations.jl into jd/pnm_pf_update

additional debugging

8334a72

clean up

50c5556

monre Symbol Subtitutions

f7febf1

jd-lara requested a review from Copilot May 31, 2026 15:59

Copilot started reviewing on behalf of jd-lara May 31, 2026 15:59 View session

fix issue with the Symbol calls

4c6014c

jd-lara linked an issue May 31, 2026 that may be closed by this pull request

KeyError(:MonitoredLine) or KeyError(:Line) during build! due to Symbol("PowerSystems.MonitoredLine") and Symbol("PowerSystems.Line") #1615

Closed

Copilot AI reviewed May 31, 2026

View reviewed changes

Comment thread src/operation/problem_template.jl

populate branch maps before deepcopy

ab6aa09

jd-lara mentioned this pull request Jun 1, 2026

modeled_ac_branch_types can diverge from template.branches, causing KeyError on PTDF build #1621

Closed

jd-lara added 2 commits June 2, 2026 00:41

fix reg device

9947136

formattet

c5a5df4

jd-lara requested a review from Copilot June 2, 2026 00:00

Copilot started reviewing on behalf of jd-lara June 2, 2026 00:01 View session

Copilot AI reviewed Jun 2, 2026

jd-lara requested a review from Copilot June 2, 2026 15:12

Copilot started reviewing on behalf of jd-lara June 2, 2026 15:12 View session

Copilot AI reviewed Jun 2, 2026

View reviewed changes

populate maps and deepcopy after reconciliation

a70fec7

m-bossart reviewed Jun 2, 2026

View reviewed changes

add types

d531158

jd-lara requested a review from Copilot June 4, 2026 14:59

Copilot started reviewing on behalf of jd-lara June 4, 2026 14:59 View session

fix typo

24dbed4

jd-lara requested a review from m-bossart June 4, 2026 15:04

Copilot AI reviewed Jun 4, 2026

View reviewed changes

remove subnetworks kwarg

6995326

jd-lara mentioned this pull request Jun 5, 2026

KeyError(:MonitoredLine) or KeyError(:Line) during build! due to Symbol("PowerSystems.MonitoredLine") and Symbol("PowerSystems.Line") #1615

Closed

jd-lara linked an issue Jun 5, 2026 that may be closed by this pull request

modeled_ac_branch_types can diverge from template.branches, causing KeyError on PTDF build #1621

Closed

m-bossart approved these changes Jun 5, 2026

View reviewed changes

jd-lara merged commit 5ffb2a5 into main Jun 7, 2026
10 checks passed

Conversation

jd-lara commented May 30, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

github-actions Bot commented May 30, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

m-bossart Jun 2, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

m-bossart Jun 2, 2026

Choose a reason for hiding this comment

Uh oh!

jd-lara Jun 4, 2026

Choose a reason for hiding this comment

Uh oh!

m-bossart Jun 2, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

m-bossart Jun 4, 2026

Choose a reason for hiding this comment

Uh oh!

jd-lara Jun 5, 2026

Choose a reason for hiding this comment

Uh oh!

jd-lara Jun 5, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

github-actions Bot commented May 30, 2026 •

edited

Loading