Expect DC current fields in kA

[luke-kiernan]

Warning

These tests fail until PowerOpenAPIModels is regenerated against Sienna-Platform/SiennaSchemas#94 and released. The conversion is driven by the models' declared units (IC.has_declared_unit). Merge together with the rest of the A→kA set (Sienna-Platform/PowerSystems.jl#1836).

Summary

No source changes, just tests. The parser keeps reading PSS/E's currents in A (set_value!(component, :dc_current, d["if"], "A")), and its unit-aware set_value! converts to whatever unit the schema declares. Once the schema declares kA, the parser stores kA, and the PSS/E-native dict stays in A. So we need an extra factor of 1000 in a few places.

The A→kA set

Sienna-Platform/SiennaSchemas#94 → PowerOpenAPIModels regeneration → Sienna-Platform/PowerSystems.jl#1836, this PR, and consumer updates in PowerFlows.jl, PowerOperationsModels.jl and PowerSystemCaseBuilder.jl.

🤖 Generated with Claude Code

[luke-kiernan]

Open question for review: what value means "no DC current limit"?

The stack currently uses two different "infinities" for these currents, and moving the fields from A to kA makes them diverge further:

Where	Value	Unit after this set
SiennaSchemas / PSY defaults for max_dc_current, max_dc_current_from/_to (and PSY's OpenAPI import fallbacks)	1e8	kA (was A)
PowerFlowFileParser: RAW IMAX = 0 → PTI_INF = 9999.0 (src/pm_io/psse.jl:2328)	9999.0	A, then 9.999 kA once stored
PowerFlows PSS/E export: PSSE_INFINITY = 9999.0 → IMAX = 0 (src/psse_export.jl:2)	9999.0	A

So:

1. Defaults. Should they stay 1e8 (now kA, effectively unbounded either way), or become 1e5 so they still equal 1e8 A? 1e8 A was suggested as PSS/E's conventional infinity, but nobody has confirmed that.
2. The 9999.0 sentinel. A RAW "no limit" (IMAX = 0) is parsed to 9999 A, which becomes 9.999 kA once stored. That's a finite and fairly tight limit, not "unbounded". Should the parser and exporter use the same sentinel as the defaults? And is IMAX = 0 really "no limit" in PSS/E?

Nothing in this set changes these values. Whatever is decided, it's about a one-line change each in the schema, PSY, PFFP and PF.

[rodrigomha]

The tests-only approach checks out: every current field goes through set_value!(…, "A"), which converts to the declared kA. Two requests:

1. RAW IMAX = 0 should store 9999 kA, not 9.999 kA.
src/pm_io/psse.jl:2336-2337 puts PTI_INF = 9999.0 into the dict, which holds PSS/E currents in A. set_value! then divides by 1000, so "no limit" becomes a finite 9.999 kA limit.

• Use 1000.0 * PTI_INF for the two max_dc_current_* lines.
• Keep PTI_INF as it is, because SMAX uses it in system p.u. and PowerFlows reverts it on export.
• PFFP doesn't depend on PSY, so this value has to be kept in sync by hand with PSY's INFINITE_CURRENT (Store DC current fields in kA, not A PowerSystems.jl#1836).
• Please add a test that a RAW with IMAX = 0 stores 9999 kA. Nothing covers this path today.

2. Stale docstring.
src/openapi/dc_branch.jl:225-226 says dc_current and max_dc_current_* "are already natural (Amperes …) and pass through unscaled". They are now converted from A to kA.

[rodrigomha]

See above

[rodrigomha]

All addressed, thanks. Approving. The remaining failures, including the new IMAX = 0 test (9.999e6), clear once PowerOpenAPIModels is regenerated with kA.