ARVREL v0.1.0-beta.3

ARVREL v0.1.0-beta.3 — Transformer Differential Public Beta

ARVREL is a vendor-neutral Windows virtual protection relay laboratory for IEC 61850 Sampled Values engineering, education, process-bus troubleshooting, and protection-algorithm research.

This is a public beta technical preview, not a certified protection IED.

Release highlight — two-winding Transformer Differential IED

v0.1.0-beta.3 adds a dedicated two-winding transformer protection workspace while preserving the existing feeder protection laboratory.

Implemented transformer elements:

• restrained 87T with generic Is1 / K1 / Is2 / K2 dual-slope semantics;
• 87T-HS unrestrained high-set stage;
• 87N-HV and 87N-LV restricted earth fault stages;
• H2 inrush and H5 overexcitation security;
• paired HV/LV Sampled Values engineering and synchronization checks;
• CT ratio, transformer rating, polarity and supported vector-group compensation;
• deterministic CT waveform-distortion evidence;
• context-gated external-fault / CT-saturation security for 87T, optional high-set supervision and winding-specific REF supervision;
• practitioner evidence for external-fault arming, CT-saturation suspicion and active security holds.

The transformer implementation is vendor-neutral. It does not claim proprietary MiCOM CTSat/NoGap, SEL EFD, GE/Multilin, Siemens, or other manufacturer-algorithm equivalence.

Public tester self-test

A new deterministic 10-scenario Transformer Self-Test lets users verify the packaged protection core before connecting live Sampled Values or opening a field capture.

The Transformer IED workspace may therefore be opened while the main source remains Internal Demo or no SV streams are present. This does not bypass the real process-bus runtime guards: applying Live/Replay transformer protection still requires two distinct HV/LV streams.

The self-test verifies:

1. compensated through-current stability;
2. internal restrained 87T operation;
3. 87T-HS operation;
4. H2 blocking;
5. H5 blocking;
6. external fault with delayed HV CT distortion is secured by P13;
7. distorted internal fault remains trippable and is not overblocked by P13;
8. 87N-HV operation;
9. 87N-LV operation;
10. REF securely blocks when an independent neutral-current input is unavailable.

Expected public-test result:

PASS · 10/10 · transformer-public-beta-v1

The result viewer exposes copyable expected/observed evidence suitable for a reproducible issue report.

Transformer engineering model

The live/replay transformer runtime uses paired TransformerMeasurementFrame evidence rather than treating a transformer as a single feeder stream.

Key engineering behavior:

• HV and LV streams are selected independently;
• both streams must have compatible sampling/frequency context and acceptable synchronization evidence;
• vector-group clock displacement is applied through the transformer engineering plan;
• CT ratios and transformer nameplate data establish per-unit scaling;
• zero-sequence removal is limited to the differential compensation path where required by the connection model;
• REF retains local uncompensated residual quantities and requires an independent neutral-current input;
• calculated phase residual is never silently substituted for the REF neutral CT.

Automatic compensation is intentionally rejected for unsupported/ambiguous transformer configurations rather than guessed.

External-fault / CT-saturation security

P13 deliberately separates waveform evidence from a protection decision.

Waveform distortion by itself never blocks transformer protection.

A security hold requires a restraint-leading external/through-fault context before qualified CT distortion can supervise the affected protection path. The public self-test includes both sides of this invariant: an external-fault saturation case must remain secure, and a distorted internal fault must remain dependable.

P13 is disabled by default until explicitly enabled by the practitioner.

Existing laboratory capabilities retained

• live Npcap IEC 61850 Sampled Values capture;
• PCAP and PCAPNG replay;
• SCL-assisted stream binding, mapping, scaling, APPID, VLAN, svID, datSet, confRev, quality, freshness, and smpCnt trust checks;
• 4I+4V one-cycle RMS phasors, symmetrical components, residual quantities, waveform and phasor instruments;
• feeder 50P-1, 51P, 50N, 51N, 67P, 67N, 27, 59, and 59N virtual protection elements;
• practitioner settings and read-only active algorithm source;
• research shadow workspace with deterministic policy validation;
• relay LCD, pickup/trip annunciation, latched trip causes, event trace, settings identity, and evidence export;
• coherent waveform/phasor hold during SMV discontinuity recovery.

Packages

Official release assets are expected to include:

• ARVREL-Setup-v0.1.0-beta.3-win-x64.exe — per-user Windows installer;
• ARVREL-v0.1.0-beta.3-win-x64-portable.exe — single-file portable executable;
• ARVREL-v0.1.0-beta.3-win-x64-portable.zip — portable package;
• ARVREL-v0.1.0-beta.3-legal-notices.zip;
• SHA256SUMS.txt;
• NuGet dependency report and CycloneDX SBOM when generated by the release workflow.

The executable packages include the compiled ARIEC61850 dependency. Source-development builds still use the sibling repository layout.

Requirements

• Windows 10 or Windows 11 x64;
• no additional dependency for the deterministic Transformer Self-Test;
• Npcap only for live capture;
• an authorized, isolated laboratory network for live process-bus testing;
• appropriate SCL context for trusted SCL-bound stream operation.

Npcap is not silently installed or relicensed by ARVREL.

Recommended public test order

1. verify package SHA-256;
2. run the 10-scenario Transformer Self-Test and retain PASS · 10/10 evidence;
3. use PCAP replay for repeatable paired-SV evaluation;
4. use Live Npcap only after replay/internal behavior is understood;
5. file reproducible issues without publishing proprietary station data.

See docs/TRANSFORMER_PUBLIC_TEST.md.

Safety boundary

The public beta is virtual-output only. It does not provide physical relay contacts, operational GOOSE trip, MMS control, switching authority, IEC 61850 conformance certification, IEC 60255 type-test, calibration, or protection-grade real-time guarantees.

The deterministic self-test is software regression evidence. It is not relay-test certification or proof of a real substation protection scheme.

Do not use ARVREL as the sole basis for operational protection settings, commissioning acceptance, or switching decisions.

Licensing

ARVREL source is available under GPL-3.0-or-later. An alternative commercial license may be negotiated for proprietary redistribution, OEM integration, contractual support, or other non-GPL terms. See COMMERCIAL-LICENSING.md.

Third-party components retain their own licenses.

Known beta limitations

• community binaries are not currently claimed as Authenticode-signed and may trigger Windows reputation warnings;
• live performance depends on Windows scheduling, Npcap, adapter drivers, publisher behavior, and host load;
• the beta is not a hard real-time platform;
• transformer automatic engineering is intentionally limited to supported two-winding vector-group cases; unsupported or ambiguous configurations must not be guessed;
• REF requires an independent neutral-current input for authority;
• broad clean-machine, multi-adapter, multi-MU and diverse vendor stream validation continues during the beta period;
• feeder 46, 47, 49, 81U/O, 32, 37, 50BF, 79, 25, 86, 74TCS, and 60 remain unimplemented;
• negative-sequence and memory polarization for feeder 67N remain deferred.

Reporting

Use GitHub issues for reproducible non-sensitive defects. Include the copied Transformer Self-Test evidence when the issue involves the new transformer workspace. Use the private security-advisory workflow for vulnerabilities. Never publish proprietary substation captures or SCL files.