Skip to content

VALIDATION_CHECKLIST.md

Latest commit

 

History

History
129 lines (106 loc) · 7.93 KB

VALIDATION_CHECKLIST.md

File metadata and controls

129 lines (106 loc) · 7.93 KB
Raw

ArIED 61850 Validation Checklist

Record the application commit, ARIEC61850 commit, Windows version, .NET version, test data provenance, and evidence source for every completed validation set.

Build and package

  • scripts/verify-source-clean.ps1 passes.
  • Restore and Release build pass against the documented ARIEC61850 integration baseline.
  • Required Smart Control and SCL workspace contracts are detected by CI.
  • Portable x64 package is generated with the intended version.
  • Package contains the required license, notice, copyright, trademark, support, and third-party documents.
  • Package does not contain source, captures, logs, local paths, confidential project data, or obsolete license files.

SCL workspace

  • SCD, CID, ICD, IID, SSD, and XML file filters open the intended test files.
  • XML loading rejects external entities and prohibited DTD behavior.
  • Multi-IED files produce the expected IED and AccessPoint context.
  • Duplicate endpoints are not added more than once.
  • An IED without an IP address remains visible as configured offline context.
  • Opening SCL does not require a live network connection.
  • Live connection and discovery occur only when the user requests an online action.
  • Configured model context remains distinguishable from the observed live MMS model.
  • Design-to-live comparison reports missing, extra, and mismatched model elements without rewriting the source SCL.

IED identity

  • The card name changes from the temporary endpoint label to the resolved IED name after discovery.
  • Logical Device instances remain separate from the IED name.
  • Synthetic domains such as IED_ACTRL resolve to the expected IED_A and CTRL split.
  • Multiple MMS domains from one IED produce one stable identity.
  • Diagnostics records whether identity came from explicit model metadata, LD instance removal, common prefix, or bounded heuristic fallback.
  • No customer, employer, or proprietary device identifier is required by the validation fixture.

Per-IED workflow

  • Every card has independent connect, disconnect, start, stop, configure, and remove actions.
  • Connecting or discovering IED B does not block monitoring of IED A.
  • Stopping IED A does not change IED B.
  • A failed device does not destroy the other device sessions.
  • The selection wizard opens only for its owning IED.
  • No signals are automatically selected on first discovery without user review.
  • Cancel restores the previous selection.
  • Saved selection returns for the same resolved identity.
  • Saved selection returns by endpoint when identity matching is unavailable.
  • Project save and open preserve the intended device and selection context.

MMS discovery

  • Logical Devices and Logical Nodes are discovered in stable order.
  • Data Objects and Data Attributes retain functional constraint and type context.
  • Values, quality, and IED timestamps are mapped to the correct references.
  • DataSet directory order is preserved.
  • RCB inventory identifies buffered and unbuffered candidates where exposed.
  • Unsupported file service or optional metadata is reported without discarding an otherwise valid model.
  • Reconnect creates a fresh native client and rebuilds session-scoped discovery state.

Reporting and monitoring

  • Existing static DataSet and RCB coverage is attempted first.
  • Actual DataSet members map to the correct selected references.
  • Member order remains consistent with report values.
  • Partially uncovered static groups receive a dynamic recovery attempt where supported.
  • Points without usable static coverage receive a temporary DataSet or URCB attempt before polling where supported.
  • Dynamic-monitor cleanup removes temporary resources when permitted by the IED.
  • Only still-uncovered or unverified points enter cyclic MMS polling.
  • Report-covered points use report-primary acquisition with only the configured verification or fallback behavior.
  • GI provides an initial image when supported.
  • Reservation, resource, DataSet, or RCB write rejection produces a clear diagnostic.
  • Acquisition identifies the actual source, such as configured report, temporary dynamic report, or polling.
  • A point is not removed from polling merely because it was placed in a candidate report group.

Process data and events

  • Discovery progress affects only the busy IED and does not block the rest of the application.
  • Connection, monitoring, report activity, unread event, and diagnostic states are understandable without relying only on color.
  • Live Monitor contains IED timestamp but does not present local receive time as process time.
  • Event Log contains semantic value transitions and does not create a process event for quality-only changes unless explicitly designed.
  • Event Log and CSV retain the originating IED and signal reference.
  • Breaker and switch double-point values decode correctly.
  • Boolean start, operate, trip, and indication edges are correct for the synthetic model.
  • Analog magnitude and unit are correct.
  • Missing quality is shown as unavailable and is not promoted to Good.
  • Copy and export output remains bounded and sanitizable.

Smart Control

  • The selected reference resolves to the intended control Data Object root.
  • ctlModel is read from the live IED before dispatch is enabled.
  • Direct normal, SBO normal, Direct enhanced, and SBO enhanced models follow the required sequence.
  • Oper, SBOw, and optional Cancel types are retrieved from the live model.
  • The named ctlVal field is located without positional guessing.
  • DPC, SPC, INC/ISC, BSC, and APC values bind only when representable by the live type.
  • Origin, ctlNum, timestamp T, Test, interlock, and synchrocheck values remain consistent through one sequence.
  • Enhanced models wait for positive or negative CommandTermination.
  • ControlError, AddCause, and LastApplError are surfaced.
  • Open and Close require the staged confirmation action before dispatch.
  • Cancel clears the staged command without dispatch.
  • Selection timeout, Cancel, association loss, and competing-client ownership produce deterministic outcomes.
  • Process feedback mapping is independently verified.
  • No generic .Oper, .SBOw, or .Cancel write fallback is available.
  • A descriptor that is not operationally ready leaves dispatch disabled.

Large-model behavior

  • 10,000 points do not create 10,000 tasks or timers.
  • The signal scanner is not retained as a permanent main-workspace grid.
  • Poll scheduler CPU remains stable when most points are report-covered.
  • Repeated reports use indexed reference lookup.
  • Event bursts do not create one Dispatcher operation per update.
  • Latest-value updates are coalesced by point key.
  • Event and diagnostic collections remain bounded.
  • WPF row and column virtualization remain enabled.
  • The UI remains usable at the tested Windows scaling levels.

Recovery and cleanup

  • Cable removal affects only the owning IED session.
  • Reconnect creates a fresh native client and invalidates stale association-scoped control state.
  • Saved signal selection remains available after disconnect and reconnect.
  • No duplicate live points, report plans, or polling entries appear after restart.
  • Temporary DataSets and report state are cleaned up when supported.
  • Application shutdown completes without re-entering the WPF closing lifecycle.

Claim boundary

  • The evidence source is identified as static review, automated test, loopback, simulator, isolated laboratory IED, or approved commissioning environment.
  • Results do not claim formal conformance, universal interoperability, cybersecurity approval, functional safety, switching authority, or equipment isolation.
  • Shared evidence is synthetic or sanitized and legally redistributable.