Change and Version Management for SCL Files, IED Settings and Network Configurations

A practical IEC 61850 configuration-management workflow covering identity, ownership, change risk, deployment, rollback and as-built reconciliation.

An IEC 61850 change is safe only when SCL, IED settings, network/time configuration, drawings and tests describe the same released system. Updating one relay locally without reconciling the master SCD creates configuration drift; updating a dataset can affect many subscribers even when only one publisher file changed.

This guide defines version identity, ownership, semantic change review, approval, deployment, rollback, as-built reconciliation and audit for protection/control systems.

1. Treat the system as one configuration item

  • Requirements, SLD, protection philosophy and cause-effect/GOOSE matrix.
  • Vendor ICD/IID files and namespaces.
  • Master SSD/SCD and generated target CIDs/device files.
  • Relay protection/control settings and logic.
  • Switch VLAN/QoS/multicast/PRP/HSR and clock/PTP configuration.
  • HMI/gateway database, point list and graphics.
  • Firmware, tool versions, licences and dependencies.
  • FAT/SAT/regression scripts, captures and results.

2. Version identity

ObjectIdentity fieldsVerification
SCLFile name, Header id/version/revision/history, namespace, hashValidator and semantic diff
IEDModel, serial, hardware, firmware, options, loaded file/hashDevice readback/record
SettingsSetting group/file revision/checksumCompare/readback and functional test
Network/timeDevice, firmware, config revision/hashExport/readback and traffic/time test
Test evidenceRequirement/case/config/tool/calibration revisionTraceability matrix

3. Ownership and engineering rights

  • Owner/design authority owns protection/control requirements and approves safety impact.
  • System configuration authority owns master SCD and system addresses/subscriptions.
  • IED engineer owns device capability/settings within allocated rights.
  • Network/time engineer owns switches, redundancy, multicast, QoS and clocks.
  • Cyber authority owns access, credentials, secure configurations and release controls.
  • Commissioning authority controls loaded site state and test restoration.
  • No tool or person may overwrite another domain without an agreed exchange/reconciliation process.

4. Change request content

  • Problem, requested outcome and initiating event.
  • Affected bays/functions/devices/files and current baseline.
  • Protection, safety, operations, network, cyber and regulatory impact.
  • Exact proposed semantic/configuration difference.
  • Outage and backup protection/control plan.
  • FAT/SAT/regression cases and pass limits.
  • Deployment sequence, responsible roles and communication.
  • Rollback trigger, package and maximum recovery time.

5. Semantic impact analysis

  • IED/LN/data model/process association changes.
  • Dataset membership/order and ConfRev.
  • GOOSE/SV/report control blocks and addresses.
  • Subscriber ExtRefs and reverse dependencies.
  • Protection settings, logic, timers and outputs.
  • VLAN/priority/multicast, bandwidth and redundancy paths.
  • PTP profile/domain/grandmaster and quality behavior.
  • Cyber access, services, keys/certificates and logs.
  • Drawings, HMI/gateway point mapping and operator procedures.

6. Risk classification and approval

ClassExampleMinimum assurance
Editorial/no functional effectControlled description textValidator/diff and document review
Local non-criticalAlarm threshold/display mappingBench/FAT plus affected SAT
Protection/control functionalSetting, ExtRef, interlock or controlIndependent engineering review and end-to-end regression
System/common modeDataset template, switch/PTP/firmware/tool updateMulti-discipline approval, representative/full regression and rollback rehearsal

7. Controlled engineering workflow

  1. Copy/branch the approved baseline; never edit the only master.
  2. Apply the authorized change with qualified tool versions.
  3. Run XML/schema/OCL and project semantic rules.
  4. Generate semantic diff and dependency/impact report.
  5. Independent-review critical functions and approve.
  6. Generate target files; validate they match the master.
  7. Test in representative lab/FAT with final settings/network/time.
  8. Create signed/hashed deployment and rollback packages.
  9. Deploy under outage/backup plan and perform SAT.
  10. Reconcile as-left state, approve as-built release and close change.

8. Deployment sequencing

  • Identify publisher/subscriber compatibility during partial deployment.
  • A dataset/ConfRev change can require coordinated subscriber updates.
  • Maintain independent backup protection during unavailable/ambiguous states.
  • Control simulation/test/output blocks and operator authority.
  • Load network/time infrastructure before/after IEDs according to tested sequence.
  • Pause at defined verification gates; do not continue on unexplained alarms.
  • Keep rollback media/files and skilled staff at site.

9. Verification after deployment

  • Read back device identity, firmware, settings and loaded configuration hash.
  • Verify affected reports/controls/GOOSE/SV/time and physical outputs.
  • Test each redundant path and alarms.
  • Confirm HMI/gateway/event mapping and control authority.
  • Clear test/simulation/subscription/output blocks.
  • Record network/time baseline and check no unintended traffic.
  • Obtain operations/protection/commissioning acceptance.

10. Emergency change

Emergency action can shorten approval lead time but must not eliminate identity, backup, test and reconciliation. Define who may authorize, minimum pre-change backup, immediate functional check and mandatory retrospective review.

  • Record reason, time, personnel and exact before/after state.
  • Preserve the original package and rollback.
  • Use temporary setting/change with expiry where appropriate.
  • Maintain protection coverage and notify operations.
  • Recreate/test/reconcile in the master system promptly.
  • Audit that the temporary change did not become an undocumented permanent baseline.

11. Audit and drift detection

  • Periodically export/read device, switch and clock configurations.
  • Compare semantic state and hashes with approved baseline.
  • Investigate every unexplained firmware/settings/subscription/address difference.
  • Review access and configuration logs.
  • Check field drawings, labels and asset identities.
  • Trend recurring emergency/temporary deviations.
  • Test backup restore and repository disaster recovery.

12. Change-impact examples

Proposed changeHidden dependenciesRegression scope
Add member to GOOSE datasetConfRev, all subscribers, frame size, test toolsEvery subscriber plus timeout/performance
Relay firmware updateData model, timing, SCL, settings conversion, cyberIED round-trip and all critical functions
Switch VLAN/QoS updateGOOSE/SV/PTP paths and queues across baysConnectivity, load, timing and redundancy
PTP grandmaster/profile changeAll MUs/relays, holdover and event chronologyOffset, failover and protection security
Protection setting changeTrip matrix, BF/backup coordination and HMIStudy boundaries and physical trip chain
IED replacementFirmware/options, address, CID, calibration, outputsComplete bay primary-to-breaker SAT

13. Rollback package

  • Last approved SCD/CIDs/settings and switch/time/gateway configurations.
  • Compatible firmware/installers, tool versions/licences and vendor instructions.
  • Device identity/serial/port/credential provisioning under controlled access.
  • Step-by-step rollback sequence respecting publisher/subscriber compatibility.
  • Backup protection and outage plan during restoration.
  • Verification cases and success/failure decision gates.
  • Known irreversible conversions or database schema changes.
  • Responsible roles, support contacts and estimated restoration time.

14. Release-completion checklist

  • Approved change request and semantic impact report.
  • Validator/diff/independent review complete.
  • FAT/SAT evidence closes every affected dependency.
  • As-left hashes/readbacks match released files.
  • No test/simulation/output blocks or temporary access remain.
  • Operations procedures, alarms, drawings and asset register updated.
  • Rollback package and clean backups verified.
  • Master SCD/repository is the final as-built source of truth.

References and further reading

Engineering note: A released version is a tested system state, not a filename. It must bind SCL, settings, networks, time, firmware, evidence and the installed asset.

LearnSwitchgear

Search the engineering library