Existing data centers change over time. Loads increase, equipment ages, controls are modified, emergency repairs alter configuration and documentation can drift away from reality. Recommissioning provides a structured way to verify whether the facility still performs as intended, but testing a live critical site requires much greater operational control than testing a new empty building.
Establish the baseline first
Before any active test, review current one-line diagrams, mechanical schematics, control sequences, alarms, load levels, maintenance history, open defects and previous incident reports. Field walkdowns should verify that documentation matches the real installation.
Understand current redundancy
Do not assume the original design redundancy still exists. A failed UPS module, unavailable generator, isolated battery string or degraded chiller can change the safe testing envelope.
Recommissioning should begin only after the team understands the actual current state.
Start with non-intrusive evidence
Historical BMS, EPMS and DCIM trends can reveal control instability, capacity problems and repeated alarms before active testing begins. Thermal imaging, vibration analysis and other condition techniques may also identify risks without changing system state.
Use phased testing
Testing should progress from low-risk verification to more intrusive functional tests. For example, confirm alarms and status points before simulating source failures. Test one redundant component before testing full-system response.
Each phase should have clear go/no-go criteria.
Maintenance windows
Active tests should be scheduled when load, staffing, weather, customer activity and available technical support are appropriate. Avoid performing major resilience tests while another system is already under maintenance unless that degraded state is deliberately part of the approved scenario.
Protect the rollback path
The test plan should define how the facility returns to a stable configuration if the expected response does not occur. Rollback equipment, spare parts, vendor support and switching authority should be available before the test begins.
Test controls and monitoring
Older facilities often have control changes that were never fully validated. Recommissioning should verify current BMS sequences, interlocks, alarm thresholds, sensor calibration, communication failures and fallback modes.
Rebuild documentation as part of the process
Any discrepancy between field condition and documentation should be corrected. Recommissioning is an opportunity to restore configuration control as well as prove equipment performance.
Typical recommissioning priorities
- Verify utility-to-generator transition.
- Test UPS operating and bypass modes.
- Confirm cooling redundancy and controls.
- Verify critical alarms and monitoring.
- Test maintenance bypass arrangements.
- Review battery and generator readiness.
- Confirm emergency operating procedures.
- Update as-built drawings and settings.
Key takeaway
Recommissioning is reliability verification for a facility that has changed with time. The goal is not to recreate new-building commissioning blindly, but to use evidence, phased testing and operational controls to prove that the current data center remains resilient without exposing live services to unnecessary risk.
References and Further Reading
- ASHRAE Guideline 0-2019, The Commissioning Process.
- ASHRAE commissioning resources for new and existing buildings.
- ISO/IEC TS 22237-7:2018, Data centre management and operational information.
