Infrastructure Engineering · Lesson 23 of 24 · 25 minutes

Build an integrated commissioning evidence loop

A completed component test answers a component question. Integrated continuity requires an initial state, expected sequence, instruments, acceptance criteria, stop conditions and recovery boundary that span the service chain. Compare two clocks, then examine the measurement boundary behind an apparently precise efficiency number.

Learning objectives

• Build observable cause-and-effect and acceptance records.
• Compare restoration time with electrical and thermal intervals.
• Compute interval energy and first-order PUE standard uncertainty.

Predict → run → explain

1. Read the declared boundary and units. Predict the selected result before calculating.
2. Run the starting case and inspect the ledger, assumptions and model limits.
3. Apply the challenge, run again, and explain the difference. Export only the last completed inputs and results.

Restoration s = sum of declared stage s; thermal/electrical margin s = corresponding interval s − restoration s; annual IT MWh = IT MW × hours; facility MWh = IT MWh × assumed PUE; standard PUE uncertainty = PUE × sqrt(uFacility² + uIT²), assuming independence.

Declared starting inputs

{
  "stageSeconds": [
    2,
    12,
    3,
    8
  ],
  "electricalSeconds": 600,
  "thermalSeconds": 37.5,
  "itMW": 20,
  "hours": 8760,
  "pue": 1.35,
  "facilityRelativeUncertainty": 0.01,
  "itRelativeUncertainty": 0.01
}

Challenge

Which teaching constraint fails when staged restoration takes 23 s?

{
  "stageSeconds": [
    2,
    12,
    3,
    23
  ]
}

Total restoration becomes 40 s: a −2.5 s thermal margin even though electrical margin remains 560 s.

A synchronized commissioning evidence wall — static equivalent

1. Reveal requirement → state → event → expected response.
2. Run a labelled restoration timeline alongside electrical and thermal limits.
3. Attach instrument records to each acceptance criterion rather than using decoration as evidence.
4. Route discrepancies back through correction and retest before baseline acceptance.

Cause-and-effect matrix, timed event ledger and acceptance table with unresolved fields.

Change one stage duration; inspect event timestamps and compare required versus observed values using clearly synthetic records.

Narrative steps are not physical seconds. Essential states and quantities remain available without animation.

Assessment and answer criteria