MEP Functional-Test Estimates: Count Test Cases and Witness Cycles, Not Only Equipment

Original illustrative engineering schematic. The worked calculation below uses explicit example inputs.
A commissioning allowance stated as hours per AHU can hide the work that actually controls the programme. One unit may have several required operating cases, a shared plant test may involve multiple systems, and a witness may attend an existing test rather than require a second demonstration. A useful estimate counts defined test executions before converting them into labour.
This original example covers one bounded functional-test work package, not the complete commissioning service. Installation inspections, point-to-point checks, calibration, startup, testing and balancing, training and seasonal tests must be assigned separately unless explicitly included. All durations and crew sizes are invented planning inputs, not published productivity norms or instructions for testing live equipment.
1. Build the matrix from approved requirements
Start with the approved sequences, commissioning plan and acceptance criteria. Give each test case an ID linked to its equipment or system, required initial condition, expected response and evidence to retain. A case is not simply a checkbox saying the equipment runs. It is a defined scenario with an outcome that the responsible reviewer can evaluate.
The DOE commissioning guide distinguishes verification checks from dynamic functional testing of systems and their interactions. It also describes testing through different operating conditions. Use that distinction to avoid charging the same task twice under startup and functional testing, while ensuring neither package silently assumes the other will demonstrate the sequence.
Identify each included asset or system and the exact approved test-procedure revision.
List required operating cases and state whether any are combined in one coordinated session.
Assign execution, observation, evidence recording, restoration and acceptance responsibilities.
Record prerequisites such as completed controls checks, available plant, suitable load and approved access.
Identify deferred, seasonal or integrated tests as separate tasks with their own dependencies, not as unexplained omissions.
2. Count case executions, not only procedure templates
Assume six AHUs each require four defined functional cases, and two pump systems each require three. The procedures have already been reviewed for safe execution. The same approved AHU template can be reused, but each asset still needs its own required execution and results. The two pump systems are separately testable, so their case executions are also distinct.
AHU case executions = 6 × 4 = 24.
Pump-system case executions = 2 × 3 = 6.
First-pass executions = 24 + 6 = 30.
The number thirty does not imply thirty mobilizations or thirty identical durations. Four AHU cases may share one setup and restoration sequence. Conversely, one case that depends on a different season or unavailable load may require a later visit. Keep three quantities visible: procedure templates, required case executions and practical site sessions.
For this example, all six AHUs and both pump systems receive the stated first-pass tests. No sampling of contractor execution is assumed. If the actual project permits sampling, obtain the approved population, selection rule and escalation requirements. Do not reduce the execution count merely because fewer tests will be witnessed.

Illustrative planning basis: six AHUs with four cases each plus two pump systems with three cases each produce 30 first-pass cases and 50.4 contractor person-hours. Observing those existing tests does not itself add a contractor repeat cycle. These are planning assumptions, not test procedures or standard labour norms.
3. Convert session durations into person-hours
For each AHU, assume 0.75 h of session setup, four cases at 0.40 h each, and 0.35 h for restoration and the session record. Two contractor technicians attend throughout. For each pump system, assume 1.00 h setup, three cases at 0.50 h each, and 0.50 h for restoration and records, with three contractor personnel attending throughout.
One AHU session = 0.75 + (4 × 0.40) + 0.35 = 2.70 elapsed h.
AHU contractor labour = 6 × 2.70 × 2 = 32.40 person-hours.
One pump-system session = 1.00 + (3 × 0.50) + 0.50 = 3.00 elapsed h.
Pump contractor labour = 2 × 3.00 × 3 = 18.00 person-hours.
Base contractor functional-test labour = 32.40 + 18.00 = 50.40 person-hours.
Person-hours measure effort, not calendar duration. A two-person team spending 2.70 hours on a unit contributes 5.40 person-hours. Dividing the whole package by an arbitrary headcount does not prove completion time: the two work types need different participants, and prerequisites, working windows and shared equipment may prevent parallel execution.
The assumed session durations include the specified preparation and final restoration, so do not add those same minutes again under a generic attendance allowance. Other travel, document preparation, coordination meetings or waiting remain outside this numerical example and need explicit lines when applicable. Record which role attends which part if the real team is not present for an entire session.
4. Distinguish observing a test from repeating it
Assume one independent witness attends the existing first-pass sessions for two AHUs and one pump system. This is a stated example attendance plan, not a recommended sampling rate. The witness sees eight AHU case executions and three pump-system executions. Because the contractor was already performing those sessions, concurrent observation does not automatically create another contractor test cycle.
Concurrent witness attendance = (2 × 2.70) + (1 × 3.00) = 8.40 person-hours for one witness.
Contractor first-pass labour remains 50.40 person-hours.
Witness cost ownership and appointment are recorded separately.
If the contract instead requires a separate repeat demonstration of the same two AHUs and one pump system, the contractor must perform additional work. Assuming the full original session durations and crews again, additional effort is (2 × 2.70 × 2) + (1 × 3.00 × 3) = 19.80 person-hours. The alternative contractor total becomes 70.20 person-hours. This is a changed attendance arrangement, not an automatic addition to the concurrent-witness case.
EPA’s commissioning guidelines distinguish coordination, review of pre-functional evidence and development of functional procedures with contractor assistance. That division is a useful reminder to establish who performs and who reviews each task. A line labelled commissioning attendance should identify the party and event instead of concealing duplicated or missing effort.
5. Model retesting as a separate scenario
Now test an alternative to the base case: four AHU case executions across two units need repeating after corrective work. Assume each case still takes 0.40 h and each affected AHU needs another 0.50 h for return-session setup, restoration and records. Two technicians attend these return sessions. Separately, one pump case needs 0.50 h plus 0.75 h for return-session setup, restoration and records, with three personnel.
AHU retest effort = [(4 × 0.40) + (2 × 0.50)] × 2 = 5.20 person-hours.
Pump retest effort = (0.50 + 0.75) × 3 = 3.75 person-hours.
Retest-scenario addition = 5.20 + 3.75 = 8.95 person-hours.
Base plus this retest scenario = 50.40 + 8.95 = 59.35 person-hours.
Corrective installation or programming work is not included in those retest hours. Neither is another independent witness visit. Add such work only under its own stated scope and responsibility. The scenario is not a measured failure probability, contractual entitlement or recommendation to expect a particular defect rate. Where evidence is insufficient, show the sensitivity rather than disguising it as a standard percentage.
DOE’s FEMP process includes retesting, monitoring and revised documentation after improvements. In the estimate, connect each return test to its defect or change record and the conditions needed to prove closure. A report marked corrective work complete does not establish that the affected function has passed.
6. Carry the calculation into the programme and handover
Issue the thirty-case first-pass matrix with its asset IDs, procedure revisions and prerequisites.
Separate base contractor effort, witness attendance, separately required demonstrations and any explicitly chosen retest scenario.
Schedule the actual qualified participants and available test windows, not only a total labour allowance.
Retain measured conditions, instrument references, results, unresolved issues and evidence of restoration to the approved operating state.
Reconcile completed, failed and deferred cases before calling the package complete. A completed attendance sheet is not acceptance evidence.
A bounded estimate is useful only when its records become useful handover evidence. Nexora’s maintenance-handover guide explains how asset identity, operating conditions and accepted baseline records stay connected after the construction team leaves.
The HVAC Estimation, BOQ and Job Costing Tool can organize the resulting quantities and labour build-up alongside budgets, commitments and actuals. Its verified scope is commercial estimation and job costing, not automatic generation of test procedures or commissioning approval. Keep the approved test matrix as the supporting evidence behind the estimate.
Continue the calculation review
Sources and calculation basis
Primary references checked on 23 October 2026. The worked examples and diagrams are original illustrative calculations, not manufacturer selections or project approvals. Apply the stated limitations and the requirements governing the actual project.


Comments