Pitot Traverse Uncertainty in a Short Rectangular Duct
A TAB technician must measure airflow in a 24 in × 18 in rectangular supply duct serving a critical cleanroom. The only accessible straight run is 3 duct diameters downstream of an elbow and 1.5 diameters upstream of a transition - well short of the recommended 7.5/2.5. The technician plans a 25-point pitot traverse using a Type S pitot and a digital manometer reading to 0.001 in. w.c. The measured velocity pressures at the grid points range from 0.008 to 0.061 in. w.c. The design airflow is 6,200 cfm. The cleanroom owner requires the reported airflow to be within ±5% of design. The technician is asked to report a single cfm number.
1. Assess whether a 25-point traverse in this location can support a ±5% claim, and explain the uncertainty contributors.
Read the worked answer
A ±5% claim cannot be honestly supported here.
Contributors:
- Velocity profile distortion. At 3D/1.5D the flow is neither fully developed nor symmetric; a fixed 25-point grid samples a skewed profile and systematically misestimates the mean. This is a bias (systematic) error, not random scatter, so averaging more points does not fix it.
- Low velocity pressure. The lowest reading (0.008 in. w.c.) is near the manometer resolution. At 0.001 in. w.c. resolution the relative quantization error at 0.008 is ~12.5%. Velocity pressure at these levels is also sensitive to the square-root relationship: V ∝ √Vp, so small Vp errors become velocity errors.
- Type S pitot coefficient. A Type S pitot has a coefficient near 0.84 that varies with velocity and must be calibrated; using an assumed 1.0 introduces a large bias.
- Density correction. Cleanroom air at controlled temperature/pressure differs from standard; uncorrected density shifts the result.
Conclusion: the combined bias dominates and likely exceeds ±5% regardless of point count. The technician should not certify ±5% from this location.
2. Recommend how the technician should proceed and what the report should state.
Read the worked answer
Recommended approach:
- Find a better measurement plane. Locate a run meeting at least 7.5D downstream / 2.5D upstream; if none exists, add straightening vanes or use a factory-calibrated flow measuring station or a duct-mounted averaging pitot array installed per manufacturer's required straight run. Document the actual straight run used.
- Calibrate the Type S pitot against a reference (wind tunnel or known coefficient) and apply the coefficient and density correction. Record the calibration date.
- Use a manometer with adequate resolution for the lowest Vp; a 0.0001 in. w.c. instrument for the low points.
- Report uncertainty explicitly. State the measurement location, straight-run limitations, instrument model/resolution, calibration traceability, and an estimated expanded uncertainty (e.g., "measured 5,900 cfm, estimated uncertainty ±8%, not suitable for ±5% acceptance").
- Do not fabricate a compliant number. If the acceptance criterion cannot be met, the report must say so and recommend the corrective action (relocate the test plane or install a flow station) to the engineer.
The report should flag the limitation clearly so the cleanroom owner understands the cfm is an estimate, not a certified value.