Choosing the method for the situation
Pitot traverse in straight duct remains the reference method. Hood measurement suits terminal devices where duct access is impossible. Anemometer grids work at large open faces. Each carries different uncertainty, and mixing methods across a system produces a balance that does not reconcile.
Traverse errors
Too few points, points placed on a grid that does not suit the duct shape, or a traverse taken too close to a bend or damper will all produce a confident number that is wrong. Straight duct upstream and downstream is not a formality.
Hood errors
Hoods add resistance and can change the flow they are measuring, particularly at low-pressure terminals. Back-pressure compensation exists for this reason and needs to be applied. A poor seal against the ceiling tile leaks more than most technicians expect.
Instrument and recording discipline
Calibration certificates, ambient corrections for density, and consistent recording of system conditions at the time of reading are what allow a second engineer to reproduce the result. Without them the report is an opinion.
