Lab-to-lab Variation in Testing Fans

B3 - Measurement & Test

AMCA International conducted a round robin series of test on three fans to investigate lab-to-lab variation in air performance and sound test results. The purpose of the round robin was to advance the science of testing fans in accordance with ISO and AMCA standards, specifically ISO 5801, ISO 13347, AMCA 210 and AMCA 300, and to advance our knowledge of test result uncertainty such that tolerances for certification programs and acceptance tests are fair and realistic.
Three fans were part of the round robin, and all three were tested on multi-nozzle chambers. Centrifugal and tubeaxial fans were tested using a chamber at the fans’ outlet, and both were powered by dynamometer. A vaneaxial fan was tested on a chamber at the fan’s inlet and was powered by a calibrated motor. All three fans were sound tested in a reverberant room.
What we found is that the agreement between labs is actually very good. The determination of air power, power consumption and sound power through the measurement of pressure, temperature, torque, rotational speed and sound pressure is quite consistent from lab-to-lab, leading to good agreement in test results if the fans are well-behaved. If the fans are not well-behaved, meaning there is a significant amount of swirl at the fan’s outlet or the fan’s vibration is excessive, lab-to-lab variation can be quite high.
During an analysis of the air performance data from a high swirl fan we were able to tease out from the test results a correlation between air performance and the ratio of the outlet area of the fan to the area of the test chamber. This correlation is well known, but the results appear correlated to fan to chamber area ratios at ratios much higher than had previously been accepted. Sound power data, of course, was strongly correlated to fan vibration. An interesting note is that the uncertainties published in the aforementioned standards do not take into account errors associated with fan outlet area or fan vibration.
The CFD portion of this work focuses on gaining insight into the above mentioned correlation between air performance and the ratio of the outlet area of the fan to the area of the test chamber. To mimic the above experiments a three factoral DOE is designed to cover the fan type, the area ratio and the test chamber area shape as variables. Full description of preferred solver setup, post processing method and simulation quality metrics is provided. Relative comparison of experimental vs. CFD data is presented. Finally, conclusions about performance predictions are offered.