Product documentation is an essential component of
the support offered by Multi-Wing. Our wind tunnel has been in use
for more than 30 years and has produced trustworthy figures upon
which Multi-Wing’s selection software, Optimiser, is based. When it
comes to airflow, pressure development and consumed power, all the
curves and calculation is derived from measured data. The system of
fan laws enables the user of Optimiser to examine the impeller
performance in any possible combination of blade profile, diameter,
pitch angle, speed of rotation etc.
When it comes to noise, an analogue approach is not feasible. Noise
does not behave as predictably as the other performance dimensions
described by the fan laws. Furthermore, it would be unworkable to
measure all combinations that the modular Multi-Wing system enables
at any working point on the curve. This is a problem that has
challenged acoustic experts over many decades. Today, most impeller
manufacturers use a few empiric-based equations.
The method used by Multi-Wing is called Allen-Bareneks (or sometimes
Madison’s or ASHRAE), which is proven to be acceptably accurate and
versatile in any situation where the axial impeller is used. The
simplest version of the equation is described as:
Lw=Lw,s + 10 log (Q) + 20 log
(Ptot) [dB]
Where
Lw is the sound power level
[dB]
Lw,s is specific noise level for a given profile type
[dB]
Q is airflow [m3/s]
Ptot is the total pressure
[Pa]
Considering that the noise generated from a rotating impeller is a
complex phenomenon, this equation considerably simplifies its
description. The total noise is a result of contributions from tip
vortex, impeller leading/trailing edge, wake and possible stalling
related turbulence, all of which are logarithmically added together.
On top of that is the influence from the inlet and motor, which is
unknown from an impeller producer’s point of view. Although the
method should be used as a guideline only, experience shows that it
is a sufficiently accurate tool in most engineering situations.
Some limitations should nevertheless be considered. The equation
functions optimally close to the highest efficient point of the
curve, especially on the high-pressure side of the efficiency peak
an additional noise of up to 4-5dB is likely. In low-pressure
applications, often noise sources other than the impeller itself
dominate when the air speed increases.
Of course a relevant issue is what specific noise should be applied.
The Optimiser operates with three different categories of profiles;
the airfoiled, the broad bladed, and the sickle shaped impellers. An
estimated base noise over the frequencies from 63Hz to 8KHz with a
total noise of:
Airfoil 29 dB
Broad bladed 27dB
Sickle bladed 26 dB
Further reflection is needed when comparing small, fast spinning
impellers with large, slower running impellers at the same duty
point and approximate efficiency. If both impellers are of the same
blade type, the noise result will be identical. However, it is
probable that the bigger, pressure capable and slower impeller can
deliver the performance at a lower tip speed. Both noise results are
likely to be within the expected accuracy of the method, but the
bigger impeller will have an acoustic advantage to the small one.
