Of Airflow, Balancing, and Engineering Progress, or, What’s Inertia Got to Do With It?

I got so excited at the end of the last issue, I just couldn’t wait to show you the picture I promised (see Figure 1).  While I grant you only an engineer could get excited about the kind of stuff I talk about when I get technical, it’s my hope that the lay public in my audience finds the technical stuff at least interesting enough to avoid exclaiming, as my fiancée has on occasion been known to, “snore.”             

Anyway, the illustration represents the latest iteration of the “Eureka” moment I spoke of in the last issue.

Way back when, I was designing the central A/C system for Mary Tyler Moore’s 5^th Avenue co-op, and I ran across an issue where the space available to supply cooling to a corridor/art gallery existed only above a doorway which entered the corridor near the center of one of its long walls.  I designed the branch duct feeding the outlet to keep the velocity down low enough to preclude rumble in the duct, and when I looked for a grille in which to terminate the duct, I ended up with a grille about twice as wide as the supply duct.  That’s because grilles always have to be larger than their supply ducts in order to keep noise within limits since all the bars/louvers in the core of a grille subtract from the free area available for discharge airflow.

So, some distance upstream of the outlet, I designed a duct transition to gradually expand from the width of the duct to the width of the outlet, figuring I was home free.   Piece of cake, right?

Well-l-l, not exactly.  I’ll get back to this a little later.

At about the same point in my career, I started to think there had to be a better way to feed linear diffusers than multiple branch ducts feeding “pancake” plenums from a trunk duct stepping down in area after each branch duct.  My first whack at it was a tapered plenum like the one shown in the drawing, but with no dividers within.  I’d figured that as air exited the diffuser at its upstream end, the continuous taper would keep the pressure in the plenum high enough to have the diffuser deliver air for its entire length.   It didn’t quite work out as I’d expected.

At first it worked well enough, I suppose, but then I did a project where the Architect asked if he could use decorative grille as the outlet instead of a manufactured linear diffuser with integral volume dampers and directional controllers (which are what the little squiggles and lines shown in the duct collar in Section A—A represent).  The “decorative grille” turned out to be a latticework constructed of 1/16” square brass laid on a bias to form 1” square diamond-shaped openings for the length of an 8 foot long by about 3” high outlet.  Just about all of the air spilled out of the first two or three feet of the outlet, and spilled is the operative word.  The room the diffuser served was about 20 feet long, with the diffuser on its short wall at the ceiling, needing to throw the air at least 10 feet into the room.  The reality was that air found its way about 3 or 4 feet into the room.    Hmmm.

To get back to the grille in the corridor/art gallery, the Architect phoned me to tell me that the grille was noisy.  Huh?  It couldn’t be!   Well, as it turns out, it wasn’t really what I’d call noisy.

I mean the air was not whistling out of the grille.  Nah.  It sounded more like a freight train going by at 60.

What the devil was going on?  So I put my hand in front of the grille and discovered all the air was roaring out of the center portion of the grille, with virtually nothing coming out of that portion of the grille “expanded into” by my carefully designed duct transition.

The nerve! It didn’t know it should change direction to come out of the perimeter as well as the core?   Then I remembered that some guy named Newton came up with an observation about something he called inertia, which in this case meant that air would continue to go in the direction it started in until pushed or deflected to a different direction. 

In ductwork this is accomplished by a doohickey called a turning vane.  So I learned to design these things in to get the flow’s direction correct, but in a linear diffuser the problem remained that air needs to get to the far end of the diffuser.    Hmmm.

And that brings us back to where I was before I brought us here.

One of the ways to design ductwork is what’s called the equal friction method, which, when applied most simplistically, has all portions of a system designed with a constant friction loss per unit of length.  It would be more properly done by each branch of a duct system having the same overall loss. If such is done, the system becomes self balancing, and stuff like volume dampers become unnecessary.

Since this is nearly impossibly difficult to do on a complete system, volume dampers remain necessary, but a linear diffuser is one device, not a complete system.

So I came up with area inversely proportional to length as shown in the drawing, and because I’m an engineer instead of a physicist, I didn’t do a tap dance with Reynolds numbers, boundary, layers and free stream velocities.  (Huh?)