Y JOINT With Constant Cross Section Area
I want to create an acoustic tube wave guide splitter that is basically one input tube of larger cross section area joining two tubes that have identical, but smaller cross section areas (half of the input tube).
The criteria that makes this tricky is that in order to reduce wave reflections we need to keep the transition as smooth as possible, ideally keeping the cross section area of the joint constant. The area of the output pipes is half of the area of the input pipe. So the sum of the output pipe areas is the same as the area of the input pipe.
Another criteria is to keep the angle between the output pipes small like about 30 degrees or less.
The issue is that if I make a simple Y joint as demonstrated on this board already, the cross section area changes through the joint.
Is it possible to create something like this preferably in COMSOL or if not then in SolidWorks?
2 Answers
Zoltan, Here is an example of a Y joint that is Smooth and meets your A1 = 2*A2 criteria. I posted it earlier but was dissatisfied that there is no simple way to control the "Constant cross section" criteria. It is the shortest joint based in your 30 degree specification. A very interesting problem trying to workout what happens to the flow in the transitions section. Good luck.
Thanks Riend, I will check it out.
In mean time I have simulated a compromise Y joint. Actually not even a Y joint but a penta joint (3D distributed), where one large input tube joins 4 output tubes of smaller but equal diameters. (Same principle is applicable for a Y joint as well).
The trick here is to compensate for the acoustic reflections originating from the increase of cross sectional area at the joint. This I have achieved by tapering the input tube into a cone with decreasing diameter towards the joint. This tapering will generate a slight positive reflection back to the input.
When the forward wave reaches the widening cross section at the joint, it will create a negative reflection, which will get superimposed over the positive one and cancel one another (although not perfectly).
This solution works well (with only 0.1% of the forward wave power getting reflected), but only at specific frequencies, which in this case actually suits my need, because we use this at a single frequency.
However, a joint with a constant cross section area is a better solution, because it will not be frequency dependent.
