karimwassef
Active member
Maybe rectangular channels would work too, but not as well.
Large volume laminar flow machine
- Thread starter karimwassef
- Start date
karimwassef
Active member
Scaled and completely covered except for display and overflow. I think this is a reasonable balanced size
<a href="http://s1062.photobucket.com/user/karimwassef/media/1_zpsyhu2toge.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/1_zpsyhu2toge.jpg" border="0" alt=" photo 1_zpsyhu2toge.jpg"/></a>
still a little wasteful to dedicate 2 x 150 gal = 300 gal to the curved sides and 2 x 22 gal = 44 gal on the linear ends... 344 gal to curve.
~910gal
360 gal display (40% of total)
135 gal to channel + 345 gal curve = 480 gal (52%) ... wow
70 gal overflow (8%)
<a href="http://s1062.photobucket.com/user/karimwassef/media/1_zpsyhu2toge.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/1_zpsyhu2toge.jpg" border="0" alt=" photo 1_zpsyhu2toge.jpg"/></a>
still a little wasteful to dedicate 2 x 150 gal = 300 gal to the curved sides and 2 x 22 gal = 44 gal on the linear ends... 344 gal to curve.
~910gal
360 gal display (40% of total)
135 gal to channel + 345 gal curve = 480 gal (52%) ... wow
70 gal overflow (8%)
Last edited:
karimwassef
Active member
Does anyone have experience running free CFD simulation tools like OpenFoam or OpenFlow? I think that's the next step and it's been a decade since I did that.
ericarenee
New member
herring_fish
Crazy Designer
Gorgok You might be right. We will have to see what actually happens when the experiments are run.
As you paint a picture, whether it turns out to be all correct, partially right or even all wrong, nothing that comes up will be a surprise because of all of the input from you and the rest of us. Keep it coming.
We have already added a partial or total cover based on ericaree's idea.
Karimwassef I am hoping that you will not need to use the baffles and eggcrate because I am guessing that the water will be comparatively laminar coming off of the paddle's smooth pushing. Then (in effect) the pressure should drop as the water goes around the curve. That should smooth things out even more.
"¦but
As you paint a picture, whether it turns out to be all correct, partially right or even all wrong, nothing that comes up will be a surprise because of all of the input from you and the rest of us. Keep it coming.
We have already added a partial or total cover based on ericaree's idea.
Karimwassef I am hoping that you will not need to use the baffles and eggcrate because I am guessing that the water will be comparatively laminar coming off of the paddle's smooth pushing. Then (in effect) the pressure should drop as the water goes around the curve. That should smooth things out even more.
"¦but
karimwassef
Active member
The injection channel is 1' and the tank is 3'. Without channeling, the water will not flow around linearly. It will start to curve, but most will travel along the path of least change against the tank front with a width of 1'. That will create waves and turbulence against the other 2ft of width, losing the laminar bulk motion.
This refers to the curved baffles.
This refers to the curved baffles.
Last edited:
karimwassef
Active member
The eggcrate will be necessary to stop as many of the living creatures from getting into the mechanism. It's one thing for a 1" blenny to wiggle back there. It's another for a 10" naso to try to make the baffles his new hiding place.
This is especially true since the flow will not be continuous. There will be quiet periods where curiosity will make explorers venture into the darkness of the mechanism.
When we show this works, we still need to work out the logistics of maintenance and cleaning.
The straight baffles may not be necessary but that depends on many variables still in design.
This is why I want to simulate the flow. I want to start with the circular sides and then start pushing them in to become curved corner straight edges. As the corner curvature increases, the flow will become more disrupted locally and flow will drop.
The baseline should be the flow with the ideal circle. Flow from other designs can then be compared to it. That establishes the design trade off. I would like to have less than 1' on each end dedicated to the mechanism, but I wouldn't want to lose more that 20% of the flow.
This is especially true since the flow will not be continuous. There will be quiet periods where curiosity will make explorers venture into the darkness of the mechanism.
When we show this works, we still need to work out the logistics of maintenance and cleaning.
The straight baffles may not be necessary but that depends on many variables still in design.
This is why I want to simulate the flow. I want to start with the circular sides and then start pushing them in to become curved corner straight edges. As the corner curvature increases, the flow will become more disrupted locally and flow will drop.
The baseline should be the flow with the ideal circle. Flow from other designs can then be compared to it. That establishes the design trade off. I would like to have less than 1' on each end dedicated to the mechanism, but I wouldn't want to lose more that 20% of the flow.
Last edited:
herring_fish
Crazy Designer
Yea you might be right.
karimwassef
Active member
I agree on erica's idea before. I missed the concept of a sealed injection channel with a slot and raised sides to contain the wave. Sorry erica.
On the paddles, I think their design will also need to take the reversal wave into effect. Maybe raised curved tops with horizontal edges that create downward pressure.
On the paddles, I think their design will also need to take the reversal wave into effect. Maybe raised curved tops with horizontal edges that create downward pressure.
JamesHolt
Active member
Think CNC they are screw driven (basically) and can move faster than you can see
herring_fish
Crazy Designer
I agree. that is where it needs to be beefy. In the picture (at bottom) that I am re-posting(from #113), you can see that the longer the paddle is (see big red arrow), the more it wants to rotate and that makes it push up on one side and down on the other(see the little red arrows).
This is something like what I had in mind. Notice that there are wheels on the top and bottom to take the stresses. Also there are rollers on the inside at both sides and both ends.
Now I offer this picture for a "go by" if you wanted to build two carriages your self. I would go with round tubes, not square but you get the idea.
I have wheels drawn in below but you could also use ball bearing carriages but don't use Igus style carriages. It uses great space age, low friction, plastic. They work great for other applications. I use them but not for this. Stick with wheels or ball bearings. That my be a little more money but that is one of the places that the stress will be.
If your budget can handle the Cadillac ball bearing carriage, you can get the seller to spec out what you need. I would probably build my own if I could.
herring_fish
Crazy Designer
So the front glass of the display is 54 inches? The baffles are 9 inches apart and a total stroke, in the back, of 36 inches and a displacement of 18 inches in the display? ...and 6 cycles per second?
Last edited:
Gorgok
New member
I think those that really can haul are screw and ball drives, which isn't quite the same as just a screw drive... though i didn't see if these had screw and ball drives as well. Still they can be really fast with enough power.
That roller carriage is really easy to build. A pair of cheap rollerblades would get you enough wheels for probably less than you can buy the wheels for. The shaft size is a standard, you just need to get some bolts with a shank that size and thread a plate for them. Optionally include some adjustment standoffs for the bottom side and you can grip your rails with perfection.
If you want to get really fancy instead of the 'standard' size wheels get some goalie wheels, which are smaller so the contraption takes up less space. I do think having urethane wheels is better than straight bearings or something for this application though as you will hear it moving with metal to metal contact much more than with rubber on metal.
karimwassef
Active member
Thanks guys. Some things I DIY. Others I choose to assemble based on higher quality off the shelf parts.
The stresses here are likely outside of my comfort zone for a DIY.
It's putting the cart before the horse though. Before looking at the carriage and drive, I want to confirm the flow meets the requirement and the overall design is small enough relative to the display space. Then, I can determine the stresses on the paddle joints and the power needed to drive the carriage. That's why I need to complete the CFD analysis.
The stresses here are likely outside of my comfort zone for a DIY.
It's putting the cart before the horse though. Before looking at the carriage and drive, I want to confirm the flow meets the requirement and the overall design is small enough relative to the display space. Then, I can determine the stresses on the paddle joints and the power needed to drive the carriage. That's why I need to complete the CFD analysis.
herring_fish
Crazy Designer
So is this roughly what you want to build? Additionally, is it 24 inches tall and about 3,000 pounds?
karimwassef
Active member
The front is 8' = 96in long. The ends add another 1' + 3.5' for a total of 13.5' tank.
Yes, the height is 24".
karimwassef
Active member
Here is the desired design.
I think this is practically implementable and within some reasonable cost constraints.
It starts with a standard glass tank built at 10' x 4.5' x 2' (I don't like acrylic tanks. It was my biggest problem with the circular tank.)
The mechanism is essentially an acrylic construction inside the glass. The pieces are not glued to the tank (acrylic doesn't bond to glass), but think of it as an acrylic tank and mechanism inside the glass tank.
<a href="http://s1062.photobucket.com/user/karimwassef/media/1_zpszxx5ajyn.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/1_zpszxx5ajyn.jpg" border="0" alt=" photo 1_zpszxx5ajyn.jpg"/></a>
I think this is practically implementable and within some reasonable cost constraints.
It starts with a standard glass tank built at 10' x 4.5' x 2' (I don't like acrylic tanks. It was my biggest problem with the circular tank.)
The mechanism is essentially an acrylic construction inside the glass. The pieces are not glued to the tank (acrylic doesn't bond to glass), but think of it as an acrylic tank and mechanism inside the glass tank.
<a href="http://s1062.photobucket.com/user/karimwassef/media/1_zpszxx5ajyn.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/1_zpszxx5ajyn.jpg" border="0" alt=" photo 1_zpszxx5ajyn.jpg"/></a>
karimwassef
Active member
I'll work up the 3D model. Still trying to find a simple free 2D CFD software to analyze it.
karimwassef
Active member
Tried drawing it up, but the 6" overflow won't work. Not enough room for the bulkheads.
This is the current view at 5' x 10'
<a href="http://s1062.photobucket.com/user/karimwassef/media/1_zpsga7km8vj.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/1_zpsga7km8vj.jpg" border="0" alt=" photo 1_zpsga7km8vj.jpg"/></a>
Top left is the assembled system.
Top middle removes the water.
Top right is the glass tank only with the overflow.
Bottom left is the acrylic section with PVC round corners and the injection mechanism. This view also has the high top to contain splashes and surface waves around the paddles.
Bottom middle and right are the acrylic only views.
This is the current view at 5' x 10'
<a href="http://s1062.photobucket.com/user/karimwassef/media/1_zpsga7km8vj.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/1_zpsga7km8vj.jpg" border="0" alt=" photo 1_zpsga7km8vj.jpg"/></a>
Top left is the assembled system.
Top middle removes the water.
Top right is the glass tank only with the overflow.
Bottom left is the acrylic section with PVC round corners and the injection mechanism. This view also has the high top to contain splashes and surface waves around the paddles.
Bottom middle and right are the acrylic only views.
karimwassef
Active member
With dimensions
<a href="http://s1062.photobucket.com/user/karimwassef/media/2_zpss4gt47lm.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/2_zpss4gt47lm.jpg" border="0" alt=" photo 2_zpss4gt47lm.jpg"/></a>
<a href="http://s1062.photobucket.com/user/karimwassef/media/2_zpss4gt47lm.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/2_zpss4gt47lm.jpg" border="0" alt=" photo 2_zpss4gt47lm.jpg"/></a>
