Karim's 1500gal dream reef

[dave.m]

I'd go with option 2. Like most aspects of this build, I believe you are over-thinking things way too much. No doubt you are having a lot of fun with the process, but you will not really know how any of this will work until you actually build something and get it up and running. Thinking you are going to create the perfect system all in the first go is not realistic IMHO.

Dave.M

 

[karimwassef]

I don't expect it'll be perfect on the first go. But there are many things that can be changed later and a few that absolutely cannot.

My goal is to make sure the design is not pigeonholed.

For example- I can build a greenhouse, but if I build it facing south, I won't have sufficient light. The greenhouse must face south. Not really fixable.

The wind in TX flows south to north so my big doors and vents are in the right orientation to leverage that flow for cooling. I could build it with east west doors, but it would cost $10k+ to fix later.

I can go with standard glass, but that would block 40% of light especially the high blues, violets and UV. The result would be brown ugly corals most people associate with solar lit tanks. Also fixable later for $10k+.

I could go with air ventilation cooling only, but the calculation show that the greenhouse would become a sweltering 105F 100% humidity zone with fogged up glass and dead bleached coral. Nothing a $10K HVAC can't fix, but I'm laying out 200ft of underground pipe for an air cooling loop under the greenhouse slab. That would be impossible to do later.

I'm working out the tank dimensions to maximize grow volume at 12' x 8' x 27". I could go smaller and then start over in a few years or be perpetually sorry I didn't do it right.

There are some things that I could have done later, but would have required draining the tank. If I put the right "hooks" in up front, I have the flexibility to evolve over time. A good example is the powerhead flow vs vacuum pump flow... I don't know which is better so I'm going to build the tank so either will work.

It's the difference between permanent and modular flexibility.

I guess what comes across as over thinking is really just me future-proofing the most expensive and least changeable elements.

When I think of how I wish I had designed my current tank, I kick myself every day for locking in some key elements that would have made it spectacular.

But you are correct: in the case of locking out the dark zone - it really doesn't have much a penalty to change later... so you're absolutely right on that one. Well, except a big tang that gets locked in the dark zone and dies of hunger... or the decimation of the shrimp population. All fixable later without massive expense though.

I do like to share ideas and conversations that come up... you may have opinions or thoughts and I'm a believer that the collective is smarter than the individual... so I ask

 

[dave.m]

Asking is free. :thumbsup:

Dave.M

 

[karimwassef]

By the way - I've been working on solving the flow problem for the coupled sealed tanks and it turns out to be quite complicated. I even asked on the physics forums and we're still not there:
https://physics.stackexchange.com/q...g-after-the?noredirect=1#comment766362_341609

 

[Bonvivant]

next set of experiments

<a href="http://s1062.photobucket.com/user/karimwassef/media/IMG_0503_zpswurgwmj8.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/IMG_0503_zpswurgwmj8.jpg" border="0" alt=" photo IMG_0503_zpswurgwmj8.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/IMG_0506_zpsfkhpwwqb.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/IMG_0506_zpsfkhpwwqb.jpg" border="0" alt=" photo IMG_0506_zpsfkhpwwqb.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/IMG_0507_zpsyeiyscxl.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/IMG_0507_zpsyeiyscxl.jpg" border="0" alt=" photo IMG_0507_zpsyeiyscxl.jpg"/></a>

<a href="http://s1062.photobucket.com/user/karimwassef/media/IMG_0509_zpszhtogmck.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/IMG_0509_zpszhtogmck.jpg" border="0" alt=" photo IMG_0509_zpszhtogmck.jpg"/></a>

I'll post a video in a little bit.

Here's the conclusion. Coupling the tanks works to create a reverse surge that pulls water up one tank as it flows down and out the other. But as expected, it's slower (lower flow) than a regular surge. I'll use the videos to collect some data. The design still needs 4 inches / sec flow, so this may or may not work.

Also, the inertia of the slower up/down flow is so large that there's little to no overshoot, so no oscillation either.

What abut using a vacuum pump instead of shop vacs? They are quite strong. I use them for veneering.

 

[Bonvivant]

What about using a sealed piston/plunger with an actuator to accomplish basically the same thing. Im thinking almost like how a stringe works. I'm no where near an engineer but it seems that it would be a safer solution. For example, if one of those vacumn valves fails, bad things are going to happen.

 

[karimwassef]

I can't maintain a seal for a plunger in a reef tank. It takes 2x the space and has reliability and cost concerns with large mechanical systems.

If a valve fails, nothing happens other than air flows in and water flows into the tank and down to the sump.

 

[karimwassef]

Vacuum pumps are very slow. I'm looking into them but they're like mighty turtles- they move mountains very slowly.

The one I have is oil lubricated so the air can have oil in it that might end up in the water. I'll keep looking but that may be the right vacuum source for smaller systems.

 

[karimwassef]

I'm trying to model the flow in and out of the pumps as a circuit with resistors and capacitors... just need to figure out how to relate it back to the design elements.

<a href="http://s1062.photobucket.com/user/karimwassef/media/Designs/0_zpsqiwqc0rh.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/Designs/0_zpsqiwqc0rh.jpg" border="0" alt=" photo 0_zpsqiwqc0rh.jpg"/></a>

 

[karimwassef]

Found a 12cfm vacuum pump but need more specs to see if it works.

It's used for milking and wine degassing so maybe the oil lubrication isn't an issue.

https://www.amazon.com/gp/aw/d/B06Y51XSBM?psc=1

 

[taricha]

By the way - I've been working on solving the flow problem for the coupled sealed tanks and it turns out to be quite complicated. I even asked on the physics forums and we're still not there:
https://physics.stackexchange.com/q...g-after-the?noredirect=1#comment766362_341609

Eh, this is my stab at it (hs phys teacher - so probably wrong)
But I get 25.4 seconds to equalize out.
Your 2" pvc air pipes won't regulate anything- too big, so the flow is controlled by the water pipes - which I assumed were 2" as well so I could have something to work with.
Made some assumptions that I don't feel too bad about.
Bernoulli, Torricelli, separation of variables and get....
Time to equalize = (A1/A2) * sqrt(h/2g)
Where A1 is area of container, A2 is area of water pipes, h is height difference in fluid levels, and g is gravity.
Pics for details...

 

[karimwassef]

Wow! Awesome. Thanks for working all that out!
I think you've worked it out for two bucket with a connecting water pipe on the bottom. In that case, the water moves exclusively from one bucket to the other and can actually oscillate since the differential equation would have both an exponential decay term and a wave term.

In my case though, the water empties into the bottom display tank and is picked back up again into the other tank as the air equilibrates pressure above it.

I think I figured it out today, but I cheated a bit. I modeled it as a charged and dishcharged capacitor connected with a switch on top and connected to ground on the bottom. The open bottoms of the buckets connected to an open tank is like the circuit ground. Charge is volume. Voltage is pressure. Current is flow rate.

The tricky part was figuring out how to convert back into fluid flow. When I ran the standard surge, I realized that the coefficients are the same, so I used that to work out the final form.

Got stuck with pipe resistance creating head pressure and the iterative nature there, but got close enough, I think. My time constant tau varies as a function of the design variables:

<a href="http://s1062.photobucket.com/user/karimwassef/media/Designs/1_zpsrf273yyy.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/Designs/1_zpsrf273yyy.jpg" border="0" alt=" photo 1_zpsrf273yyy.jpg"/></a>

 

[karimwassef]

I think it was about 9 seconds for an 8' tall tank settling with an empty tank to the middle state at 4'. That's going through 5 x 4" PVC pipes at 2ft in length.

 

[karimwassef]

Unfortunately, even a single elbow is equivalent to 14' (yes-feet) of pipe and consumes 4' of head!! Two elbows stops the flow...

http://www.engineeringtoolbox.com/resistance-equivalent-length-d_192.html

So... no elbows. Needs a straight shot to work.

 

[taricha]

I think you've worked it out for two bucket with a connecting water pipe on the bottom.

Nah, the pipes connecting on the bottom are theoretical. Don't actually exist, just there to visualize the volume transfer enforced by your actual air pipe on the top.

In that case, the water moves exclusively from one bucket to the other and can actually oscillate since the differential equation would have both an exponential decay term and a wave term.

Yeah, but when I looked at the size of the oscillate term (for numbers you were working previously 2'x2' cylinder buckets drained by 2" pipe) it was a few parts in 10^5, so I ignored the oscillate term.

In my case though, the water empties into the bottom display tank and is picked back up again into the other tank as the air equilibrates pressure above it.

Maybe you underestimate air pressure. that connected top air tube is going to really strictly enforce water out of one bucket simultaneous with water into the other bucket. The connected air tube makes equal pressure in both buckets at all times a good assumption.
UNLESS: you have an incredibly restrictive air flow tube, that is actually governing the flow because it's more restrictive than the water flowing out through the bottom. Is that what you are trying to model?

I think I figured it out today, but I cheated a bit. I modeled it as a charged and discharged capacitor connected with a switch on top and connected to ground on the bottom. The open bottoms of the buckets connected to an open tank is like the circuit ground. Charge is volume. Voltage is pressure. Current is flow rate.

Concern with the model:
Capacitor discharge rate is proportional to the size of the charge. hence the exponential decay solution.
Fluid discharge rate is not propotional to volume. It's proportional to Sqrt(volume) - actually sqrt(height) and height is proportional to volume, but you get the point.
Are the two functions similar enough shapes to allow your answer to work? Maybe? I dunno.

Got stuck with pipe resistance creating head pressure and the iterative nature there, but got close enough, I think. My time constant tau varies as a function of the design variables:

Your model has a lot of detail "hair." Torricelli has no hair and is a best case simplified scenario. All the hair will make the flow rate less than what Torricelli model calculates, and should be considered. I just am not sure you didn't build a lot of "hair" on a wrong dog.

I think it was about 9 seconds for an 8' tall tank settling with an empty tank to the middle state at 4'. That's going through 5 x 4" PVC pipes at 2ft in length.

How wide are the 8' tall tanks? Let's check and see how things compare.

Just to restate. I think you may need to build your model on Torricelli's Law. That is after all derived just for figuring out how to empty a fluid container through a drain. Then any restrictive air flow, or turbulence, or long pipes will somewhat slow the flow process.
Torricelli, total time to empty container

 

[karimwassef]

I do use Torricelli's law. It's in the second column of formulae.

velocity = sqrt(2gh) where h is then corrected for resistance of the pipes.

my simplified super-hairless time constant tau

<a href="http://s1062.photobucket.com/user/karimwassef/media/Designs/0_zpsdptjkxkx.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/Designs/0_zpsdptjkxkx.jpg" border="0" alt=" photo 0_zpsdptjkxkx.jpg"/></a>

... this is basically the same as yours. It turned out to be ~6.5 seconds with 5 x 4" diameter PVC pipes. The roughness corrects it to 6.7sec.

In the case of two tanks, the tau term in the exponential is actually halfed which shows up in the exponential (but I still kept the original surge tau).

The results above are for tanks that are 2' x 4' x 8'. The starting condition is that one tank is full (8') and the other is empty (0') so the change in height is 4' in and out of the display bottom tank.

The hair actually is very small until I added elbows... then it broke. So I'm redesigning completely to avoid any and all elbows.

I think if you use the same dimensions and pipes, you'll get the same?

 

[karimwassef]

the capacitors flow is based on charge and charge's analog in the fluid system is volume... which with a constant cross-sectional area is linear with height. So capacitor charge is linearly equivalent to water height.

 

[taricha]

Making more sense to me now.

I do use Torricelli's law. It's in the second column of formulae.

velocity = sqrt(2gh) where h is then corrected for resistance of the pipes.

my simplified super-hairless time constant tau

<a href="http://s1062.photobucket.com/user/karimwassef/media/Designs/0_zpsdptjkxkx.jpg.html" target="_blank"><img src="http://i1062.photobucket.com/albums/t496/karimwassef/Designs/0_zpsdptjkxkx.jpg" border="0" alt=" photo 0_zpsdptjkxkx.jpg"/></a>
... this is basically the same as yours.

yep, it's exactly the same result I got for two tanks coupled together

It turned out to be ~6.5 seconds with 5 x 4" diameter PVC pipes. The roughness corrects it to 6.7sec.

In the case of two tanks, the tau term in the exponential is actually halfed which shows up in the exponential (but I still kept the original surge tau).

The results above are for tanks that are 2' x 4' x 8'. The starting condition is that one tank is full (8') and the other is empty (0') so the change in height is 4' in and out of the display bottom tank.
....
I think if you use the same dimensions and pipes, you'll get the same?

Almost! We disagree by a factor of 4.
Straight Torricelli draining one full tank that you've described (from 8' to 0') comes out to be 3.23 seconds.
You say when you couple two together and each goes to 4' the result will be ~6.5sec (twice as slow).
My derivation gives me that when the two are coupled together it takes half the time - 1.62 seconds.

 

[karimwassef]

I realize that there are two variables working together here. My formula has a starting condition where one tank is full and the other is empty. That means that the time constant tau it calculates is actually based on the 8' height differential at the start (same as the full surge). That's a tau (time constant) of 6.5 sec.

I likewise have a 1/2 tau for the two tanks vs surge. It settles twice as fast as the tau above (6.5 sec) would predict = 3.3 sec.

That's not actually the full time. It's the time constant of the exponential which drops from peak to 1/e ~ 37% left. That gets to the knee but 2 x tau is closer to balance ~ 14% left. Since I'm using 2 x tau as my "end-state" time. That's actually 2 x tau / 2 = tau = 6.5 seconds again.

Looking back, the way I framed the question wasn't very clear and my answer (for comparison) wasn't clear either. Sorry for that

The fact that the "time it takes to level" is unclear is a bi-product of the exponential nature of the solution and I was using my own assumptions for what that means.

 

[karimwassef]

Looks like we have the same result, just different numbers used. We also approached it in two different ways so that gives me confidence that it's right.

Back to the "hairs" - do you agree that two elbows or more would stop the flow?