Flow calculations

I am trying to compare how well a MCw6002 or Whitewater would work with my setup. Unfortunately, I don’t know the formulas to make the comparison; If somebody could please give me a copy of them it would be really nice. Also, I know they exist because people like Cather often quote calculations they have made.

In my understanding there isn't a formula to precisely calculate the flow through a loop.

You can estimate it based on information of the individual components of a system, but you cant calculate it.

You can measure it with a flowmeter though.

you take the flow resistance chart for the water block and compare it to the PQ chart of the pump.

Probably better off looking at procooling for some resistance curves of systems, I know you can find some for a typical WW system.

ls7corvete said:

you take the flow resistance chart for the water block and compare it to the PQ chart of the pump.

Probably better off looking at procooling for some resistance curves of systems, I know you can find some for a typical WW system.

Click to expand...

But if you do that what about resistance from the rad? the res? the tubing?

9mmCensor said:

In my understanding there isn't a formula to precisely calculate the flow through a loop.

You can estimate it based on information of the individual components of a system, but you cant calculate it.

You can measure it with a flowmeter though.

Click to expand...

Even so, If ya known the formula I want it.

I am sorry. I cannot tell you it or even admit it exists or that it is a secret guarded by the Masons.

9mmCensor said:

But if you do that what about resistance from the rad? the res? the tubing?

Click to expand...

Thats why I said he is probably better off looking at teh curves for a whole system over at procooling

ls7corvete said:

Thats why I said he is probably better off looking at teh curves for a whole system over at procooling

Click to expand...

doh. Got hasty and responded after I read only the first sentance.

Anyway, those will still be only approximations do to the differences in systems (tubing, routing, cleanliness of the equipment, etc).

Will this help?

Pump Head-Pressure-Specific Gravity

Head and Pressure Conversion

Liquid Head in feet = psi x 2.31/ Sp. Gr

Pressure in psi = Head in feet x sp gr/ 2.31

Liquid Head in feet = psi x 144/ W

Pressure in psi = Head in feet x W/ W

Where W = Specific weight in pounds per cubic foot of liquid being pumped under pumping conditions:
For Water W = 62.32 lb per cu ft at 68 degrees F.
There are some other conversions and things HERE that may be useful to you about mid-page.

In my opinion the only chance for you is ready-to-read review. The best woul be with some blocks runing with specific pumps.
Ther eare formulas and simulators but they're all far from reality. What's the point of using one if results differs 10 times.

Jeff Moser said:

Will this help?....

Click to expand...

Nope. The gentleman essentially needs to calculate the total head loss in his cooling loop and find the point on the Head vs. Discharge curve of his pump, in order to determine his flow rate. It must be an iterative process that ultimatly converges on the "right" answer. Recall that head loss, both frictional and local, is proportional to the flow velocity squared. Many trial calculations must be made (making certain starting assumptions- likely velocity). Parameters must be adjusted and aditional calculations made, until the answers for the system makes sense.

Will I tell you how to do it? It would take five pages, and I haven't the time. My advice is to look at what type of systems have worked for other people.

cpufan said:

Nope. The gentleman essentially needs to calculate the total head loss in his cooling loop and find the point on the Head vs. Discharge curve of his pump, in order to determine his flow rate. It must be an iterative process that ultimatly converges on the "right" answer. Recall that head loss, both frictional and local, is proportional to the flow velocity squared. Many trial calculations must be made (making certain starting assumptions- likely velocity). Parameters must be adjusted and aditional calculations made, until the answers for the system makes sense.

Will I tell you how to do it? It would take five pages, and I haven't the time. My advice is to look at what type of systems have worked for other people.

Click to expand...

I'll expand on this a fraction...

You could start by assuming a reasonable flow rate/velocity. You will then need or need some estimate of the pressure drop vs. flow rate curves for your waterblock and rad at least. Figure out the pressure drop. Estimate/measure the lengths of tubing and 90-degree bends in your system. Calculate the pressure drop through these items (numerous websites showing how to do this or will even do it for you). Also consider reservoirs, and barbs (as most barbs have sudden contractions/expansions) this causes pressure head loss, again find correlations and calculate from web/text references. Figure out/estimate other losses. Add up all head losses. Get pressure head vs flow rate graph for your pump. Find location on graph where the pressure you calculated intersects the curve, and read off the new corresponding flow rate/velocity.

Repeat this process until the assumed flow rate matches the flow rate you read from your pumps graph at the end of the process.

Possibly useful links (Maybe?)

This thread gives the easiest to follow basic method I've found. Obviously it makes some assumptions, but then so does any method (or you'd need to model it with a finite element / finite volume CFD code and even then you still make assumptions).
In terms of giving a fairly useful P vs Q curve for the system I think you might like this link:
Groth's page (Ignore the page title)
which is an interactive utility to give Pressure vs Flow curves depending on what waterblock you pick, what rad, what diameter and length of hose etc. It includes the 6002 and the WW. I'm not sure how accurate it all is but selecting a WW with no hose or rad gives data looking close to Cathar's, so hopefully its somewhere near...
Have fun...

Albigger said:

You could start by assuming a reasonable flow rate/velocity. You will then need or need some estimate of the pressure drop vs. flow rate curves for your waterblock and rad at least. Figure out the pressure drop. Estimate/measure the lengths of tubing and 90-degree bends in your system. Calculate the pressure drop through these items (numerous websites showing how to do this or will even do it for you). Also consider reservoirs, and barbs (as most barbs have sudden contractions/expansions) this causes pressure head loss, again find correlations and calculate from web/text references. Figure out/estimate other losses. Add up all head losses. Get pressure head vs flow rate graph for your pump. Find location on graph where the pressure you calculated intersects the curve, and read off the new corresponding flow rate/velocity.

Repeat this process until the assumed flow rate matches the flow rate you read from your pumps graph at the end of the process.

Click to expand...

Exactly!

Note that it is somewhat tedious to calculate the losses due to friction for each iteration, since the frictional loss coefficient (a.k.a. friction factor) is dependent on velocity.

The most popular way to calculate frictional losses is to use the Darcy-Weisbach equation, from the middle of the 19th century. It conveniently expresses the frictional loss in terms of the velocity head. The D-W friction factor is dependent on the fluid velocity, tubing diameter, and the kinematic viscosity of the fluid.

A convenient chart called the Moody diagram is used to determine the friction factor. It was developed during that late 19th century. The D-W friction factor can be read directly if we know the Reynolds number and relative roughness of the tubing. The Reynolds number is calculated using the assumed velocity for the current iteration.

Once the D-W friction factor has been determined, the frictional headloss in the length of tubing can be calculated using the D-W equation.

I'm sure I have either confused or totally lost all of you by this point.

I have not compared the D-W friction factor for typical water cooling system velocities. Perhaps just assuming a typical average D-W friction factor would be sufficient.

Won't this factor differ depending on type of tubing? If so - results can only be guessed when there are no measurements.

Ven0m said:

Won't this factor differ depending on type of tubing? If so - results can only be guessed when there are no measurements.

Click to expand...

Yes. The relative roughness is defined as the average roughness height in the conduit divided by the diameter. The average roughness height for plastic tubing can be found in manufacturers data or published charts on such things.

cpufan said:

Exactly!

Note that it is somewhat tedious to calculate the losses due to friction for each iteration, since the frictional loss coefficient (a.k.a. friction factor) is dependent on velocity.

The most popular way to calculate frictional losses is to use the Darcy-Weisbach equation, from the middle of the 19th century. It conveniently expresses the frictional loss in terms of the velocity head. The D-W friction factor is dependent on the fluid velocity, tubing diameter, and the kinematic viscosity of the fluid.

A convenient chart called the Moody diagram is used to determine the friction factor. It was developed during that late 19th century. The D-W friction factor can be read directly if we know the Reynolds number and relative roughness of the tubing. The Reynolds number is calculated using the assumed velocity for the current iteration.

Once the D-W friction factor has been determined, the frictional headloss in the length of tubing can be calculated using the D-W equation.

I'm sure I have either confused or totally lost all of you by this point.

Click to expand...

Very insightful. I do not have the background you seem to, but know the basic procedure.

It actually wouldn't be that bad with calculating the friction factors each iteration, just set up an excel spreadsheet. The pain would be as you said manually reading off friction factors from the moody chart and entering them in, and reading off the pump's curve as well. (I suppose the really ambitious could code up some points from the graphs, and perform a quadratic fit, but this begins to get tedious)...

cpufan said:

I have not compared the D-W friction factor for typical water cooling system velocities. Perhaps just assuming a typical average D-W friction factor would be sufficient.

Click to expand...

Assuming a typical friction factor may work. I would also assume (perhaps incorrectly) that many of the tubings in use have a fairly limited range of relative roughness, so that this number may be a typical average number.

EDIT: as noted though, would need to have experiment(s) to correlate to in order to have any real value.

Guys I think you are making this about ten times harder than it has to be. If you want to measure flow on your unique setup go to home depot get two five gallon buckets. Then time how long it takes to fill one (use the other for source) make your calculations and there you go.

turbotoy said:

Guys I think you are making this about ten times harder than it has to be. If you want to measure flow on your unique setup go to home depot get two five gallon buckets. Then time how long it takes to fill one (use the other for source) make your calculations and there you go.

Click to expand...

No one said it was hard to determine the flowrate on one unique setup. (Be careful if you do this to make sure the inlet/outlet are at the same elevation)

Also this is somewhat not representative of a closed loop system because in the arrangement you propose the outlet water comes to rest and the inlet water has to be accelerated wheres in a closed loop the energy wasted by the moving exit water would get recycled into the system. Etc... Etc...

However his question was to calculate which would be better in his system, if he had both blocks already of course he could just test them.

EXACTLY Albigger. I would not even need to BUY the bucket as I already have so many buckets. But, yea I don't have the blocks, this is a prelim on the purchase.