Size of pipe to achieve required flow rate? 1

[SL1000]

Hi please can you help me with this

I have two tanks being fed from a brine header (pump fed) which typically operates at 2 Bar and I am told to assume this is reasonably constant(however there are a multitude of take-offs from this and some wonderous piping arrangements!)

I will take actual pressure readings if required over a working day to establish the conditions.
If I trace the pipe work and note the various sizes of pipe back to the header how do I calculate the requirements to increase flow into both tanks to 100m3/hr or rather how would I go about calculating the piping needed to give me 200m3/hr as the line is split into the two tanks.

Some assumptions and basics may help here.

I will need to check but there seems to be a 5" line feeding the split which then goes to each tank with a 3" pipe. In each line is a 4" flow meter and a 3" butterfly valve.

I need to know if it possible to increase the size of the pipe and leave the instrumentation as is or will the valve need to increase in size?i.e. what is the effect of the valve pressure drop as a proportion of the flow restriction. The vavle is on the tank top in each case.

I would appreciate it if any replies started pretty simple for me to grasp the concepts and then build in any complexities as we go.
Thanks in advance

 

[parsec]

Sounds like you need a good fluids reference. Probably Crane Tech Paper 410 or Cameron Hydraulics would be a good place to start. You are on the right track. Since your system is pump fed, you also need the characteristic curve of the pump. Actual data is helpful, but you need to know how the pump will react to the extra flow. It may or may not be capable of putting out the extra flow. One easy way to tell is to determine if the butterfly valve at the top of the tank is throttled or is it full open. If it is throttled, you probably have some extra oomph in the pump.

Assuming the pump is OK, your next step is to calculate the system curve at the new flow rates.

The head available for pumping will be : Suction pressure (may be plus or minus)plus Total Dynamic Head of the pump at the specified flow rate minus the pressure in the tank. So in addition to the piping from the pump to the tank, you need to know what the suction piping and source look like. (i.e how long, what size, etc.)

Once you have all of this information, you can do a pressure drop calculation for the system, including suction and discharge. Be sure to tally all of the flows from the pump (if there is more than the two tanks). This will let you size the new piping.

A good rule of thumb is to keep your piping velocities to between 2 to 5 feet per second. This gives you a reasonable pressure drop. For 200M3/hr (about 733 GPM) this puts you in the range of 6 to 8 inch for the main and 4 to 5 inch for each branch. A butterfly valve is a low pressure drop device so you may be in luck and not have to replace it. Chances are it is currently throttled, in which case it is probably the major pressure drop in the system. They are probably used to balance the flow between the two tanks. They wioll still probably need to be throttled in the new configuration to achieve the proper balance. The flow meter may be OK, it depends on its sizing charecteristics.

Hope this helps

 

[SL1000]

Thanks parsec, some useful information here. I'll need to read this again when I get the chance, thanks for that.

One thing I didn't mention was that the valves are fully open during the filling cycle. The process is being used to mix filter cakes with brine, circulate until the required viscosity is achieved and then discharge to a bigger holding tank. The filter cakes are dumped to the two tanks from two filter presses.

The main problem is that the current flowrate achieved into the tanks is around 60m3/hr depending if both are on at the same time. I fear that whatever mods I make may result is me 'robbing' other lines taken from the brine header.
I am not sure how easy it will be to establish exactly what the 'other' flows from the system are. I may need to look at the other processes to see if I can establish some sort of limits or prioritise the flow for a given filling regime if this is a problem. Any thougts?

I am unsure without refencing the text you mentioned how to calculate the required pipe size for a given flow rate and given pressure. Can you suggest an appropriate formula if this is possible so I can get a feel for it (even if not really applicapable to this problem, I can then calculate given ideal conditions) or do I really need to start looking at pump curves? Another issue is different line sizes in the system, or do I just add up the frictional losses accordingly. I think I should be able to have a go at the two being in parrallel once I know a few more basics.

regards
Simon

 

[parsec]

Simon

I reread your question and I think I now understand more clearly. You brine header is at 2 Bar which can be assumed constant. That means forget about the pump curve. I initially thought your mixing tank was at 2 Bar. Well, if there is sufficient capacity in your header, and it appears that it may be difficult to tell otherwise, then the problem is simpler. You now know the pressure that feeds the tanks.

The equation is:

DeltaH=KV^2/2g
DeltaH is head differential (including pressure drop and elevation) in feet of liquid
g = gravitational constant
V Velocity (ft/s)
K dimensionless constant (equals f*L/d, which is friction factor times pipe length divided by pipe diameter)

Fittings and valves typically have established K's.
DeltaH would be 2Bar minus tank pressure (atmospheric?) so you have about 1 atmosphere as your driving force.

Ssee the reference material for the rest of the items.

It is probably a good idea to measure and plot header pressure over the course of a day to see fluctuations. It is clearly possible that you could rob other flows during the time when you are filling these tanks, unless the header is extremely large. As I suggested earlier, an appropriate pipe size would be about 6 to 8 inches for the combined and 4 to 5 inch for the split to each tank, depending on the length of piping and the number of fittings (i.e. - elbows, etc.). SO, unless the header is about twice that size, (16 or 20 inches) there may be trouble when you make this mod.

When you run the numbers, check the longer piping run to a tank. In order to achieve proper balance, you may need to throttle the butterfly valves during filling. Chances are that the tanks are pretty close to each other and the piping to each tank is probably pretty close in length and fittings so it should split evenly.

 

[SL1000]

Thanks again I'll be getting round to this in the next week or so, this info will no dooubt prove very useful and save me some time getting started. The header is of the order you have mentioned possibly 24".
The tanks are close together and each hold around 120m3.

cheers
Simon