Pump Sizing 2

[stuckoutthere]

Good Evening,

I have been trying to work out the size of pump required for an upcoming pressure test project. I followed the below but think the pressure drop I have calculated is far to large and is now a sticking point for me deciding on pump size. I just want to see if anyone on here can help me.

Details
Hose - Thermoplastic, 2" Bore
length - 15000m
Pressure required to be tested to - 10000psi
Test Fluid- Water Gycol, 20% @20c, density - 1030kg/m3, Dynamic Viscosity - 1.65x10^-3, Kinematic Viscosity - 1.6x10^-6m/s2
Flow velocity - 15m/s2
Flow rate - 38m^3/hr

1. Tried to find the roughness factor for the tubing. Could not find one for thermoplastic hose but found one for PVC(0.03MM)

2. Worked out the reynolds factor.

3. Calculated pressure drop. Calculated 807Bar

Have I went about this correctly or am I missing something. How do I go and select the correct pump size now? If someone can help me can you show how you do it(calculations/step by step) and then I can check and understand.

Thanks

 

[BigInch]

You are not letting anything out of the 15,000 m long pipe, so you don't need pipe flow equations or pipe roughness to fill a pipe or vessel. You need volume of pipe or vessel you will fill and the flowrate of the pump you will fill it with. Volume/flowrate = time to fill.

You don't need 807 bars to fill the pipeline, only to fill it at a rate of 38m3/hr. You can use a lower rated pump, and take more time to fill, but at the flowrate that your lower pressure pump will put into that 2" diameter pipe. I'd tend to want to break that 15,000 m into smaller segments, but you might have trouble reconnecting it. I don't know if breaks and reconnects are allowed in your scenario.

You will want to fill the pipeline with a pump with a high flow rate capacity to keep fill time reasonable. Then when full of water/glycol and you have removed all air, switch pumps to a high pressure, low volume pump to kick up pressure. It will rise rapidly, once the pipeline has been filled with water. Obviously you will need that pump for topping off to an ultimate pressure with a capability to reach 10,000 psi, or a head of around 2500 feet (if the 15,000 m long pipeline is horizontal), but it won't need much of a flowrate to do that final pressure kick.

I hate Windowz 8!!!!

 

[LittleInch]

You have a strange set of circumstances here. - a small diameter long distance pipe will transport very little fluid without needing a huge pressure drop. 15/sec - I assume its not m/sec2 - is very high and will generate very large differential pressure.

Where did 38 m3/hr come from? Looks quite precise, but if this is the regular flow rate your pipe is seriously undersized for the distance.

From what I can gather here, you're trying to fill this pipe with water / glycol to test it - correct?

The flowrate / velocity needs to be much lower - more like 3-5 m/sec. As you're filling it, presumably after sending a pig along what I guess is an undulating cross country line, the pressure drop increases every second as more liquid enters the line. therefore you will need either to pre-pressurise with air and bleed it off at the other end or control flow rate d/s the pump and gradually open your control valve to maintain a steady filling rate.

To be honest it sounds like you've bitten off a bit more than you can chew and pressure testing to 10,000 psig (I didn't know RTP went that high) is no laughing matter, but maybe its only the pump you don't know about.

So basically I would say no, you've got it about right, but the data you provide shows you will need a huge pressure to flow that volume when your pipe is full. Plastic pipe is abut 5 micron roughness usually - really very smooth.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[stuckoutthere]

Thanks for the answers.

I wouldn't say I have bitten off more than I can chew but I will admit I am still learning the finer details. But if your not learning your not doing anything.

The 15m/s^2 I gained from the spec sheet for the hose. It stated max flow velocity as that. I am willing to try and change all variables to achieve this. I have noticed though that I entered the wrong flow rate. That will have confused matters slightly but the formula I used was:

Flow rate = 1/4*Pi*(pipe diameter)^2*Velocity

The line has to be tested to x1.5 working pressure that is where the 10000Psi comes from. It will be pigged, then fluid pushed through before being capped and brought upto pressure.

Thanks

 

[LittleInch]

Ok, I don't mind teaching, but you need to look at what is being said and respond. 15m/sec is the max flow you can get in the hose without damage. However the consequences of that velocity are a huge differential pressure / unit length compared to a more normal 3-5 m/sec as the friction is more or less proportional to velocity squared (9 vs 225). Its a bit like saying the max speed of a downhill skier is 100 mph. However to maintain 100 mph for 15 minutes he needs a mountain 150,000 ft high....(I haven't checked the figures here, but I think you get the drift.)

Back to your issue.

The volume of your 15km of 2" ID hose is about 30m3.

3m/sec in said hose is 22m3/hr.

It therefore would take about 1 1/2 hours to fill your 15km at 3m/sec - sounds good to me. Try that for size and see what pressure you get when the line is nearly filled to size your filling pump and make sure you throttle the pump at the start or pressurise the pipe before filling to create an acceptable back pressure to stop the pump running away with itself....



My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[stuckoutthere]

LittleInch thanks for the responses.

I'll have a read over your post again, look at my calculations, re-calculate and I may be back to ask further questions but again thank you for your help.

 

[SNORGY]

I treated this as an interesting diversion from my own real work...but I will claim it as 2 PDH hours in the "Contribution To Knowledge" category.

Since you describe pushing a pigto ensure flooded piping throughout its length as you fill behind the pig, that answers my initial concern that you can't (for a strictly horizontal or a downward-sloping line) assume that you will have a flow profile necessary for Darcy-Wiesbach to hold true. However, if you can neglect the added force to push the pig ahead of the filling fluid (which I believe you can, after some distance far less than 15 km), then you are able to apply it.

The above stated, I made some simplifying assumptions:

(1) Constant speed, constant volumetric flow rate, positive displacement pump with a pumping efficiency of 80%, all applicable throughout the entire filling cycle.
(2) Towards the end of the filling process (which I took as the end of Kilometer #15) is where the frictional losses and, hence, pumping pressure at Kilometer #0, are maximized.
(3) I assumed that I would ultimately be limited on either filling pressure or rate or power, so with Assumption #1 in mind, I arbitrarily selected a filling pressure of 1000 psi = 6895 kPa.

Using the data provided, I derived an expression for the fill rate Q by combining the Darcy-Wiesbach equation and the Colebrook-White equation; this allowed the friction factor f to drop out completely and yield an expression for Q as a function of filling pressure P. Now knowing Q and P, I could compute the pumping power and the filling time. I set an upper limit on the pumping power, which controlled the filling rate (again referring to Assumption #1). I concluded that you could fill the pipeline in abot 3 hours with a 32 bhp (24 kW) pump at a maximum pump filling pressure of 1000 psi (6895 kPa).

I then checked these results by backing out the friction factor from Colebrook-White (I used Goudar-Sonnad, which is close enough and spared me the iteration), plugging it into Darcy-Wiesbach and calculating the friction losses, and things appeared to be correct. I then checked what the head loss was for water in 2" SCH 40 pipe at 45 USGPM (the calculated filling rate) in the Cameron Hydraulic Data Book, which reports 3.82 ft per 100 ft, so for 49213 ft of pipe the head loss for water would be 1880 ft, corresponding to a pressure drop of 814 psi. From this, I concluded that my results feel like they are in the ball park.

So, my approach for a problem like this would be to choose a filling pump power and pressure and then calculate (and accept) the filling time it provides. You can run several scenarios to optimize time with power, with pipe length, with pressure, etc. but, after all of this,the approach by LittleInch was right all along.

 

[SNORGY]

As was the approach by BigInch...

 

[SNORGY]

And my spelling / typing of "pigto" and "abot" (perhaps others...) are not what I'd like, but in fairness, I broke my hand riding a horse last weekend.