How to Calculate Flow Rate When Withdrawing Product From a Fixed Pressure Source 2

[MechEng176]

Hi folks,

I'm working on an application where I'm tying into a live pipeline and we're diverting some of the flow to storage tanks. The flow from the pipeline goes through a metering skid which has all the bells and whistles like PSVs going to closed drain drums, strainers, check valves, PD flow meters, orifice plates/globe valves, etc. before entering the tanks.

Since we're pulling very little of the mainline flow, the pipelines pressure won't drop. I'm looking for a relation where I use the mainlines pressure (and possibly flow rate) as well as the ID (4 inches) of the pipe connecting the mainline to the metering skid to determine the flow rate at which the product will be leaving the mainline.


Thanks again!

 

[msquared48]

Well, the Bernoulli Equation, and Q = VA, and Q = Q1 + Q2

Mike McCann
MMC Engineering

 

[MechEng176]

the thing is, we need a way to verify that the metering skid will see Xm3/min flow rate when the mainline is running at 500psi BEFORE we go to site. We have to purchase and build a lot of different spools and changeover to connect the mainline to the skid to the tanks. If the 3'' pipe we currently have won't be sufficient, then we will have to get 4'' piping. And we need to know this well in advance so we can show up to site with the proper equipment.

 

[MechEng176]

I think I may be on to something here but just not sure about the way I went about doing it:

We have a pipeline that is transporting oil and it should be at around 300psig. We are taking oil out of the mainline and holding it in storage tanks, the majority of the pipeline flow will still continue downstream in the pipe, but some will flow towards the tank because of the path of the pipe leading to the tank is at 0psig (I can neglect head pressure because the 300psig that the line is at is more than what the head pressure will be in a 7m high tank). The liquid is oil, the density is around 800kg/m3.

So in my case, using Bernoullis I'll have

Ppipeline = Patm + 0.5*Density*V^2

So V = SQRT(Ppipeline - Patm) / (0.5*Density)

V = (2,070kPag - 103kPag) / (400)
V = 4.9175m/s
Q = 4.9175m/s x Area

where Area is Area of a 4'' Diameter circle in meters to get Q in m3/s.

Q = 4.9175m/s x 0.00811m2
Q = 0.0399m3/s
Q = 2.4m3/min


hmmm... I was actually looking for a flow rate of 3.5m3/min. Really hope I made a mistake and got a conservative estimate.

 

[LittleInch]

I don't think the formula you're using is applicable. If you have 300 psig in your line, then your potential flow rate through a 4" hole to effectively 0 psig on the other side is enormous, far far more than 2.4 m3/min

The key to your potential flow is in the middle of your phrase "..orifice plates/globe valves, etc.." These items will introduce pressure drops in your system, as well as the pressure drop in your 4" pipe that will then equate overall to your available pressure drop (approx. 300 psig) at the flow rate you want.

3.5m3/min = 210 m3/hr. That's rather fast for a 4" line (~ 7m/sec), never mind a 3" line. This must be a huge oil line if 210 m3/hr classifies as "very little of the mainline flow"....

I would go for a 6" line myself, but to work out how much flow you're going to get you need to work out the pressure drop coming form the piping, orifice plate and globe valves. Also make sure your meter can handle the flow rate.

You say 500 psi in one post and 300 psi in another. Which is it?

Connecting this sort of thing to a pipeline is no simple thing and I would ask what sort of control and shut off arrangement exists on the tanks you're filling and what sort of pressure control system you're using. Tanks and their piping are normally designed for much lower pressures than 300 psig.

Has this design been checked and approved?


My motto: Learn something new every day

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

 

[davefitz]

The system you have described should have a flow measuring element- it will not add too much cost to the system already described. The lowest cost such device may be an "elbow tap", but many other options are available.

"Whom the gods would destroy, they first make mad "

 

[MechEng176]

The mainline is actually 36'' and we are pulling 10% of it's flowrate to fill one 400bbl frac tank, I guess that isn't very little of the mainline flow. We're doing this as a test to see if the pipelines control room will pick up a leak. The hard pipe from the mainline to the metering skid, and metering skid to the tanks is all 4''. The metering skid itself is all 6'' and has an automated valve downstream of the flow meter that opens according to how much flowrate I tell the on-skid AccuLoad system that I want. In the beginning of the test, I will tell the system that I want 3.5m3/min flow for a total of 50m3. Then the valves downstream of the flow meter will open until the flow is 3.5m3/min. After that valve is the globe valve to drop the pressure so that when it goes into the pipe going into the tanks, the pressure will be lower.

I know how to calculate piping friction drop, in this case I'll simply estimate 30m of 4'' with 5x 90 degree bends. But how would I go about calculating the pressure drop through a globe valve? We got rid of the orifice plate because adjusting a globe valve is much easier than replacing an orifice plate.

This specific line that we're tying into has an average operating pressure of around 300psi. The 500psi was for another line but we were doing pretty much the same thing. The design of the metering skid has been approved and used before and the 6'' PD flow meter on it goes to around 230m3/hr. The skid itself has been used before and is rated to 9,930kPag, that's when the onskid PSVs pop to the closed drain drum on the skid. The skid also has automatic shutoff inlet/outlet valves if it sees that the pressure is getting too high. The tanks are open to ATM and all the piping connecting the mainline to the skid and skid to tanks is high pressure treating iron (the piping the use for fraccing wells) and is good for 10,000psig. The weakest link are the tanks which have a 1lb PVSV as well as a vent running down the side so they are basically at ATM pressure and I don't really see a way they could over pressure.

 

[LittleInch]

So if I've got this right, you have a 4" tap from your main line to a metering / flow control skid comprising filter, PD meter good for 230m3/hr and a flow control valve with presumably instrumentation to control to a set flow. D/S of this you have a manual globe valve.

I don't understand the purpose of this manual valve. the flow control valve will control on flow and given you have basically an open ended pipe into your tank, the pressure D/s of this valve will be essentially atmospheric plus your pipe friction losses which won't amount to much, even at 7 m/sec.

I'm guessing here, but I assume what you're doing with the globe valve is gradually increasing and decreasing flow at the start and stop of your experiment instead fo using the control valve. I would suggest you add a manual 100% open to close dial on your control valve and send the signal through your control system as a secondary input with the flow set point control into a low selector block. The control system takes the lowest reading and sends that to the valve positioner. You can then open the valve slowly until the flow control takes over.

If you really want to leave it as it is, then you should be able to obtain a Valve CV versus per cent open chart for your valve and use that. There are many equations which will give you pressure drop versus flow for a set Cv figure. either way the aim should be to open up your globe valve to maxim and let the main control valve control on flow.

You need to check the flow conditions of your main control valve to ensure it is OK for this pressure drop and flow, but so long as it's Ok, I wouldn't bother with a second control valve (your globe valve).

your flow in the 4" pipe is a bit high for a tank entry. I would slow that down to about 3 m/sec max otherwise you risk erosion damage in the tank and potential for high static charge.

My motto: Learn something new every day

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

 

[BigInch]

If you have infinite supply upstream at a fixed pressure, the hydraulic boundry conditions you need to solve the problem switches to the downstream end and you have to figure out what flow you can get through the downstream stuff given the upstream pressure.

I hate Windowz 8!!!!

 

[MechEng176]

LittleInch,

Here's the story behind the globe valve. First, there used to be an orifice plate on the skids exit to the tanks. Basically, orifice plate followed by 6feet of the skids 6'' piping, then a 6'' onskid valve which had the changeover to the temporary hard pipe (hammer union fraccing/treating iron). That temporary hard pipe went to an 8'' tank manifold (MAWP of 1,440psig) from where it was connected with 4'' ANSI 300 hose (double steel braided, 300# flanges welded on, 1 fixed flange and 1 floating flange) to the tanks. The skids piping is 6'' and the orifice plate that came with the skid had a 1'' hole. When the flow through that got too high (1.5m3/min), the 6feet of pipe D/S of the orifice plate started screeching and shaking and it was so bad I could hear cavitation when I put a metal screwdriver on the pipe next to my ear.

To stop the 6feet of pipe from howling, we changed out the 1'' orifice plate for a 2'' one. Now we've swapped out the orifice plate for a globe valve. The glove valve is not there to to control or apply back pressure to the flow meters in any way, the green Fisher control valves that run on instrumentation air D/S of the flow meters are used for that, they slowly open and when the flow meters register the set flow rate they stay whatever XX% open they are. This way there is nothing manual about the process, I just enter the flow rate I want, then hit start and the control valves start opening. It's pretty neat actually, only thing is, we used N2 bottles for instrumentation air and the regulator on the skid that we hook up the bottles to has a slow flow rate so the valves open really slowly. If I understand the designers logic, he said all we need the orifice plate (now glove valve) is to drop the pressure because the fluid is now going to tankage. I placed a gauge on the tank manifold during the last test and when the mainline was at 500psig, the skid was also at 500psig (this reading was taken U/S of the flowmeter), and at 1m3/min flow rate the tank manifold only had 20kPag (3psig)of pressure.

Thanks for the tip on the flow speed into the tanks through 4''. Never even thought of the spark potential. The tanks are all bonded and grounded and all have 2 vents and 1 PVSV, we inject N2 into one vent to keep the tanks inert. This gas then leaves through the second vent to a scrubber unit. But it never hurts to be too safe. From what I've been told, we don't need the scrubber unit by law or anything but sometimes nearby residents on the pipelines Right of Way complain.


BigInch,

Thanks for that tip. I used a friction drop calculator (

http://imgur.com/VTs9MKV)

to determine the pressure loss through a 4'' line at the 3.5m3/min flow rate. Since the pressure drop is 74psig, and less than the lines pressure, I'm all good.

Also, this can't be ignored any longer, are you two (LittleInch and BigInch) related? Like Junior and Senior or something like that?


Thanks again.

 

[BigInch]

Mini-me


Learn from the mistakes of others. You don't have time to make them all yourself.

 

[LittleInch]

Ha Ha.

Dealing with your first point - I think the orifice and your globe valve are not required and you should / could just remove them.

The difficulty is that sometime people get confused by the purpose of a control valve. All a control valve can do is create a pressure drop across it when a fluid is passing through it. That's all. It has an effect on flow because it introduces a spot variable change in pressure which, when added up with all the other fixed pressure losses due to a certain flow velocity (pipe, other mechanical fittings etc) means that the overall system flow can change without changing the fixed elements ( you wouldn't want to start introducing extra lengths of pipe to decrease flow when you can use a control valve instead). In your case you and presumably the designer before you referred to this as a "FLOW Control Valve". Wrong. It's just a control valve which happens to control on flow, but could control on pressure or any other input.

If you remove the globe valve, the control valve which is controlling on flow will still control on flow, but would simply have a larger differential pressure across it. Unless it is not sized for it this is perfectly OK. Your globe valves / orifice plate are only acting to decrease the pressure drop across the control valve. As the flow rises, your globe valve will gradually increase its differential pressure and will, in the end, fix the upper flow limit with your control valve on the skid wide open. Maybe that's why the orifice plate was put there, but in terms of reducing pressure for the tanks, the (flow) control valve will do that for you. Basically your globe valve is a waste of space and could limit your max flow.

Hope that made sense.

Flow velocity into tanks can be an issues, and more so when anyone tries to dip a tank or hasn't properly earthed and connected each element attached to and within the tank.

your other question - Me and BI go back a long way, but we're not brothers. He got me hooked into eng tips about a year ago and as I'm a bit cr*p at picking usernames just "borrowed" his. I know a lot, but he knows more....

My motto: Learn something new every day

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