Natural gas pipeline velocity

[RGME]

Hello all,

I have a situation where the mainline block valve in a 36" natural gas pipeline is closed for a period of time. During this time, the pressures on either side of the valve drift apart. e.g. Upstream pressure = 800psig. Downstream = 600psig. If the mainline valve is now opened what is the maximum gas flow rate across the valve? What would be the best equation to use to estimate this?

Also, what is a good rule of thumb to use for maximum allowable gas velocities in the pipeline? I have read in many cases that 100ft/sec should not be exceeded, but I believe this is to avoid excessive noise. Is this true?

Thanks

 

[zdas04]

With 200 psi dP across the valve, you are a long way from choked flow, and the transient velocities should all be in the manageable range. I wouldn't be concerned about that.

It seems to me that there are as many "rules of thumb" for velocity in pipelines as there are engineers.

The rule that makes the most sense to me is to minimize pressure drop because it is expensive to boost the pressure back up to recover the pressure lost.

In lower pressures (under abourt 300 psig) I usually go with a design range frm 11 to 100 ft/sec, and then calculate the expected dP/mile--for pipes under 12-inch I like to stay under 15 psi/mile. For 20-inch and bigger the number is usually closer to 5 psi/mile. These numbers come from the cost of compression vs. the cost of running bigger pipe.

For mainline kinds of pressure I like to stay under 5 psi/mile for any size pipe.


David Simpson, PE
MuleShoe Engineering

http://www.muleshoe-eng.com

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

The harder I work, the luckier I seem

 

[sailoday28]

Upstream pressure = 800psig. Downstream = 600psig. If the mainline valve is now opened what is the maximum gas flow rate across the valve?

If there are long lengths of pipe up and downstream of the valve, then fast or slow opening of the valve will not come into play during the initial transient.

Assuming the gas composition is the same on both sides of the valve, up and down stream temperatures will affect the initial flow rate.

If the natural gas can be considered a perfect gas with constant specific heats, then a method of characteristics analysis will yield the flow.

 

[RGME]

Thanks for the responses,
I am not concerned about the effect of velocity on the pipeline. However, there is an inline flowmeter close to the mainline valve (annubar) with an upper limit on flow. I do not want the transient flows to damage the flowmeter. Do either of you know which flow formula best represents this situation? Thanks.

 

[sailoday28]

As a first approximation, assuming perfect gas, constant specific heats- Treat both the initail high pressure and low pressure gases as slugs.
I would assume a sudden opening and that the higher pressure gas expands isentropically into the lower pressure side of the valve.
The interface of the slugs is at the same velocity and pressure.

A denotes sound speed U velocity
if interface
lif left side of interface
rif right side of interface
o initial condition

Using the method of characteristics moc), the high pressure gas on the left side depressurizes as follows

2/(gamma-1)*Aohp=2/(gamma-1)Arif + Uif (1)
1= Alif/Aohp + (gamma-1)/2 * Uif/Aohp (1a) moc

Similarly the low pressure slug
2/(gamma-1)*Aolp=2/(gamma-1)Alif - Uif (2)

1= Alif/Aolp - (gamma-1)/2 * Uif/Aolp (2a) moc
Pressure at the interface is determined from

ASSUMING BOTH SLUGS AT SAME INITAIL TEMP, Aolp=Aohp
Adding equations 1a and 2a, the velocity portion drops out.

2= Alif/Aolp + Arif/Aohp (3)

Alif not equal to Arif but Plif=Prif

(Prif/Pohp)=(Arif/Aohp)^g (4)
where g=(gamma-1)/(2gamma)

(Plif/Polp)=(Alif/Aohp)^g (5)
where Prif=Plif (6)

RESULTS in 2 = (Pif/615)^(1/g) + (Pif/815)^(1/g)

Solve for Pif resulting from fast opening.
In turn using above equations, velocity at the slug interface may be obtained.

Remember, this is a simplified wave analysis.

Regards

 

[sailoday28]

CORRECTION
With reference to my simlification of g in equation 4

g= 2gamma/(gamma-1)

Equatin 4 comes from the isentropic relation
P/Po=(T/To)^[gamma/(gamma-1)]
And the relation of sound speed to temperature
T/To= (A/Ao)^2

Sorry

 

[BigInch]

You can find flow across any valve by using the

Q = 963 * Cv * ((Pin^2 - Pout^2)/G/T)^0.5

Q = Standard CFH
Cv = valve flow coefficient at a given percent open
(get this from the valve manufacturer or from standard approximate curves for your type of valve, ball, gate, etc.)
Pin = upstream pressure psia
Pout = downstream pressure psia
G = Specific Gravity (ref: Air = 1)

Do not use a ball valve if you think transient flows will be a problem, use a plug or globe type. Check the Cv vs % Open curve with the speed of valve opening. If manual, its likely not a problem.

Velocities for short distances across a valve can be much higher than the average velocities out on the pipeline, no need to limit them to such low values. 50-75 ft/sec should be alright, as long as you can stand the pressure drop across the valve and noise.

 

[zdas04]

BigInch,
Why would you say that plug valves are better suited to transient flow than ball valves? When I look at a flow-percent vs. open-percent curve for either one they look quite similar and ball valves are lower maintenance. I've found that for applications that truly need throttling it is best to avoid any of the inherently on-off valves (i.e., ball, gate, and plug).

David

 

[sailoday28]

I assume that the 36" valve is presently installed.

 

[BigInch]

I think you will find that a plug valve (usually) has a different number of turns to reach a given stem position resulting in a given %open than a ball valve does. Although the Cv-%Open curves appear similar, Cv vs Stem position (number of turns) are different. Thus, ball valves will usually open/close too fast for a given actuator travel or number of turns on a handwheel. Since most flow & pressure change occurs over the range of 0 to 3% position, the very small actuator motion needed to do this with a ball valve will in fact create more transients than a corresponding plug. Fast acting ball valves are a primary cause of transients, as they deliver alot of both flow and pressure. The shutoff region pressure changes do not concern me too much, because although pressures may change rapidly for a millisecond, not much corresponding flow sustains the wave, so the wave is quickly dissipated. In any case, RGME made no mention of his need for throttling service, so I did not recommend a valve more suitable for a throttling control application. Valves that reach high flow capacity quickly are the ones that cause transients. I also assume that no throttling is needed at that location, as he mentions the size is 36" and I don't recall ever actually seeing a 36" control valve. I think I have seen a 24" control valve and that was an absolute monster. I think I reasonably assumed that he needs only a valve for on/off & tight shut-off service. I have never heard of an aversion to tight closure when trying to control transients.

Even if we were talking about a control or throttling application, since control valves are never assumed to completely close, they are always backed up by at least one completely closing block valve, if not one completely closing block valve on each side of the control valve, so effectively you will wind up with a unit installation that can be completely closed off in whatever case you choose to look at. Transient design does not seek to never allow a pipeline to be closed, in fact usually the opposite; to control the transients at complete closure/opening. Thus, I do not understand your aversion to using valves that close completely when considering transient flows. Complete closure is usually the end result (and critical case) that you must design for anyway.

 

[WebWizz]

Further to the recommended velocities: try

http://www.engineeringpage.com

> line sizing

Then choose process fluid (or utilities if applicable), for this line sizing calculation a combo box choice can be made - if you take natural gas the program shows a recommended velocity of 30 m/s (appr 98 ft/s) in the results.

Also handy for other fluids.

Regards.

 

[Ashereng]

The OP stated mainline block valves. I too assumed this was on/off service.

Another option for 36" gas pipeline service is slab gate/through conduit valve. We have installed several on our pipeline project, as well as ball valves. Some manual, most actuated.

36" is a big valve for me. I do not see many of these.

With respect to transients caused by ball valves, we have found that this may be addressed adequately with a good positioner.

"Do not worry about your problems with mathematics, I assure you mine are far greater."
Albert Einstein
Have you read FAQ731-376 to make the best use of Eng-Tips Forums?

 

[RGME]

The 36" valve is used for on/of service. Currently it takes approx 40secs to travel from closed to open. I have calculated that if I open the valve against a 100psi differential that the initial transient flow rate will be very high (1.2 BCFH) which is much higher than my flowmeter annubar can handle (139MMSCFH). Looks like we'll have to remove/replace the flowmeter. thanks for all the input.

 

[Ashereng]

RGME,

On large valve stations, there is usually a bypass. Do you not used that initially before opening the main valve?

"Do not worry about your problems with mathematics, I assure you mine are far greater."
Albert Einstein
Have you read FAQ731-376 to make the best use of Eng-Tips Forums?

 

[RGME]

Ashereng,
You are correct. There is a 16" bypass. The bypass valves are manually operated and this is an unmanned facility. The mainline valve is automated. I am trying to determine the largest pressure differential against which the 36" valve may be opened. Looks like it's not going to be very high.

 

[BigInch]

The largest dP you can open the valve on will depend on the max torque delivery of the actuator. On a valve this size, have the actuator assembled to the valve at the mfgr shop and test it for satisfactory operating torque at max dP.

I think you might want to install a smaller valve in front of your meter that will give you some control capability. Normally meter installations have a pressure control valve either up or downstream of the meters themselves. Meters don't like to be exposed to transients and tend to be inaccurate if the flow varies too much.

 

[sailoday28]

RGME (Mechanical)
the initial transient flow rate will be very high (1.2 BCFH)

How did you calculate the initial flow rate?

Regards

 

[Ashereng]

RGME,

Can you automate the bypass valve, and use it to equalise the pressure prior to opening the 36" main?

"Do not worry about your problems with mathematics, I assure you mine are far greater."
Albert Einstein
Have you read FAQ731-376 to make the best use of Eng-Tips Forums?

 

[RGME]

sailoday28,
I assumed two gas reservoirs at different pressures (800psig and 700psig) connected by a 100ft long 36" pipeline and applied the Darcy formula to calculate the flow rate. I understand this may not have been the correct approach to take but I thought it would get me into the ballpark. Right now I'm having someone run a transient analysis to check my number.

Ashereng,
The 2006 budget will not cover this, but maybe 2007....

 

[green1959]

I have worked with a client in the natural gas industry who limited gas line velocities to a max of 35 m/s purely to prevent corrosion inhibitor being stripped off the internal surfaces.

This fits with the 30 m/s mentioned in a previous post.