JT Effect in Gas Pipelines 2

[CSmith]

Does pressure drop due to friction in a gas pipeline result in a Joule-Thompson temperature drop?

I believe it does, but my colleague disagrees. I'm asking because he has a very accurate gas pipeline modelling spreadsheet and isn't sure if JT effect should be included in it or not. Expert opinions to resolve the matter would be appreciated!

Cheers
Clinton

 

[TD2K]

There is a JT effect with dP in a pipeline but it's also affected by heat input (or losses) to the surroundings.

 

[quadswift]

Is your question as simple as it seems or do you have some kind of a problem?

My understanding is that there has to be throttling or an orifice for the J-T effect to occur.

More temperature will be lost due to the water temperature affecting the product on a sub-sea pipeline.

There are systems where leaks can be detected on gas pipelines using temperature sensors to identify cold spots due to the J-T effect at leak points. Again this uses the pressure through the orifice (leak site) principle.

 

[denniskb]

Clinton forwarded me this thread as I am the colleague that asked him the difficult question so let me expand on it.

Every time I find reference to JT cooling, including text books, there is a discontinuity such as a control valve, orifice or membrane associated with it. (Do a Google search and see this for yourself)

Detailed pipeline simulation theory, however, includes JT cooling, as well as heat transfer to the surroundings, yet there is clearly no dicontinuity here.

All of this is built into my detailed pipeline modelling spreadsheet and has been tested in the real world against operation pipelines and serious money simulation software and it works very well.

The challenge is to explain "why" JT cooling and not adiabatic expansion is appropriate to pipelines.

If JT cooling is appropriate, then, when I deal with gas expansion and compression (e.g. compressor station blowdown and pressurising) should I use the JT effect or adiabatic theory to determine the resulting temperatures? The difference can be very significant when dealing with pipeline operating pressures of 15 MPa (2200 psi).

Probably the experts have it right but I would like to understand why. Dennis Kirk Engineering

http://www.ozemail.com.au/~denniskb

 

[zdas04]

I've looked for J-T cooling effects due to the friction dP along pipelines and if it is there it is far too subtle for normal oil-field instruments. In fact, the only times I've ever been able to measure J-T cooling was in choked flow (i.e., sonic flow through a nozzle, orifice, relief valve, or choke - where the upstream pressure was more than about twice downstream pressure). Ditto for heating due to friction forces--maybe the two factors are so close that they cancel each other out?

Whether the flow is adiabatic or not is a whole other question. I have checked temps on pipelines many times and always get an approach to soil temp after a reasonably short (meters, not kilometers) thermal entry length. That heat transfer is far too rapid for an adiabatic assumption to be valid until you have passed the thermal entry length. Normally when you are concerned about dT effects it is because of a big difference (e.g., 150F out of a compressor going into -10F ambient and then into 50F ground temp). After a reasonably short distance underground (certainly less than a km) you'll reach an equilibribum where any temp change due to J-T cooling, heat of friction, or the body temperature of microbes will be lost to the heat sink that is the earth.

 

[ggordonstewart]

Heat transfer in pipelines can be very complicated due to the variance in the soil that occurs along the length of the pipeline. If you are trying to determine the distance between line heaters so that the temperature of the gas never falls below the hydrate temperature, then you have be be pessimistic in your estimation of soil type and pipeline depth. If the pipeline is long enough, the gas will eventually reach soil ambient temperature. As the pressure drop per unit length is small, the Joule-Thompson cooling will also be small and will usually be hidden by the heat loss to the soil. Since the difference between dry sandy soil and wet clay (for example) is so great, it is hard to find enough accurate soil information along the total length of the pipeline to accurately predict temperature change or its source. G. Gordon Stewart, P.Eng.
Gas & Oil Process Engineering Consultant

http://www.ggordonstewart.com/

ggstewar@telusplanet.net

 

[Jepster]

JT effects in "typical" gas pipelines is negligable compared to the mass heat transfer to the surroundings especially with connate water (saturated gas) or included formation water production. The only real places you will see this effect is where you take a decent pressure drop in the system or if you create a false oriface such as a forming hydrate plug.

Jep Bracey
Flow Assurance Specialist
Marathon Oil

 

[malancha]

We have recently carried out subsea pipeline depressuring studies (through the blow down valve attached to the pipeline) using OLGA and JT effect has been found to be present not only downstream of the blow down valve but also on the upstream side of the valve (in the pipeline). I have checked the basic energy equation along the pipeline and JT effect can be present in the pipeline even without a restriction. So theoratically J-T effect should be considered for accurate pipeline temperature prediction. I have been given to understand that such effects have also been found to be present in long distance gas pipelines e.g. Zee pipe in Europe.
I would like to listen further from experts.

 

[Jepster]

Malancha,
You are correct about it occuring at certain localized zones in a pipeline during blowdown. This is especially true of long subsea tiebacks and how the blowdown is controlled (1 inch versus a 2 inch choke). What happens is that as the fluids go 2 phase you can get a localized JT effect and some of this is also driven by the hydrostatic head of the fluids during blowdown causing delta P different than at the base of risers.

Jep Bracey
Flow Assurance specialist deepwater development
Marathon Oil Co.
jtbracey@marathonoil.com

 

[denniskb]

I have been following the responses to this question with interest and thought I needed to get back into the fray to try and clear up some things.

Most of my work on gas pipelines has been with long onshore transmission pipelines in Australia carrying dry gas so I had not anticpated the responses about very low ground temperatures, offshore pipelines and wet gas but they do add an extra dimension to the topic.

My pipeline models include the JT cooling effect, as all good pipeline models do, and the results match the major pipeline modelling software packages and also very closely match real operating pipelines so the cooling is definitely real and can be accurately predicted.

JT cooling does matter on long pipelines where the gas temperature can fall up to 10degC below the ground temperature and this is despite the heat transfer through the pipe wall.

With reference to the response from zdas04 my experience is that the gas approach to soil temperature usually takes around 30 to 50km and if the pipeline is working really hard up to 120 km. I'm sure this happens much quicker with offshore wet gas pipelines.

malancha's response is very interesting and I think I can clear up a few things here. I believe that the cooling effect upstream of the valve is mainly due to adiabatic expansion and not JT cooling (though this will be present as well). Consider a pipeline 10km long staring at 5000kPa and depressuring it to 500kPa. At the end of the blowdown the pipeline contains only the gas that was in the 1km at the far end of the pipeline. Remember your basic thermodynamic studies, if you take a volume of gas in a container and expand it to a larger volume and lower pressure it cools down - if it happens reasonably quickly then this is adiabatic expansion. You need to be very careful in calculating the JT cooling effect across the valve as the gas arriving at the valve has already dropped significantly so the two cooling effects are additive. You can get some extremely cold temperatures from this when depressurising compressor stations.

Jepster adds yet another temperature effect. I'm sure that the localised cooling is a result of liquid condensing in the pipeline and running down to low spots where it then evaporates back into the gas flow. The cooling is the result of the vaporisation energy being drawn out of the liquid. I have seen this before when we vaccuum dry pipelines and water pockets in the bottom of valve bodies can freeze (even in our hot ambient temperatures) and potentially fracture the valve.

OK so now I want to restate my original question which is "WHY JT COOLING AND NOT ADIABATIC COOLING" as there is clearly no flow discontinuity that is normally associated with JT?

I am putting together a well documented calculation paper on gas pipeline flow modelling (steady state only) which tries to pull together and explain all of the competing effects on pressure and temperature and will make the paper available when it is finished. Hopefully it will make the topic easier to understand for those that are new to the field or those that need a better understanding of it. Any further suggestions are welcome - especially an answer to my still outstanding question "JT or not JT".
Dennis Kirk Engineering

http://www.ozemail.com.au/~denniskb

 

[malancha]

denniskb,
one way to find if the JT effect is existing is to use some compositional model with a Z-factor of 1 which will make it an ideal gas. As we know that JT coefficient for ideal gas is zero, comparing the temperature profiles predicted for similar gas compositions but with different Z-values can provide an idea if JT effect contributes to the cooling effect. Probably this can be done by changing inert/non-condensable component compositions.
Unfortunately packages like OLGA does not offer an opportunity to try things in this way. If you are developing a new model, this technique can be tried. Further I will be interested to receive your paper when ready.
Thanks.
A.Bhattacharya
Principal Process Engineer
Fluor Limited
E-mail; abq@india.com

 

[jcaiken]

Pardon my ignorance but I understood that JT cooling was the consequence of adiabatic (constant enthalpy or isenthalpic) expansion.

When depressurising pipelines don't forget the effect of the work done in accelerating the gas from the stagnant end to the depressurisation valve. This results in a measure of entropic cooling.

 

[vzeos]

I believe jcaiken is correct, the JT effect refers to an isenthalpic process, i.e., a process that is completely irreversible in which there is no external work performed by the fluid. The frictional pressure loss in a pipeline is isenthalpic as is the pressure loss through regulators, valves and fittings. Check out a Mollier diagram for natural gas or methane, the isenthalpic temperature drop associated with the pressure drop is independent of the pressure reduction element.