Pipeline Blowdown Thermodynamics

[denniskb]

Wow I just came across thread124-202829. I'm not sure if I am glad I missed it or not but I wish to add to the subject and tie in an earlier posting of mine regarding why JT cooling is used in long compressible pipeline flow calculations "JT or not JT".

I take the approach of dividing a problem into its components, solving each then summing the results together to obtain the final outcome.

Pipeline flow and blowdown are not that different so lets consider blowdown and to make things easier lets assume that the pipeline is insulated from its environment so it consists of the contents and the pipe only.

First, when the pipe contents blowdown the thermodynamic change in the fluid inside the pipeline is basically isentropic (so 100% isentropic efficiency). This is a very significant contributor to temperature change.

Next, since the fluid is moving there is also friction which generates heat energy and warms the gas inside the pipeline. This happens little at the far end of the pipeline where the velocity is low and more at the discharge end. This is typically a small contributor to change.

Then, there will be heat transfer between the fluid inside the pipeline and the steel pipe wall. The thermal mass of the steel is typically high compared to the contents so if the blowdown is slow enough this can significantly warm the gas and cool the pipe.

This takes care of the changes inside the pipeline. Despite the large number of variables and the time and distance affects the calculation methods to solve the above are all available. Note that I did not say they were easy because this is not at all simple to solve.

Be warned that while the pipe temperature drop and the falling pressure stress may not make normal brittle failure a risk it can easily get cold enough for a crack to propogate along the pipe length without arresting itself.

For now lets just assume that the pipeline contents get pretty cold.

The next issue is the blowdown valve and vent piping.

First, consider the tyically short branch pipe up to the blowdown valve. Since this is usually a much smaller size than the pipeline the fluid velocity is high and so is the heat transfer and also the thermal mass of the pipe steel is now low compared to the fluid passing through the pipe. This pipe can be expected to quickly reach the temperature of the fluid exiting the pipeline.

Next, the blowdown valve has a large pressure drop, typically choked flow for most of the blowdown time and so high friction. The resulting thermodynamic change is isenthalpic (or JT and so 0% isentropic efficiency - see Aspen HYSIS note on other posting). But remember that the fluid entering the valve is already cold and now it gets even colder. Now brittle fracture of the pipe steel may be a real risk particularly since severe vibration may also be present.

Last, after the valve one of my initial assumptions becomes a problem. If the vent piping is still assumed to be insulated from its environment then the whole vent system will fall to this now very cold temperature as again the velocity is high and the thermal mass low. In fact this section is most likely to pick up heat from the surrounding air and warm the pipe up as you move further away from the valve.

Again the calculation methods for this are all available though again not easy to solve.

I have been looking at this overall blowdown problem for years and assembling the components necessary for a reasonable solution as I see many engineers struggle with the problem. My main driver is the very real risk of brittle fracture occuring because the designer either did not have the right tools to predict it or they entered a 50% isentropic efficiency into HYSIS without knowing better.

Until a better method is available I would strongly recommend that 98-100% isentropic efficiency be used in HYSIS modelling.

I hope that this posting is informative on this topic rather than the beginning of another very long thread like the one referred to at the beginning.

You might note that "JT or not JT" does not appear to be addressed above but in fact the answer is there and the correct answer is "yes" JT cooling is thermodynamically correct in pipeline flow modelling. I am in the process of proving this at the moment and will post something when I have done so.

I hope this helps with a better understanding of the basics of pipeline blowdown and perhaps prevents a few failures.

Regards
Dennis Kirk-Burnnand

Dennis Kirk Engineering

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

 

[BigInch]

Dennis, I wouldn't think you'd need to prove it. Its been a known fact to pipeline engineers for at least the 33 years I've been doing this. But I won't stop you if you think you need to do it.


"If stupidity got us into this mess, then why can't it get us out?" - Will Rogers (1879-1935) ***************

http://virtualpipeline.spaces.live.com/

 

[denniskb]

BigInch,

I use JT cooling in all of my pipeline simulation calcs.

My "JT or not JT" query was about WHY is isenthalpic (which is normally associated with violent pressure drop which is not present in pipeline flow) the correct process rather then isentropic (which is better suited to these slower changes). I put this to you - "how do the molecules in the gas stream know which value the process engineer used in setting the HYSIS isentrpoic efficiency somewhere between 100% (isentropic) and 0% (isenthalpic)?"

What I am trying to prove is that the basic process is indeed isentropic but that the end result for pipeline flow is isenthalpic. Conversley for pipeline blowdown this is not the case. If I am right then we can in fact predict how the molecules will behave during both events and the isentropic efficiency will be an outcome of the calculations and not an input (by an engineer who is guessing the value to use).

Regards

Dennis Kirk Engineering

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

 

[25362]

What about the reduced conditions of 0.8>T_r<5 and P_r<11.8 to ensure that the fluid cools down on expansion?

These values were taken as the extremes of the approximate "inversion curve" for all fluids in pipelines as they appear in Perry's Chemical Engineers' Handbook.

 

[BigInch]

Then your interest is directed towards uncontrolled blowdowns?

"If stupidity got us into this mess, then why can't it get us out?" - Will Rogers (1879-1935) ***************

http://virtualpipeline.spaces.live.com/

 

[sailoday28]

If the pipe line is perfectly insulated, the flow is adiabatic. If friction in zero, the flow process is isentropic. If, in addition, the change in PE, KE are negligible, then JT is correct.

Regards

 

[dcasto]

so you mean to tell me when the local utility company drilled into my 2100 psia 99.998% pure ethylene line and depressure it it to 14.6 psia, the whole line ruptured from being at -121F? Nope, it survived, sorry the truck didn't but the operator escaped injury too. took us 3 days to thaw out the ground and cut out an 8' section (round number). Then went back on line.

Then, when we set up our portable flare and blow down 50,000 pounds per hour from 2100 psig to 0 psig through 100 feet of carbon steel lines, they will rupture? Done that one more than 50 times.

 

[BigInch]

That's why I didn't get too excited about this.

dcasto, doesn't the fire usually take care of thawing the frozen ground?

"If stupidity got us into this mess, then why can't it get us out?" - Will Rogers (1879-1935) ***************

http://virtualpipeline.spaces.live.com/

 

[MortenA]

Dennis any news on this?

When you normally flow in the pipeline is would say adiabatic isenthalpic (as far as i understand it=JT) when disregarding heat exchange with surroundsing (NOT recommended but OK for simplicity) - but during blowdown - SOME of the gas (say the far end) is just expanding (adiabatic isentropic) - and some of the gas is moving, experiensing a pressure loss where the work is "transferred" to the gas (adiabatic isenthaplic) - or did it get it mixed up?

Best regards

Morten

 

[denniskb]

MortenA,

Yes the gas at the far end has minimum velocity so expands in a near isentropic process yet the gas near the exit has velocity and expands in a near isenthalpic (JT) process. And yes again, the difference between the two is the work associated with the pressure drop due to velocity (i.e. friction heating). Strangely there is no mention of this in the Bernoulli equation?

For natural gases (which I normally deal with) the result of both processes is cooling with the greater temperature drop being that from the isentropic process. If the line is short and the blowdown fast (low heat transfer to the environment) then the gas exiting the pipeline can have been be cooled by the isentrpoic process then as it passes through the blowdown valve it gets hit again with isenthalpic cooling.

Typically the gas inside the pipeline undergoes some heat transfer with the pipe wall (particularly as the velocity increases) which cools the pipe and warms up the gas, though the gas will usually be much colder than the pipe. As the pressure drops the thermal mass of the pipe wall can far exceed that of the gas contained so the gas is warmed more than the pipe cools.

When the gas passes through the blowdown valve the thermal mass of the blowdown pipe becomes much less significant and the pipe can drop all the way to gas temperature. This is the bit that many engineers get wrong and is where the risk of brittle fracture is high. At this point it is no longer possible to ignore heat transfer from the environment because that is the only source to warm the pipe and most of us have seen that the blowdown line does warm up away from the blowdown valve. The trouble is knowing how far downstream of the valve do you need to be before you can change from low temp steel back to plain carbon steel.

All the steps needed to complete this series of calculations is available they just need to be compiled together in a useable form then perhaps some of the guesswork evident in the responses above will no longer be necessary.

Dennis Kirk Engineering

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

 

[MortenA]

With regards to the Bernoulli equation i would assume that this is not accounted for since its developed for liquids - and for liquid flowin "in a steady pace" not for the expansion case (sorry i have no better words in english for it).

An extreme example that im currently looking at is the thermodynamics of a gas storage cavern.

With regards to heat transfer and pipewall temepratre:There is an article re. this is july 2004 issue of Hydrocarbon processing.

However - in general terms i think it depends a lot of the ambient medium. If the pipe is in air, water or soil. Air and water gives fairly good outside heat exchage - and thus a warmer pipe - wheras a buried pipe may see a poor outside heat conductivity that leads to the pipewall temperature being much nearer the gas temperature. This is also true if the pipe is insulated or even just coated - some coating like neoprene is not that terrible an insulation.

I use pipeline studio for these calculation and is does Ok - but it has problems if you gas gets cold enough so that condensation should occur - this cannot be handled by Pipeline studio that only handles singlephase (gas or liquids)

Best regards

Morten

 

[sailoday28]

FYI as I have previously posted.
denniskb (Mechanical) 28 Oct 08 11:06 TAKEN FROM YOUR ORIGINAL POST.


Pipeline flow and blowdown are not that different so lets consider blowdown and to make things easier lets assume that the pipeline is insulated from its environment so it consists of the contents and the pipe only.

First, when the pipe contents blowdown the thermodynamic change in the fluid inside the pipeline is basically isentropic (so 100% isentropic efficiency). This is a very significant contributor to temperature change.

IF THE FLUID UNDERGOES AN ISENTROPIC PROCESS THERE IS NO HEAT TRANSFER TO OR FROM THE SURROUNDING PIPE

Next, since the fluid is moving there is also friction which generates heat energy and warms the gas inside the pipeline. This happens little at the far end of the pipeline where the velocity is low and more at the discharge end. This is typically a small contributor to change.
WITH FRICTION THE PROCESS IS NOT ISENTROPIC. HOWEVER WITH NO HEAT TRANSFER THE PROCESS IS THEN ADIABATIC.

Then, there will be heat transfer between the fluid inside the pipeline and the steel pipe wall. The thermal mass of the steel is typically high compared to the contents so if the blowdown is slow enough this can significantly warm the gas and cool the pipe.
NO IMPACT FOR ADIABATIC PROCESS

This takes care of the changes inside the pipeline. Despite the large number of variables and the time and distance affects the calculation methods to solve the above are all available. Note that I did not say they were easy because this is not at all simple to solve.

Be warned that while the pipe temperature drop and the falling pressure stress may not make normal brittle failure a risk it can easily get cold enough for a crack to propogate along the pipe length without arresting itself.

For now lets just assume that the pipeline contents get pretty cold.
FOR ADIABATIC PROCESS AND NEGLECTING CHANGE IN ELEVATION THE STEADY STATE ENERGY BALANCE IS
H +KE = CONSTANT. AS VELOCITY INCREASES (KE INCREASE) H WILL DECREASE. WHERE H IS SPECIFIC ENTHALPY.
FOR J-T PROCESS H IS CONSTANT.

REGARDS

 

[MortenA]

Sailor (capitals or not)

Your statement re. pipe mass and is not allways true. I have just compeltede a study of blowdown of gas pipelines from 150 bar->0 of intermediate size (12"-14") and short length 24-89 km and here pipe mass does not mean a lot but heat transfer to soil (buried) does matter.

But i agree with your terms.

I use PipelineStudio for this.

Best regards

Morten

 

[sailoday28]

Morten
I have only commented in caps, on the original post..
I don't understand what you mean by"Your statement re. pipe mass and is not allways true."

regards

 

[MortenA]

Sailor may be i misunderstood you or took something out of context:

"Then, there will be heat transfer between the fluid inside the pipeline and the steel pipe wall. The thermal mass of the steel is typically high compared to the contents so if the blowdown is slow enough this can significantly warm the gas and cool the pipe.
NO IMPACT FOR ADIABATIC PROCESS"

But your statement is true I guess if you extend your system to cover soil/air/water

Best regards
Morten

 

[sailoday28]

Morten. What I mean is that if the process is adiabatic, with no heat transfer to/from the fluid, the metal mass has no impact on the flow.

If adiabatic with regard to the outside of the pipe, yes then the metal mass will contribute to the process.
Regards

 

[denniskb]

Let me pick up on the postings since my last reply on 2Dec.

MortenA - your gas storage cavern application is a good example of how wide a net we need our thermodynamics, fluid flow, heat transfer solutions to address. Probably the biggest difference for this case would be the near lack of flow velocity. Yes phase change adds many levels of complexity to the problem and while Hysis is able to detect and calculate the thermodynamics of phase change the resulting pipe metal temperatures are very suspect. "Blowdown" for the Imperial College of London is the only accurate model I am aware of for this.

sailoday28 - the only place that a purely isentropic or even isenthalpic process can exist is in a computer simulation and we know how to calculate these. What we are about is real world processes and how to accurately calculate these. The approach I use is to break the problem down to individual increments, each of which we know how to solve, and then summ the results to approximate the final outcome. For example i)Isentropic Expansion, ii)Pressure Drop due to Friction, iii)Friction Heating, iv)Heat Transfer to Pipe Wall. I carry out the series of calculations over a discrete time interval and length of pipeline then take the outflow from one section as the inflow to the next section.

MortenA - The thermal mass balance between the fluid inside the pipe and the pipe wall depends largely on the fluid pressure/density and the pipe size/wall thickness. This balance changes dramatically as the fluid pressure drops during a blowdown.

Both - as noted the adiabatic calculations change depending on whether the control volume is drawn inside the pipewall or outside the pipewall. Adiabatic is again an ideal case suitable only for computer simulations.

Hopefully with a constructive assessment of the thermodynamics, fluid flow and heat transfer issues we might eventually put a workable solution in the hands of the engineers that are tasked with making these difficult assessments and from previous postings making sometimes unsafe decisions.

Dennis Kirk Engineering

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

 

[MortenA]

Dennis - I know about Blowdown even though i haven't had the chance to use it/study it myself. I wouldn't use HYSYS (as it is called - not HYSIS) either since its still mostly a SS software with regards to pipelines. We have a full license for the dynamics but im not convinced that it can be used for this purpose yet. As long as the temperature remains within the gas phase then im comfortable using Pipelinestudio. This software (based on TGNET) includes a full heat balance for the various walls that you define and the surroundings.

Best regards

Morten

 

[sailoday28]

Dennis-You state real world processes.AND Next, the blowdown valve has a large pressure drop, typically choked flow for most of the blowdown time and so high friction. The resulting thermodynamic change is isenthalpic (or JT and With a large change in velocity, how can this be reperesented

With a large velocity change, how can you state that JT applies?

 

[denniskb]

sailoday28

The very high velocity inside the throat of the valve (or other choke point) drops off immediatedly downstream of the valve (so kinetic energy becomes less relevant) and the JT effect occurs between the valve upstream and downstream sections.

At any instant the ENTIRE pressure drop across the valve is due to friction through the valve and through the shock wave. This is the key to what makes a process JT cooling. Of course there will also be heat transfer from the valve to the gas and ambient heating but with everything happening fast and in such a localised point these will be negligible.

This is also the reason why JT cooling is included in pipeline flow simulations because at steady state flows the ENTIRE pressure drop in the pipeline is due to friction. In this case you must also include for ambient heat transfer because the change happens slowly over a long distance.

Going along with my methodology of breaking the problem down to manageable calculations and summing the results I believe that the first two processes that happen in pipeline flow are in fact isentropic expansion followed by friction heating (resulting in isenthalpic (JT) expansion).

I have been able to demonstrate this by calculation (typically within 0.1°C) though I have a couple of issues yet to resolve and I was hoping that this forum may help me resolve these.



Dennis Kirk Engineering

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