Thermal Expansion in liquids 4

[bif]

A recent Hazard study has highlighted the possibility of thermal expansion problems in exisitng acid and alkali pipework, all at ambient temperature.
Is there any information available regarding thermal expansion, due to ambient temperature differences, in liquids between 2 block valves.
I can't help but think there is minimal risk, but need to clarify the situation

 

[phex]

do some calcs, shouldn't be too difficult. find maximum ambient temperature at plant's location. then go on calculating thermal expansion of process fluid at max temp. with new volume, you can calculate pressure in piping system. check your valves against the resulting pressure.

taking this one step farther, calculate backwards... first think about how much additional pressure your valves/piping can stand. then calculate backwards to get the maximum temperature your pipework will stand. check if this temp. has been reached in plant's location the last 2 or 3 decades. if not - you're on the save side.

just my 2 cents.
chris

 

[TD2K]

You can do the calcs definitely but unfortunately, blocked in liquids will give you very large increases in pressure for relatively small changes in temperature (we are talking degrees, not many tens of degrees).

You can do a keyword search here but a rule of thumb is that blocked in liquids will increase 40 to 100 psi for every deg F temperature rise. What typically saves piping systems is that most valves are not truly bubble tight and it only takes a very small leakage rate to relieve the pressure. Some valves are however bubbletight and those are the ones where blocked in lines can leak, typically at flanges.

 

[Scipio]

Crunched a few numbers on this one, here's an example of how extreme a worst-case scenario can be - never tried this before, so numbers could be a bit flakey;

Pure water, 1000 kPag @ 20°C, has a molar density of 56.139 kgmol/m³ (according to HYSYS). Compressibility factor of water at this condition is 0.0073.

Raise the temperature to 30°C, molar density falls to 55.722 kgmol/m³. This means if the original system was had 1 m³ of water exactly, or 56.139 kgmol of water, the required volume to contain all of the water at the higher temperature is 1.0075 m³.

Now, assume water is in an infinitely rigid container of fixed volume 1.0000 m³, not subject to any thermal expansion. The pressure required to increase the molar density back to 56.139 kgmol/m³ at 30°C is 26350 kPag, assuming I'm using the molar density right.

Reality sets in, and you've got a little entrained gas, maybe one or two small pockets of vapour at instrument takeoffs and boltups, a little flexibility in piping, some bubbling across valves, it doesn't take much to eat up that extra 0.0075 m³ because of the higher compressibility of vapour.

 

[Scipio]

Knew I forgot something in my last post, compressibility of water at elevated pressures takes a big chunk out of that final pressure, for instance compressibility at 30°C and 2000 kPag is 0.014236, which should be enough to stop the pressure raising much beyond that point, regardless of movement in the surrounding piping - never looked into compressibility of liquids before though, so I'm not sure if they are treated the same as compressible gas (which I'm most familiar with).

 

[TrevorP]

I agree with TD2K. This has been covered several times before, so do a keyword search. There was one interesting thread where the person asking the question set up a test rig in the end, and got about 60 psi per °F, i.e. pretty close to the theoretical value (See Thread391-15161). However, as TD2K suggests, there is usually some leakage between the seats. Whether you want to rely on this depends on a) how likely thermal expansion is (i.e. whether the valves are normally open, and blown clear before closing them, even in emergency shutdown situation), b) the risks associated with a leaking gasket at the valve, c) whether the valves are located indoors, in a climate controlled room, or outdoors, and d) how "bubble-tight" the valves really are.

 

[quark]

TD2K and Scipio!

Good posts by both of you and deserve a star each atleast.Though they are not matching numerically, after considering the compressibility factor I think they should.It is indeed a good technical explanation. Next time when I do hydrotesting of pipelines on a hot summer day, I will check this.

TD2K!

Is your sunsign Cancer?

 

[25362]

Thread378-55232 deals with a similar subject.

 

[25362]

See thread378-55232 on a similar subject.

 

[quark]

Sorry Guys for my misbehaviour, but I cannot resist.

TD2K!

I would like you to have a look at this thread378-55233 and help me and the original poster out.

Regards,

 

[Montemayor]

bif (Chemical):

If your HazOp is as serious and organized as the ones I have participated in, then any resultant action taken on this item MUST include a thermal relief valve that vents to a safe and controlled area (or volume).

You state this is acid and alkali fluids in existing pipework. These are, I presume, classified as HAZARDOUS fluids by OSHA. As such, you are obligated and mandated to provide safe and controlled relief of these fluids. This is not an option. You must provide relief for thermal expansion of these fluids. As a result, there can be no debating what the action item for this situation must be: it must have thermal relief - and designed in a safe and controlled manner.

I am telling you this from actual plant experience. This is a serious subject that must be handled in a serious manner. Do not rely on what others have mentioned about most valves not being truly bubble tight. This is worthless information where the HazOp is concerned because all it takes is for the valves to seal 100% one time and cause a pipe rupture with subsequent release of acid or alkali. You should be designing for the worst credible pressure scenario - and what I just described is both credible and possible. An analogous and similar fluid flow situation that often is addressed in HazOps is that of a check valve stopping the backward flow of a fluid. The fact that 100% of all check valves do not seal effectively is enough evidence to establish the universal HazOp rule that NO check valve will be assumed as halting backward flow. Similarly,I apply the bubble-tight feature of any block valve to your situation, but in a reverse logic: I will never assume that 100% of all block valves will leak sufficiently to relieve thermal expansion. I believe that most HazOp groups in the USA would agree with me on this point.

From a practical, engineering point of view why would anyone fail to incorporate a relatively cheap, 3/4" or 1" thermal relief valve (that doesn't require OSHA calculations nor documention!) to prevent what I just described? That isn't being very practical and much less safety-minded.

I believe you are wrong when you describe the situation as minimal risk. There is certainly a credible risk; and the consequences of someone getting doused with hazardous fluids is far too serious to classify as "minimal". I would implore you to install the required, cheap thermal relief valves and avoid such a scenario.

Art Montemayor

 

[ChemE001]

Well said Art. Thermal relief is cheap, does not require strict documentation, except for highy volitale fluids, and makes good practical sense for process safety management. Don't leave the plant without it.

 

[Scipio]

Bear in mind while a thermal relief valve can be cheap, sometimes it can be fairly expensive properly venting it, if you're doing a refit on an existing plant and would have to bring the plant down to tie the relief valve into a flare header, for instance. Art's absolutely right about assumptions on valve's leaking, however, another option to a thermal relief valve is installing an accumulator to absorb the extra volume. Has the advantage of not needing a discharge point.