11 replies to this topic

[sahas]

Does the design of a thermal expansion relief valve falls under ASME code ?

If so, what are the restricted lift limit requirements for an thermal expansion valves in case we would like to limit capacity of the liquid relief valve ?
sahas

[Art Montemayor]

sahas:

To my knowledge, there are no "design" codes imposed on a thermal relief valve. You should be explicit in stating what you mean by design: do you mean the process design or the mechanical design? Since you mention ASME, I assume you mean the mechanical design: the metallurgy, the orifice type & size, the connections, the stress analysis, etc.

Additionally, the employment of a thermal relief device falls under a common-sense safety engineering decision - not a "code". If you block-in a liquid and apply heat to it, it will expand hydraulically and cause extreme stress on the equipment and piping that contain it - to the point of rupture. Engineers are paid to ensure that this event never occurs in equipment that they are responsible for designing, checking, and operating.

Thermal relief valves, in doing their expected action, actually allow a miniscule amount of liquid to escape from the contained system. If you make the appropriate calculations (this has been mentioned many times on this forum) you will find that the amount of hot liquid needed to relieve the system is in the size of 1 to 2 cubic centimeters - a volume so small, it isn't worth the time or effort to measure or contemplate. The important thing for an engineer to accept is that this liquid - albeit very small in quantity - is incompressible and can cause exponential pressure increases if left on its own in the blocked system. You must relieve it to save the system and possibly human well-being.

U. S. government regulation of this event doesn't even consider the calculation of the valve orifice to be of sufficient importance to document - as long as there is a working, effective orifice ready to relieve. Therefore, I can't imagine what is so important about restricting the lift of such a device. I would certainly not recommend anyone tampering or trying to "adjust" a relief device by restricting the lift - unless they are the manufacturers of the device or they are willing to take total responsibility for the consequences. These thermal relief valves are normally 1/2" to 3/4" in line size; why would you restrict the relief down to fractions of a cubic centimeter? The attempt to reduce such a small discharge could result in a total plugging of the orifice and render the device ineffective. Without further basic data to justify it, I would not do this.

[pleckner]

Just want to add something to Art's great response.

If you can exceed design pressure of the pipe/equipment, then Code requires that you protect against this. Thermal relief is absolutely a credible relieving scenario in many instances.

There should be no concern about limiting capacity through the PSV because if it does open it will immediately re-seat and stop any additional flow as soon as the pressure is relieved.

[Adriana]

Hello everybody,

I just wondering if you could give some help in trying to measure the relief load due to thermal expansion. I have a full liquid hydrocarbon system to protect. The set pressure is 93barg.
I know from Art's answer that I am going to have hot liquid to relief, how can I measure this quantity? I mean from the simulation point of view?

Many thanks,

Just want to add something to Art's great response.

If you can exceed design pressure of the pipe/equipment, then Code requires that you protect against this. Thermal relief is absolutely a credible relieving scenario in many instances.

There should be no concern about limiting capacity through the PSV because if it does open it will immediately re-seat and stop any additional flow as soon as the pressure is relieved.

[Art Montemayor]

Adriana:

This is a curious request and it generates other questions. The reason it is curious is because hardly anyone has any interest in the amount (or flow rate) emitting from an activated thermal relief device. We engineers know, beforehand, that the amount is going to be relatively miniscule. So why bother in identifying it? Is that what you mean by “the relief load due to thermal expansion”?

What makes it even “curiouser” – as Alice would say – is that almost every time an engineer has brought out the question or query, he/she leave the impression that the reason for the question is a complete mis-understanding of what is being relieved and why. This impression is left because once the action of a thermal relief is thoroughly explained, the OP normally never responds and just fails to even say “thank you”. Witness the lack of response on the part of the OP in this thread.

Our first question is: why is there an interest or need in measuring the amount of hot liquid relieved? Could it be that the fluid in question is highly poisonous or toxic?

Our second question is: what does a simulation program have to do with this action? Are we at a stage where we have to depend on a simulation program to make our decisions or derive our calculations for us? To put it quite simply, the use of a simulation program to calculate the amount of liquid discharged through an orifice is an over-kill – and probably a waste of expensive and inappropriate software. All that is required is a knowledge that the liquid’s specific volume enters into play here and that, by your definition (“I have a full liquid hydrocarbon system”) – you haven’t any place within the system to put the excess volumetric amount produced by thermal effects. Therefore, the difference in the liquid’s volume – as measured at the set relief temperature/pressure - will tell you the amount to relieve. And this amount is what I’m predicting is going to be relatively very small and normally of little or no interest. You know the identity of your liquid and therefore should also know its specific volume at the different temperatures & pressures.

Can you tell us why you have an interest in this amount or give us all the basic data that describes your needs? Perhaps we can help you out.

[JoeWong]

Andriana,

I am in agreement with Mr. Montemayor comments. Hydraulic expansion relief load is small and relieve in very short period. A DN 20 × DN 25 (NPS ¾ × NPS 1) relief valve is sufficient.

If you are really interested, your may read seciton 5.14 in API Std 521, Jan 2007.


Mr. Montemayor,

I have some personal opinion. Appreciate you feedback and advice.

Perhaps, if the fluid is subcooled (no flashing) at both relieving pressure (121% of 93barg) and PSV backpressure, we may adopt the method proposed in section 5.14, API Std 521, Jan 2007 to estimate the liquid relief load and calculate the PSV orifice base on liquid relief in API 520, pt I.

However,
If your fluid is at subcooled condition when pressure is at 110% of set pressure and may flash when it passing the PSV, the PSV for thermal relief may be undersized. We may need to consider the following steps:

1. Calculate liquid relief (with flashing when across PSV) considering thermal expansion of captured volume up to 110% of the relief valve set pressure.

2. Calculate relief load (stepwise) and time between liquid and/or two-phase relief of the remaining captured volume after 110% of the set pressure is reached (i.e. valve fully open conditions) until vapor-liquid disengagement occur. Vapor-liquid disengagement can be checked using Bubbly-viscous and/or Churn-turbulent model proposed in DIERS method.

3. Calculate time required for the captured volume pressure to increase from 110% of set pressure to 121% of design pressure, assuming continued thermal expansion with insufficient void volume to allow for vapour-liquid disengagement.

A PSV shall be sized (large enough) for step 1 & step 2 liquid and /or two phase relief so that time required in step 2 less than step 3.

Appreciate you advice.




regards,

JoeWong

[pleckner]

Joe,

You're really hung up over that procedure for two-phase flow you found in that article, aren't you? That procedure was proposed for fire relief of a liquid filled, blocked-in vessel. For pipelines, we typically are concerned about the sun heating the pipe or in a heat exchanger, the hot fluid heating the cold fluid. And this appears to be what the original post was all about. Once the pressure increases (and it will with just a slight increase in temperature), the valve will pop and fluid will relieve. I don't care if it is two-phase or not. It will be a very small quantity and only needs to pass a very small amout to stop the scenario from ever happening again. Remember, this is not a boiling case we're talking about.

Andriana,

This is a double post. You properly started a new thread so remember in the future, please don't attach onto an existing thread unless you are addressing the original post.

Saying this, I gave you a response in your new post. Art is correct but if you think you might have a flow requirement greater than what a minimum sized PSV would handle, you can attack the references I gave you.

[Adriana]

Montemayor,

Thank you very much for your response. I am currently preliminary sizing the relief valves in this plant. Therefore, I am interesting in the relief flowrate through this valve, I can’t accept at the first hand that is relatively small amount, I wanted to know how small it is.
I already did the calculations and I got 86420 kg/h to be relief and following the calculation procedure for liquid relief in Section 3 of API 520, I obtained a 1/2in^2 orifice size which gives me a better idea on how small is the orifice.
Now from the API 521 Section 5.14, you can read “The required relieving rate is not easy to determine” however they specify a common size. From here, I realise that the size that I get is under the recommended value, so I am going to specify it as an orifice H or DN20 as a minimum size for this service.

I am process engineer with more than 5 years of experience, but not in sizing relieving device, that is the reason of why I put a basic question on this forum. However, through your answer and JoeWong’s answer I am in a better understanding of this situation. Thank you very much for your time.

[pleckner]

Adriana,

Something is bothering me about your numbers.

First, the minimum size relief valve usually expected for thermal relief is a 3/4" x 1" (DIN 20 x DIN 25) and this has an orifice size of less than 0.1 sq. inches; much less than even a D orifice in API terms. However your saying you calculated a 0.5 sq. inch orifice (quoting: I obtained a 1/2in^2 orifice size)!

Can you share some of your calculations? Please tell us exactly what equation in API RP520 you used and the parameters you put into the equation. Can you elaborate with some detail how you calculated your flow rate? I'm concerned that you might have missed something or mis-interpreted something.

And on the assumption at this point that you do indeed need to have a 1/2 sq. inch orifice, a G orifice would be adequate, the H is too big.

We await your reply.

[JoeWong]

Phil,

Indeed you are right that i should not pay too much enphasis on external heat input (due to fire) to pipeline / piping trapped volume. Code is not called for. Thanks for waken me up.

Infact, sometime i also have doubt on the application of PSV on pipeline (versus a filter). I will raise a new post for discussion.




regards,

JoeWong

[gvdlans]

Although in general thermal reliefs will be small, there are exceptions. I once did a project related to cross country pipelines with liquid propane and butane and there the amount relieved due to solar radiation turned out to be significant.

A quick calculation using data from www.webbook.nist.gov for pressurized propane shows that if a 16" pipeline is heated up from 15 to 25 degrees Celcius, about 3.5 m3 or 1750 kg of propane needs to be relieved per kilometer of pipeline. Assuming a solar radiation heat input of 1 kW/m2, I calculate a relief rate of 0.42 kg/s or 1500 kg/h per kilometer of blocked in pipeline.

[pleckner]

gvdlans:

I don't disagree with what you said. Long pipelines do indeed require larger PSVs than the old stand-by 3/4" x 1". And this is exactly why I am asking Adriana for more details. This may be a similar situation and with Adriana being inexperienced in PSV analysis, I think it is important we know where the calculations are coming from. The last post from Adriana to Art is a little confusing and somewhat contradictary. It could just be a problem with the English language but I want to make sure it is not technical.