7 replies to this topic

[Bill B]

I have several situations wherein the relief temperature calculated from API 521 using the ideal gas law is much higher than the wall temperature. I found a post from Wendy a few months back, with responses from Phil L which indicated that I should perform the RV calcs based on the very high T1. In that situation F' used in the sizing equation (5) [A= F'A'/sq rt P1] is of course a nonsenical number. So which of the following approaches should I use:

1. Use API 521 eq (5) above, with F'=.01, which is the default minimum.

2. Use that equation, with F'= .045, which the standard states is the number to use when the mimimum value is unknown.

3. Use the API 520 gas equations (3.2 - 3.4) using the very high T1. In this case I need a recommendation for calculating the mass flow rate (of course normally based on BTU rate / Hvap).

Thanks for the help. I am stumped on this one.

[djack77494]

So which of the following approaches should I use:

Bill,
Please clarify your posting. It appears as if you are attempting to size a PSV for a Vapor filled vessel which is exposed to fire conditions, but I got a little confused by your reference to Hvap. May I suggest that you provide the vessel design temperature and pressure, plus the fluid properties normally and at relieving conditions?

Without more details, I would suggest that you attempt to calculate F' using equation (6). You should be able to readily calculate your relieving temperature using equation (7b). The only variable that I really don't know how to handle is Tw. For that I might develop a reasonable basis. For example, if you had an ASTM A515 Gr.79 vessel having a rupture stress of 20,000psi, than I might develop a basis that states that the vessel should be able to resist rupture for 1 hour (in the expectation that effective metal cooling with water would be achieved in that time period). Figure 2 would then lead you to select a maximum wall temperature of 1100F (if I've read the figure correctly). Now you can calculate F' and then A. I believe you are allowed to reduce A by considering the environmental factor, F, if applicable. Hopefully you'll soon receive an expert response to your query, but you might want to try the suggested approach and report back as to what your result were.

Good luck,
Doug

[Bill B]

Doug, I appreciate your comments. Actually I had done what you had suggested, which led to the original post. While I have several cases that are stumping me, here is one:
1. Ammonia gas on shell side of exchanger.
2. Normal operating Pressure = 56 psia, Normal temp = 610 °R.
3. Relief pressure, plus 21% accumulation = 317.2 psia.

T1 (relief based on ideal gas) = Tn x P1/Pn = 610 x 317/56 = 3680 psia.
since T1 < Tw (= 1560 °R), F' in equaltion 3 cannot be calculated.

One choice is to "pick an F1" and do the calculation, or use some other approach. I am just not sure what to do. Seems like bottom line is that vessel will long gone before the relief pressure is attained.
Thanks again.

[djack77494]

T1 (relief based on ideal gas) = Tn x P1/Pn = 610 x 317/56 = 3680 psia.

Units should be degrees R here.

Bill,
I think your interpretation is correct. What that means is that as your tank is heating up due to an external fire, the metal (wall) temperature exceeds its allowable limits (well) before the tank's pressure builds to the relief valve setpoint. Assuming you are not free to lower the PSV setpoint, then a relief valve would not be effective in mitigating the risk potential in the event of a fire. Do not be surprised by this since we really should have no expectation that the PSV would be effective. For a gaseous system, we don't have the large heat sink that we would have if we were boiling off a liquid, and so using pressure relief for over-temperature protection is unlikely to be successful.

[Art Montemayor]

Bill:

I interject myself into this thread after carefully following what I have visualized as a classic case where API depressurization is often called for as a safe resolution to a potentiallyl dangerous hazard.

I have confronted this type of situation before and the solution was to install a blowdown manifold where I could manually activate depressurization of several gas-filled vessels after a local fire alarm sounded into the main control room. Activation of the depressurization system was done from the control room and also monitored from there.

As I recall, I set the depressurization for a 100 psig level. In order to lessen the surge and capacity effect of the depressurization, I ramped the blowdown in stages. My vessels were initially at 550 psig and I stage the depressurization in two steps.

Doug is absolutely correct and I believe you suspected this to be the unfortunate case where you have no practical control on the ability to contain a total collapse of the vessel since the hoop stress would be subjected to a wall material stress strength weakening - leading to a total vessel collapse and rupture. The vessel would fail due to material failure not because of over-pressure. This is the reason the API included depressurization as a safety action in their standard 521. Don't forget to calculate the relieving gas temperature due to the Joule-Thomson effect and how it will affect your blowdown piping.

[JoeWong]

PSV is a final safeguarding device to prevent equipment or system from catastrophic failure. However, this device may not protecting equipment and/or system from FIRE attack. In many events, external FIRE for gas filled vessel will probably fail the vessel before the internal pressure reach its PRV set pressure.

Providing PSV for fire attack mainly

- to avoid continuous increase of internal pressure minimizing the impact when the protected equipment is failed
- to "buy" time
- to meet code requirement

We shall pay more attention on other active and passive protecting system to safeguard the system :

(i) Depressuring (as stated by Mr. Montemayor)
(ii) Provide Rupture / bursting disc instead of Pressure Relief Valve (for system that will lift PRV)
(iii) External cooling (deluge & spraying)
(iv) Fire Extinguishing agent (with external cooling)
(v) External Insulation (fireproofing)
(vi) Sand Cover Storage
(vii) Diversion wall for isolation and proper drainage
(viii) Reliable & Compatible Fire detection system

In estimating PSV relief load, temperature, etc, one of the assumption is "the material will not fail prior to pressure reaching the maximum accumulated working pressure".

[dgoyal]

I got here some more confusion , and some more question coming in my mind
what will be relief temperature
if relief temperature comes 600 deg c from api calculation that is it necessary to design the flare header and connecting equipment like flare KOD also for this temperature

hoping to have some view on this.

Deepak

[JoeWong]

Deepak,

First of all, the API equation will lead to a very high temperature especially for system with low operating pressure and high PSV set pressure gas only system. This is one of known issue and constantly debate here and there...

However, it is normal not to design the flare system for fire temperature :

i) there are a great heat "absorption" capacity of flare network.
ii) blowdown system is normally initiated during fire (plant wide or within a zone). Thus cold blowdown would release cold vapor which potentially drop the temperature quickly.
iii) A gas is relieve into flare system, there are JT effect to bring temperature down
iv) there are other fire protection means to being the temperature down
v) the processing equipment is NOT designed to take temperature due to fire (can be as high as 1200 degC). Read more "Should maximum recommended wall temperature (Tw) for carbon steel vessel used as design temperature ?"
vi) it is NOT cost effective...