PSV Sizing for Fire Case - Relief Temp above Design Temp 2

[joeycranston]

In reviewing API 521 (4th ed. - 1997), Section 3.15.1.1 Effect of Fire on Wetted Surface of a Vessel says that one should refer to Section 3.15.4 "if elevated temperature is likely to cause vessel rupture." The following section on Effect of Fire on the Unwetted Surface of a Vessel says one should refer to section 3.15.4 AND 3.15.5.

The referenced section 3.15.4 is titled, "Protective Measures Excluding Insulation." It suggests water sprays, depressuring, fire-proofing, earth-covered storage, and diversion walls. The section referenced only for unwetted walls (3.15.5) is for External Insulation.

Why isn't the external insulation section referenced for wetted wall vessel failures? The suggestions of fire-proofing and earth-covering are essentially equivalent. Was this an oversight or am I missing something?

 

[jistre]

I believe it's because typical insulation is not a fire protective material. A plant fire can be very intense and the insulation isn't designed to protect against the heat and forces imparted by the fire. In addition, high pressure water or foam streams used to fight the fire may physically dislodge insulating material. In unwetted vessels, this is less of a big deal because during a fire the fluid inside expands at a relatively slow rate because it's already a gas and the heat doesn't move into it very well from the vessel wall. A wetted vessel, on the other hand, can easily transfer the heat from the vessel wall into the liquid phase and then boil it off, creating a large relief rate. If the relief calculations use the insulation as justification to lower the heat transfer rate into the liquid, then the valve will be sized for that vessel always having insulation. If that insulation then gets blown off the vessel during a fire, the relief valve will likely be very undersized, won't relieve the vessel properly, and puts it at risk of rupture.

I think. I just recently am trying to get my head around API 521, so any other thoughts or steering are welcome.

 

[MortenA]

General: the 1997 ed. is not the lastest API 521 i think. Theres now also an ISO version. The (ISO 23251 2007 ed.) text reads:

5.15.5.1 (assume its the same as 3.15.5.1 but im not sure): Credit for thermal insulation is typically not takenbecause it usually not meet the fire protection insulation requirements given in 5.15.5.2 through 5.15.5.4. If these requirements are met, a reduction in fire heat input can be obtained by using the environmental factor F...

So (as jistre also says) the destiction is because often insulation does not meet the requirements for fire protection. But if it do meet these requirements then you are allowed to take credit.

Best regards

Morten

PS: Get the new copy not much use learning an old code.

 

[Propacket]

Hi guys. This is my first post.
Whenever I apply fire scenario for wetted and unwetted vessel case, unwetted case always gives the highest temperature sometimes even higher than the carbon steel maximum allowable temperature. Reason is simple. In wetted vessel, majority of heat input is used to vaporize liquid. so temperature increase is lower than that of the unwetted vessel where all the heat is used to increase the vapor temperature. Chances of failure of wetted vessel are small. That's why API has not mentioned wetted vessels.

 

[joeycranston]

Propacket, Thank you for replying to the post; especially since this is your first post. I appreciate it. The circumstance to which I refer would be a high boiling liquid (such as a heavy oil). In such a case, there would be little or no vaporization to mitigate the temperature rise of the liquid. The temperature of the liquid in the vessel required to lift a safety valve may be high enough to compromise the maximum allowable working pressure of the vessel because the metal would weaken.

 

[Propacket]

joeycranston

What ever liquid you are dealing with , you should first know the boiling point of the liquid at relief pressure. If you have a heavy oil, boiling point could be very high. Depending upon the type of oil,i have seen heavy oil boiling points of 700-800 degC exceeding carbon steel maximum temperature (593 degC).In such cases, there will be no vaporization and the vessel will fail before the PSV lifts.
Since there is no vaporization, we can dismiss this case and consider only the unwetted case.

 

[longesthorn]

joeycranston

My first time posting as well.

For PSV sizing, taking credit for insulation (assuming the API requirements are met), would only reduce the heat input rate (Q) from fire (lowering the F factor in the equation), reducing the relief flow rate if vaporization occurs. For a vessel containing a heavy oil with a bubble point temperature (at relieving condition) that exceeds the max vessel wall temperature (593 oC for CS as Propacket stated earlier), taking credit for insulation would not prevent vessel failure due to high temp, and the case would be dismissed for PSV sizing purposes.

There is a theory that heavy hydrocarbons (heavy gas oils, resids, etc.) undergo thermal cracking reaction under prolonged exposure to fire (similar to a Visbreaker); producing light hydrocarbon vapors which could be used for conservative PSV sizing.

 

[joeycranston]

longesthorn, Thanks for the reply. I think you are missing the point of the question. I'm not concerned with PSV sizing because, as you've rightly stated, the PSV is not going to lift. This section of the API standard is for the situation where the vessel is likely to fail due to an external fire. I agree that insulation will not necessarily prevent a vessel rupture but it could extend the heat-up time enough to allow adequate emergency response. I assume this is why the standard suggests fire-proofing. I do not understand why insulation would be excluded. Even a safeguard like water-spray has a finite amount of useful time after activation.

 

[longesthorn]

joeycranston, After re-reading the thread, I think I understand your original question a little better. You asked why, according to API 521 ,external insulation can be considered an alternate means of protection for unwetted vessels, but not for wetted vessels.

The way I read API 521 (the 1997 version) Section 3.15.5 is that credit for insulation reducing heat input from fire can be taken for both dry vessels, and vessels with a wetted surface areas. When taking credit for insulation on a vessel (refer to Table 5, and Equations 4, 6 and 9) the Environmental factor (F) is lowered; thus reducing the heat input rate, as well as vessel wall temperature rise. Ultimately, if you look at the equations, the the impact is a reduced relief rate.

API 521 ("Pressure Relieving and De-Pressuring Systems") is a guideline specifically for examining the causes of overpressure as well as to determine relief rates. It evaluates fire in terms of overpressure protection and relief system design. If a fire is likely to cause a vessel to fail from excessive temperatures, even if the PSV is sized adequately, other means of protection should be evaluated. This applies to both vessels in "dry" service and vessels containing high BP liquids. Section 3.15.4 mentions the alternative methods, but outside of depressuring, guidelines for applying these methods are be beyond API 521's scope.

There are other codes and standards that address overall plant Fire Protection, not just for overpressure protection. These documents should cover determining Fire-proofing needs, evaluating emergency response times, and fire-proofing material requirements and applications (including the use of insulation). I would recommend referring to API Publication 2218 - Fireproofing Practices in Petroleum and Petrochemical Processing Plants. This document provides a more comprehensive guideline for fire-proofing (particularly for passive methods). If I'm wrong, please let me know.

 

[joeycranston]

longesthorn - You are exactly right. I have not seen API RP 2218. I'll get a copy and review it. Thanks.

 

[MortenA]

@ongesthorn,

You write "..When taking credit for insulation on a vessel (refer to Table 5, and Equations 4, 6 and 9) the Environmental factor (F) is lowered; thus reducing the heat input rate, as well as vessel wall temperature rise." But for the wetted case the wall temp will not come down since the wall temp=boiling point - that wont change due to insulation. The question was re. with relevance of API 521 when the boiling point was above the temp where the metal has lost so much strength that it will burst anyway.

 

[longesthorn]

MortenA, what I meant by reducing the "vessel wall temperature rise" was that the rate of increase of the vessel wall's temperature will be reduced if credit for insulation is take. This doesn't mean that the vessel won't reach its structural failure temperature, it'll just take more time (both unwetted vessels and vessels with BP liquids).

For the situation being discussed, API 521 suggest other means of protection since PSV cannot protect against the vessel from failing due to excessive temperature. "Fire-proofed" insulation, and other means of protection can be used as a safeguard to ensure adequate time for emergency response, however guidelines for its application is outside of API 521's scope.

 

[dcasto]

Lets put it more simply. If you have a vessel that operates at 50 psia and 460 R and you apply heat to it and the vessel reaches 1380 R (920F), the pressure would be at 150 psia. If the vessel is rated to 250 psia and the PSV was set at 250 psia (i know, compressibility factors and such) the vessel would fail.

You can either reset the relief valve to say 100 psia, added deluge water or add a blast wall or just decide the failure of that vessel will not add enough risk to worry about.

If the gases inside are deadly you might then take a track to mitigate the risk of failure, but if they are that deadly, the relief had better be piped to a safe point and you'd better get real particular with the seals on valves and connects because they will be close to failing.

 

[joeycranston]

dcasto, Thank you for replying. Your example is a good one. The key difference with regard to insulating the vessel is the time to reach 1380 R. Without insulation, it may take 15 minutes. With insulation, it may take 2 hours. From an emergency response standpoint, these are dramatically different situations. I can feel quite comfortable that, if after 2 hours there is still a fire burning, I have sufficient evacuated people from the area. However, the same is not true for a 15 minute scenario. As others have said, API 521 is only considering ways to keep the vessel from rupturing and not considering the affect on people. So, I assume that is why insulation is not included.