Fire Case Unwetted Surface Area - Vessel Top Head 1

[loilfan]

For an open pool fire on a vertical, uninsulated, air-filled vessel do you include the full area of the top head (ellipsoidal) when determining the unwetted surface area? It's at ground level and is less than 25' tall. I have a few different sized vessels between 8' to 12' diameter.

Since the flame provides convective and radiant heat, it seems to me that we should be taking the full area. Just curious if this is too conservative and if anyone uses a portion of the top head area for their calculation instead.

 

[The Obturator]

If I read and understand API-521, then the answer is, Yes, you should allow for the dished end. You have a gas filled vessel with 100% wetted area.
Re. API-521 Para 4.4.13.2.4.3 (Page 42) explains this. It further reads on the next page "The surface area potentially exposed to a fire should be used when determining the fire relief requirements of gas filled vessels".

Per ISO, only the term Safety Valve is used regardless of application or design.

 

[don1980]

Before you proceed with these PRV sizing calculations, consider the following sizing method which will provide results that are equally as technically valid and as safe as the results you'll get from the API sizing equation for vapor-filled vessels. As an added plus, the method I propose has some entertainment value, whereas cranking data through the API calculations has no entertainment value....in fact it has negative entertainment value.

My alternate proposal is to take a full-size P&ID (any P&ID will do) and flip it over to be back (blank) side. Divide the the page into 14 grids, and enter the 14 API orifice sizes (D-T) in the grids, one orifice in each grid. Then stand back, say 10-15 feet away, and throw a dart at the page. The resulting size will be equally as technical validity and safe as the result you get from applying the API equation....I promise I'm not making this up.

There's another alternate method, and this is actually the one I recommend. Disregard sizing for fire exposure for this vapor-filled vessel. Instead, size the PRV based on the other remaining credible scenarios. If there are no other credible scenarios, then the PRV size is arbitrary - simple choose a PSV size based on the smallest allowable nozzle (usually 1" or 1.5") for this particular vessel. PRVs are ineffective at protecting vapor-filled vessels from fire exposure. For safety, focus on other protective measures (fire resistant insulation, water spray, and automatic depressurization) which actually help reduce the risk, rather than wasting time sizing a device (PRV) which is ineffective at providing meaningful protection. Without some amount of liquid (some amount of actual wetted surface area) to remove heat from the vessel wall, that vessel is going to rapidly fail due to excessively high wall temperature. And the size of the PRV (whether you have a 1D2 or 8T10) won't have any significant effect on the time-to-failure. PRVs protect vessels from fire exposure by releasing high amounts of heat from the vessel. This occurs when heat is transferred from the vessel wall to the inner liquid (boiling liquid, Hvap) at a high rate. The liquid boils, and this heat is removed from the vessel through the PRV. The high heat transfer rate, from the vessel wall to the boiling liquid, helps to cool the vessel wall. And this buys you some time to get the fire extinguished before the vessel fails due to excessively high wall temperature.

When you have a vapor-filled vessel, the heat transfer from the vessel wall to the inside vapor is insignificantly small, because vapor has an insignificantly small heat capacity as compared to that for a liquid. This means that the wall temperature is going to rise unabated, until it reaches the yield temperature, and the vessel fails. Yes, the vapor inside the vessel will thermally expand. And depending on the starting T & P, the resulting pressure rise may or may not reach the PRV set pressure before the vessel fails. But even if it does, and the PRV opens, the amount of heat released is insignificantly small. The vapor isn't absorbing enough heat to have any real effect on the wall temperature. The PRV opens to burp out some hot vapor, and then it re-closes, while the vessel wall temperature (and the temperature of the vapor inside the vessel) continues to rise. With a vapor filled vessel, the PRV has no significant affect on the rising wall temperature.

When you have a vapor filled vessel, focus your attention on real protective measures that have real safety benefits. When you size a PRV for fire exposure, for a vapor filled vessel, you're just pretending to address the problem, and you're hiding a safety problem from the plant personnel who need to be considering alternate protective measures which have real benefits.

 

[RVAmeche]

To followup with Don's response - API 521 agrees with his methodology (of using alternative means, not throwing darts). While it provides an equation for fire relief of gas-filled vessels, it also recommends alternative measures (cooling water, depressuring systems, etc) to keep the vessel from failing.