API 521 and jacketed vessels 1

[pianoman1]

I'm working on some pressure relief systems for jackets that are ASME BPVC Sec VIII stamped at 55 psig, 300F, but the tanks themselves are not code pressure vessels. I'm looking at three scenarios - closed outlet valves, open pool fire and pressure regulator failure. Failed pressure regulator seems straightforward, evaluate the cooling water supply piping and see if friction losses keep the flow rate below the flow capacity of the SRV at the relieving pressure. Regulators all have a max Cv to use in this calc. Closed outlet valve is also easy, the pressure regulator will still limit pressure downstream with no flow. I don't believe double jeopardy applies as a failed pressure regulator would be unlikely to be considered a 'latent failure', I think it's the kind of thing you'd notice in operation for tanks on the operating floor in a system that is batch-run on a daily basis.

The open pool fire is easy if I can demonstrate that cooling water will continue to flow through the jacket at a rate that absorbs the heat without getting to the boiling point. However, if there's more heat than the flow can absorb, I'm dealing with two-phase flow in downstream piping, probably slug flow, trying to predict if the resulting backpressure will exceed the SRV setpoint, if two-phase flow needs to be considered for the lifted safety, etc., and I'm not sure how to deal with this. The tanks are vertical SS, less than 1000 gallons, typical contents are isopropyl alcohol (IPA) at 80% with water. I assumed the tank was empty of product for worst case.

Anyone been through this with jackets? API 521 has no specific detail on jacketed tanks. Thanks for any feedback.

 

[don1980]

I deal with this question (how to size a PSV to protect jackets from fire exposure) quite often. People get stumped by it for a very good reason - it's because a solution doesn't exist. A PSV is necessary for code compliance, but it's wasteful and pointless to size the PSV for fire exposure. Proceed with sizing the PSV based on the other scenarios, and if none exist then install a thermal PSV. Generally, liquid thermal expansion is a credible risk, and it's one that you can safely defend against by using a PSV.

Why can't a PSV defend a jacket from fire exposure?
[ul]
[li]Due to the narrow passages (no chance for vapor-liq disengagement) this is a classic 2-phase scenario. But when you size the PSV for 2-phase flow, you'll generally find that the PSV inlet is larger than the jacket nozzles.[/li]
[li]The heat rate from fire exposure is extremely high, due to the high surface area to volume ratio. This means the jacket contents come to a boil quickly, and the jacket is emptied quickly. Then you have a vapor filled jacket will will fail quickly due to excessive wall temperature. So, the jacket will fail regardless of whether the PSV is sized for fire exposure or not.[/li]
[li]A fusible plug can be used to empty the jacket before it fails, but the jacket will still fail. So the question is whether or not there's any safety value in installing a fusible plug. That's a case-by-case decision. My opinion is that it's generally not needed because no one is going to be standing near the vessel while it's surrounded by fire. When the jacket fails, the amount of energy released is relatively low. There's little risk that jacket failure will even dislodge the vessel, much less sending it flying into the air.[/li]
[/ul]

 

[pianoman1]

Excellent response, thank you for this. I have one question, tho - in my case I assume that cooling water will still continue to flow through the jacket, it sounds like you're assuming the jacket would be closed in during a fire. If the heat added is only enough to flash, say, 20% of the mass flow rate, is it worth taking a swag at the PSV using two-phase flow? I agree that the hazard is very low, as you said, if the jacket ruptures it'll just spill water - onto the fire - where no-one will be standing.

 

[don1980]

I based my comments on a closed-in system because that's the default assumption when designing for fire scenarios. When the system isn't closed, there's rarely a risk of overpressure. In other words, we assume the system is closed because there's no credible cause of overpressure if we don't. I think this applies to this jacket case too. If vapor is formed while the system is open, the resulting increase in pressure will push the fluid through the piping more rapidly. If the jacket isn't insulated, then it's likely you'll get complete vaporization between the time the fluid enters and exits the jacket. Fire exposure to a jacket is a close analog to a fired steam boiler. So, will complete vaporization cause overpressure? That very unlikely. But if you wanted to verify that for yourself, calculate the friction loss in the outlet system to see how much backpressure that puts on the jacket.

 

[pianoman1]

I see, I can easily evaluate the downstream piping for steam flow at the makeup mass flow rate and see if it backs up the pressure enough to cause the safety to lift, then evaluate the safety at vapor conditions. I can also evaluate the jacket at closed-in conditions, although I don't think it's credible they'll operate that way if the place catches fire.

Thanks again, Don.

 

[pianoman1]

Don - you mentioned that "The heat rate from fire exposure is extremely high, due to the high surface area to volume ratio." Does this mean that the constant 'C' used in determining the heat absorption rate in the API 521 calc Q=C*F*A^.82, which is normally 21,000 for areas with adequate drainage, should be one of the higher values from Table A-1 to get a more realistic heat absorption rate?