Hot Oil Drain Tank with N2 Blanketing System PSV Sizing

[isidore72]

Dear All, I am looking for pointers to design for PSV for Hot Oil Drain Tank with N2 Blanketing System.
I am designing the PSV based on Fire case and Wetted vessel scenario, and I have made the following basis and would like to know if they are correct.
During External Fire Scenario, hot oil tank and its contents (both Hot Oil and N2) that would be subjected to heat. How should I approach the sizing scenario for the PSV.
a) Should I size the PSV based on a wetted vessel (as it contains HOT Oil), and take into account the ONLY hot oil vapour generated due to heat, how about the N2 Gas Expansion?
b) I attempted to size it based on unwetted vessel , only taking into account of the N2 relieving, but I however think that this might not be appropriate and also I found that the Relieving Temperature exceeds the specified Wall Vessel Temperature (1100F) due to big difference between the N2 Blanket Operating pressure(1.08 BarA) and High PSV Set Point (4.5 BarG).
Or I should size for both scenario and take the worst case scenario? Please advise.

I have summarized my key findings in the attachment below for both cases with the suggested orifice area for reference.

Please find the link as the following for attachment Link

 

[met11]

What is the hot oil? A heat transfer oil?

If it will vaporize at the relief set pressure, then you can get your relief rate from the rate of vaporization using the equations in API 521. If it doesn't vaporize (i.e. high boiling point), the relief valve will protect against the initial expansion of the gas which would be negligible, but it would not be adequate to protect the vessel. You'd have to look at alternatives such as emergency depressurization, fusible plugs, deluge etc.

 

[jari001]

No attachment came through.

 

[don1980]

First, determine the boiling point of the oil at the relieving pressure, and assess whether the vessel will remain intact at that temperature. If it will, then proceed with calculating a heat input from fire. If it won't, then there's no point in proceeding with sizing the PSV for fire.

When calculating the heat input from fire, API 521 recommends using the wetted area. Heat transfer through an unwetted wall is very low by comparison.

 

[isidore72]

Dear All,

I have re-attached the attachment via Link here.

To be more spcefic, the Hot oil used is Dowtherm Q Heat transfer Oil.

For Wetted Vessel Case (i.e. Hot Oil vaporization case)

I performed flash using Hysys to get the boiling point of Hot Oil 419 C at the relieving pressure i.e. 12.5 =barA [ PSV set P of 9.5 barG with 21% accumulation. ]

Relieving Temp is below Wall Vessel Temp, TW which is assumed to be 1100 F (593 C) at this point.

For Unwetted Vessel Case (i.e. Nitrogen gas expansion case)

I have 1 major concern over the Unwetted Case calculations.

a) Exposed surface area of Vessel, A' - In reality my vessel is a wetted vessel with a column of vapour space of N2 sitting on top of the Hot Oil. Should I estimate the exposed surface area only of the N2 portion for my calculation, which is say the top half portion of the vessel only?

The Relieving Temperature of 1556 C exceeds the Wall Temp, hence I had to specified the F' value to 0.01 and proceed on with the rest of the calculation, as per API 512 4.4.13.2.4.3.

For both cases, I get surprisingly similar orifice size which is 2H3 (with very different relief loads of course, one is N2 and the other being Hot oil vapour.

Thanks
Isidore

 

[met11]

You definitely shouldn't get a similar orifice size, I wouldn't even consider the unwetted case as a valid case. Operating pressure is so much lower than the design pressure that the vessel would melt long before the set pressure is reached.

For the wetted case, I don't have Hysys so I can't check. But your relief pressure in your post doesn't match the pressure in your document (9.5 barg vs. 4.5 barg in document). Assuming a relief pressure of 5.445 barg, the Dowtherm Q would start to vaporize around 367°C per the information online. You can get a heat of vaporization from the vapor pressure data (my old University buddies Clausius and Clapeyron had an equation to get close). This is all pretty straightforward, and summarized in API 521. Just assume the vessel is full for calculating the wetted area, which would be worst case.

 

[jari001]

my old University buddies Clausius and Clapeyron had an equation to get close

I love it! lol

 

[isidore72]

Dear met11,

Thanks for your interest.

I apologies for the confusion over the SetP, I was sizing for a multiple tanks here, just let's say Let's just assume 9.5 barG is the one for further discussion.

With respect to your suggestion that unwetted case is not a valid scenario, my calculation tend to suggest Case B where Nitrogen Gas Expansion Case would require larger orifice albeit a lower relief load than that of Case A Hot Oil Evaporation.


This case even apply for my other calculation for Cyclopentane and N2 System too, whereby Cyclopentane has an even lower bubble point than Hot Oil, where the rate of evaporation of Cyclopentane is higher than Hot Oil - I know there' s a whole lot of factors contributing to this, just looking for a thumb rule.

Thanks,
Isidore

 

[met11]

"The Relieving Temperature of 1556 C exceeds the Wall Temp, hence I had to specified the F' value to 0.01 and proceed on with the rest of the calculation, as per API 512 4.4.13.2.4.3."

The unwetted case may be giving you a large orifice because you've arbitrarily set the F' to 0.01, using a vessel wall temperature higher than the melting point of carbon steel.

 

[mbt22]

In other words, as met11 has said a couple of times, a relief valve is not the right control for this gas filled vessel. Flame impinges on vessel, walls heat up, internal temperature and pressure increases (probably limited by heat transfer rather than in equilibrium), strength of wall drops as temperature increases, vessel bursts. All before the design pressure is reached.

For a fire case, the source of the extra volume is vapourising liquid. The vessel sits at the boiling point of the liquid at relief pressure until you run out of liquid (and so to a first approximation the gas isn't getting hotter!).

Matt

 

[vladrath]

mbt22 and met11 are correct case B is invalid.

When you boil and relieve your hot oil your tank will experience tempering and stay at a equilibrium temperature. Any nitrogen relieved will be equivalent to a volume of oil as the oil is pushing out the N2, basically.

So just size for hot oil relief.

I would only employ the gas equation if I was dealing with a vessel that was entirely filled with a gas.