8 replies to this topic

[BillyEv]

Therminol 66 vaporizes from a liquid to a vapor across the relief device, is this considered two-phase flashing flow per API520 5.11?  

 

Process Fluid: Therminol 66

Equipment Protected: ASME Heat Exchanger, Tube Side

Relief Scenario: Fire Case resulting in vaporization

Set Pressure: 160 psig

Relief Pressure: 208 psia

Relief Temp: 1002 °F

Critical Pressure: 353 psia

Critical Temperature: 1056°F

 

Should the valve open due to the fire scenario due to pressure increases greater than the relief valve set point, the Therminol 66 temperature will remain unchanged as the liquid passes thru the relief orifice, and the pressure across the orifice will drop to superimposed back pressure, which is atmospheric (minimum) at the start of the relief event.  The fluid exiting the orifice will be in the superheated zone of the Pressure-Temperature curve where the Therminol 66 can theoretically only exist as a vapor.  The built-up back pressure at 100% open full is less than the allowable backpressure and the back pressure is not high enough to condense the vapor. 

 

Therminol 66 vaporizes from a liquid to a vapor across the relief device, is this considered two-phase flashing flow per API520 5.11?

 

[latexman]

Yes, it is two-phase flashing flow.

[BillyEv]

Yes, it is two-phase flashing flow.

Figured as much.  A colleague said it was not, when I questioned him on a calc that he previously submitted for a similar application (but that valve was never installed).  To me the code is pretty dang clear. I used ASPEN Hysys to create a hypothetical component with fluid properties that closely match what is published on Therminol 66.  I then used ASPEN Hysys to generate the fluid (and vapor) properties required for the omega parameter and other required inputs to calculate the orifice size.

 

My primary concern is the valve size (orifice size) calculated when I use the formulas contained within API 520 Appendix C.2.2.  The Crosby JLTJBS valve (and orifice) is larger than most colleague expect for this application and the concern about oversized valves has been previously raised due to the shock stresses caused on the pipe system from a quick open, quick burp, quick close.

 

I'm going to create a dynamic PSV relief scenario within HYSYS for my system in hopes that the relief valve size will be able to be reduced somewhat from a simple vapor creation assumption / calculation I used initially.

[latexman]

I can imagine when the fire scenario starts there might initially be liquid thermal expansion activating the PSV with no flashing.  Then, the Therminol 66 heats up more/closer to a sat'd liquid state, and there will be a time when liquid thermal expansion activates the PSV with flashing in the tailpipe.  After a long time and some loss of mass upstream of the PSV, the Therminol 66 vapor pressure might activate the PSV, and there could be two-phase venting if there isn't sufficient disengaging volume upstream of the PSV, or vapor only venting, if there is.

 

Could your colleague and you be envisioning different stages of the same scenario?  I can't really tell from what was said.

 

Btw, I use Aspen Plus v.12.1, and Therminol 66, a.k.a.THERM66, properties are already available in Databank APV121.PU.  Are these not available to Aspen Hysys?  Maybe these have not been harmonized yet.

[BillyEv]

I can imagine...

Your imagined fire scenario is basically what I envision to occur and basically what my Colleague described what would occur. I think my colleague has tunnel vision about the lack of two-phase flashing flow as he already submitted a calc for a relief device for a valve that was never installed.  He calculated the vaporization rate in lbs/hr then calculated the orifice size using techniques in API-520 for fluid flow, then selected a liquid trim Consolidated (Dresser) valve. 

 

I discovered sizing issues with his calc as I was tasked to size four or five more relief devices for similar applications.  For the two-phase flow omega factor, I assume worst case at the initial stage of vaporization where the relief flow is primarily saturated liquid, vs toward the end of a relief event where vapor venting is more likely as the fluid mass trapped evaporates.

 

I'll check out the Therminol 66 properties in Aspen Plus v.12.1 Databank APV121.PU.  I did not find anything in Hysys V11.  My hypothetical was very close to the  published relief properties  Eastman Chemical furnished me (minimal info) initially.  Their data lacked some properties I needed, available in HYSYS.   The hypo was close enough for my needs at the time I used it to generate physical properties at different relief pressures/temperatures.  

Edited by BillyEv, 16 August 2022 - 07:35 AM.

[Bobby Strain]

You can use the HEM model which will likely reduce the calculated orifice size. Also, you can add fireproofing insulation to the heads and piping. This will then reduce the relieving temperature to that of the shell side relieving temperature.

 

Bobby

[breizh]

Hi,

 

You may want a set of properties for Therminol 66 together with safety reminders about HTF and calculation methods to size relief valve. 

 

https://5.imimg.com/...herminol-66.pdf

 

Breizh

[breizh]

Hi,

Consider this resource to support your work .

Breizh

 

[Chemitofreak]

What kind of heat exchanger is it ?

 

Shell and tube heat exchangers that have large bolted body flanges (e.g.,channel heads) tend to self-relieve during fire exposure because the heat relaxes the long flange bolts. If this is your case then you can get away with a RV. 

 

Is therminol your hot fluid?

In case of fire even if there is a liquid thermal expansion, it will not be a governing scenario.

 

What is the compressibility of therminol vapour at relief condition, if it is less than 0.8 then I would suggest you to do HEM Enthalpy calculations.