Liquid Thermal Expansion

[msg12]

My application is a direct fired heater in a hot oil system which has a PSV on the outlet of the heater coils for thermal expansion due to blocked in coils. I would use the API 521 4.4.12.3 (6th ed) equation for liquid thermal expansion (Q=BH/500GC) to size the PSV; however, my contractor is using the equation for pressure rise in piping (4.4.12.4.1) and saying that due to this calculation, the final pressure calculated is less than the design Pressure of the coils so no PSV is required. This doesn't really make sense to me because that pressure rise equation doesn't take into account the duty that the heater is putting into the coils, as well as they assumed a final temperature as the design Temperature of the coils as if it couldn't get any hotter than that?

Please help

 

[Duwe6]

". . direct fired heater in a hot oil system."

How about Overfiring Scenario during 'blocked flow'. That is a goodly ammount of vapor to relieve. [reciently review HTO Furnace PRV calc's, and that was the 'driving' scenario]

Who cares about the "design pressure in the coils"? The pipes are the strongest element in a direct-fired furnace. What is the MAWP of the boundary valves, the flanges [if any] and instrumentation? Contractor's reasoning sounds to me like "since I am driving a sedan, there will be negligible damage when/if I hit a tricycle" -- absolutely true, but it is far from the 'driving' scenario.

 

[zdas04]

I'm not sure what the variables in your equations are and I'm not going to take the time to pull out API 521. The real question is "is there any credible way that can result in the vessel becoming blocked in while liquid full?". If not then you can be confident that over pressure from thermal expansion is not credible. If you can block in a liquid-full system then it is another thing altogether.

The coefficient of thermal expansion for water at 20C is 0.000207/C. For water
dV=Vol*β*ΔT

The bulk modulus(i.e., the amount of pressure required to reduce the fluid volume by 1%) of water at 20C and 1 bara (change is slow enough with pressure increases to disregard) is 2200 MPa so
dP=dV*BulkModulus/0.01/Vol

This resolves into
dP=100*dT*β*BulkModulus

Plugging values for a 1C temperature increase you get:
dP*100*1C*0.000207/C*2200MPa =42.5 MPa/C

If there is a credible scenario that can result in the system being shut in liquid full, then thermal expansion will exceed your PSV set point pretty early in the process. The good news is that getting rid of an infinitesimal amount of liquid will resolve the over pressure (i.e., tiny PSV's are all that is required for this case).

David Simpson, PE
MuleShoe Engineering

Law is the common force organized to act as an obstacle of injustice Frédéric Bastiat

 

[don1980]

Adding to the good comments from Zdas04, be aware that the coils will still fail if the burners continue while the coil-side is blocked in. The PSV will prevent hydraulic expansion failure during the early moments of this scenario, but the PSV won't prevent the eventual failure of the coil due to excessive temperature, if this scenario continues for very long.

 

[TD2K]

I've seen PSVs sized for thermal expansion on the outlet of heaters and I'm not very fond of them. It's true that the volume of liquid to be relieved isn't very high using the Q=BH/500GC but if you aren't going to assume that the heater trips, the temperature of the hot oil will continue to rise, what is the boiling point of this hot oil? Do you start to get a flashing liquid through the PSV or worse, a vapor/liquid relief load? Then on top of that you have the coil configuration in the heater to think about, if you start to form vapor, will that free flow to the PSV or does it try to push a 'plug' of liquid out ahead of it?

I'm assuming you have low flow trips on the heater, just review your system and satisfy yourself you want to take credit for those.