13 replies to this topic

Hi,
I am trying to size a PSV for tube rupture case in a sour gas heater.
Gas in shell at 0.7barg (3.5barg design) and Saturate Steam in tubeside @ 41.5barg (49 desing).

After reading section 5.19 (API 521 2007) I am trying to find the right equations to use. Could you please give me any enlightment? Thanks.

Another question is, as I have HP Condensate in tubeside outlet, what my fluid would be?
2phase mixture? and how could I messure L and V flowrates? If liquid flashes I dont know how to find it.

Thank you very much for your help beforehand.

[CMA010]

You basically have 2 orifices, so you can calculate the flow as such. The HP condensate flowing out of the tube will flash. Based on the set pressure of your PSV (+10%) and the maximum pressure of the HP condensate you can calculate the percentage of flash.

[JoeWong]

Hi,
I am trying to size a PSV for tube rupture case in a sour gas heater.
Gas in shell at 0.7barg (3.5barg design) and Saturate Steam in tubeside @ 41.5barg (49 desing).

After reading section 5.19 (API 521 2007) I am trying to find the right equations to use. Could you please give me any enlightment? Thanks.

Another question is, as I have HP Condensate in tubeside outlet, what my fluid would be?
2phase mixture? and how could I messure L and V flowrates? If liquid flashes I dont know how to find it.

Thank you very much for your help beforehand.

This will be a multiple complicated relief scenarios (check out here) and your exchanger may expose to severe surging and vibration during tube rupture. Don't think a straight forward PSV attaching to the shell will eliminate the potential of catatrophic failure of heat exchanger which potential results discharge to atmosphere. I think it should be manage using other ways...

Any chances to manage it in other way like...

i) Use low pressure steam so that you eliminate the tube rupture scenario
ii) Steam supply from high pressure steam header (OP at 41.5 barg) but the condensate return to low pressure condensate system e.g. 3.5 barg condensate system
iii) Enhanced strength welding
iv) High integrity protection system.
.
.
.

[pleckner]

I don't think you have a 2-phase relief situation here. The steam will fill the shell and pop the PSV. Any condensate in the tubes at the time of break will quickly flash off to steam. With the heat transfer interrupted, you won't be condensing much steam afterwards.

But I agree with Joe that a PSV might not be the answer. A shell & tube heat exchanger is a relatively small vessel and pressure surges can be experienced. If your company standard allows, consider the use of a rupture disk instead. The first answer you will get is no-way, if it bursts we loose everything. But in reality, a ruptured tube is not common and rarely causes catastrophic failure. It may not even cause the PSV to open as it is more likely there would be a small leak of steam into the shell, which you should pick up in operations.

The tube rupture scenario calculation is an orifice equation (see CMA010 response) but you use choked flow for the steam rate. By the way, assuming all steam through these orifices adds to the conservatism.

[JoeWong]

I don't think you have a 2-phase relief situation here. The steam will fill the shell and pop the PSV. Any condensate in the tubes at the time of break will quickly flash off to steam. With the heat transfer interrupted, you won't be condensing much steam afterwards.

Phil,
I have the similar way of thinking like you initially before i submit my earlier response. "The condensate will quickly flash off to form steam...". But just to make sure above statement, i have a quick check...and found that condensate at 41.5 barg may not 100% flash to steam at 3.5 barg. There still remain 70% of condensate in steam at shell side...Thus i immediate change my response.

I don't think you have a 2-phase relief situation here. The steam will fill the shell and pop the PSV. Any condensate in the tubes at the time of break will quickly flash off to steam. With the heat transfer interrupted, you won't be condensing much steam afterwards.

Thank you very much to all you guys for the help.
My thought at the end of the evening yesterday agreed with quote above, so I went for the steam case. However, after reading last Joe's comment, I am not sure anymore.

Phil,
I have the similar way of thinking like you initially before i submit my earlier response. "The condensate will quickly flash off to form steam...". But just to make sure above statement, i have a quick check...and found that condensate at 41.5 barg may not 100% flash to steam at 3.5 barg. There still remain 70% of condensate in steam at shell side...Thus i immediate change my response.

The thing is we cant change process conditions so no low pressure steam or condensate as you purposed, and client wanted a PSV in the very last moment. I think just changing design Pressure in Shellside should be enough to avoid the scenario though (to make it higher...)

Regarding the liquid flashing or not, Joe, how could I check that? as I dont know how much condensate I would have at the tube rupture moment.

Thanks againg for your opinion guys. Let me know what you think.

[JoeWong]

The thing is we cant change process conditions so no low pressure steam or condensate as you purposed, and client wanted a PSV in the very last moment. I think just changing design Pressure in Shellside should be enough to avoid the scenario though (to make it higher...)

I am surprise that the statement "...and client wanted a PSV in the very last moment.". Looking the condition, overpressure protection system is required regardless whether client or other requesting.

One way is to increase the shell side design pressure to match the criteria of tube rupture not credible. The heater would be "protected" from overpressure of high pressure steam. But you have to think a little further the extension of this high design pressure on process line as well as upstream and downstream equipments where there is a potential of blocked in. It is complicated.

Regarding the liquid flashing or not, Joe, how could I check that? as I dont know how much condensate I would have at the tube rupture moment.

As you can read from HERE, the relieving scenario could be very complicated.Quantity of condensate in the shell side would varies. Similarly quantity of condensate in steam relieve through the PRD would varies.

As for PRD sizing, you may consider to use the Leung omega method to estimate quantity of condensate through the orifice (ruptured tube).

[pleckner]

Yes, I shot off my response assuming all condensate will flash and obviously that was dead wrong (I'll know better next time).

Guesstimate a volume of condensate by assuming the tubes are full with liquid and use the tube ID, the tube length and number of pases. Use this value to determine how much of the shell will be filled with liquid. If it is 80% or less, then I don't think this will be a two-phase relief. Once the rupture happens heat transfer is basically done and the steam should stay as steam. However the point not brought up is the fact that the relieving vapor is not just steam but a mixture of steam and sour gas.

Come up with your numbers and get back to us in this Forum.

[CMA010]

Hmm, replied to quickly. Doubt there will not be a two-phase relief. You'll have to calculate the steam flow as mentioned earlier. If there is a control valve upstream you might use it as flow limiting device.

[JoeWong]

...If there is a control valve upstream you might use it as flow limiting device.

I guess the control valve opening may not smaller than the the tube ID...thus, control valve may NOT be a flow limiting device in this case.

[fallah]

[/quote]

I guess the control valve opening may not smaller than the the tube ID...thus, control valve may be a flow limiting device in this case.
[/quote]

In these cases usually FCV installed in the condensate line which exits from the piping pot.I think if FCV installed in upstream, its opening can not be smaller than the tube ID.
Regards

[CMA010]

...If there is a control valve upstream you might use it as flow limiting device.

I guess the control valve opening may not smaller than the the tube ID...thus, control valve may NOT be a flow limiting device in this case.

Only one way to find out, calculate. Well, that's what we usually do....

[sarodeys]

Dear sir
calculation for excess flow through tube ruptre, PSV location is on reboiler
tube have high pressure
or plz give any suggistion

[knapee]

Dear Sirs

I would like to make some other idea in this post. First the PSV location generally prefer to be located on the upstream side of exchanger. In this way, the overall exchanger is protected within its allowable overpressure.

Second, the isolation valve could be another choice to installed around the heat without
a relief snide the isolation valve.

Please correct my point if I make any thing wrong.

Thank you.