Heat Xchanger tube rupture

[moemantha]

I am trying to size a tube rupture PSV on a heat exchanger, where there are two phases on the high pressure side and one phase on the low pressure side. Can anyone help or refer me to an article that explains this?

 

[TD2K]

That's an ugly case. It will be interesting to see what others suggest.

I'd start off by assuming the 2 phases are fairly well mixed on the tube side so the same vapor/liquid % flows through the broken tube as exists on the high pressure sid. For the vapor, I'd treat it as a compressible fluid passing through an orifice. For the liquid, I'd treat it as a liquid at the high pressure side conditions, ignore any flashing effects in the, and only flash it to vapor and liquid after it is in the low pressure side. You'll have some interations to do until for the inlet % of vapor and liquid, you get a flow rate that passes through twice the area of a tube rupture (accounting for each side of the ruptured tube). The temperature of the low pressure side (is it hotter or colder than the high pressure side?) may be a factor in estimating the final amount of vapor/liquid to be relieved.

Some companies do not allow you to take credit for the effect of the tube feeding one side of the tube rupture. This gives you a conservative answer.

I'd also look at different % between vapor and liquid on the high pressure side (assuming the high pressure side isn't evenly mixed) just to make sure that I don't have a case requiring a much higher required relieving load.

If the PSV is on the shell of the exchanger, then it can be sized for the final amount of vapor and liquid to be relieved. If the PSV is however on associated piping, you'll need to consider it having to relieve the equivalent volume of fluid between the exchanger and the PSV that is displaced by the high pressure fluid leaking into the low pressure side. I say need because there may be other potential relief paths to vessels or other areas that you can take credit for though quantifying these flows is again difficult.

Finally, this is an area you might want to try and review under the AIChE's DIERS work on multiphase relief. The approach that I've described (briefly) is how we have done these but this is not a straightforward calculation.

 

[TD2K]

Sorry for the somewhat disjointed posting above (I know, 'what's new?').

I had just finished typing it up and was trying to improve the 'readability' when I got pulled into a 3-way telephone conference call. As I didn't know how long posts can sit on hold (and I didn't want to try to retype it) I hit the 'submit' button rather than risk losing it all.

 

[saidsaz]

if its a new design might be feasible to use API 2/3 rule ..and forget about it ..

if u really need to size a psv ..i wonder how complicated it will be doing it in a simulation software ,,assuming rubtured tube an orifice....