I can find lots of information on providing relief protection on the cold side of a liquid/liquid heat exchanger to prevent overpressure if the liquid is blocked in and subsequently warms up, but what about vacuum protection for the hot side? If the hot side is blocked in and flow is maintained on the cold side, a partial vacuum condition could occur on the hot side due to thermal contraction of the liquid. Has anyone provided vacuum relief in this scenario?

Thanks,

Mike

The blocked-in senario you refer to usually is discussed in terms of tube-side protection. If we were to assume your question refers to the tube-side, then no, the tubes will be able to handle a vacuum condition. And besides, and I've been beating this one to death, there is no code that requires the exchanger tubes to be protected from failure, just the shell (assuming this is to be designed as a pressure vessel in the first place).

Just for information, the shell-side of a heat exchanger should be protected against vacuum and one way is with a vacuum breaker. However, if a vacuum could develop in the shell, I would just design the shell for full vacuum and be done with it.

Pleckner - Thanks for the response. I should have mentioned that this is a plate and frame heat exchanger. I am not so worried about the heat exchanger itself as I am about expansion joints that could be blocked in with the heat exchanger. Any thoughts?

Thanks again,

Mike

In a liquid filled system, we are concerned about thermal expansion because just a tiny amount of expansion will cause a huge pressure increase. If the dynamics went the other way, I seriously doubt you would experience any realistically detectible drop in pressure due to liquid contraction; it just wouldn't matter. Remember, even an atmospheric API vessel will take an inch of water pressure or so and a few ounces of vacuum. These components you are talking about had better be designed better than this.

So, without seeing any numbers to the contrary, that's my take; which maybe why I haven't really seen this as a problem in the past.

Thanks again. I think that closes the case for me. That explains why I couldn't find any discussion on this topic anywhere else!