The system in question is a three stage ejector system pulling vacuum on a flash tank. The vacuum in the tank needs to reach a design pressure of 36 mmHgA for the product to meet specification. The vacuum system, when in service, was unable to pull down below 240 mmHgA. This meant that no product could be made until the cause of the vacuum issues could be identified and corrected.
A physical inspection of the equipment uncovered no major findings, the ejectors and condensers were in good condition. The last stage ejector was no loaded by itself against its suction isolation valve and pulled a vacuum of 37 mmHgA. When the suction valve was opened to test the second stage ejector, the vacuum degraded immediately to 74 mmHgA suggesting that a source of load was present, most likely an air leak. The second stage ejector was then turned on. The second stage ejector achieved a vacuum of 9 mmHgA, which is in line with the expected 10 mmHgA no load pressure, but after 5 minutes the pressure degraded to something closer to 60 mmHgA with stability issues. The first stage ejector was then turned on and the pressure at the inlet of the system was measured to be 185 mmHgA. At no load the pressure should be closer to 1.0 mmHgA with the three stage ejectors.
The time it took for the second stage ejector’s no load pressure to degrade from a good pressure of 9.0 mmHgA to an unacceptable pressure of 60 mmHgA was a red flag that condensate flooding was suspected. Once the second stage ejector was placed in service, it took 5 minutes for the condensate to back up into the second intercondenser and flood it. This observation along with the varying third stage suction pressure indicated both an air leak and condensate back up into the second intercondenser. The system was flooded with water and hydro-tested. One floor below the ejector installation, a stream of water was found to be shooting out of the condensate drain leg due to a failure at a joint in the fiberglass piping.
In this instance, condenser flooding was determined to be caused by an air leak. Air was leaking into a hole in that condenser’s drain leg. The drain legs in a vacuum system are under vacuum down to the liquid level in the hotwell. This means that liquid doesn’t leak out of a failure but air is pulled in. This often happens at the water / vapor interface point part way up the drain leg or near the liquid level in the hotwell. In this case, the air that was being pulled into the system was rising up towards the vacuum system. This was opposite the flow of condensate draining down into the hotwell. This air bubble in the intercondenser’s condensate leg prevented the free drainage of condensate causing it to back up into the condenser.
Thermal imaging of the condenser confirmed the condenser was holding a liquid level. As condensate floods up into a tube bundle, the cooling water inside the tubes sub-cools the condensate. A condensate temperature near the cooling water supply temperature allows for flooding to be easily seen via thermal imaging, assuming the condenser is not insulated.