3 replies to this topic

[dehn0045]

In my process there is a vacuum system that removes non-condensables from the system. To pull the vacuum there are liquid ring rotary compressors. To prevent contamination of the compressor oil there is a knock-out condenser. This condenser is oversized (the process outlet is very close to the outlet cooling water temperature which is very near the inlet cooling water temperature -- no measurable difference between all 3 values). The problem is that I am seeing severe contamination of the oil in the first vacuum compressor in the line. The only thing that I can think is that there is liquid entraining from the top of the knock-out condenser. Does anyone anyone know any way that I can predict whether I should expect liquid entraining out of the top of this condenser.

Notes: The exchanger is set up as an 8 pass shell and tube exchange, cooling water on the tube side. The boiling point of the condensate is roughly 250 degF at process pressure. The process outlet temperature is near ambient -- 80 degF. I have already checked out the reflux drum level and the liquid drain does not appear to be plugged. I don't believe self-venting flow is an issue as the condensate flow rate is only about 200 lb/hr.

 vacuum_system.xls   25KB   100 downloads

[eilpar]

Your observation shows that there is hydrocarbon carryover to the comressor inlet header. What is the hydrocarbon your condensing? Is it very light? Is its vapor pressure at 85 C not too low? If so, some of it may go along with the non condendables (NC) depending on its partial pressure. I have observed that in crtitical cases chilled water/ brine is used to make the NC hydrocarbon free.

Another question is on the possibility of bypass in the KO condenser? Please have a look at the fabrication drawing. If the top of tube bundle-shell gap is not small there may be leakage of gas from the inlet to the outlet.

Regards

PAR

[Art Montemayor]

dehn0045:

I presume that you are using lube oil (or something similar as your seal fluid in the Liquid Ring Vacuum Pump (LRVP). I also assume you are doing this to avoid the high vapor pressure of water as a seal fluid at the vacuum level you are working at. This is not a problem – as long as you keep other liquids out of the oil. You are doing the right thing in showing concern. Please refer to the Rev1 version of your workbook. You also did the right thing in generating a workbook sketch with full description of your system. I wish everybody that put a query on these Forums did the same. Note how easily I was able to quickly spot the areas that I know – from practical field experience dealing with LRVPs - should be modified or corrected to operate in an expected manner.

You should not expect any liquid to entrain out of the condenser with the few vacuum vapors you are pulling. You are doing the right thing in attacking this problem NOW – before it gets out of hand. You are operating a unit that was probably piped up by a contractor with a penchant to “save” money by combining several pipes – especially the equalizer line and the pump’s minimum recycle flow. You should never – but never – put any obstructions in the flow of the equalization line or in the condenser’s gravity condensate flow. These lines are much too sensitive to be molested by other flows competing for the same pipe volumetric flow. I strongly recommend you do the following:

1. Modify you system the way I have shown it in Rev1. Note that I show in bold RED those lines that you must modify as priority.
2. 80 ft/sec is a very fast velocity for a vapor in a vacuum line coming out of a condenser or a 2-phase separation step (like the condenser and the overheads separator that I recommend you install). The small, overheads separator that I recommend is there to protect the LRVP during surges or process excursions. I consider this as protection and vital to keeping any foreign liquid out of the LRVPs. The moment you contaminate the oil in LRVPs with a higher vapor pressure liquid, you are going to witness vacuum failure. The LRVP is only as good as the low vapor pressure liquid that makes the seal in the “liquid piston” it uses to “pump” the non-condensables. When the vapor pressure of the seal liquid goes up, the LRVP starts to fail.
3. Make sure that your ACTUAL velocities are low in the overhead lines. By this, I mean that you should take into consideration that as the vacuum gets lower and lower, the velocity of the non-condensables that you are “pulling” get higher and higher due to their increased specific volume (the inverse of density). This is a common oversight on the part of many engineers and the effect of the higher specific volume causes a higher pressure drop and more entrainment. This is exactly what you don’t need in a vacuum system. Always keep your vacuum lines “short & sweet” – with a conservative, low velocity. Make them all self-draining, without any liquid traps or low spots.
4. Obtain a copy of my Excel workbook, “Producing and Maintaining a Vacuum” – which you can download for free in our Forums. Just use the SEARCH function and you will find it.

I hope these comments serve to correct the flaws in your installation and make your engineering life a lot sweeter and more fun.
 Vacuum_System_Rev1.xls   54KB   144 downloads

[dehn0045]

PAR -- Thank you for your comments. I have done relatively simple Aspen simulations of the process flows and have determined that based on VLE the oil should not become contaminated UNLESS there is condensed liquid being entrained into the suction of the vacuum pump.

Art -- Thank you for such a detailed response. You recommended less than 20 to 25 ft/s gas velocity for this type of application. I assume that this is based on some literature or practical experience. If you could recommend some literature on this topic I would be very interested.