Online Cleaning of Compressor Impellers
Online Cleaning of Compressor Impellers
(OP)
Hello. I was wondering if there is an approved method of cleaning centrifugal compressor rotors while in service. Our main 8MW compressor is once again experiencing a build up of material on the 4th stage impeller due to a leak in the 3rd to 4th stage intercooler. The corrosion inhibitor is precipitating out of the water stream and is mixing with soot particles in the air stream to form a hard, chalky residue on the blades. What is unfortunate, is that we just completed this repair less than 2 weeks ago. A new leak has apparently formed in the tube nest and is creating the same problem. This is a 40 year old Ingersoll Rand/Dresser rand unit that is acting as a large exhauster for an engine test cell. The airflow though the unit ranges from 7 to 30lbs/sec, which is actually fairly low compared to its design potential, so I am thinking velocity through the unit is probably not terribly high.
This is the final stage and is discharging to atmosphere, so any residue it discharges will be going down the line and to the stack.
Getting the volute off was not easy as it was corroded on very tight to the gearcase spigot. We initially attempted to remove the inlet cone from the volute, but it is a two piece affair, and once the outer cone was off, the amount of corrosion discovered on the inner bolts, housing and alignment screws convinced us just taking the whole volute off was the simpler route.
If there was a way to dissolve this build up while in service, it could allow us to continue in operation during critical testing timeframe commitments. We are 2 weeks off from our annual shutdown, so it looks like we will go with the manual route for now. New tube nests are under construction and are scheduled to arrive about 2 weeks after our shutdown has finished (naturally). Unfortunately, due to changes in the piping codes since the original tube nests were constructed, the new tube sheets have to be an inch thicker, which will necessitate raising the cooling water supply and return piping network by an inch. Fun times.
I manually cleaned the impeller and it took 5 1/2 hrs to get it all off.
This is the final stage and is discharging to atmosphere, so any residue it discharges will be going down the line and to the stack.
Getting the volute off was not easy as it was corroded on very tight to the gearcase spigot. We initially attempted to remove the inlet cone from the volute, but it is a two piece affair, and once the outer cone was off, the amount of corrosion discovered on the inner bolts, housing and alignment screws convinced us just taking the whole volute off was the simpler route.
If there was a way to dissolve this build up while in service, it could allow us to continue in operation during critical testing timeframe commitments. We are 2 weeks off from our annual shutdown, so it looks like we will go with the manual route for now. New tube nests are under construction and are scheduled to arrive about 2 weeks after our shutdown has finished (naturally). Unfortunately, due to changes in the piping codes since the original tube nests were constructed, the new tube sheets have to be an inch thicker, which will necessitate raising the cooling water supply and return piping network by an inch. Fun times.
I manually cleaned the impeller and it took 5 1/2 hrs to get it all off.
RE: Online Cleaning of Compressor Impellers
https://www.waterlinepublication.org.uk/articles/t...
Since complete removal of dissolved O2 in the cooling water is not possible, injecting a solution of chelates such as EDTA etc may not be of much use here. In any case, injection of a liquid mist stream into the feedstream may result in high vibration. Talk to the water treatment folks also. Risky proposition.
Online injection of a wash solvent into the compressor is common practice in high pressure cracked gas compressors in ethylene production plants, but not reliable - high vibration is a common problem.
Ask for seal welded tubes for the tube to tubesheet joints in the new compressor intercooler tube bundle if it is not too late. What materials are specified for cooling water side (tubesheet) and for the tubes ?
RE: Online Cleaning of Compressor Impellers
The tubesheets will be plain carbon steel with Cupro/Nickel tubes as per the original as-delivered drawings. They are expanded to fit the tubesheet. It was frustrating to find out the OEM for the coolers had disposed of the original engineering drawings. Not unreasonable after 40 years I suppose, but frustrating. We had an engineering firm come by last summer and do 3D laser scanning of the existing units to create new drawings to fab the replacement coolers. The drawings turned out very nice. Hopefully they don't slide in 80% of the way then go "thunk" and stop moving. They should eliminate the leaks for quite awhile, hopefully at least for the next 8.5 years until I retire. "No leaks" is of course the ultimate solution to all of this problem.
My original idea was to introduce a sidestream of distilled water steam through a 3/4" NPT port on the pipe leading into the 4th stage to hopefully keep the debris fluid enough to allow the impeller to shed it in operation, since the entire compressor is in a below ambient pressure condition acting as an exhauster.
RE: Online Cleaning of Compressor Impellers
https://www.bing.com/search?pglt=43&q=morton+e...
One or more proximity probes (journal bearings) or accelerometers with 1/rev tachometer with appropriate measurement instruments should be used to monitor during the initial injection tests, periodic injection, or continuous monitoring.
Why create a new problem, possibly a wreck, when dealing with the current fouling problem?
Walt
RE: Online Cleaning of Compressor Impellers
RE: Online Cleaning of Compressor Impellers
Oddly enough, vibration has fallen by from 50 microns/sec to 42 microns/sec over the past few days, taking it from the lower alarm threshold to about the upper 1/3 of normal operating range. We will go back in anyway.