Consistent Mechanical Seal Failure on Armstrong 4300 pump 1

[LongArmJim]

Hello,

I am a tech for a pump motor and controls company in ny and have been having issues with one set of pumps. The pumps (2) in question is an vertical Armstrong 6x6x11.5 4300 series being driven by a 25hp 1800 RPM electric motor on VFD. There is an aluminum coupling with keys between the motor and pump with 3 Allen head bolts. The pump is used for cooling the engine of a generator in a rooftop CoGen plant. The water/propylene glycol mix leaves the engine coolant circuit at 200*F and is about 177*F when it hits the pumps. There are two pumps on the condenser side of the loop, same pump, same motor, same seal that have 0 issues.

I have been back to replace these seals 2 times already this year. The only evidence that the seal is leaking is the residue left behind after the water evaporates away. Its a very flaky white residue. The seals begin to leak and the leaks gets worse and the amount of residue increases (obviously). When i remove the seal there is no evidence of cracks, excessive wear on the faces, or damage to the shaft. With a rag and some brake clean seal faces can be cleaned and appear to be slightly used but not damaged. The flushing liquid does get filtered before flushing the seal. The seals in question are 1.625" Crane Type 8B2 Armstrong #975002-334

I can provide pictures of the seal on the pump, and after cleaning the faces, (residue is still on the seal cartridge) if necessary.

Thanks for taking the time to read this, and any help would be greatly appreciated.

 

[itsmoked]

Pictures ALWAYS help. Use the

to paste them in here.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[Rputvin]

What kind of time frame are you seeing between failures? You said twice this year, but if you've got run time that might help show some trend.

The company I work for is actually dealing with a similar issue. We've had a customer smoke the seal on one of our pumps a few times. Our working theory was low flow was failing to remove heat and allowing the surfaces to overheat.

We sent the failed seals to our mechanical seal manufacturer, they performed an analysis and determined rust and particulate in the water was plating the face. We're looking at moving our seals over to a more robust version that has better heat dissipation and handles contamination a bit better.

Multiple failures can point to a service condition issue, but it can also point to an installation issue. ANY contamination on the seal face, even oil from your hands, can be enough to allow a leak, which then propagates into a real issue over time. If the techs installing the seal are pressing it in by hand without anything between their hand and the seal it's likely going to fail out of the gate, or very soon.

 

[LongArmJim]

Thanks Rputvin

I may have to send out the next one and see what is going on. They run monday-friday for about 10 hours when the Co-Gen plant is running, only one pump at a time. I think they are on a daily alternation schedule. I agree that any contaminants on the seal face would result in premature failure and I am very careful to install clean seals, make sure no contact with faces happen, and i make sure there is nothing in between them. And like i mentioned, these are the only 2 pumps i have giving me this issue. I was thinking it could be a flow issue with seal flushing, as running pressure differential is only about 18 psi. There is shelco cartridge filter on the flushing line that is clean. I know a cyclonic separator needs a 20 psi differential to work properly. Im wondering if there can be a similar issue with this filter.

 

[quadtracker]

Always make equipment safe to work on before maintenance begins. Is it possible that the runout on the rigid coupling is high. Runout should be no more than .001" Too much runout will cause small deviation in between seal faces causing leakage. The leakage may be small weep or fine spray across the gland and the wet fluid evaporates leaving chemical solids behind. The coupling runout can also be caused by a bearing failure in the electric motor some pumps use the motor bearing to support the pump shaft and impeller. Check motor bearing runout with a dial indicator on the motor shaft with the coupling removed. Grab the motor shaft with your hand and wiggle it side to side and rotate it there should be no more than .001 " runout. To check coupling runout reassemble pump and check runout with dial indicator below coupling on pump shaft by rotating coupling by hand.

 

[quadtracker]

Just a thought! I also noticed in your picture the split line on your rigid coupling has a gap. Normally , in my experience, the coupling split line is closed tight when torqued with a .020" to .030" crush gap. But not all manufacturers are the same.

 

[georgeverghese]

Would imagine that seal rubbing surfaces may be running at least 20-25degF higher than 177degF on these hot side pumps; ie close to 200degF. So, with this PPEG/water mix concentration, do you have sufficient backpressure to keep this seal flush liquid away from boiling point ?

 

[quadtracker]

18 psig plus your system pressure is more than enough to keep the fluid from boiling. Think of your automobile it has a thermostat that probably keeps your engine coolant at 180 F to 195 F pending its rating. You know were the gage normally runs but on a very hot day the gage goes up, the thermostat is wide open allowing max flow through the rad. The rad cap is set at about 10 psi to relieve the pressure but it does not open because the coolant has not turned to steam and expanded above 10 psi. However if your fan belt brakes or fan quits the coolant temp rises which produces more than the 10 psi and boils out of the rad. Check the steam tables and it will show when water turns to steam at a given temperature at a given pressure.

 

[Hoxton]

Flaky white residue? Suggest you get it analysed. Also get an analysis of the coolant, check it is to engine manufacturers specification.

Is it an open (header tank) or closed (pressurised) system.

We solved a lot of similar problems by going to pre mixed coolant (uses de min water)

AND TOPPING UP WITH PRE MIXED COOLANT. NOT TAP WATER!mg

 

[LongArmJim]

Wow! Sorry for the lack of response on my end, I was away for a little and put work to the side!

@quadtracker- yes, these armstrong couplings will not touch, they are tightened to spec evenly and measured with a feeler gauge for the an equal gap. And i just had a situation with one that was not closing properly and it lead to a vibration and a noise. neither are present with this pump. Next time i go to the site i will have to try your suggestions. The vibration analysis I did on the motors does not indicate any bearing wear though. Thanks for your help! Ill have to bring my dial in next time. there is play in the pump shaft, when the pump is un coupled but it is nice and solid when everything is bolted together.

@georgeverghese this was my initial though

@Hoxton ive never had the seal analysed, how ever the coolant was analysed. its a small concentration of propolyne glycol according to my seal vendor it is nothing the seal shouldn't be able to handle. It is a closed loop pressurized system the coolant ran thru the generator engine.

Thanks!

 

[quadtracker]

If your runouts are good you may have to go into the pump and check throat/throttle bushing clearance around the pump shaft, it may be washed allowing too much pressure from pump into stuffing box chamber. And or the impeller may not be balanced any longer for 1800 rpm.