Adding oil pumps to a ONAF transformer 1

[jkristinn]

The proposal of adding oil pumps to a bank of 1 phase ONAF GSU transformers (100 MVA bank, 220/13,8 kV, conservator type) is under discussion. The reason is not to increase the power capacity of the transformers (as the capacity of the generators connected to the bank will not change), only to lower the hot-spot temperature with the goal of extending the transformer life and decreasing the failure probability. The transformers are new.

The OEM of the transformer has been contacted and the price for the job is not a deal breaker. The OEM has proposed in-line pumps with axial impeller and states that even a failure in the pumps will not interrupt the normal oil flow by convection, so the ONAF rating is not in jeopardy. Our concern is the probability of a pump failure blocking the cooling system flow from the respective cooler (each 1 phase unit has 4 coolers), e.g. in the case of a bearing failure, can anyone comment on the possibility?
Any other concerns we should have?

 

[ScottyUK]

Axial pumps generally have a open impeller design and would allow oil to circulate past the impeller if the pump was stationary. You should really ask this in forum407 where the experts reside, but in my opinion it would take something close to catastrophic failure of the pump to actually block the cooler pipe. Even if that occured the three remaining banks with operational pumps would probably equal the capability of four naturally-circulated radiators.

Make sure you include decent quality isolation valves. One day you'll need them and that won't be the day you want to find out that they aren't up to the job. Have the pipe shop make you a spool piece to replace the pump and get the cooler back in operation under natural circulation if you don't carry a spare pump on the shelf. ;-)

 

[rcw retired EE]

Modern pump designs have improved reliability. We had a 1950's vintage transformer fail in spectacular fashion when problems with the pump bearings allowed the bronze impeller to shave off metal pieces that became embedded in the windings. Pumps require more maintenance checks than convection coolers. (We should have caught the bearing problem).

Scotty has some very good suggestions.

 

[jkristinn]

Thank you both for your replies.

Scotty: The isolation valves seem to be solid. The idea of an spool piece is a very good one, thank you!

Rcwilson: You have just added to my worries: metal debris from a failed pump, the impellers and motor body are cast aluminum (Al Si10 MG DIN 1725). I don't know which material is used for the ball bearings. I would be happy to learn more about what maintenance you suggest (although I'm sure that the OEM will have a suggestion on the topic).

 

[rcw retired EE]

Go with the recommended maintenance. A manufacturer with experience selling and guaranteeing transformer cooling pumps would know better than I what to look for after a few years operation.

We never pulled that pump out of the transformer cooling pipe, because it required draining the whole transformer since the isolation valves did not work as Scotty warned. It was a 50 MVA, 13.8 kV voltage regulating transformer feeding an aluminum potline operating at 95-100% load 24/7. We did tap changer maintenance about every 12-18 months based on # of operations, switching the potline to the standby regulator during maintenance.

Root cause was assigned to the pump failure. But a competing theory was that my crew had improperly topped off the main tank using a "dirty oil" hose that was only for draining the nasty oil from the tap changer tank. That may have dumped some metal shavings from the worn out tap changer mechanism into the main tank. There wasn't much left of the transformer. It is possible that when the lid was sent high in the air the explosion also ruined the pump, making it look like the source of the metal particles. It was fortunate that the spare regulator could be switched into service quickly and another plant had a surplus unit that worked as a replacement. Since production got back on line and replacement cost wasn't large, the root cause investigation was minimal. (We would not get away with that in a modern well run facility).

Other 50 MVA regulating units had more modern pumps that ran quietly 2-3 years continuously. The vacuum tap changers on those regulators allowed much longer maintenance intervals. We checked amp draw and vibration readings during operation, like any other pump. I don't recall any pump issues.

 

[prc]

Adding oil pumps to an ONAF transformer is not a good solution. It will not reduce hot spot, but only reduce the top oil temperature. To get full effect of oil pump the cooling should be directed type ie ODAF.It is true that open impellar type oil pumps will not give much obstruction to oil flow during ON conditions.Need not worry about bearing dust if it is a good quality pump.We need to worry about metal dust getting in to winding more in oil directed cooling where the output from the pump will be going direct to inside of winding.

A better soution is to add more fans, or fans with higher air flow or change the direction( from horizontal mounting to bottom of radiators) OEM can help you in this.

 

[Jebb]

I know Doble offers a "Particulate Metals in Oil" test to detect metal in the oil from pump related issues. I've never used it so I don't know how accurate it is. Our company is not a big fan of forced oil cooling because it just adds another failure level. Scotty's ideas about additional valves and a spool are good ones. Very few think about future maintenance, including the OEM's......

 

[waross]

Check valves. You can run without them as long as all pumps are running. If you are sure that a pump failure will be promptly recognized and repaired you may not need them. Evaluate the back flow with a pump stopped or the spool piece in, vs the flow from just three pumps with the fourth path closed off. You may get better flow if you valve off a failed circuit and avoid the backflow. Or, check valves will automatically close off back flow and leave you with three active paths.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[bacon4life]

I am not sure if this unit it large enough to worry about static electrification, but you may want to verify that the OEM has checked that the oil flow velocities will be low enough to avoid static electrification. Turning the pumps off at low oil temperatures also reduces the risk.

http://www.epri.com/abstracts/Pages/ProductAbstract.aspx?ProductId=TR-113741

 

[prc]

In this unit one need not worry about static electrification at all. Static electrification occurs when oil flows over pressboard inbsulation at velocities above 2m/sec. In this case since oil flow is not directed one,(oil is not made to flow over the winding or insulation surfaces) chances of static electrification are nil. Here the oil from the bottom of cooler will simply flow in the gap between tank and winding ie purpose of pump is to increase thye oil flow in the cooler.

The major issue is converting ONAN/ONAF unit in to OFAF has limited effect in reducing the hot spot in winding as the oil inside the winding will move just as in case of ONAN case,but with ODAF oil will flow over the winding surfaces removing heat quickly and reducing hot spots.