motor stopping rapidly during a fault 16

[electricpete]

We had a 500hp hermetically-sealed centrifugal chiller motor trip on instantaneous current 15 seconds after starting.

Normally I would think instantaneous trip is due to electrical fault, not mechanical origin, however...

Personnel present reported that the machine stopped very abruptly <1 seconds after the trip, whereas it normally takes 10-15 seconds to coast down during normal shutdown.

It occurred to me maybe there is a higher different load torque imposed by the machine at the time of trip during shutdown. When I talked to the chiller mechanical engineer, he said the torque should be roughly the same.

Is it possible that certain fault currents can act similar to dynamic braking to stop motor quickly?

(I kind of doubt it... suspect other simpler explanation... just wanted to ask about this possibility)

 

[electricpete]

I just realized a fatal flaw in my explanation. This is
4kv system hi-R grounded thru 240-ohm giving max ground fault current of 10A.

For all three phases to trip we must have had phase to phase. I will have to think about where that may have occurred.

And if ground fault was the first fault and later escalated to phase-to-phase... why didn't the ground fault trip at all during the entire event? Ground fault is identified as 51G so I think there is some time delay.

I will review the settings and provide some more info if I can.

 

[electricpete]

last post one correction:
For all three phases to trip on instantaneous we must have had phase to phase

 

[RAMConsult]

Pete, Thanks for the update! You're right, this is one of those curious events that occurs from time to time. I have a feeling that your 'contaminated refrigerant' scenario is a good starting point:

1. When was the last time the refrigerant quality was checked?
2. Was this the first start after any type of maintenance on the refrigerant circuit?
3. Is it possible to determine the direction of refrigerant flow; i.e., is it motor first then compressor or vice versa?

One possible scenario is the slug of water knocked off two birds with one stone- it caused an end-turn to end-turn flashover then traveled to the compressor where its incompressability caused the rotor to stall. The stalled rotor would have forced a kickback through the shaft causing the motor rotor to instantly reverse and unscrew from the compressor shaft. The problem I have with this scenario is I would expect to find plenty of internal damage in both the motor and compressor, yet you say that both shafts turned freely. Suggestion- have a sample of the green corrosion analyzed, it may yield a clue as to the source of the contamination, also watch as the windings are removed to see if there was a long term build up moisture in the slots. Please keep us advised.

 

[electricpete]

Thanks for the comments.

green corrosion I am fairly common is copper oxide from copper and occurred after the fault. This seems well supported by the fact that the green corrossion is localized only at the fault.

We also have red rust corrosion accross the entire stator core, rotor core, and stator frame which I assume is something like iron oxide.

 

[electricpete]

RAMc - I agree that analysing the residue may provide some clue which may challenge my assumptions. I don't really have the resources to go after it that deep.

 

[jbartos]

Suggestion: The insulation thermal runaways can become very fast, i.e. faster than a ground relay can trip. Therefore, the line-to-line protector will trip first.

 

[busbar]

Green copper-oxide formation may have been a result of the original fault. Often refrigerants will form hydrogen fluoride and hydrogen chloride in arc presence. Six weeks of air exposure may have introduced moisture to form hydrofluoric and hydrochloric acids, which may have reacted with winding enamel.

 

[windie]

Wind turbines often have a clutch or &quot;mechinical fuse&quot; between the induction generator (-ve slip motor) and gearbox. I have been told that this is to protect the gearbox in the event of an electrical fault that leads to the rapid deceleration of the generator.

It seems to me that this can occur (and pete's experience backs this up), but I am still not sure of the exact mechanism.

 

[vcoul]

Electricpete,
I take it that the chiller being discussed is a low pressure R11 machine. Has a sample of the refrigerant been taken and sent out for analysis? Moisture content, total acid number and high boiling residue are 3 things that may indicate ongoing issues such as low side leaks (vacuum drawing in moisture laden air) or excessive oil in refrigerant.
Total acid number will indicate a long term moisture problem.
High boiling residue...has someone been adding oil to this machine due to &quot;oil loss&quot;? Oil is noncompressible and seems more likely than a slug of water.

 

[electricpete]

vc - I am told that samples prior to the fault showed no anomalies. I'm pretty sure we check moisture and TAN. I have not heard of boiling residue (can you explain a little more).

No sample was drawn after the fault. By the time we inspected and started thinking about moisture scenario refrigerant was gone and opportunity was lost.

 

[vcoul]

electricpete,
high boiling residue is the percentage of (usually oil) liquid left from a weight of refrigerant analyzed by a lab set up to do this. If you evaporate all the refrigerant off during a test like this you would have 0% HBR. Usually depending on how well loaded the machine is you might find 3-5% HBR. A heavily oil foulded machine will have poor boiling action, poor heat transfer and if severe enough will carry over waves of oil rich refrigerant. Look in the sight glass if you have one. Badly fouled refrigerant looks like dark beer.

 

[jbartos]

Suggestion: Will the motor be repaired or scrapped? If the motor is being repaired, there will be some observations available, once the motor is disassembled.

 

[YorkMech1]

I normally just read and absorb as much wisdom from the learned ones that post here as I can.
In this case, since it involves a piece of equipment I am more than fairly familar with, I feel compelled to ask the following as well as interject some thoughts.

1- It seems to me that someone failed to mention that this unit evidently has one of the old forced commutation VSD's on it called the "Turbo-Modulator".

2- When applied to a "Mono-Shell" or "Screwed-Compressor" unit, an additional device was installed @ the output called a "Vacuum-Switch".
During a catastrophic Fault, the output SCR's were "ALL" fired Which would result in almost immediate stopping of the motor thru some serious braking. To "Prevent" the compressor shaft(s) from unscrewing at any 1 the 4 screwed components which "Can & Will" result in extreme damage to the compressor, a Vacuum Switch was installed to disconnect the motor from the Drive in <1 milli-sec. or 1/10 of a Hz.
It is apparrent that the 1000volts that is stored in two capacitors and triggered to dump the vacuum switch was never applied, for whatever reason.
However....if this unit does not have the "Turbo-Modulator" installed as I stated, then my entire post is a mute point.

 

[electricpete]

Thanks for the comments. This is a York motor/compressor unit. Siemens is OEM for the motor.

Motor is turned on and off by cycling a breaker which applies 4kv direct to the motor. No turbo-modulator, no vacuum switch in the circuit.

I am not much familiar with compressors and certainly not familiar with the features you describe. For our unit, normal motor torque tightens the screwed coupling. Sudden motor stopping from dynamic braking or other anything else would tend to unscrew the unit to my way of thinking.




=====================================
Eng-tips forums: The best place on the web for engineering discussions.