Cause of chiller water-cooled SCR and 300HP compressor motor failure

[PQ101]

The 600Ton Chiller#1 compressor motor failed. Discovered SCRs show sign of meltdown. Compressor motor was meggered and read Zero ohm on 480V 3-phase delta winding. Shortley after the back-up chiller#2 experienced same type failure. During both events chiller controls did not cause system shutdown. The voltage values on 480V main incoming panel are well within the nominal 480volts. Regarding chiller controls failed to shut the system down, I'm investigating that right now.

Q- Appreciate if somepne could explain the relation between damaged SCRs and damage to compressor motor windings.

 

[jraef]

SCRs implies you have a soft starter on that chiller. SCRs are in essence controllable diodes; they block in both directions until gated on, then conduct in only one direction. To make a complete phase control device you need two inverse-parallel SCRs, so a 3 phase soft starter has at least 6 SCRs in 3 inverse parallel pairs. If the SCRs experience stress, such as severe voltage spikes or excessive current for prolonged periods, the insulating (blocking) properties of the semiconducting materials inside of the SCRs cease and they become full time conductors.

If you have only one shorted SCR in a 3 phase starter there is no immediate danger because there is no complete circuit path. But as soon as you start gating the other SCRs on, there is a severe current imbalance in the motor and a high DC content in the power feeding it. This causes the motor to heat up disproportionately to the overall current being seen by the protective devices. So it is entirely possible for the motor windings to overheat without exceeding the FLA of the motor, therefore there is no tripping. MOST commercially available soft starters have a "shorted SCR detection" circuit built in for this purpose. It detects, on shutdown, that one (or more) of the SCRs is still conducting and sends a signal to the control board that disables any subsequent starting until this is fixed. There are a few chiller mfrs who insist that soft starters are easy to build and so make their own, maybe yours does not ave a shorted SCR detection circuit.

If TWO SCRs in opposing phases short, then the problem is more immediate because now you can have continuous power flow through one motor winding. If the motor is not spinning, this means there is no impedance in the circuit and only the winding resistance is limiting the current flow, meaning it is essentially a short circuit. So most of the time, the Short Circuit Protective Device will catch that (fuses blow or a circuit breaker trips). However, if the motor is still spinning when the 2nd SCR shorts, the motor may still continue turning slowly in a "single phasing" condition and because there is now some impedance, the current ends up too low to trip the SCPD again and it continues on unabated until the motor fries. This is especially dangerous in integral motors where you cannot see a shft spinning, such as in many chiller designs.


"If I had eight hours to chop down a tree, I'd spend six sharpening my axe." -- Abraham Lincoln
For the best use of Eng-Tips, please click here -> faq731-376

 

[PQ101]

Q- If SCRs fail-open, would it cause single-phasing on compressor motor? In this case some of the SCR heat sinks indicate sign of excessive overheat, and actual meltdown inside of the chiller enclosure.

 

[itsmoked]

Yes. Partial single phasing. Not good.

The motor would not want to start either, and would draw high current trying.

The system should really have some protections from all this. Any SCR issues should promptly cause a shutdown.
I have some doubts that this was caused by the soft starter.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[PQ101]

I noticed the chiller package is not provided or equipped with internal circuit breaker (on the skid enclosure) as shown on York manual. The 480V to the chiller is supplied thru main switchgear via 480V fused-disconnect, and there is no shunt trip or protection to trip the main. I'm not sure why the chiller supervisory system did not sense SCR overheat or to cause system shutdown-on any other parameters or predetermined setpoint limits to prevent the damages. The SCRs are water-cooled, have their own condenser, and any coolant restriction can be an issue. The issue is why the SCR went thru meltdown process, why control system did not cause chiller shutdown, and if motor winding failure was SCR related.

 

[PQ101]

Sorry, no fused disconnect. Only local service disconnect within the visual distance

 

[PQ101]

The service disconnect for the chiller indicates sign of overheated contact on B&C phase.

 

[jraef]

How did I now it was going to be a York?

Because when I said "There are a few chiller mfrs who insist that soft starters are easy to build and so make their own...", I was thinking of York. When I was in that business, York blew everyone off who was trying to sell them Soft Starters because they felt it was simple enough to make themselves. Of course, it's never as simple as people think it is and this is probably one of the many little details they left out, whereas people in the business have learned from previous mistakes and include them now.

So the answer to your question of "Why..." can only be answered by York, because as an industry, commercially available soft starters would have had several levels of protection against such an event.


"If I had eight hours to chop down a tree, I'd spend six sharpening my axe." -- Abraham Lincoln
For the best use of Eng-Tips, please click here -> faq731-376

 

[LionelHutz]

What do you mean no fused disconnect?

The level of damage sounds like a failure of the upstream protection devices to provide a decent level of protection.

If the motor failed first then the SCR's could have been trying to turn on into a short circuit. They won't survive that for very long.

Was the motor running or stopped during the failure? Large SCR's tend to fail shorted so if the motor was running at the time of the failure then it likely failed first and damaged the SCR's after. It the motor was stopped then it could have been something like a voltage transient causing one SCR to short which then did bad things when the motor attempted to start.

Basically, you don't provide nearly enough info to determine the cause.

 

[PQ101]

The question is: Will damage to SCR during chiller normal operation, cause damage to the compressor motor winding? yes or no? If yes, explain the possible damages or scenarios.

Chiller was operating fine at the time. When it stoped operating, operator noticed the chiller is inoperable-damaged. Chiller #2 picked up, sometimes later there was nother problem, same nature. Both chillers were damaged. The issue if shutdown is whole different story. Protection logic should have caused chiller shutdown if the conrols were designed with good degree of reliability. As for 480V service, thes are no internal circuiot breaker on the skid (as part of solid starter enclosure or anywhere near York operating consol on the skid. from ther it goes to 480V disconnect switch gainst the back wall, and from ther to outdoor tyoe switchgear breaker, and to utility 2000KVA 12.47kV-480V padmount transformer.

 

[itsmoked]

Yes, if an SCR opens the motor will be single phased. It will draw excessive current and slag down if not detected and responded to. Guaranteed!

Any controller that is running a motor over about 5HP should have phase loss protection that shuts down the motor preventing its being damaged.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[edison123]

Get some single phasing protection for the motor pronto. You're going to need a CB too in there somewhere.

Muthu

http://www.edison.co.in

 

[jraef]

from ther it goes to 480V disconnect switch gainst the back wall...

If that is a fused disconnect, or if the breaker in the outdoor switchgear is sized correctly, the requirements are probably met. There is no requiremet that there be another disconnect mounted to the machine, as long as the primary one is not feeding anything else and is "within sight" or otherwise provided with a locking mechanism.

Assuming it is fused or protected by that breaker, the issue you asked about comes back to what happens with a bad SCR. First off, the likelihood of an SCR failing "open" involves the actual meltdown or vaporization of the SCR itself; when they fail, they usually become full time conductors, not insulators, so they essentially "fail closed". When this happens to multiple SCRs in opposite phases of one starter, the motor can single phase. The circuit may or may not draw enough current to open a fuse (or circuit breaker for that matter). That's why you cannot rely upon the SCPD to protect against single phasing.

But the curious part of this story to me is that your second soft starter failed as well in the same manner; the odds are astronomical that this would happen in succession out of the blue. That points to the common power source as the initial cause of the problem and the SCR damage as being secondary to the original issue. I would be looking there rather than chasing phantoms.


"If I had eight hours to chop down a tree, I'd spend six sharpening my axe." -- Abraham Lincoln
For the best use of Eng-Tips, please click here -> faq731-376

 

[PQ101]

Chiller #2 experienced damage 2 weeks after unit #1 failure. Inspection discovered common issue: sign of SCR heat-sink meltdown on both units. Monitoring data indicates RMS voltage on 480V to chiller panel was decreasing periodically to 450volts and as low as 445volts. Motor nameplate rated at 460volts +/-10%. Right now not concerned about protection mode of the chiller, circuit breaker, fused disconnect, etc.

Q- Will voltage fluctuation on incoming service cause SCRs to overheat and fail, and in theory be considered as possible factor to cause damage to compressor motor winding?

 

[itsmoked]

If you are actually seeing heatsink meltdown then I'd say your problem is with whatever system is supposed to be keeping the heatsink at the design temperature. If there is water cooling I would be looking at that. If there is forced air.. Test that.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[PQ101]

Thanks everyone. I'll take it from here.