480V Chiller drops out when other chiller is started 1

[killowatt]

I have three pairs of 480V chillers. Each chiller is started using VFD. When the second chiller in that pair is started, the running chiller drops out. The voltage drop is not excessive and the transformer supplying these pair of two chillers are large and adequate capacity. I have checked the inrush currents of chillers. What else could cause these chiller to drop out? Please help. Thanks in advance.

 

[rbulsara]

Does VFD gives any display of reason/flag as to why the chiller is dropping out? like UV?

Usually modern chillers have lots protection built-in, it should give some alarm.

 

[buzzp]

If there is no indication or alarms and it simply stops running, you may want to look at your control power for your contactor (if one is used).

Is the VFD switched out of the circuit after the motor is up to speed to across the line?

 

[busbar]

 
Seems like some hermetic-compressor chillers have notoriously proprietary motor-protection/-control modules with ill-defined setpoints, nebulous features and nonexistent test procedures. Some repair instructions by their producers are of ridiculous, expensive, Detroit-mentality “replace the module and see what happens then” purview.

 

[alehman]

Do these chillers use "soft starters"? These are basically a VFD used only for starting, and bypassed once the motor is up to speed. If so, I wonder if the running chiller is being confused by waveform notching caused by the starting chiller. Drives packaed on such equipment sometimes are fairly inexpensive and may be lacking filters.

I've found some modern chillers to be overly sensitive to minor power disturbances. In some facilities the chillers shut down on disturbances that don't bother any other building components on unprotected power including computer systems and building automation. They usually have a substantial restart delay as well.

 

[rbulsara]

just a note for some posters:

Now-a-days VFDs are common on chillers and they are used for varying the speed of the chillers (centrifugal compressors)and hence control the cooling load also. Of course they also start as soft starters, but are not taken out of service after start.

 

[buzzp]

Rbulsara,
Seems like the poster has one chiller on a drive to match the load while the other is running across the line or off. He has 3 pairs of chillers with two paired up with one another. He did say they are started with a VFD but was not clear if the matching pairs are on one or two VFDs, seems like one vfd for both chillers, hence 3 drives with 6 chillers.

 

[rbulsara]

buzzp:

You have a good question. I was not really trying answer his question as there is not enuff info to go by. However re-reading the original question, I suspect the 'pair' means the two chillers fed from a common tranformer. He may have 6 VFD's but 3 trasformers.

We shall wait for kilowatt's feedback.

kilowatt:
Pleae post more info as to size of transformers, chillers HP etc..starting currents, any recorded voltage drop etc..

 

[jraef]

Actaully, it may be a totally valid control scheme. Notwithstanding the need for additional info to be sure, this looks to be like the kind of system I have employed on pumping and refrigeration systems.

The VFD is used to load-match the system using motor #1 only. Then when it gets to 100% speed, the inherent full-speed running losses in the VFD are avoided by switching on the motor #2 at full speed, usually Across-the-Line or with a soft starter of some sort. Once it is at speed, motor #1 is switched off ASAP to avoid a load surge (I usually tailor a decel ramp on motor #1 to match the accel of motor #2). Then if the load demand continues to increase, motor #1 is added back in with the VFD to match the increasing load demand again.

The reason this is done with separate controllers instead of just bypassing the VFD with a contactor is because the nature of the load might make it problematic (or impossible with that VFD mfgr) to perform a "synchronous transfer" from the VFD to contactor bypass. If the motor slows even a fraction, you will generate a huge transfer spike and may damage something. I know of only one VFD mfgr that has proven they can do a synchronous transfer successfully, although several have said it can be done in theory. By bringing on motor #2 (which needed to be there anyway) X-Line and then turning off motor #1, the potential spike problems are avoided without a complicated and/or expensive addition to the VFD.

His system may not have experienced sifficient load demand above the capacity of one motor for him to have witnessed the 1st motor coming back on through the VFD.

"Venditori de oleum-vipera non vigere excordis populi"

 

[buzzp]

I agree there is simply not enough info to steer him in the right direction as to what the problem may be. Jraefs second paragraph summarized what I was trying to say.