VFD operation on high inertia loads

[jakegordonII]

We are attempting to drive a solid bowl centrifuge (very high wk-2 with normal 7 to 10 minute wind up time when starting accross the ine). Motor is 100 HP. any experience and learnings. We are having trouble with the inverter faulting out due to braking issues.

 

[jOmega]

Mike,

Thank you for the addition insight into the world of centrifuges.

If OSHA codes are adhered to, personnel safety should be a non-issue... but we both know that's not the real world. And so, there will be a time when someone will get entangled with the machine. And 5 minutes to stop is much too long. When it comes to personnel safety, the machine is usually sacraficed as necessary to get the thing stopped.

Some methods I've encountered over the years, include an eddy current clutch-brake with the brake very much oversized to produce the necessary braking torque.

On DC drives, I've seen the regen capacity doubled over the motoring capacity (bridge sized for twice as much current.

And on a few occassions, I've used Plug reversing under current limit control (VFD). This requires adding an encoder or zero speed switch on the centrifuge to sense when zero speed is achieved. No line regen occurs, and dynamic braking not required. But the motor-- now that's a different story. Since the inertial energy is dissipated as heat in the motor, it must be sized larger (over framed) or force-ventillated to get rid of the heat.

I don't have my copy of NFMA-79 here with me but I'll look on Monday to see if there isn't a breaking requirement. Seems I've seen a spec for centrifuges somewhere along the way.

Enjoyed the dialog.

jOmega

 

[jOmega]

Make that NFPA-79

Sorry for typo

jW

 

[ischgl99]

I've enjoyed the dialog also, it's been informative. I would be interested in seeing if there is a braking requirement. I still work with centrifuges, but not for the manufacturer anymore and do get asked about braking on occasion.

I agree the machine should get sacrificed to save the person, but that is not easy to do with present designs. To stop the machine quickly, the brake would have to act directly on the bowl, otherwise the bowl can come loose due to the fast stop. The main issue is the manufacturers will not supply it until the customer demands it and is willing to pay for it. There are limits to what can reasonably be done, and you can't completely design out stupidity. I know of mechanics that pride themselves on being able to start dismantling a machine before it has stopped and also know of a plant where 2 men were killed doing just that. Going back to the train analogy, if you jump in front of it, no safety feature is going to save you.

I look forward to what you find in NFPA.

Mike Bensema

http://www.dutchmenservices.com

 

[jraef]

Interesting thread. Too bad the original poster is not contributing any more.

For what it's worth, the sugar industry stops centrifuges all the time because it is a batch process. 5 minutes stop braking time as opposed to 45 mintues stop time coasting adds up to increased productivity.

The standard for that process has been water injection, but now they are adding line regen drives to make it quicker. US Drives has done quite a few, probably as a result of targeted marketing to that industry. ABB, Baldor, Siemens, GE, Toshiba and Robicon are all manufacturers that I know of who offer line regen options. Schneider is probably one of the few big suppliers that does not.

I've also seen them used on Gold Concentrators, a type of centrifuge which is also a batch process operation. Quando Omni Flunkus Moritati

 

[jbartos]

Suggestion: Visit

http://www.ab.com/manuals/ms/193-2.16.pdf

etc. for more info

 

[ischgl99]

jraef, how long does the centrifuge normally run when processing a batch of sugar? Most of my experience is in the dairy and wastewater industries where the machines are only shut down when they need to be, many times weeks later.

With these regen drives, is the decel time fixed or does it vary based on the load? With a VFD, if it is progammed to stop with water injection and for some reason the water is shut off, it will decel too fast and possibly trip. Is this the same problem when using a line regeneration? Mike Bensema

http://www.dutchmenservices.com

 

[jraef]

mbensema
Batch time varies from process to process acording to the production stage (there are several). Generally speaking though, the highest batch rate I have seen is 10 - 14 batches per day on the first stage. Without braking they might get 9 or 10. With braking they get 14 or even more. Subsequent concentration stages run longer and so get fewer batches per day.

With regenerative braking the VFD rarely ever tripps off because any excess energy is fed back into the AC line. Another new technology being used is having several VFDs on a common DC bus so that the regen energy from one is being used to spin up the others. I have no direct experience with that, I just know it is being done.

The water not only speeds the braking, it is part of the process as well so the batch is ruined if it fails.

Quando Omni Flunkus Moritati

 

[jbartos]

Suggestion: Visit

http://www.vfds.com/teco-m.htm

for: Boost Fixed and Accel For high starting torque or to accelerate high inertia loads

http://www.csemag.com/index.asp?layout=story&webzine=cse&publication=cse&articleid=CA230289

http://www.hq.usace.army.mil/cemp/e/Et/mitergat.pdf

etc. for more info

 

[ischgl99]

Thanks for the info jraef, you learn something new everyday. 14 batches a day is a lot of machine cleaning, glad i'm not a millwright at that plant. Mike Bensema

http://www.dutchmenservices.com

 

[greye]

hi there about your centrifuge braking why dont you use a hydraulic drive very simple and no problem braking to come down to plough speed +-70rpm works like a dream i have done two latest one using load sensing with speed feed back
for spot on speed control at start ramp to charge speed ramp to dehydration speed and braking to discharge speed.
charge speed is 520rpm dehydration speed is 900rpm and discharge speed is 70rpm with latest drive i have used proportional control for the pump programme from the dcs.

greye hunter

 

[Skogsgurra]

My first assignment (ASEA) back in 1964 was a centrifuge for sugar refineries. We had just switched from WL sets to thyristors. A DC motor and a four quadrant thyristor rectifier was ideal at that time with close to no-loss regeneration. And I start to think that it still could be ideal. DC motors are still made, you know.

 

[DickDV]

It's not clear to me that the initial poster intended to brake his centrifuge at all. He mentions braking issues with the drive but, it just might not be due to braking.

One characteristic about Schneider/Square D drives is that they have a dead-slow processor. This results in very poor motor torque response especially when operating in sensorless vector mode.

The unfortunate result of this slowness is that, even on coast-to-stop applications, the motor torque continues PAST the desired speed setpoint and then the drive has to correct by braking the load back down to setpoint. A need-less DC Overvoltage Fault results.

If this is the type of problem that is presenting itself here, then the answer is a faster processor drive like an ABB ACS800 or similar units from others. These units will process motor torque in a more timely fashion and pull off as the speed setpoint is approached. No regen is required and no trip occurs even in the total absence of braking capacity.

Come to think of it, the Reliance GV3000 is another drive in the dead-slow catagory and displays the same troublesome behavior on high inertia loads.

 

[greye]

have you given any thought to using hydraulic drive.
closed or open loop system very simple systems and braking not a problem if you need further info let me know.

greye.