Jump to content


Photo

Torque Follower Mode


  • Please log in to reply
10 replies to this topic

#1 jraef

jraef

    Posting Freak

  • Moderator
  • PipPipPipPipPipPip
  • 683 posts
  • Gender:Male
  • Location:USA, California

Posted 22 September 2006 - 02:02 AM

I had a question thrown at me today and it left me wondering. Here's the scenario:

2x 300HP 460V motors driving a VSI Crusher (Vertical Shaft Impactor), both motors connected to a common shaft. I have done several of these using one VFD feeding the 2 motors and that has worked fine up until now. A new wrinkle has come up with one of my customers today. He has reported that the sheaves (pulleys) are wearing at different rates on the crusher and the motors, causing the belts of one to slip more that the other, which of course accelerates the wear on the belts as well, until eventually one of the motors is pulling all of the load and the other is idling. When they replace the sheaves and belts, all is well again, but this process is frustrating to them to say the least. A VFD manufacturer has told them that the solution is to use 2 separate VFDs in Open Loop Vector Torque Mode, set to deliver the same amount of torque from each motor so that when the sheaves wear, the motors will continue to share the load equally even though one may be spinning faster than the other.

I actually have done something similar to this in the past (and even proposed it to this customer when he first bought the system, but he declined it because of the cost). The tactic I used was to have one VFD in speed mode, then send the output torque to the second VFD, which was programmed as a torque follower. The caveat to that was that I used the old PDL drives which had a high speed fiber optic peer-to-peer link between the drives, so the torque signal for VFD-2 to follow was available dynamically and with very little error. I felt at the time that trying to use a 4-20ma output from VFD-1 would introduce an error percentage in the output signal, as well as at the input A/D converter and the risk of picking up noise en-route. That, combined with the fact that the torque output accuracy was +- 5% was going to allow for a significant delta in the motor torque outputs, leading to the master pulling the slave.

Now they are thinking of buying into this vendor's concept and I am less than convinced. My feeling is that there is no way that just programming the torque to be the same is going to work, in fact this vendor's product only specs the torque accuracy as +-10%! That means that at any given moment there could be a 20% delta in torque between the 2 motors, hardly "matching" if you ask me.

I want to make a counter proposal to this customer even though I no longer sell VFDs, I just don't want to see him spend a lot of money on junk. My problem is, the PDL drives are no longer available and I can't readily see if other VFD manufacturers offer peer-to-peer communications. Am I overly concerned about error?
"He's not dead, he's just pinin' for the fjords!"

#2 Gigger

Gigger

    Intermediate Member

  • Full Member
  • PipPipPip
  • 13 posts
  • Gender:Male
  • Location:New Zealand

Posted 22 September 2006 - 04:11 AM

QUOTE(jraef @ Sep 22 2006, 02:02 PM) View Post

I had a question thrown at me today and it left me wondering. Here's the scenario:

2x 300HP 460V motors driving a VSI Crusher (Vertical Shaft Impactor), both motors connected to a common shaft. I have done several of these using one VFD feeding the 2 motors and that has worked fine up until now. A new wrinkle has come up with one of my customers today. He has reported that the sheaves (pulleys) are wearing at different rates on the crusher and the motors, causing the belts of one to slip more that the other, which of course accelerates the wear on the belts as well, until eventually one of the motors is pulling all of the load and the other is idling. When they replace the sheaves and belts, all is well again, but this process is frustrating to them to say the least. A VFD manufacturer has told them that the solution is to use 2 separate VFDs in Open Loop Vector Torque Mode, set to deliver the same amount of torque from each motor so that when the sheaves wear, the motors will continue to share the load equally even though one may be spinning faster than the other.

I actually have done something similar to this in the past (and even proposed it to this customer when he first bought the system, but he declined it because of the cost). The tactic I used was to have one VFD in speed mode, then send the output torque to the second VFD, which was programmed as a torque follower. The caveat to that was that I used the old PDL drives which had a high speed fiber optic peer-to-peer link between the drives, so the torque signal for VFD-2 to follow was available dynamically and with very little error. I felt at the time that trying to use a 4-20ma output from VFD-1 would introduce an error percentage in the output signal, as well as at the input A/D converter and the risk of picking up noise en-route. That, combined with the fact that the torque output accuracy was +- 5% was going to allow for a significant delta in the motor torque outputs, leading to the master pulling the slave.

Now they are thinking of buying into this vendor's concept and I am less than convinced. My feeling is that there is no way that just programming the torque to be the same is going to work, in fact this vendor's product only specs the torque accuracy as +-10%! That means that at any given moment there could be a 20% delta in torque between the 2 motors, hardly "matching" if you ask me.

I want to make a counter proposal to this customer even though I no longer sell VFDs, I just don't want to see him spend a lot of money on junk. My problem is, the PDL drives are no longer available and I can't readily see if other VFD manufacturers offer peer-to-peer communications. Am I overly concerned about error?



#3 Gigger

Gigger

    Intermediate Member

  • Full Member
  • PipPipPip
  • 13 posts
  • Gender:Male
  • Location:New Zealand

Posted 22 September 2006 - 04:29 AM

Hello Jraef,

I am very interested in your problem,as I am also involved with crushers, having two motors on a common shaft,namely the Barmac crusher which is made in New Zealand by metso Minerals.My company supplies the soft starters & VSDs for there applications.

I have found when I have experenced this problem, it has always been caused in the tensioning of the two sets of belts ie: each motor set. If one set of belts is tighter than the other this will cause that motor to work harder than the other and hence the belt wear & pulley wear on the slack set of belts will wear far faster than the other set.
Also this can be caused if one pulley has a different type of V groove to the other pulley, and the two sets of belts are different type belts,this will also cause one motor to work harder and increase current and wear.

I think the problem can be over-come by adjustment, rather than putting a second VSD on the applcation.

Can you let me know the out come when you get a result.

Regards,

Gigger

#4 mariomaggi

mariomaggi

    Senior Member

  • Full Member
  • PipPipPipPipPip
  • 230 posts
  • Gender:Male
  • Location:Milan, Italy
  • Interests:electric vehicles http://www.evlist.it , 3-phase AC sources to test PV inverters, high-power AC/DC power supplies up to 1500 VDC to test photovoltaic inverters, bidirectional AC/DC power supplies to test and simulate big batteries, infrared windows for thermography, renewable energies, special electric motors, special inverters, energy savings, power quality

Posted 22 September 2006 - 06:21 AM

Dear jraef,
I agree with Gigger:
QUOTE
I think the problem can be over-come by adjustment, rather than putting a second VSD on the application.


If both motors and transmissions are similar, you could have other possibilities to check belts, i.e. using a microphone to test noise, a vibration sensor, or a laser to check alignements, or an handy tension checker to measure the force on each belt.

When conditions of belts are not perfect, you could have other influences, i.e.: belts exposed to sun rays could be hotter than usual, then results an additional friction.

Regards
Mario


Mario Maggi - Italy - http://www.evlist.ithttps://www.axu.it


#5 joze

joze

    Junior Member

  • Full Member
  • PipPip
  • 8 posts

Posted 24 September 2006 - 10:36 AM

jraef:

Why use analog signal when most drives have some sort of industrial comm integrated (or at least optional)? Torque follower (VFD-2) could easily communicate with the first drive (VFD-1) and you should not expect any errors in communication.

If possible, tell us more about torque following application. Thanks.

Best regards,

Joze



#6 jraef

jraef

    Posting Freak

  • Moderator
  • PipPipPipPipPipPip
  • 683 posts
  • Gender:Male
  • Location:USA, California

Posted 24 September 2006 - 10:03 PM

QUOTE(joze @ Sep 24 2006, 03:36 AM) View Post

jraef:

Why use analog signal when most drives have some sort of industrial comm integrated (or at least optional)? Torque follower (VFD-2) could easily communicate with the first drive (VFD-1) and you should not expect any errors in communication.

If possible, tell us more about torque following application. Thanks.

Best regards,

Joze



Joze,
That is what I am looking for. Although most drives have communications capabilities built-in, they are for the most part passive devices or nodes on a network that must have some other active comm device, such as a PLC or PC, handling the communications. Very few have what is called peer-to-peer communications capabilities where one drive can talk directly to another.

I have found out that the new Siemens Simatics drives are in fact capable of that, I am still looking for other recommendations.

As to the mechanical solutions mentioned by others, that would be ideal of course, but we all know that real world installations (and the people working in them) are less than ideal. That is why this customer is looking for a different solution.
"He's not dead, he's just pinin' for the fjords!"

#7 marke

marke

    Posting Freak

  • Moderator
  • PipPipPipPipPipPip
  • 2,604 posts
  • Gender:Male
  • Location:Christchurch, New Zealand

Posted 26 September 2006 - 06:21 AM

Checkout the Control Techniques SP series drives.

Best regards,

#8 niallnz

niallnz

    Intermediate Member

  • Full Member
  • PipPipPip
  • 27 posts
  • Location:Christchurch
  • Interests:Electrical Engineer, working with industrial controls, motors and drives etc

Posted 26 September 2006 - 10:55 PM

Hi jraef

Since you asked for info on which drive manufacturers can do what you're asking, I'll jump on the band waggon tell you that ABB does master/follower drives that use high speed fibre optic communications. I've set up a pair on a crusher with great success.

But as Mario and Gigger say V-belt addjustment is very important, I've been down the road of thinking that a DC drive on an extruder was faulty, when in reallity the v-belts were not the new matched set of 8 that the customer said they were. I think I wasted about a day and a half on that one.

Cheers Niallnz

#9 jraef

jraef

    Posting Freak

  • Moderator
  • PipPipPipPipPipPip
  • 683 posts
  • Gender:Male
  • Location:USA, California

Posted 27 September 2006 - 02:16 AM

Thanks.
I looked in some ABB literature on the ACS-550 and didn't see anything listed, so is it in the 600 or 800 only?

As far as the v-belt situation is concerned, on this machine the sheaves are oriented vertically with respect to one another.

IPB Image

The way it was described to me is that the upper sheave runs hotter than the lower one because the heat from the lower travels up the shaft and adds to the upper one. This makes the upper belts run hotter also, which hastens their wear. There is also a lot of dust in the air as the machine operates and it adheres more to the hotter belts, further accelerating the wear on them and the sheave itself. So even if they have them adjusted perfectly, they don't stay that way for long, and the personnel at the site are not qualified to adjust belt tension on a regular basis, so it only happens when the mechanics make their rounds.

Here is a typical application I have already done with a single PDL drive (shown out in front under the kiosk) running 2 motors, just to give you an idea of how nasty the environment is.

IPB Image
"He's not dead, he's just pinin' for the fjords!"

#10 niallnz

niallnz

    Intermediate Member

  • Full Member
  • PipPipPip
  • 27 posts
  • Location:Christchurch
  • Interests:Electrical Engineer, working with industrial controls, motors and drives etc

Posted 27 September 2006 - 04:22 AM

Hello Jraef

You are correct the master/follower that I mentioned is only available on ACS600/800 drives. Whilst the ACS550 is a good drive, if it was given too much capability no one would by the ACS800.

Cheers Niallnz


#11 joze

joze

    Junior Member

  • Full Member
  • PipPip
  • 8 posts

Posted 27 September 2006 - 11:18 AM

Hi,

if we are going "vendor specific" here is my suggestion:

Telemecanique (brand of Schneider Electric) ATV71. Built in CANopen and application functions which you can use. Also, if you already know PDL you probably know that Schneider & PDL go way back together.

Best regards,

Joze




1 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users


    Google (1)