Induction motor speed running away

Join Date
Aug 2020
Location
Washington
Posts
7
I have an auger powered by an induction motor under flux vector torque control driven by a PowerFlex755. According to my log, the command velocity was ramping up at a reasonable rate while the actual velocity was lagging behind. As soon as the command velocity began ramping down again, the actual motor velocity ran away up to the drive speed limit and shut down. I have included a a screencap of the log showing the velocities. https://imgur.com/a/MWtlVRm

How/why would this happen? PID instability? I'm at a loss. Thanks
 
Last edited:
Hi James. Thanks for the response. The speed reference is sent via Ethernet from a ControlLogix processor and is geared to a pump speed. I am ok with the ramp times not quite matching up, but my real concern is that the command and actual accelerations move in opposite directions as soon as the drive is commanded to slow to a stop. If you take a look at the linked plot, the feedback speed (red trace) is increasing while the speed reference (blue trace) is decreasing.
 
Is the drive in velocity mode or torque mode? If you are in torque mode and the resistance goes away, your motor will behave exactly as you described.
 
Is it possible the load is over-driving the motor? It could be you need some braking torque, and without a DB resistor, there's nowhere for the energy to go.

The drive would go into bus regulation mode, and try to scrub the excess voltage by speeding the motor up.
 
Is the drive in velocity mode or torque mode? If you are in torque mode and the resistance goes away, your motor will behave exactly as you described.

The drive is in torque mode. It suppose it is possible, but unlikely, that the resistance decreases. The motor turns an auger that pushes a powder into a tank.

The odd part is that the motor feedback acceleration increases at the exact moment that the command velocity begins to decrease (when a pump switches off).
 
If the VFD is in torque mode, then you are providing a torque reference, not a velocity command, right?

It has been a few years (like 12) since I worked with VFDs in torque mode and runaway velocity was something that could be complex to prevent in some applications.

Maybe modern drives have better controls, but back in the olden days, we had to switch the VFD to velocity mode in order to truly control speed and to do that we had to briefly disable the drive, then re enable it.
 
A few thoughts on this
Fist you never said what the application is.
From your description and your screen cap I would guess that it a web tensioning application.
Normally with a torque control application you set the max speed of the vfd and control the allowed torque (Torque Command or torque limit). The torque limit is usually received in the vfd as the channel 0 or the main analog input
In your case you have the max speed set as the default max 60 hertz
And it looks like the vfd receives the torque limit through the Ethernet command.
In the start up with very low torque command it slowly ramps up to the desired torque and the motor speed will go up to meet the web speed. The motor speed will be whatever it needs to maintain the web speed. Keep I mind if your web speed drops down, increases or even stops the motor speed will also follow the web speed but torque will remain constant as long as it in torque limit.
You say when you stop the vfd the speed goes up and the torque goes down. That would be explained if the web broke or ran out and then you try to ramp down the torque on the motor. While the load (tension on the web ) is now at zero but the torque command is at some level ramping down, the vfd will try to output the torque you ask for but it has nothing to work with so the output hertz / motor speed will go up trying to give you what you ask for. Your actual speed command is the max hertz you set in the vfd
I generally like to look at the line speed usually from the main drive and set the max speed / hertz to just above the level that way when the tension on the web drops the speed will only increase a small amount until you can get control or shut it down.
In your case I would recommend that you set both the accel and dccel ramp time in the vfd to 0. The ramp time in the vfd is for speed command only it has no effect on the torque command. In this case it is hurting you.
You are controlling the ramp of the torque limit with the PID control so the ramp in built in to the PID control.
keep in mind that a torque command in a vfd is not a true command, it is a torque limit, if the torque limit is set high but the actual load / tension on the web is low the vfd will increase the output hertz until the torque limit is reached.
You need to quickly shut down the vfd or set the torque limit to 0 when the web breaks or when the load drops quickly.
To help explain this try to run the vfd without the wed / load the motor speed / vfd output hertz will increase to the max limit on the vfd. You will not have any speed control or torque control
A good vfd should be able to maintain 100% torque at 0 speed and have full control over the motor. You don’t need a Flux Vector vfd to do torque control they are designed for accurate speed control or even positioning.
I have used them to replace servo drives without any problems
 
Is it possible the load is over-driving the motor? It could be you need some braking torque, and without a DB resistor, there's nowhere for the energy to go.

I don't expect it to, but I will look into it.


If the VFD is in torque mode, then you are providing a torque reference, not a velocity command, right?

All of the logic is built in velocity and it sends a jog (MAJ) command with a velocity reference, but it must be getting converted to a torque somewhere. I can monitor the commanded torque on the drive and see that it fluctuates wildly, but the velocity feedback usually tracks pretty close to command velocity in the logic. To be fair, I don't quite understand torque mode or why it is being used in this application.


A few thoughts on this
Fist you never said what the application is.
From your description and your screen cap I would guess that it a web tensioning application.

The motor turns an auger that pushes a powder through a track into a tank. The VFD speed is geared to the speed of a pump in the tank. The VFD is commanded to stop when the tank is full.
 
if that's the case why are you trying to control torque you should be controlling speed
as simple speed follower setup feed speed follows the pump speed. i don't see any advantage to torque control here
 
I have an auger powered by an induction motor under flux vector torque control driven by a PowerFlex755. According to my log, the command velocity was ramping up at a reasonable rate while the actual velocity was lagging behind. As soon as the command velocity began ramping down again, the actual motor velocity ran away up to the drive speed limit and shut down. I have included a a screencap of the log showing the velocities. https://imgur.com/a/MWtlVRm

How/why would this happen? PID instability? I'm at a loss. Thanks

Looking at your screen capture, I would say that the load torque drops much faster than your speed reference is dropping. But the motor continues to speed up as it crosses your speed reference.

I agree that this should not be happening, so something else is in play

If you are sending a velocity reference, it should not matter that you are in flux vector control, (which gives you more torque at low speed, lower voltage and current at high speed, etc but should not overspeed your motor)

The acceleration and deceleration times can cause your speed to lag your setpoint by a significant amount. That can also happen when you reach current limit on the drive output ... but I don't think that's what is happening.

I've never used 'jog (MAJ) command with a velocity reference' ... how does that work?
 
I've never used 'jog (MAJ) command with a velocity reference' ... how does that work?

Like this apparently: https://imgur.com/a/Q3CMQvU

I've been digging deeper. I thought the velocity commands were being sent over Ethernet but it appears that is only used for feedback to the PLC. In reality the velocity reference is passed to an analog motion module (1756-M02AE) where it compared to the motor feedback to compute a torque reference, which is then sent to an analog input on the VFD. I still don't know why torque control is being used but this equipment has been running like this long before I got here.
 
I didn't realize this was a motion application. Again, my experiences are antique now, but many (perhaps all?) of the motion control applications I worked on had the servo drives set up for torque mode and the motion controller closed the position loop using a position feedback and drove it with a command reference to the servo amplifier. Your VFD being set up in torque mode sounds like the same situation.

In other words, my previous comment can be ignored.

Is this an application that worked well and now doesn't? Or is it new (or relatively recent) and now exhibits this problem?
 

Similar Topics

Is there any issue with running a VFD-controlled 3-phase induction motor in torque control mode in the braking region continuously (back-driven by...
Replies
14
Views
3,977
Why it is said Maximum torque of induction motor is inversely proportional to frequency square ? Maximum torque formula for induction motor is...
Replies
13
Views
4,354
As you know there are three region for induction motor's operating. What I'd like to ask if motor is regenerating energy during plugging which...
Replies
2
Views
1,529
I can not understant two points about magnetization current in induction motors. 1-It is said magnetization current is almost constant in...
Replies
3
Views
2,236
Hi! Would it be possible to replace an induction motor (VFD controlled) with a stepper motor and drive? Are there any advantages by doing so...
Replies
8
Views
2,832
Back
Top Bottom