VFD Sensorless vector vs V/Hz control for starting high torque load

[philly]

I have always heard that for loads that require a high starting torque a VFD should be set to a sensorless vector mode as opposed to V/Hz mode. What specifically about the sensorless vector mode makes the drive capable of producing a higher starting torque then V/Hz mode? What is the limiting factor with the V/Hz mode that may not allow a high torque load to be started.

We have a load application that requires up to 250% starting torque to start the load. Drive and motor are both rated and have been tested for this starting requirement. Currently this drive is in a V/Hz mode and sometimes has trouble starting this load. We are working on setting the drive up in a vector contorl mode (FOC with encoder) to help with this issue?

What are the limiting factors of V/Hz and the advantages of vector control for this starting application?

 

[TxEngr]

Philly,

I noticed that no one responded to your first post and you?re still seeking info.

I?m not the best person to answer your question here but can provide some input since I installed a 350 HP motor and drive on a lime kiln a few years back. In that application, the drive was oversized for 100% torque at zero speed (constant speed application) and I noticed that your drive was also sized for this. The drive was an ABB ACS800 and was run in torque control. The autotune was performed with the motor uncoupled from the gearbox and upon coupling, only minimal tuning was done. We probably would have been fine with the ?out of the box? settings. I had an ABB engineer in to help with the setup but his work was minimal.

It seems that you?ve done everything right in your installation. The one question I had was if the motor was uncoupled from the gearbox for the autotune or was the gearbox disconnected from the kiln? That might impact your autotune. Secondly, check your gearbox to make sure you don?t have a lot of ?slop? in it. That could be affecting the operation. The only thing I could suggest is the purchase of the ?enhanced? AB drive (the ABB one ).

Jaref maybe able to provide you a good answer to you question here.

TxEngr

 

[Jraef]

In non-engineer speak...

In a Scalar drive (V/Hz only), the drive spits out a Voltage a Frequency according to the speed you have selected. But the VFD has no idea whether or not that had the desired effect on the motor and load; it does it's thing and hopes for the best. If the load turns out to be more than the motor can handle, the motor slows down, the slip increases and the motor pulls more current and creates more torque, but this is a "sloppy" process at best. So essentially the drive provides an output, the load creates an "error" in performance, but the drive doesn't really do anything about it. The end result is not always effective.

In Vector control, the VFD uses feedback from the motor to see the error, then to determine the exact vector of voltage and frequency to produce exactly what is needed to correct the error. It has a high speed processor to crunch the numbers and quickly tweak the output voltage and frequency pattern to maximize the torque and / or tighten the speed regulation. Because of this, you can operate a motor at Breakdown Torque at any speed, even zero if necessary.

Vector control can be accomplished with external feedback from encoders called Closed Loop Vector or Field oriented Control (FOC), as you are doing, or with what is called "Sensorless Vector Control" (SVC) or "Open Loop Vector Control". Both names are technically incorrect; there is always a sensor and it is always closed loop. But the difference is in that what is called SVC uses very sensitive current sensors inside of the drive to watch exactly what is going on, and comparing it to a mathematical model it has created of the motor during setup (the "tuning" procedure). Both methods work fine, but the SVC has one limitation; it cannot watch the motor performance if the motor is not moving, i.e. zero speed; like what you would want with a hoist. FOC, where you have the encoder feedback, can do that.

So for you, Starting Torque will always be highest and repeatable with Vector control of any type compared to Scalar control. If you need it at Zero Speed because you are needing to release a brake and you want full torque first, then you need FOC. If not, SVC will probably work fine and you will not need the encoder.

 

[philly]

Jraef

Thanks for the great explanation. I consider you an expert on this topic

One of the problems I am having with my powerflex700 drive is that when it is trying to start its kiln motor load from rest it is stalling. The motor is rated for 575A but has a 60sec overload rating of 1200A to produce 250% torque. The drive itself has an 820A rating but it also had a rating capable of providing this 1200A for short time periord for starting this load. What is happening now is that in both V/Hz mode and in FOC mode (with encoder) when we try to start drive from rest the drive does not go above the 820-850A value and then shortly trips out on overcurrent. It appears that the drive is not supplying the appropriate amount of current to the motor to provide the required amount of starting torque. We do have a problem getting the autotune procedure to produce repeatable test results, however I'm not sure if this is the problem or there is another setting in the drive that needs to be looked at. Could this be an issue with the autotuning?

In Vector control, the VFD uses feedback from the motor to see the error, then to determine the exact vector of voltage and frequency to produce exactly what is needed to correct the error. It has a high speed processor to crunch the numbers and quickly tweak the output voltage and frequency pattern to maximize the torque and / or tighten the speed regulation. Because of this, you can operate a motor at Breakdown Torque at any speed, even zero if necessary.

Can you not operate the motor at its breakdown torque in the V/Hz mode. Wont the load torque cause the motor to draw more current, increase slip and operate at the breakdown torque in the V/Hz mode? Can you not get the same amount of torque in both control modes?

So for you, Starting Torque will always be highest and repeatable with Vector control of any type compared to Scalar control. If you need it at Zero Speed because you are needing to release a brake and you want full torque first, then you need FOC. If not, SVC will probably work fine and you will not need the encoder.

I understand the concept here, but why is V/Hz not capable of providing this starting torque? Does it have limitations that dont let it output full torque at low speeds? Is it because of voltage drop at low speeds and with a scaler control mode the drive doesn't increase voltage to maintain torque?

What is the main difference for setting the drive up in "Torque Regulation" vs "Speed Regulation" mode?

 

[winnie]

If you supply the motor with the correct frequency and the correct terminal voltage, you can get any torque you want at any speed. This includes producing full breakdown torque at zero speed, or (if you are willing to oversaturate the machine) even higher torque than normal breakdown.

I learned on slightly different definitions than Jraef, but in practise they come down to the same things.

As I learned it, 'space vector modulation' is a method used to generate the PWM output of the inverter. It is a separate process from any feedback used to determine _what_ the desired output is. In theory, you could design a system that uses space vector modulation in an open loop fashion, and get the same sort of performance that you expect from a V/Hz machine.

In practise, inverters built with vector modulation are also built with the feedback methods that are capable of optimizing output, and are simply called vector drives.

In theory, you can also go in the other direction, using feedback loops to adjust voltage and frequency in order to get the desired performance at any given torque and speed, without using vector modulation techniques. I am in fact doing just this for some lab tests. We have a motor model that calculates approximations to the correct frequency and voltage, we have feedback loops, and we generate output by defining the frequency and voltage being developed by the inverter, without using vector modulation.

A simple V/Hz control method doesn't have any of these feedback loops. So instead of adjusting voltage and frequency so as to get maximum torque out of each amp, you apply a given frequency to the motor, with a voltage that should give normal operation at that frequency, and you hope that the motor does the right thing. Hit the wrong voltage, and you get the wrong magnetic saturation, and you need too much slip to get your desired torque, and you end up needing more amps for each foot pound of torque.

-Jon

 

[philly]

If you supply the motor with the correct frequency and the correct terminal voltage, you can get any torque you want at any speed. This includes producing full breakdown torque at zero speed, or (if you are willing to oversaturate the machine) even higher torque than normal breakdown.

When you are referring the the correct frequency and terminal voltage, are you referring to simply keeping the V/Hz ratio constant. So are you saying that with V/Hz mode you can produce full torque or even breakdown torque at zero speed as long as you apply the correct voltage and frequency whichi I'm assuming dictates the V/Hz ratio? Why then is V/Hz typically not used for supplying full toruqe at zero speed?

A simple V/Hz control method doesn't have any of these feedback loops. So instead of adjusting voltage and frequency so as to get maximum torque out of each amp, you apply a given frequency to the motor, with a voltage that should give normal operation at that frequency, and you hope that the motor does the right thing.

When you use the term "Get the max torque out of each amp" are you saying that each amp provides the most torque when the V/Hz ratio is at the motors designed value?

Hit the wrong voltage, and you get the wrong magnetic saturation, and you need too much slip to get your desired torque, and you end up needing more amps for each foot pound of torque.

I'm not sure if I follow what you are saying about having more slip to get the desired torque. With any motor isn't the amount of toruqe that the motor produces a function of the amount of slip? In other words as torque goes up so to must the slip? With that said how then would you need more amps for each foot pound of toruqe. Can you explain this one.

 

[Jraef]

Are you sure you don't have some sort of current limit feature enabled? I'm not familiar enough with the Powerflex, but different mfrs call it different things, and other features may have the same net effect. For example, "Stall Prevention" or "Jam Protection", things like that. It may be prudent to do a Line by Line look at every parameter and read the descriptions thoroughly for anything that is enabled, either by someone doing the commissioning or by factory default.

Also yes, it could be an issue with autotuning if it is having trouble establishing the motor model it needs.

 

[winnie]

Sorry, I should have been clearer. By 'correct' voltage, I don't mean holding the V/Hz ratio at the design value.

A particular torque can be produced by numerous different combinations of slip and magnetic field strength. The higher the magnetic field strength, the less slip needed to produce a given torque. One particular combination of magnetic field strength and slip will give best efficiency, and one combination will give lowest current requirement (usually two points are very close together, but they can be different if a higher voltage gives lower current but uses more power to produce the same output).

The nominal V/Hz ratio pretty much supplies the 'optimal' flux for the nominal full load of the motor. If you want to operate at higher than 100% load, then you want _higher_ than nominal flux for best efficiency, similarly for really low torques, you actually get better efficiency if you use _less_ than nominal 100% flux levels.

By 'get the most torque out of each amp', I mean that you need the correct V/Hz ratio for the desired torque, and in general this V/Hz ratio will be _different_ than nominal, and different than that provided by V/Hz control.

As I said above, the torque results from both the slip and the flux. So you can get the same torque with different (related) combinations of flux and slip. You can increase the flux level, and get the same torque with less slip. Generally high slip is associated with high losses, but producing flux is also associated with losses. As a vastly simplified analogy, consider a system where the output is the _product_ of inputs A and B, but your _loss_ (inefficiency) is the _sum_ of A and B. You could get your desired output C with an infinite number of combinations of A and B (simply solve A * B = C, but only one value of A will minimize the sum A+B at the same time as producing the desired product.

-Jon

 

[TxEngr]

I'm assuming you're doing the autotune with the motor uncoupled from the gearbox then trying to run the load connected. It is possible that it could be an issue with the autotune, but it could also be that either you have the wrong data for the kiln load or they have left too much mud in the kiln and your load is both unbalanced and excessive. Does the kiln have a pony motor that can be used to get things going to see if the system will run once it's going. If so, try that and look at the overall tuning of the drive while running. Make sure the kiln is unloaded and then try restarting it.

I once had a kiln that couldn't start up loaded without the use of the diesel pony drive but unloaded it would work just fine. A bigger motor and new drive fixed the problem.

 

[Besoeker]

Vector control can be accomplished with external feedback from encoders called Closed Loop Vector or Field oriented Control (FOC),

If you need it at Zero Speed because you are needing to release a brake and you want full torque first, then you need FOC.

At zero speed, the encoder gives no feedback.

 

[Jraef]

At zero speed, the encoder gives no feedback.

???
An absolute encoder does. Are you saying I should have used that specific term?