Sensorless vector/vector controlled

peoplehouse

Member
Join Date
Jan 2009
Location
California
Posts
117
Hello group, can anyone explain in laymens terms the difference or preference in using(vfd)sensorless vector and vector controlled? I'm a little unclear on the terminology and effects.
Thank you,
Jeff
 
Vector controlled is with an encoder. This is still the best method for speed control with varying loads.
Sensorless is the new kid on the block. It uses feedback from the motor windings as an encoder. To do this, the controller needs to be autotuned to know the exact properties of the motor.
I haven't done this, but assume there is tweaking of various parameters to get everything right. Someone with elevator/conveyor experience probably knows this best.
 
Last edited:
Thanks Keith, I've never done any PI or PID functions with a drive, if I'm reading your reply right is that when you would want to use vector controlled? Then sensorless would be constant torqe, voltage or current?
Thank you,
Jeff (confused) H.
 
'Vector' control in asynchronous AC motors generally refers to flux vector control. There are technically two flux vectors; one for the rotor and one for the stator. The flux vector for either is the magnitude and direction of the magnetic flux. So flux vector control is the active and independent control of both the stator and rotor magentic flux vectors of a motor.

Why is this important? If you can control these two quantities you can accurately control the torque the motor produces. And if you can do that you can accurately control speed with relatively high bandwidth over a wide range of loads.

The 'sensorless' part refers to the rotor position sensing device, usually and encoder. In order to very accurately control the rotor flux vector you need to know how the rotor is moving. This is what the encoder does. It provides direct information to the flux controller about where the rotor is. Since the flux controller only needs to know how much the rotor has moved from scan to scan an incremental encoder works just fine.

Sensorless vector system attempt to emulate the function of the encoder by looking at motor stator voltage feedback and determining how the rotor has moved. To do this the flux controller needs to know some information about the motor. like stator and rotor resistance and inductance as well as other items. However, this feedback get very 'noisy' as the motor speed decreases. More correctly the noise in the signal becomes a larger precentage of the total signal as speed decreases. At zero speed the whole thing kind of falls apart since all of the voltage feedback is basically noise.

So if you want to run at low speeds accurately or want to generate full torque at zero speed you need to use actual rotor feedback (an encoder). At 'higher' speeds you canus sensorless system. 'Higher' means different things for different drives.

I don't think that the use of a PI loop disqualifies a semsorless vector system. I think it has more to do with how slow you need to have accurate control. Some newer drives have very good sensorless algorithms and will give you very good performance down to slow speeds.

Keith
 
To add a little and why you may need it.
A planar ie volt freq type VFD will only control voltage. It has no way to sense what the rotor is doing with respect to the stator and cannot control slip. Where this becomes very important is at very low rpm
for example motor has 60 rpm slip and is a 3600 rpm motor at 60 hz.
at 1 hz then the motor shaft should be rotating at 60 rpm but because of 60 rpm slip it is really doing zero.

The vector type drives sense both the stator rpm ie at 1 Hz 60 rpm and the rotor rpm. If the rotor is not doing 60 rpm it adjusts the output to get the rotor to turn at correct speed. In a planar VFD ie V/F this would be comparable to torque boost.

Sensing is done either with an encoder or other rpm sensor
OR it is done without an rpm sensor ie sensorless where VFD is able to compare stator rpm to rotor rpm electronically as explained above.

Dan Bentler
 
To try to further put this into "laymen's" terms, an open loop, V/Hz, scalar drive simply takes the drive input speed command and converts that into a frequency and sends it to the motor. The motor takes the frequency and converts it back to speed based upon how many magnetic poles it is wound with. The only thing the drive knows about the motor (and this is the key piece of info) is the nameplate data you entered in the software and the current the motor is drawing. As a result, the drive cannot manage motor speed or torque very accurately and you would only use this type of control in applications where that is ok.

Vector control or Flux Vector Control adds an encoder to the motor shaft so the drive now has an accurate speed feedback signal. The drive can now control speed directly instead of frequency and the result is precise control of speed. In addition, since shaft speed and drive output frequency are both known, the torque in the motor shaft can be calculated with high accuracy permitting very good control of torque even down to zero speed.

Sensorless vector control is very much like Flux Vector Control except that the speed feedback signal is generated by a motor model residing in the drive memory rather than by an encoder. Development of that model is done when commissioning the drive and the accuracy of control is largely dependent upon the accuracy of that motor model and the calculation update time of the microprocessors. The control result can be almost as good and fast as Flux Vector but there are subtle differences that, for a few critical applications, make sensorless vector unsuitable. The big benefit with sensorless vector is lower cost: no encoder, no encoder card in the drive, and especially, no special motor with backside mounting provisions for an encoder.

One last point. The quality of sensorless vector performance is dependent upon the quality of the motor model and processor speed. There are huge differences in drives in this regard which, naturally, causes huge variations in sensorless vector performance. Some of the poorest have specs virtually equal to open loop drives while some of the best have specs remarkably close to Flux Vector. You have to read the fine print to be sure you are getting what you need.

Hope this adds something to what has already been said on this subject.
 
Thanks Keith, I've never done any PI or PID functions with a drive, if I'm reading your reply right is that when you would want to use vector controlled? Then sensorless would be constant torque, voltage or current?
Thank you,
Jeff (confused) H.
Sensorless has one primary definition/function - a fake encoder. Sensorless by itself doesn't suggest anything else - Use of an Encoder does. Look at the parameters associated with encoder operation, then understand that Sensorless is just a poor man's encoder.
A VFD varies voltage and frequency. Current can't be controlled directly - the size of the motor and load dictates current. We vary voltage to control current.
The simplest VFD varies frequency on a linear scale. 0-10 volts in, 60 Hertz, means 5 volts will run the motor at half speed - 30 Hertz. That's what we expect with open loop control. The true speed will vary depending on mechanical conditions such as load. We can't possibly know that without an encoder.

We have to vary the voltage (and as a result, the current) too. That's where all the tricks are. Constant torque, vector control, etc. There are plenty of algorithms to choose from. Some don't need an encoder. They just vary the output curve based on what's expected for your size motor.

Forget about sensorless and PID for now. That's what's throwing you. Study 3 basic control strategies. V/I, constant torque, and vector control. Nothing else will make sense until you have this foundation.
 
Last edited:
Thank you all for the replies and links. Now I know enough to become truly dangerous!
Seriously, thank you for taking the time to help me out.
Jeff
 
Just one further addition, Jeff, since you asked about PI or PID. These terms are associated with a closed loop control system. When you have the simple V/Hz drive I described above, there is no speed feedback so there is no closed loop. Therefore, no PI or PID in the speed loop.

As soon as you have a speed feedback signal as in sensorless or flux vector drives, the speed loop closes up and you now have to set P, I, and sometimes D gains to get the speed controller to work right.

Notice that in each case, I said speed feedback and speed control loop. That's internal to the drive. There is such a thing as an external control loop as in pump control where you are measuring pump output pressure and controlling pump speed to keep the pressure constant. Any drive whether V/Hz or vector can do external control loops and, again, where you close up the loop with a feedback signal, you have P,I, and D gains to deal with.

Or, to summarize, internal PID is for speed loop control on precision drives and external PID is for control of non-drive things like temperature, pressure, position, etc.
 

Similar Topics

Does anyone has an idea in what range the speed of a sensorless vector control driven motor will deviate from the real measured speed?
Replies
4
Views
2,381
I already experienced that autotuning of a motor in sensorless vector control isn't always possible to reasons such as very long motor cables, or...
Replies
0
Views
1,282
Lately I had to commision a some drives for fans. The motors were 22 Kw, the drives Siemens G120. Flying restart is implemented to avoid trip when...
Replies
1
Views
3,372
I routinely use a "standard" VFD for controlling air flow of multi-stage centrifugal blowers. A prospective customer recently indicated "we were...
Replies
5
Views
3,170
I have an issue after programming a Lenze 8200 Vector drive using Lenze Global DRive Easy software. I have connected using a Can Bus interface...
Replies
2
Views
570
Back
Top Bottom