Single VFD running two motors - One motor temperature much hotter. 2

[edison123]

Single VFD running two 400 HP AC motors driving on-shore oil drill via gears.
One motor is original winding and one was rewound with 10% more copper area and hence 10% lower DC resistance.
Both the motors had the same inductance since the turns are the same.

Upon recommissioning, the rewound motor with lower resistance was running much hotter (60 deg C hotter) than the original one.
Then client VFD team did “ID run that made the motors compatible” whatever that means.
And the 60 deg C temp difference between the motors just vanished.

Can the drive gurus here explain this? Thanks.

 

[LionelHutz]

Probably changed the V/Hz ratio.

 

[mparenteau]

An ID run is a procedure whereby the VFD "learns" the motor(s) characteristics via an algorithm. I'll defer to jraef on the rest...

Mike

 

[jraef]

An ID run is a procedure whereby the VFD "learns" the motor(s) characteristics via an algorithm. I'll defer to jraef on the rest...

Mike

Correct, ID Run is what ABB calls their “autotune” procedure and is necessary when running motors in any form of vector control (SVC, FVC, FOC, DTC). But you can’t do any of those control methods when you have more than one motor, you are only able to do V/Hz control, because no two motors have the same electrical characteristics even if they appear to be the same.

So if the user is trying to do SVC or something higher with two motors, that’s the problem from the outset and may have contributed to one of them failing. They need to either switch to V/Hz control, or get separate drives and run the motors in torque sharing mode as a master/follower pair. If some salesperson told them they could run two motors from one drive and use something like DTC, they led them astray.

Some newer SMALL drives will requires an ID run procedure even if using V/Hz control, because the newest transistor packs are now switching 10x faster than before, which means slight variations in motor circuits can sometimes have bigger effects. But I doubt that would be the case on an 800HP drive.

 

[edison123]

Thanks all. But why would tuning / not tuning the drive or whatever mode affect the motor winding temperature so drastically?

 

[iop95]

What is VFD model?
Both motors drive same load?

 

[jraef]

Thanks all. But why would tuning / not tuning the drive or whatever mode affect the motor winding temperature so drastically?

A drive, running ONE MOTOR in any type of vector control mode, is watching the performance of the motor and adjusting the PWM pattern going to it in order to correct any “error” between what it is being told to run the motor at, and what the sensors tell it the motor is doing. What it is tweaking is the difference between the necessary current to magnetize the windings, and the current used to produce torque. Without a motor model, which is what the tuning does, the drive’s processor is just doing a “best guess” of what to do, which then affects the differences in magnetizing current vs torque producing current. Too much magnetizing current puts the motor windings into saturation and over heats them. Too little makes the motor lose torque, slow down and draw more current. It’s a mess.

In “Sensorless” Vector Control, the processor is determining the motor position and speed by tracking current anomalies that happen as the rotor bars pass through the stator fields. With two motors, it has no idea what to do. The two motors are not giving it accurate rotor data to work with, because it cannot differentiate one motor current signature from the other. So it is usually jumping all over the place trying to find a steady state.

 

[waross]

A suggestion for your comments, Jeff.
With the lower impedance of the rewound motor, or for some other reason, the maximum allowable V/Hz ratio may be less.
That may explain one motor running much hotter than the other, that si;
The V/Hz ratio is acceptable for one motor but the same applied V/Hz ratio may be pushing the other motor into partial saturation.

Then client VFD team did “ID run that made the motors compatible” whatever that means.

Possibly an independent ID run on each motor and then the V/Hz ratio of the drive set to the value of the motor with the lowest allowable V/Hz ratio.

 

[GeekSteward]

Would not being designed for "Inverter Duty" have anything to do with the rewound motor becoming hotter?

 

[waross]

As I understand it "Inverter Duty" has more to do with thee ability of the insulation to withstand higher voltage peaks and transient spikes than are seen with an equivalent sine wave.

 

[edison123]

Both motors drive the same load via a single VFD. VFD details not available at the moment.

The new winding is VFD duty rated (which is anyway voltage related and not current related).

jeff - Isn't the very idea of V/Hz is to keep the flux density constant without going into saturation at any point?

 

[waross]

V/Hz;
The current is limited by the impedance of the winding (motor or transformer.)
The inductive reactive component of the impedance is frequency dependent.
Thus the impedance is not linear with frequency.
At lower frequencies, the resistance becomes a more significant portion of the impedance.
That is, the resistance stays the same while the inductive reactance drops with frequency.
The result is that if the selected V/Hz ratio is safe at the maximum frequency it is safe at lower frequencies.
At the nominal working frequency (50 Hz, 60 Hz.) the inductive reactance generally predominates to the point that there is little change in the safe V/Hz ratio .

As a WAG and to illustrate, let's assume a motor with 5 Ohms inductive reactance and 1 Ohm resistance.
The impedance will be 5.1 Ohms.
Now we reduce the resistance 10%, the impedance is now 5.08 Ohms. so the higher resistance motor, in this example will accept 0.37% higher voltage before saturation.
In this example, if the lower resistance motor is safe (below saturation) at 400 Volts, the higher resistance motor will be safe at 401.5 Volts.
The original V/Hz ratio would have to have been set on the very ragged edge of saturation for 1.5 volts to make a difference and start into saturation.
jraef, is it possible that the original V/Hz ratio was set so close to saturation that less than 1% increase in voltage would make a noticeable difference?
Is it more likely that the drive was not set to V/Hz mode originally and setting to V/Hz mode was all that was needed?

 

[Gr8blu]

400hp squirrel cage induction motor (which is what I assume these units are) tends to be a random-wound coil design. There are many ways that a rewind can affect the overall performance of a machine - individually, the changes are slight but the cumulative effect can be somewhat large. Changes range from cross-section to turn length, to coil forming/shaping (by hand), to the care with which the coils are inserted into the slots ... let alone matching the original winding circuit and connections.

The original "tuning" of the drive is perhaps lost to history, but a knowledgeable drives tech would have run it for each motor individually and used the "worst case" parameters regardless which motor the parameters covered. Just plugging in the rewound unit without identifying the subtle (or maybe not so subtle) differences could reasonably be expected to produce the results it did - i.e., a drastic shift in load-sharing capability.

The "inverter duty" wire used in the rewind means that the enamel - or mica tape - used as a strand insulation is considerably thicker than the normal "industrial" enamel build for machines not originally intended to operate on a drive. The modification to the strand insulation was required because the drives (at least in the early days) were prone to voltage over- and under-shoots which were both excessively fast and had high peak values. Both distortions in the waveform resulted in more stress on the insulation, particularly on the turn-to-turn and phase-to-phase insulation materials - leading to premature and (usually) catastrophic winding failures.

Just a rather innocent question - I have to assume from the original post that the machines have some sort of embedded temperature detector in the stator winding. Any possibility that the rewound machine used a different detector material (say maybe 120-ohm nickel) instead of the original (say 100-ohm platinum) - and the drives guys had to revisit their scaling factor on the sensors' resistive feedback to accurately arrive at a temperature value?

 

[waross]

The original "tuning" of the drive is perhaps lost to history, but a knowledgeable drives tech would have run it for each motor individually and used the "worst case" parameters regardless which motor the parameters covered

I agree.

 

[jraef]

As I understand it "Inverter Duty" has more to do with thee ability of the insulation to withstand higher voltage peaks and transient spikes than are seen with an equivalent sine wave.

Also better ability for the motor to cool itself at lower speeds.

As to the V/Hz issue, yes, in V/Hz mode the drive is NOT doing all the stuff I said above. My assumption was that if someone was doing an ID Run, it was because they were using vector control of some sort, and that just isn’t going to work when you have more than one motor.

If it IS in V/Hz mode, then yes, the tuning shouldn’t have mattered. So did I misinterpret this? Is it is V/Hz mode and one motor is still overheating? In that case the suggestion of doing an ID run separately on each motor and using the data from the worst one might be a decent strategy. I’ve had to do an autotune on motors with newer drives because of the higher transistor speed causing issues, but that’s usually only on smaller drives. I’ve never run into that on an 800HP drive, but admittedly, two 400HP motors on one drive is not something I come across very often any more…

I had a situation like this long ago when some ABB drives, running in V/Hz (because SVC was not a common thing yet) were running a pair of 250Hp motors on a long gang saw. Long story short it turned out to be that because of the design of the machine and the location of the drive, the cables to one motor had to go up and over something so were almost twice as long as the others. That small difference in impedance caused enough of a mismatch that the motor with the shorter cables was taking most of the load and tripping its overload relay.

 

[edison123]

Thanks jeff. The client says the VFD was in V/Hz mode and the current sharing was almost equal. After the ID run, the motor overheating is gone, and it is running well now. Still don't know how the overheating started and then was gone.

gra8blu - It's a standard random winding with just more copper and hence lower resistance. I disagree how the coils are formed (given the same pitch, there are not too many ways it can be formed) or how they are placed changes the motor inductance drastically. Yes, both motors have embedded PT 100 RTD's.