Motor in overspeed situation

[raisinbran]

We operate a 3500 foot long inclined belt conveyor. The conveyor is powered by two 2500 HP induction motors, and the motors are coupled to the drive through fixed fill fluid couplers. For some reason, during a "routine" stop, the rollbacks failed to engage and the loaded conveyor rolled backwards, overspeeding the drive system (in reverse) and causing considerable mechanical damage. The fluid couplers failed catastrophically due to high speed operation. Fortunately no one was injured.

Each motor is protected by a MultLin overload relay which operates a non-reversing vacuum linestarter. Power factor correction capacitors are installed on the load side of the linestarter and switched with the motor. The maximum allowable KVAC is 592 KVAC per motor, and the installed KVAC is 400 KVAC per motor. The CT providing information to the Multilin is installed between the motor and capacitor. Both vacuum linestarters are fed from a common 4160 volt bus.

When data was extracted from the Multilin, about 30 seconds after the "routine" stop, there was a "blip" of current detected by the CTs - one motor shows a negative KW and KVAR reading possibly indicating that the motor was acting as a generator feeding power into the capacitor bank. The other motor showing KW and KVAR readings of the same approximate magnitude, but in the positive direction. It may be academic at this point, but can I assume that both motors were acting as generators during this episode?

Typical motor start information:
I = 1500 Amps
V = 3500 Volts
KW = 2200
KVAR = 9000.

During this episode:
I = 30 Amps (approx.)
V = 4060 Volts
KW = -25 (+30 in other motor)
KVAR = -220 (+230 in other motor)
I am interested in your thoughts,
Regards,
Raisinbran

 

[buzzp]

The motors were acting as induction generators. However, this was due to the inertia of the load and nothing else. It is likely that the cap bank (and the other motor) was absorbing some of this. Do the motors operate with the same rotation and phase sequence (I would guess so)? I dont think these motors were actually driving any real load at this point simply regenerating voltages which were probably rather high while the other one was absorbing some of this power. Gravity simply kept things moving.

 

[electricpete]

I think it is signfiicant that you saw overvoltage. You may have damaged the motor insulation.

The reason for limiting vars switched during the motor is to avoid overvoltage.

Here's the basis for selecting your capacitors. If you exactly mathced the motor magnetizing vars with capacitive vars, you would have a resonant condition at the frequency of the line (60hz) which can result in overvoltage..

We want to avoid a resonant condition so we limit capacitive vars to some number slightly less than motor magnetizing vars. XC > XL .... 1 / (2*pi*Fline* C ) > 2*Pi * Fline* L

As frequency decreases during coastdown you are actually getting further away from resonance (as we decrease F>Fline in the above equation Xc increases further and Xm decreases, taking us farther from resonance.

But if you should reverse such that speed goes above normal speed, then you are back in the region where resonance is a concern and Xm matches Xs. I think that's why you got your overvoltage.


=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[electricpete]

By Xm I meant XL... magnetizing reactance of the motor.

I don't know where Xs came from... should have been Xc.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[electricpete]

Most likely you did not damage the motor if voltage did not get above the values listed but there may be some questions whether this is the peak voltage and whether the instruments read accurately during abnormal frequency conditions.

Also something occurs to me. The output votlage of a PT may be limited by saturation to some value maybe 130% of nominal in case of large overvotlage on input? I'm not sure about that.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[electricpete]

In case I wasn't clear explaining the resonance issue above, let me try again in a slightly different way.

We have L given by the motor and we select C such that Fresonant = 1/sqrt(LC) is > 60hz. (equivalent to selecting capacitive vars less than inductive vars).

This assures the actual frequency (60 and below) will always be below the resonant frequency (above 60hz) as acutal frequency decays from 60hz to 0 during coastdown.

The L and C don't change and resonant frequency remains the same during your transient (somewhere just above 60hz). If your reverse speed got above syncronous speed, you would generate an actual frequency 60hz and create the possibility of exciting the resonant frequency of your motor L and capacitor C.


=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[electricpete]

last sentence should add the word above as follow: "...you would generate an actual frequency above 60hz..."

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[electricpete]

I have been assuming the machine was disconnected from the power source during the event. In that case I see we can still read the current from the position of the CT between motor and cap, but I'm wondering about voltage.... where was it sensed from. Leads me to question my assumption: did the motor disconnect from the power source during this event?

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[electricpete]

And if not, where do you sense the voltage at?

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[electricpete]

Sorry, momentary brain lapse.... I see now it was a routine stop so of course the motors were disconnected.

Do you have voltage sensed on the load side of the switching device? That seems unusual to me (we only have PT's at the bus).





=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[jraef]

Pure speculation, but I would venture to say that although it appears you were disconnected from the line (or should have been if the vaccum contactors opened) when the backspin occurred. The PFC caps may have provided a short term excitation votlage to the motor windings, and when the motors exceeded their synchncronous speed (from the mechanical overdive), they still provided enough excitation to make the motors regenerate momentarily. But they wouldn't be able to self excite, so when the caps finished discharing, the excitation went away and the motors became simple spinning masses again.

I would seriously doubt anything happened to the motors. This was probably just a curriosity.

"Venditori de oleum-vipera non vigere excordis populi"

 

[raisinbran]

Electricpete,
The voltage is sensed on the line side of the linestarters. The purpose of including the information was to distinquish this event from a normal motor start. Multlin displays the terminology "Motor start" when the current changes from zero to some value of full load motor current. This is the definition provided in their literature. Normally, current flow would indicate a motor start, but with capacitors being switched with the motors -this may not always be true.

Jraef,
That is kind of what I was thinking, but was not sure. No, nothing happened to the motors. OEM was worried more about possible bearing/shaft damage due to rapidly spinning with an unbalanced load while the fluid coupler was coming apart. We meggared the motors, spun the shafts by hand and dial indicated the shafts - no damage was found, and the motors were put back into service.

Thanks to all for your rapid responses,
Raisinbran

 

[electricpete]

OK, KVAR and KW require a voltage and current input.

The current input was from the CT circulating current between the motor and the caps.

The voltage input was from the power system.

These two different sensing points are at completely completely isolated from each other and at different frequencies, so any phase angle information that we come up with is complete gibberish. So you can ignore the distinction between kw in, kw out, kvar in kvar out.... they are meaningless. All you know is a current magnitude and a voltage magnitude. I think this answers your original question.

I still believe it is possible to experience an overvoltage condition from the mechanism I explained above. The shorter the time between power disconnection and overspeeding in reverse, the more likely this scenario. You have a fair amount of stored energy in the caps and inductance and relatively low energy dissipation rate through resistance. Does voltage go through zero when speed goes through zero on a transient basis ... I have to think about that one. Anyway, it seems safe to say that the longer time from power interruption to reverse overspeed the less likely, the shorter the time the more likely. If no damage to your motors then it is an academic question.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[electricpete]

"All you know is a current magnitude and a voltage magnitude."

I didn't mean to imply that they are at the same place. Only that the only info you have is a current magnitude (at the motor) and voltage magntiude (at the bus).

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[raisinbran]

Electricpete,
Good point about the lack of fixed phase relationship between the voltages and currents. I stated it, but did not actually think about it in that sense. All that I know for sure is the line voltage(s) and the motor/capacitor current value(s).
Raisinbran

 

[Barry1961]

Just out of curosity how long does a normal stop take?

The reason I ask is that it may provide a indication of a overload condition. For example if the motor normally takes 40 seconds to stop when the conveyor is loaded and we know that it stopped then reversed and built up enough speed to generate the reading you gave in 30 seconds it would be possible to estimate the force required to do it.

The torque@speed transmiting curve of the coupling along with a bunch of inertias would have to be known to get an accurate picture of what happened. But just to get a rough idea of what kind of "routine" stop causes rollbacks to fail.

Barry1961

 

[raisinbran]

Barry1961,
Normal coast down time for the system is in the 5 - 10 second range, depending on the load being conveyed at the time. Regarding possible overload, the system was designed to convey 3500 tons per hour, and the load at the time of the incident was estimated to be 1200 tons per hour.
If all of the system inertias were solidly coupled to the system, it would be easier to estimate a "normal" stopping sequence. However, I do not know how to determine the effect of the fluid couplers on the total system inertia as they coast down. A large portion of the system inertia is contained in the rotors of the large motors. I do not know if that inertia is present in the system at all times during coast down and reversal.
Raisinbran

 

[mc5w]

If there is a possibility of energy regeneration such as brake failure then power factor correction capacitors need to be on their own contactor and overload relay. This will avoid overvoltages that can occur because of regenerated energy.

Starting sequence would be to start 1 motor, engage the capacitor, then start the other motor.

You might want to convert your torque converters to gradually filling them with liquid during startup and gradual withdrawal of liquid during shutdown. This would allow you to confirm that the brakes are engaged before deenergizing the motors.