We are here in Aluminum Rolling Mill with problems of Rolling Mill Chatter. The arrangement of the drive train is like this. We have two old GE motors [type MCF-8, 4000HP, 300/750 rpm, 700V] driving the input shaft of the gear box. The output of the gear box is connected to the pinion gear box and from pinion the two spindles drives the top and bottom work roll. The inboard motor is the master motor which is speed controlled and the outboard motor is in current slaving configuration and it gets the current reference from the inboard drive. Both the motors have individual converter for armature and field [The converter is Toshiba Leo pack]. The inboard motor which is the master drive gets the speed reference from the PLC and it’s speed feedback comes from direct shaft mounted Toshiba resolver. Based on the speed reference and speed feedback, the inboard motor generates delta speed and subsequently current reference for inboard motor as well as the same current reference comes to the outboard drive. The outboard drive also gets the speed feedback from the belt driven resolver. This resolver feedback is just for over speed monitoring which is done in outboard drive. Now, coming on to the attachment which shows the following:

1. Time based signal such as: speed feedback [inboard motor resolver speed], outboard speed feedback [outboard motor resolver feedback from the belt driven pulley with ratio 1:2], delta speed [speed controller output from Inboard drive], current ref [current reference generating from Inboard drive]. Seeing the signal, we can see that outboard speed feedback is not good and that is understandable as it is driven by the belt /pulley arrangement and it does not make any difference as this speed feedback is not used for control but seeing the curve I am not able to understand the relationship between delta speed signal and current reference. I am not able to understand why there is big fluctuation in current. For analyzing purpose the scale of speed feedback is 7V for 750rpm and for current reference 3V for 100%.
2. FFT analysis: In the FFT analysis we don’t see anything in speed feedback but delta speed has a frequency component which is 1X motor rpm. And since it is in delta speed the current reference also has a frequency component related to every rpm of the motor. This frequency is still OK as it does not match with the natural frequency of the drive train of our mill. [The natural frequency is between 12 to 13 Hz]. The problem of chatter arises when motor turns at 750 rpm and we have current signatures with 12.5 Hz and it coincides with the natural frequency and resonance.


Any suggestion …. or advise will be useful.We are also checking the alignment of the resolver….

Also the recent torque analysis done on the spindle showed great vaiation of the torque on the mill spindle. And for the information we also know that there are some damages on the pinion gear box.

Time_based_graph.doc (110 Kb, 23 downloads) signal recording [ FFT + time based signal]

Pawan,

It sounds like you may be dealing with a torsional vibration resonance issue, if I understand the data presentation correctly. This is not unheard of in rolling mills. Did your torque analysis include looking for and identifying torsional vibration natural frequencies? Gearbox damage/wear is often associated with the torsional vibration issues and you have also mentioned the gearbox damage.

Skip Hartman

In addition to Skip's comments, check the insulation resistance of the armatures of both motors. Leakage to ground (low insulation resistance) and shorted coils in the armature can cause torsional fluctuations as the drive tries to compensate for the winding degradation.

Insulation to ground should be greater than 100 MegOhms and the commutator should have an even color under the brushes. Burned bars indicate a short.

Sincerely,
Howard

Thanks Skip. The torque analysis which was done at both the spindles revealed the following:

case 1. 12 Hz torque variation over the complete rolling of the coil. The motor was not running at 720rpm but constant 12 Hz torque variation was seen. The torque analyzer have asked us to check the motor control and he thinks the motor control needs to be checked.

case 2: The largest amplitude in torque variation belongs to the motor speed ie 11.6 Hz. In this case the motor was really run at around 700 rpm.

case 3: For medium speed like 6.22 rps spindle speed the torque of the top spindle shows periodical spindle speed proportional disturbances. Because of this disturbances they point to mechanical defects in the drive train

case 4: For high speed 12.5rps at low nominal torque where chatter appears the torque signal shows percenatge wise high periodical variation.These variations lead even to zero crossing torque mostly for the top spindle.That means for a short moment all clearances in the spindle and couplings will be opened and closed again.One consequence is high abrasion and other is the excitation of the natural frequency in the mill equipment.

So, we are also analyzing this and trying to understand the torsional vibration phenomena + how drive control can create this type of vibration. But one thing which I have observed in all the signal recordings here that motor current variation is directly related to the motor speed whether we see any disturbance in the speed feedback or not and I am not able to understand how the drive control is changing the current to vary to that great amplitude at every revolution of motor without seeing any such frequency in the resolver feedback.

The drive speed regulator is tuned at 10 rad/sec.

Thanks once again....

Thanks Howard , we will surely be checking the armature insulation resistance during the next preventive maintenance schedule as per your guidelines.

Thank you.

The reason for the recommendation to check these things was from a troubleshooting standpoint.

Your drives monitor the motor speed by the resolver and by the motor current draw. If you have a grounded armature or shorted armature, there is a fluctuation in the current that the drive sees, at that point, and it attempts to compensate through a change in torque. The result is a torsional vibration that is directly related to speed and increases based upon load.

I have run into this a great many times in the steel industry and plastics industry.

Sincerely,
Howard

Do you have a movable gearbox on ways? This will change the torque value.

Thanks Sam.. Sorry I didn't understand your question.. but in our gear box we have high gear and low gear. For high torque rolling we use low gear ratio [3.78:1] and for low torque rolling at high speed the high gear ratio is [0.914:1].

Thanks Doctor... Mr. Howard .. Things are getting clearer to me . I will analyze more and will check the motor armature. This is a wonderful forum where we can get some clues what to look for. Any good reading material on Torsional Vibration related to our application will be helpful.