DC drive, current curve

[BVSM22]

Hello all
Can anyone here in the forum explain the reason for the steep peaks in this current curve measured in a dc drive (ASEA Tyrak8) at high load. The fault was detected after a disconnect for overcurrent, from the drive's electronic overload protection. The system could be restarted shortly after and then apparently operating normally. The current limit was then adjusted to In. DC-motor runs fine as before. Have checked trigger pulses, but not on thyristors, too dangerous during operation. Current measurement is from the drive's current transformers in the 3 phases. The peaks in the current curve is only present at high load. The customer did not want to spend time on troubleshooting at the time as the system was working satisfying and also was hard to do without.

 

[FacEngrPE]

My guess is that the thyristors under high load have overlapping on periods (trigger angle too small or before zero crossing), or something related.
Could the current limit (brickwall) be set too high?

 

[BVSM22]

Thank you for your interest.
The current limit is set according to the motor In = 527A which is much lower than the drive In = 800A, the overcurrent protection is set to motor In x 1.05.
Yes, the peaks are there when the thyristors overlap and from what I have been able to find about thyristor rectifiers, this commutation will take place with high current for a time determined by, among other things, the impedance of the network. That could fit with a weak phase as the peaks are there along with current in a certain phase, but if there should be a weak connection in a phase I would expect, at this high current, very large heat generation here. Are there parts of a network that will be able to handle that.

 

[FacEngrPE]

The current limit setting does not matter if the trigger circuit is misbehaving.
The usual case with AC is that the trigger circuit turns the thyristor on, and it turns off at the following zero crossing.
If the trigger circuit fires early, the result can be current flowing between phases. If the overlap increases much more (possible if the cause is a failing component), expect the thyristor protection fuses to open.
I have seen these event, but I do not myself do the troubleshooting. Perhaps one of the troubleshooters here can provide some hints.

ASEA has a rebuild kit that replaces the entire controller, a scheduled rebuild might be a reasonable approach depending on what troubleshooting finds. Link

 

[iop95]

Check input L1 saturation current level.
Seem that one or even two of them enter in (partial) saturation above a current value.

 

[panter]

If the AC conductors are accessible, you can see in which branch ie the thyristor is problem with ampere clamps by measuring current. That won't help you now, but it can help with defect repair in order to shorten the repair time.