MV motor starting current problems

Mill motor, 2.9MW, 6.3KV, rotor current: 775A, rotor voltage: 2770V
In this case sometimes we recorded 775A as starting current and in other times we recorded 400 A for rotor current during start and for stator the current is variable, in some case it was 85 A, and in other was larger than 400a. This motor starts with liquid starter.
Questions: what is your idea about differentiation? What is cause of current reduction? Is the liquid starter resistance variable during each case of start?
Rotor current will be directly proportional to torque and stator current. In affect we need to measure the stator current and calculate the approx rotor current. Measuring the rotor current directly with standard instruments will give errors as we are in affect measuring the current at whatever the slip frequency is.

For example, If the stator full load current was 250A (not sure what it is in this case) and the Stator start current was 400A (1.6x) then we would expect Rotor start current to be 1.6 x Rotor full load current = 775 x 1.6 = 1240A. Torque delivered would be around 1.6 x Full load torque, very typical for a ball mill or sag mill.

Correct measurement in the rotor current can be done with the likes of hall effect transducers or even shunts but not clamp-on ammeters unless they use hall affect or similar technology.

The liquid starter simply adds resistance in the rotor during starting to produce torque and depending on the electrode type and design, the start current and therefore torque should be similar with each new start. Generally the electrodes start wide open at start and slowly close up as the start progresses. Finally the rotor is then shorted as the motor gets to full speed and then runs like any cage motor. Differences in start current (eg as measured in the stator) would only be due to the electrode position at start being different each time. Liquid temperature will affect it only slightly.

The load will not vary the start current, only the resistance in the rotor and the speed of the rotor will affect that.
With no load the motor will be up to speed really quickly irrespective of the starting sequence time setting, and the measurements are probably picking up on start transients so appear somewhat random. So at no load we will not see the true performance and affect of the Liquid Resistance Starter as it is set up to deliver 1.5x FLT which we don't need for an unloaded start.

Sodium Carbonate electrolyte solution is commonly used in these starters at a recommended dose rate to give the start resistance and we just need to ensure the interlocks prevent starting unless the electrodes are in the open position.

By the way, the equivalent of DOL with a slip ring motor is electrodes shorted out which should not be possible with the control circuit. Also, this is not recommended due to extremely high inrush current with shorted slip rings.

This is in affect a simple device so we don't have any issues as long as the stator supply is solid and balanced and the slip rings of the motor are affect connected to the LRS.

Under a loaded start we will see everything will probably work OK and will have time to make the right measurements.

We can run a control system check without main power to ensure the electrodes are moving and the right sequence is in place.