Instantaneous Seeting philosopy for transformer feeder

[rockman7892]

I was curious to hear from others what their philosophy was for setting instantaneous protection (50) on relays protecting MV transformer radial feeders in industrial applications.

I have always understood the approach to be to set the transformer feeders primary relay instantaneous setting above the maximum let-through fault current on the secondary of the transformer (with multiplier to account for asymmetrical offset and additional margin) in order to ensure that the primary relay does not pickup for faults on the secondary of the transformer for coordination purposes.

I have recently heard some others suggest that the this philosophy mentioned above was more of an approach with old electromechanical relays that could not have a time delay associated with the INST setting, and that with the advent of electrical relays with available INST delay, the INST setting should be set to see faults on the secondary of the transformer with a time delay to coordination with feeder breakers.

With electronic relays is it preferable to set the INST setting to be able to pickup for faults on the secondary of transformer while still maintaining coordination with using delay on setting to coordinate with feeders? By putting an delay on this INST setting is the protection of the transformer itself being compromised by delaying INST setting for internal transformer faults?

This question mainly pertains to transformers without differential protection but I'm also curious to hear approach when diff protection is present.

 

[HamburgerHelper]

I have seen maximum low side current times 2.5 for phase inst. I suppose you could find the actual offset knowing the X/R and set it less. The margin I was told was to account for dc offset. I think it depends on how conservative you want to be but you can't have it see through the transformer.

For ground inst, you can set it set less sensitive if it is going into a delta. The only concern with that was if it is set too sensitive and it picks up on false ground current due to possibly dc saturation.

I don't think you want to set a delay on your inst. I think that you just want something that trips instanteously in the off chance it happens in your small instaneous reach. Barring that you just let your time overcurrent curves handle it.

 

[krisys]

less sensitive[/color] if it is going into a delta. The only concern with that was if it is set too sensitive and it picks up on false ground current ]

Are you intending to say lower ground instantaneous setting? Let us call it more sensitive setting.

[highlight #FCE94F]rockman7892:
[/highlight]The instantaneous setting for the transformer primary side is not only dictated by the through fault current with offset; but also the transformer inrush current during the energisation.

Normally the setting criteria is based on the higher of the above two values.

 

[dpc]

Agree with krisys. The entire point of the instantaneous trip is to be faster than the inverse-time element that will also be present in the event of a primary winding fault or feeder fault. Adding time delay to instantaneous to coordinate with LV side overcurrent protection defeats the purpose of the 50 element.

Most microprocessor relays should filter out the dc offset current, so can probably be set perhaps at 110% of the max symmetrical fault current on the low side. But I generally let the software calculate the max asymmetrical current and set it about 10% above that.

This all assumes there is some sort of inverse time overcurrent function on the transformer primary.