Capacitor safe discharge time

[Modula2]

For industrial systems, I am used to seeing power factor correction capacitor discharge resistors, in minutes. A special design appears to have a discharge in seconds or perhaps a fraction of a sec. Does anyone know of limits on quick capacitor discharge.

 

[Skogsgurra]

The long discharge time is usually when you have bleeder resistors on DC capacitors. Much shorter discharge times when you have AC capacitors (like PF correction etc). Inductive discharge is used there.

In the simplest case the inductor is the primary of a VT. Its resistance is quite low - in the kohms range - so the discharge time (seconds) is something like microfarads divided by 100 (at 1 kohm CT primary and discharging for ten time constants to reach near zero volts).



Gunnar Englund

http://www.gke.org

 

[waross]

The discharge times in minutes that you are used to seeing are minimum code requirments. This would apply to power factor capacitors that are switched by a contactor or disconnect switch and similar switched capacitors.
The actual discharge time in service, when directly connected to motor terminals or transformer terminals is as described by skogsgurra.
yours

 

[amps21]

skogsgurra,
I just commissioned a 66Mvar bank, 115kV Capacitor strings which were fuseless and had internal discharge resistors rated 5Mohms which would discharge the capacity to a very low percentile in 5minutes. So was our protection for closing, 5 minutes before a any reclose allowing time for the bank to discharge.

I did not quite understand when u said bleeder resistors for DC but inductive discharge for AC.

Thanks

 

[Skogsgurra]

amps21,

Of course, there are bleeders on AC capacitors as well - just as you said.

The possibility to use a voltage transformer primary (or a simple inductor) for a fast discharge cannot be used in a DC circuit since the reactance (wL) will be close to zero and the inductor would catch fire very quickly.

The AC makes the inductor look very high-ohm on AC so the current will be very low. When the AC is removed, what is left on the capacitor is a DC voltage and the inductor has (since w = 0) a very low impedance to DC, which discharges the capacitor very quickly.

It is basic electricity - nothing mysterious about it.

Gunnar Englund

http://www.gke.org

 

[BobM3]

Wouldn't really high discharge rates damaged the cap (internal localized heating from i squared r)?

 

[amps21]

Agreed Skogsgurra,

BOBM3- The capacitors are rated for continously carrying the charge based on the rated capcity design. I think if the losses in the capcitors were detrimental that should happen even when the capcitor is enerigised. I would agree to your view if it was c case of reenergising the cap bank without allowing it to discharge completely.

Thanks

 

[Skogsgurra]

BOBM3,

That is only true when you get down to microseconds discharge. But you are far from that - about one million times longer time.

Connecting the capacitors is known to cause heavy current spikes. That's why air cored reactors are used in series with capacitor banks to control the current transients. Such reactors not only save the capacitors, they also make the circuit breaker's life a lot easier. Nuisance tripping of electronic overcurrent relays is also avoided.

Gunnar Englund

http://www.gke.org

 

[bacon4life]

Increased losses are the price to pay for faster discharge time.

A data sheet for a 762 kVAR 8.32 kV 29.3 uF capacitor has a 1.467 mohm discharge resistor. This would have approximately 50 watts lost due to the resistor (0.065 W/kVAR). The bid specification for one of our capacitor banks calls for losses to be less than 0.2 W/kVAR. Thus it seems that losses due to the resistors are on the same order of magnitude of losses in the rest of the capacitor.


This particular unit will discharge to 50V in approximately 4 minutes.
To discharge in 1 minute the resitor would be 360 kohm with losses of ~200W.
To discharge in 1 second, you would need a 6 kohm resistor, with losses of 12000 W.

What sort of special design have you seen Modula2 that discharges resistively in fractions of a second?