Double Bank Capacitor EnergizingTransient Inrush Currents

[warickwrx]

Hi

After reading the IEEE STD 1036-2010 refaurding the use of shunt capacitor banks on power systems there still remains an issue I am unsure of. The standard provides a formula to calculate the peak inrush current in a double bank energizing scenario, I have the following question's wich need some clarification if anyone could assist. The formula only considers the impedance between the two banks.

1 Why does IEEE choose not to add the nominal bank current at 60Hz to the higher frequency inrush current at say 3000Hz surely when the 50hz peaks and the 3000Hz peak at the same time this is the worst case
2.What about the transient current that will momentarily flow between the two banks in parallel and the source impedance of say 1000Hz (should this not be included aswell)

basically would the worst case condition not occur when all three of theses currents peak at the same time ?

thanks

 

[magoo2]

Regarding your first question, nominal current doesn't apply for a bank that is already energized. It is not a steady-state calculation. From a transient standpoint, all you need to know is that the energized bank is a charged capacitor and the L and C values of the network determines the resultant current.

If you look at Figure 11 from the IEEE Std 1036 which shows the arrangement for 3 switched capacitor banks with 2 already energized and the remaining one being switched, there are 2 main inductance terms. The source inductance is shown and there is inductance between the banks. I would I think these equations are simplified because they only consider the inductance between the banks and ignore the source inductance. I believe that is why you asked the second question.

I believe A. Greenwood's book on transients goes through this calculation using Laplace transforms. I think it is also covered by an old AIEE paper by N. R. Schultz.

 

[jghrist]

If IEEE Std 1036 uses the same equations that I've used in the past, they consider that the source is a dc voltage at the peak when the switching occurs. This is a conservative simplification of the real case of a sinusoidal source.

 

[warickwrx]

When you say conservative, could you suggest another approach I have looked differential calc's closing at Vpeak and found results to be marginally larger than IEEE.

So what would be the worst condition the outrush currents from the bank already in service or the energizing of the opposing bank ?

 

[magoo2]

As jghrist said, the simplified analysis doesn't consider the source to be a sine wave of voltage. Relative to the high natural frequency components, it is set to a dc value. This was considered acceptable since the circuit was represented as an equivalent L-C network and they employed Laplace transforms to look at the major high frequency term.

I did find an old GE publication that suggests it contributed to the Std 1036 at some early point. It elaborates more on what the equivalent inductance and equivalent capacitance represents.

To be less conservative, I would use a program like PSCAD or EMTP to model it more precisely, or if you have access to an RTDS that would also work. Fundamentally, resistance needs to be represented in the model for the source impedance as well as the connections between the banks. In addition, one of the banks (or the leftmost two in the Fig. 2 attachment) would have already been energized, so would have to be considered also and I don't believe Std 1036 does that.