Cap bank issues when tripping 8850HP 2

[buzzp]

We have a 100kV/21.6kV substation. On the secondary side we have 4 cap banks, 300kVAR each, setup in a delta. We then go through 13000 ft of 636 to another cap bank (3 banks at 900kVAR each). After another 2500 ft of 636, we go to a 21.6kV/4160V transformer, which has a cap bank of 600kVAR and it is on the 4160V side. All cap banks have current limiting reactors. Please see the two attached drawings. Yes I know this is not typical to have this many banks.
When we have a fault that shuts down the (6) 4000VAC motors (total 8850HP), we will often times blow a fuse or fuses in 'cap bank 1' near the substation. On occasion, the reactors will also come apart.
One of my concerns is the delta configuration of the cap bank by the substation. It seems this configuration will provide for the highest currents when we trip the 8850HP due to the cable inductance. We don't blow fuses or trip breakers on Cap bank 2 or 3 but they are much closer to the 8850HP load. Will changing this bank to a Y help us out to give the current some place to go rather than recirculating around the cap bank?
What other solutions should I be looking at such as surge arrestors, etc.?

https://files.engineering.com/getfi...89-429f-4628-a00e-5710e536b751&file=caps1.pdf

https://files.engineering.com/getfi...81-63b1-4c51-a805-533875c0e18b&file=caps2.pdf

 

[bacon4life]

Wow, what a puzzle. It looks like the fuses were up-sized from 16 A to 50 A on cap bank 1. Are the oversized fuses still blowing? Or did the failures switch to the reactors once the larger fuses were put in?
Is the secondary side of the 100 kV/21 kV transformer grounded?
Why are cap bank 1 reactors rated for 160A when the cap banks only draw amps?

 

[cuky2000]

It does appear that went a failure occurs, the stored energy in the other parallel shunt capacitor banks, discharge high frequency outrush current in the failed circuit that probably causes the fuse to blow and may overstress the capacitors that may induce to earlier failure.
[sub]COMMENTS:
1)The dwg provided indicates that the size of the original fuse for each 300 kVar cap bank were uprated from 16A to 50A. Caution is recommended to keep the fuse size without exceeding the energy rated capability associated with the I[sup]2[/sup]t of the unit.
2) Operate capacitor banks with open phase, after tripping a fuse, overstress the cap bank and should be avoided[/sub]

Unless a transient analysis is performed, it is difficult to assure if changing the cap bank delta to wye will fix all the fuse tripping trouble.

[sub]COMMENT: Possible mitigation techniques will be adding properly rated surge arresters, tweaking the inrush/outrush reactor, install resistor or synchronous control in the circuit breakers to close and open during minimum stress (zero-crossing or peak voltage)[/sub]

 

[buzzp]

I don't have a lot of faith in the company that designed the cap banks, nor do I have faith in their drawings as far as accuracy (based on their work since I started here). The equipment was installed before my time. I don't know if 16A fuses were ever installed. I do know, they used to have 80A fuses in the cap circuit...they said they were blowing up caps and reactors and decided to reduce the fuse size to 50A. Now, it seems, we generally only blow fuses but sometimes a reactor will go too. They say we typically only see this in hotter ambient temperatures...we are in process of getting some ventilation in the cap bank panels, which sit outside (don't worry not compromising environmental ratings).
The transformer secondary shows it is grounded but I haven't verified this physically...yet.
Cuky2000 - You stated "It does appear that went a failure occurs, the stored energy in the other parallel shunt capacitor banks, discharge high frequency outrush current in the failed circuit that probably causes the fuse to blow and may overstress the capacitors that may induce to earlier failure." I am thinking the sudden loss of 8850HP is causing the initial failure, rather than a fuse blowing first. We do have a sort of functional fuse blown protection - it looks at current unbalance to detect a blown fuse. It is possible that this circuitry doesn't work in all cases. I am in process of getting indication via switch on the plunger.
Surely the Y config has to help but as you suggest, is it enough. The other possible mitigations will be considered.
Thanks to the both of you for your input...more is welcome

 

[bacon4life]

Buzzp-You might try checking the inrush estimate versus the fuse rating with

http://nepsi.com/resources/calculators/peak-inrush-current-from-capacitor-bank-switching.htm

Reconnecting the cans in a Y would reduce voltage across the caps and would cut the reactive output to one third of existing value and reduce the voltage across the caps by sqrt(3). While the lower level of voltage stress may decrease the number of failures, it does not explain your current failures.

Cuky- How does operating with a blown fuse increase voltage stress? In a delta configuration I thought a blown fuse A phase fuse would put LL voltage across the series combination of AB and CA capacitors while BC capacitor still sees normal LL voltage. In a floating Y configuration, a blown fuse A phase fuse would float the A phase capacitor while putting LL voltage across the series combination of B and C phase capacitors.

 

[cuky2000]

Hi Bacon4life,

After a fuse blown, an unbalance capacitor bank condition will occur. The remaining connected capacitors will interact with the system reactance creating a transient high-frequency disturbance with current and voltage distortions reach in harmonic content.

It should be noted that the IEEE Std 18 (IEEE Standard for Shunt Power Capacitors) limit the overstresses during contingency as follow:
• 110% of rated RMS voltage
• 120% of the rated peak voltage

We should be recognized that to visualize the complexity of the transient interaction between multiple back-to-back capacitors with reactors, motors, transformer, and T. Line inductance is too complex to see with a make eye until a transient study will be performed.

 

[buzzp]

There is a lot to think about here...thanks for the info. I will be working on an ETAP model and for short term, changing when all these banks are energized.

 

[jrostron]

As you have already figured out, the possibilities are numerous. An EMPT study should help but most likely looking at fault recorder events and wave forms will give a hint as to where to start. Many relay types record these fault events in their memory.