NGR rating calculation when used with with wye-g/delta grounding transformer 1

[LJP1]

Need help with help with NGR rating for Option#1 re: EPRI Technical Brief_Effective Grounding and Inverter-Based Generation. How to properly calculate current through NGR when connected to neutral of Wye winding of Wye-g/delta(not a zig-zag) grounding transformer. What is valid impedance model and calculation to yield correct current rating.

Background (See DG layout PDF)
A new 8 MVA distributed generation (DG#2) wind plant will be connected to an existing substation DG feeder. The new DG will share a parallel connection on this feeder with a previously existing 4.5 MVA DG wind plant (DG#1). Due to the significant total amount of parallel generation resulting on this feeder and the substation due to the addition of DG#2, Utility commissioned an engineering study to model and analyze the impact of various DG grounding configurations and their zero-sequence current sources to better understand overvoltage risk and relay protection coordination issues during a ground fault. The results of this study indicate both a risk for overvoltage greater than 1.39 PU and a risk of relay miscoordination during ground fault scenarios where the utility source and system ground reference may be lost.

Furthermore, the study results show that DG#2 is reliant on DG#1’s Yg-Yg step-up transformer and other distribution ground sources to fully mitigate these risks. Specifically, the DG#2 fails to meet an acceptable Coefficient of Grounding (CoG) of <0.80 while solely relying on the capacity of its own, single, 500 KVA grounding transformer (GT). As such, DG#2 will require a larger capacity GT solution to be designed and installed before it will be allowed to operate independently of other distribution system ground sources in-service.

Utility is requesting DG#2 install the equivalence of 2 MVA of GT capacity and install the equivalence of 1-ohm neutral grounding resistor(s) (NGR), in combination with the GT impedance, to improve its own dedicated CoG for the DG#2 wind plant and to mitigate overvoltage risk during ground fault scenarios.

GT-NGR design considerations and options:

1. One (1) 2 MVA GT with 1 Ω NGR unit:
a. Grounding transformer impedance is 6% at 2 MVA.
b. Converting to ohms at 13.8 kV; 0.06 pu * 13.8 kV * 13.8 kV / 2 MVA = 5.7 Ω.
c. The equivalent impedance = 5.7 Ω + 1 Ω = 6.7
d. The estimated the overall ampere rating will be 13800 V / √3 / (5.7 + 1) Ω ~ 1200 A.
e. Rated for 10 secs per IEEE C62.92.3 and C57.32

2. Four (4) - 0.5 MVA GT with 4 Ω NGR unit, connected in parallel
(equivalent circuit to Option 1 and shown in attached PDF DG layout):
a. Grounding transformer impedance is 6% at 0.5 MVA.
b. Converting to ohms at 13.8 kV; 0.06 pu * 13.8 kV * 13.8 kV / 0.5 MVA = 22.9 Ω.
c. Impedance on each path = 22.9 Ω + 4 Ω = 26.9 Ω
d. The equivalent impedance = 1 / ((1/26.9)*4) = 6.7 Ω
e. The estimated overall ampere rating will be 13800 V / √3 / (23.9 + 4) Ω ~ 300A
f. Rated for 10 secs per IEEE C62.92.3 and C57.32


Alternative calculation approach yields different results that seems too high of current rating through resistor.

• Let positive-sequence Thevenin impedance be Z1.
• Let negative-sequence Thevenin impedance be Z2.
• In the zero-sequence circuit, we will have the full impedance of the ground bank in series with 3 * Z_NGR. Zero-sequence Thevenin = Z0_GroundBank + 3*Z_NGR Note that in the zero-sequence circuit, 1/3 of the NGR value will be present and not the full NGR
• Let VF be the prefault voltage. This is a L-N quantity, and equal to 8 kV (13.8 kV LL)

The fault current in each sequence can be calculated as follows:
Ia1 = Ia2 = Ia0 = VF / (Z1 + Z2 + Z0_GroundBank + 3 * Z_NGR)

This leads to (assuming Z1 = Z2 = 0):
Ia = 3I0 = 3VF / (Z0_Groundbank + 3 * Z_NGR)
Ia = 3I0 = VF / (Z0_Groundbank / 3 + Z_NGR)

• Current with 2.0 MVA, 1 Ω NGR = 8000 V / (22.9 Ω / 4 / 3 + 1 Ω) = 2750 A.
• If we go with 4 x 0.5 MVA banks, with 4 Ω NGR:
o Current through each NGR = 8000 V / (22.9 Ω / 3 + 4 Ω) = 688 A

 

[argotier]

Threads like this one make me realize how much of the stuff I've seen in uni I've forgotten.. for good. Thevenin?? I've never used it in the job.

Thats why I'm just gonna comment on your first approach, which is something similar to what we do (but we do it to calculate the needed impedance for a grounding transformer given the ground fault current).

First: remember that you need to use the zero sequence impedance of the grounding transformer, not the rated one. This Zo its kinda hard to calculate... assume its 0.9~0.95 of the rated impedance. Unless you already have this value from test, of course.

Second: All transformers impedances are rated as "per phase" values. This means that you need to use one third of the value and then add it to the NGR impedance. This gets clear when you remember how you measure the zero sequence impedance of a 3ph transformer:

Case 1 above with this last correction will become:

c. The equivalent impedance = 5.7/3 Ω + 1 Ω = 2.9
d. The estimated the overall ampere rating will be 13800 V / √3 / (5.7/3 + 1) Ω = 2747 A

Same with Case 2.

Which is basically the "Thevenin formula" you found for the second approach and will give you the same results. (I've ignored the 0.95 correction factor just to facilitate the comparison with the results you obtained).

 

[LJP1]

argotier, Thank you for your time on this. A few follow up questions.

1. Based on the calculated parameters of the resistors (1Ω @ ~2.8kA-10s & 4Ω @ ~700A-10s), what experience do you have with similar sized resistors being implemented GTs for such projects? Wondering what is economically and physically practical for a DG customer to specify, purchase and install.

2. A few NGR mfgs (sales and application engineers) that I have contacted are not versed in specifying an NGR when paired with a GT application. Do you know of any NGR mfgs that have more experience and have an appropriately sized NGR product? Is their any specific NGR terminology and parameters that I should use and reference that might help bridge the gap?

 

[argotier]

No problem LJP1.

1 - I'm used to see more values like Case 2 (4 ohm - 700 A), but I've also seen some NGRs up to 4000 A (big and heavy NGRs). I'm not very fond of the idea of paralleling GT + NGR, but thats just my personal opinion because I've never seen that solution before (but one big NGR could be less expensive than 4 smaller ones, same with GT).

2 - To my knowledge, there is nothing special about a NGR that will be connected to a GT. You don't need much data to specify a NGR really, and you already have the more important one: the resistance value.

From my experience, you need the following data as a minimum to define the NGR:

Reference standard (C57.32).
Resistance in ohms (you have 1 ohm or 4 ohms).
Transient earth fault current (you have 2,8 kA or 0.7 kA).
Transient time (you have 10s).
Permanent rated current, if any.
Voltage class (consider 13.8/√3).
And environmental conditions if anything special (altitude, temperatures, pollution, etc.).

Hope it helps.

 

[LJP1]

argotier or others with knowledge,

Additional follow-up questions. Thanks in advance for your time.

Based on calculations in previous posts and summarized here:
- Current with 2.0 MVA, 1 Ω NGR = 8000 V / (22.9 Ω / 4 / 3 + 1 Ω) = 2750 A = ~2.8kA
- Current with 4 x 0.5 MVA banks, with 4 Ω NGR:

​

Current through each NGR = 8000 V / (22.9 Ω / 3 + 4 Ω) = 688 A = ~0.7kA

Questions:
- Transient Time and Current ratings - 10secs is typical spec.
How does this correlate with typically much shorter fault current and relay protection time in cycles? It all relates to energy; i.e. dissipation of watts over time, correct? So, how do NGR mfgs balance economic and physical practicalities, especially with a GT helping to mitigate the fault? Do they consider specifying a shorter time or reduced current rating?

- Voltage class -
The NGR is tied to neutral with the GT, of in-series to ground. So would the NGR voltage rating be subject to the just the voltage drop across the NGR? i.e. 2.8kV (2.8kA x 1ohm = 2.8kV or .7kA x 4ohm = 2.8kV).

- Permanent, Let-Through and/or System Charging Current ratings -
How critical are these rating parameters in comparison to the proportional affect of the transient current rating for the design and construction of the NGR? With the complication of this GT-NGR application on a distribution network, with another adjacent DG source, what would be a practical approach to estimate these parameters? Are rules of thumb sufficient, such as charging current in MV systems of 1 A per MVA capacity? If so, what are other convenient rules for NGRs?

If forced to dive deeper, for example, I believe the Permanent Current rating would be the amount of load current imbalance in the system. It maybe practical to capture phase current data at the most upstream common relay (i.e. substation low-side transformer relay) to determine the max net imbalance during different levels of generation and distribution load. However, how do I then apportion that imbalanced current out to the various ground paths of the feeders and DGs, including the portion to NGR connection? It would seem to be a pretty complicated load flow analysis.