15kV Substation Ground Grid Design - Assumptions for worst case ground fault current

[j_p_c]

Hi Electrical Folks,

I am establishing the worst case ground fault current to perform GPR and Step and Touch calculations for a new 15kV substation. IEEE 80 indicates that LLG faults should be considered in addition to SLG faults. However, my industrial application is low resistance ground. In the case of a SLG fault my currents are quite low (25-800A) and clearing time relatively fast (<0.2s), but of course if a double line to ground fault were to occur and send the bulk of its current through the earth return path then the impact on GPR and absolute fault current would be severe (~18kA of supply current). The substation is located at ~5km from the plants main 15kV switchroom in a rural area and service will consist of 3-wire supply without shield or neutral/ground conductor. For clarity, the new substation consists of a 15kV/0.6kV delta/wye transformer (HRG on secondary neutral), with its MV junction box fed from underground by 3c armoured cable from a terminal pole a distance away from the substation grounds.

I've come to my own initial findings, and wanted to confirm with the experts here. I consider the likelihood of an effective cross-country LLG fault to be so low that it should not be considered for this design application. Instead only SLG fault current should be considered. My reasoning is that if a fault were to develop on one phase at either end of this service, then the low currents and fast response time, controlled transient overvoltage and high system insulation level, would all summate to limit the fault extent to SLG. I just can't imagine a case where a SLG fault would physically evolve into a LLG fault which must flow cross country in my application.

I'd really appreciate comments on this assumption! If any other scenarios come to mind please also suggest those too.

Thanks very much,
Justin

 

[aussiejohn2]

Hi Justin,
What about vermin/birds causing a LLG at the crossarm, or a LL fault on the OHL that migrates to the crossarm and evolves to involve ground?
Regards
John

 

[cuky2000]

Statistically LLG fault account for about 5% of the total type of faults and the consequence for potential damage is usually more severe than the SLG and LL faults.
Although engineers can evaluate the cost/benefit of ignoring the risk, it is recommended to use a careful judgment since the consequence can be catastrophic for the equipment and even potential risk of causing dead.
Please confirm if the 15kV source is delta (line with 3-wires without shield or neutral/ground conductor).
As a suggestion consider the following:
a) Review the clearing time (consider that the fault is cleared by the backup protective device) usually larger than 0.2 sec.
b) Check the impact on the tripping the protective devices for a LG fault in the 15 kV side.
c) Check if the secondary neutral, ground and number of feeders could help to reduce the current injected into the earth determining if the current division factor is < 1.

I hope this help

 

[stevenal]

I fail to understand. If you are using neutral grounding resistance at the source, it will limit the ground current for any kind of ground fault. How are you getting 18kA through the ground?

 

[j_p_c]

Thanks Both of You for your time and thoughts,

1. John: yes I believe those are both real scenarios to develop LL and LLG faults, but will result in a majority of the fault current flowing through the phases. My deeper concern is a bulk return via earth alone (SLG fault at one end, followed by SLG fault at other end).

2. Cuky: thank you. The main sources are delta/wye low impedance grounded transformers (x2) and generator (x1), I mispoke before, maximum contribution from collective NGRs is 500A, not 800A. Ground conductors are not normally brought out on the long radial pole line services, though I will consider requesting one if the grid design necessitates it. Great point on backup protection, thank you, I suspect I'll be around the 0.5s mark once I factor that in. I also agree on judgement as to whether to consider LLG faults. That's what I was aiming to confirm here, is my judgement reasonable for this LRG application in disregarding the possibility of a SLG fault at one end causing the evolution of a second SLG fault at the other end which in turn must drive the bulk of the LL fault current through earth.

Thanks again,
Justin

 

[wroggent]

stevenal, I think he is looking the possibility of simultaneous LG faults, one in the substation and one back in the plant, for example.

Edit: I guess so.

 

[wroggent]

jpc, if this is an industrial application are you subject to a Code that would require a grounding conductor?

 

[j_p_c]

@wroggent: yes that's right. If this were say a high resistance ground, in certain applications I could entertain the possibility of an insulation failure at one end, after a SLG at the other. My initial thinking was I'm being overly careful in my question here, but I just wanted the support of others to cross-check.

@stevenal: yes as per @wroggent, LLG fault current all via earth return.

 

[j_p_c]

@wroggent: I'm not 100% final on this, but I don't think so. Definitely don't think there's a bonding requirement, as the only continuous parts to the pole line system are current carrying. This application is similar to a transmission line in that sense.

 

[waross]

A line to line fault will cause line to line fault current, limited by the impedance of the source and of the circuit.
If this develops into a line to line to ground fault, the portion of the current diverted to and through the ground will be limited by the NGR.
Two simultaneous line to ground faults on different phases may cause a higher ground current between the fault locations.
Note that in this case, the ground completes the circuit for the line to line fault and fault current will flow between the two ground faults.
Touch and step potentials will develop in proximity to the fault locations.


Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[HamburgerHelper]

I thought this was overkill but there are utilities that just specify the ground grid to be designed to the interrupting capability of the breakers. I have designed 40 kA ground grids when there was only like 11-12 kA of fault current available. I don't think I have ever designed a ground grid that was actually based on fault current calculations aside from counterpoise fault current splits.

I don't know what someone does though if they have a substation and for one reason or another the fault current increased and their ground grid was no longer adequate. Does anyone want to dig out all that crushed rock and add to the grid? Maybe, lay copper mesh all over the place? Maybe, you should be asking yourself what may the fault current be in 50 years instead of trying to do some precise calculation.


------------------------------------------------------------------------------------------
If you can't explain it to a six year old, you don't understand it yourself.

 

[j_p_c]

Thanks Bill and HamburgerHelper for your follow-ups,

HamburgerHelpler: that's impressive that some utilities use interrupting capability. I suppose for solidly grounded systems it bears a bit more reason as the current levels are so high. For a large enough utility I suppose it just allows them to standardize their grid designs to some extent. The whole business of monitoring grid resistance is interesting. I really don't believe any of the monitoring wells are tested at the sites I work at, but I don't have enough experience to say, but I am aware of a few approaches at least of ground grid augmentation through various means (new auxiliary electrodes, new rods or wells).

 

[HamburgerHelper]

J_P_C,

I added to my last comment after your reply but I really think that there is a lot of electrical equipment that should not be sized based on calculations. You will look like a smart dummy if you sized a transmission line conductor and find it needs to be replaced before 30 years. A lot of the cost of reconductoring is just in the labor so put up something big at the start. You will look like a smart dummy if you sized a breaker based on current fault current calculations. For distribution circuits, some utilities just standardize on a large conductor and never do lifetime loss calculations and they have the benefit of not needing x different sizes of conductors. You don't want to be smart dumb and have to send out people to put in a new conductor at a later date. If I owned a substation, I would want the grid layed down so I don't have to care about copper corrosion, changes in resistivity of the soil, increasing fault current. ect. I would want it put down and never have to think about it again and I really don't want anyone to ever get hurt because someone didn't test the grid or it was no longer adequate.

------------------------------------------------------------------------------------------
If you can't explain it to a six year old, you don't understand it yourself.

 

[stevenal]

I see. Simultaneous faults, CG at a remote location BG at the station. Line to line fault current flows through ground from remote fault to station fault. I see this as a very unlikely triple contingency; two unrelated faults occur together while someone is in a position to experience the resulting touch or step potential.

 

[davidbeach]

Not really that far fetched. The remote CG fault causes a neutral shift that causes a weak arrestor on B at the station to start to conduct, leading to complete failure of the arrestor. That first fault greatly increases the chances of the second fault, they aren't independent contingencies.

 

[j_p_c]

Thanks Stevenal and David,
Yes Stevenal's point is the direction I was headed into. I wasn't looking at the two faults as independent contingencies. But in my application with overvoltages controlled by LRG and relay response I'm willing to disregard a weak arrestor (which I suppose that weakness may in itself be a form of independent contingency). In terms of design, the arrestors in question (assumed to be healthy) can tolerate full L-L voltage for much longer than the relay response time.

 

[waross]

Be careful of assumptions. A well know aircraft company recently assumed that an AOA sensor would never fail, and then may have assumed that it would never happen twice.
Just saying. I hate assumptions in regards to safety.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[davidbeach]

Ok, it’s not an arrestor, it’s something else, but what ever it is the first fault increases the stress on everything on the other two phases. The first fault greatly increases the odds of the second fault.

 

[cuky2000]

j_p_c : Beware that for a system with impedance grounding the Coefficient of Grounding (COG)>1. This means that most likely overvoltage will occur in your system regardless of what protection scheme is used. Proper design should consider a surge arrester rated for LL instead LG to avoid equipment damage or trigger a fault.

For grounding grid design two approaches have been used:
a) Overside the ground grid
b) Support the design with calc. adding a growth factor (usually 10% to 20% of the available SC) up to the equipment rating.

It is common today to upgrade the ground grid particularly for old stations that not meet the current standard. For example, we are in process to re-built a substation to meet a new breaker rated for 80 kA. The good news was that the utility allows us to choose the calculated split factor and the clearing time-based in a detailed study that was much lower than the traditional values.

 

[jghrist]

@wroggent: I'm not 100% final on this, but I don't think so. Definitely don't think there's a bonding requirement, as the only continuous parts to the pole line system are current carrying. This application is similar to a transmission line in that sense.

If you are subject to the NEC, I can't find anything that indicates you don't need an equipment grounding conductor for a 15 kV OH line. Without the EGC, your ground fault protection might not work for a fault at the 15/0.6 kV substation.