Singlephase centre-tapped transformer need NER? 2

[veritas]

Hi

I have a 30kVA 415V to 240V single-phase trfr as per the attached diagram. The trfr feeds a distribution board in a coal mine. The secondary is 240V and I have worked out a fault between the secondary terminals to be around 3480V using PTW (4166A if I neglect source impedance).

1. How do I calculate a A-E fault? I surmise that since the driving voltage is now 110V (half of 240V) and also only half the secondary winding comes into play, the impedance is about half of 3% so the fault current will again be around 23480V? Not sure how to calculate it.

2. If the fault current is indeed around 3480A or in the region of 1000A to 3400A, should there not be a 5A NER in the earth connection? The customer has it solidly earthed with 30mA EL relays on all the distribution board circuits. Can the EL relays handle such high current? Other implications?

Thanks in advance.

 

[stevenal]

Counter-intuitive, at least to me; but the half winding fault at the transformer terminals is higher than for the leg to leg fault. See

http://www.cooperindustries.com/con...rary/BUS_Ele_Tech_Lib_Electrical_Formulas.pdf

 

[waross]

This may be code matter rather than an engineering matter.
Under the code that I work to, the center tap must be grounded when the line to neutral voltage is 300 Volts or less.
NEC may be similar.
That said, there are exceptions.
The relays should operate at any current in excess of 30 ma.
The relays are probably fe by a CT which will saturate at much less current than the fault current.
The relays should be controlling a contactor which is capable of breaking the fault current.

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

 

[prc]

I think the calculation shown in the Bussman document is bit confusing. In reality the middle grounded transformer is two transformers in parallel of 15 KVA rating with 3 % impedance on 15 kVA base. When L-L is loaded, impedance becomes 3 % on 30 KVA base. So an LN fault should result 125/ 0.03= 4166 A without considering the source impedance. With a LL fault on secondary, LV line fault current will remain same, but neutral will see double current viz 4166 x2 A. Corresponding currents in HV will be 1200 (LN) 2400 (LL).Why busman took 1.5 instead of 2 is when secondary is loaded with unbalancing, HV to LV impedance will increase due to leakage flux distortion. In large transformers such flux distortion is avoided ( a sure cause for failure !) and the enhancement factor will be 2.

 

[veritas]

Ok, I'll have to get my head around all this, especially prc's post. I've sort of just managed to understand the Bussman doc (thanks for that stevenal) and am trying to figure out why the 1.5 difference between the line-to-line and line-to-neutral fault.

In the light of all this, would it not be prudent to have say a 5A NER in the grounding circuit? Every other trfr star point at the mine has a 5A NER with 30mA to 100mA ELD's being used.

 

[waross]

Check your codes!
Under some codes, solid grounding of a 120/240 Volt supply is mandatory while at 480 Volts, solid grounding, NER grounding, or ungrounded with ground fault detection is allowed.
Don't confuse the issues;
1> Line to neutral available fault current.
2> Grounding methods.

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

 

[dpc]

In the US, you cannot resistance ground a neutral that is carrying line-to-neutral loads.

 

[wroggent]

Paragraph 3 on page 59 of IEEE 142 2007 states that in the US the practice is to apply grounding resistors at various voltage levels to limit equipment frame potentials to 100V or less. With an adequately sized equipment grounding conductor I do not see this as being a concern on a 120/240V system.

 

[waross]

In Canada:
10-106 AC systems
(1) Except as otherwise provided for in this Code, ac systems shall be grounded if
(a) by so doing, their maximum voltage-to-ground does not exceed 150 V; or
(b) the system incorporates a neutral conductor.
Do you have a similar rule in the NEC, dpc?
Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[stevenal]

No transformers in parallel here, the connection is series. If you want more detail, The Westinghouse distribution (green book) has it. It involves splitting the impedance into each secondary winding and the primary. I like Bussman since it is easier to apply. It is correct based on the assumptions given.

The leg to leg fault is balanced and zero neutral current results. The leg to grounded center tap fault results in a neutral current the same as the faulted leg current. Equipment should be sized to handle the higher current. 30 ma is a leakage pickup value, not an interruption value. I agree that no grounding resistor is called for and likely not allowed.

 

[dpc]

Agree with stevenal - the 120 V fault currents will be greater than the 240 V faults for the center-tapped transformer. This is also covered in Conrad St. Pierre's book (Chapter 14). Non-intuitive for sure. Also, for a 120 V fault, the cable impedance with reduce the fault current significantly - even only a few feet from the transformer.

 

[prc]

dpc, stevenal- Request your time to clear my doubt. I have attached the calculation in line with my thinking. I thought there can be only LN &LLN fault on LV side. Any LL fault will be actually a LLN fault as center is grounded and the current flow will be as indicated in attached sketch. As per my calculations, LLN fault current in neutral will be 2 times the neutral current under LN conditions. But Bussman puts LN fault current 1.5 times LL fault current. Can you help me to understand the logic behind this or rather where I am going wrong.
Conrad book is on grounding or short circuit current calculation?

I have no experience with such small transformers. But for more than half a century, I am engineering similar connected large power transformers for rail trackside supply application ( 132 /+25/-25 KV , middle point earthed single phase units 20-40 MVA) and thermal station start up units. Even though core is same, the windings will be arranged as two transformers in parallel. I always had difficulty to convince consultants this constructional aspect.

 

[stevenal]

Like balanced loading, the balanced fault condition sums to zero in the neutral. You correctly show -120 V on one side, +120 on the other. Reverse the current on the +120 V side, so that positive current flows from + source to fault. Whether or not ground is involved in the fault makes no difference, same as the difference between a 3 phase fault and a 3LG fault on a grounded wye 3 phase system.

Looks like you are incorrectly using the same impedance for both calculations. And flux distortion? I believe the flux will be the same everywhere in the core.

 

[dpc]

"A Practical Guide to Short Circuit Calculations" by Conrad St. Pierre. See Chapter 14. The issue is determining the impedance of the half winding. It's not a straightforward relationship with the full winding. Also, in the US, some primaries are connected line-to-line and some line-to-neutral. This is another factor.

 

[stevenal]

Electric Power Distribution System Engineering by Turan Gonen has a pretty good treatment of the issue, referencing Westinghouse.

 

[prc]

dpc- Thanks for correcting me. You are correct. In this case the secondary windings are in series. In large units we normally provide secondary windings in parallel(axially split windings). I don't have Conrad's book. But I could find the calculations in Gonen's text. But I could not get the derivation of LN impedance as he was referring to Westinghouse Utility Engineering book. I could get from internet Chapters up to 5. But this derivation is in chapter 6 Distribution Transformers.
Does any one has the link to this Westinghouse book ? Or at least this chapter on Distribution transformers. By flux distortion, I meant leakage flux that decide leakage reactance. With parallel connection the neutral will carry the adding current. But with series connection, the currents from each part of LV will be opposing and under balanced conditions, there will be no current in neutral.
I have modified the calculations as attached.
Thanks veritas for bringing up this interesting point and thanks for dpc and stevenal for the guidance.

 

[stevenal]

The Westinghouse distribution book is long out of print and copies are hard to come by. I found it doesn't really add much to Gonen's discussion.

I haven't seen what you are calling a parallel connection. For parallel connections you generally start with two isolated windings. This is common in overhead cans, where the lid can be lifted to redo the connection. Then the paralleled windings are brought to X1 and X3 bushings with no connections to X2.

 

[dpc]

ABB re-published the Westinghouse T&D book some years ago, IIRC. Older copies of the Westinghouse book are available used.

 

[stevenal]

The book referenced is the Distributions Systems Electric Utility Engineering Reference Book, not the T&D book.

 

[prc]

I have a small book "Distribution Transformer Guide" issued by ABB, Jefferson City factory in 1995. (it is an updated version of 1979edition by Westinghouse) Here this topic is discussed in detail with a worked out example. Here also they are calculating out the impedance of 415-120 connection from 415-240V impedance by a different formula. Why it is required? Why it is not given in test report? This book shows another interesting part - LN fault current is more than LL fault current if the fault is at transformer terminals. But this varies when the fault is some distance away on the cable as the cable impedance is added. At one point it becomes equal and when the fault is still away, LL current is more than LN fault current. If interested I can post the scanned pages.
Difference between series and parallel connections and change in impedances is shown in my last enclosed note .