You are on page 1of 6

Ohmly

Hi veritas
Thanks for posting that document - I found it interesting and informative.
Not sure if you are concerned, but we use two relays that have other ways of
determining the restraint current (you have 3 listed on page 5).
The SR745 uses the maximum of I(HV) & I(LV)
The BE1-CDS240 also uses either the maximum (same as SR745), or mean
restraint.
For mean restraint equation 4.2.11 is true for a two winding relay, but if it is a
three winding configuration the magntiudes are added and divided by 3, and
divided by 4 for 4 winding arrangement (most other relays still divide by 2 for 3+
winding configs)
V: Yes, good comment. The SEPAM T87 also uses the max (IHV, ILV). Will include in
my notes with next revision. I ignored them this time as the relays I used were
only of Type A, B or C.
Also equation 4.2.14 is true when the slope starts at the origin, but some relays
(eg Areva) don't do this.
The formula for a line (y=mx + b) is a bit more generic
y = diff current
m = slope setting (divide by 100% if needed)
x = restraint current
b = starting point on y axis
(we tailor the formula to suit the terms used by the relay to training
documentation)
This is a generic formula that holds true if the bias characteristic is a straight line.
For most modern relays b = 0 then equation matches yours.
This might overly complicate your document, so please feel free to dis-regard, but
I thought I would mention just in case you are interested.
V: Mate, I welcome your comments very much and will certainly take them on
board. This is an interesting point and I need to think about it some more.

V: Just as a point of interest, note that in Fig. 4.2.4 I call the horizontal axis I TOT and
the vertical IREST or IDIFF(min). Why? Well, Im actually surprised no-one has asked me
this but the reason for this is so important that I wish to dwell on it nevertheless.
Technically speaking the horizontal axes is NOT the restraint current. It is a measure
of the loading of the transformer (or generator, motor, cable, etc.) and for want of a
better term Ive coined ITOT. What happens is that the relay now uses a portion of I TOT
to derive its actual restraint current, IREST. Should the measured differential current ID
> IREST , relay operation now results.

This can be clearly seen from an electromechanical relay:

Fig. 1 Reyrolle 4C21 schematic

In the Reyrolle 4C21 shown above, the amount of restraint current is determined by
the bias plug bridge the selections corresponding to a 20%, 30% or 40% setting.
For example, suppose ITOT = 2A and it lies between TP1 and TP2. If Slope 1 = 20%
and the relay restraint characteristic is as per Fig. 4.2.4 (i.e. not an Areva relay) then
the relay will use 20% of 2A = 0.4A to restrain, i.e. IREST = 0.4A. Thus if ID > 0.4A
operation will result.
Thus the minimum differential current required for operation = 0.4A = I REST. This
explains my labelling of the vertical axes.
I hope this example clearly shows that there is a difference between I TOT and IREST. I
know that most manufacturers label the horizontal axes as IREST. I guess it is a matter

of definition but I find it can be very misleading when dealing with a relay such as the
Siemens 7SD522/523 say.
On the one hand they IREST as in my Fig. 4.2.4 and then elsewhere they use it as the
7SDs html pages show in Fig. 2. which shows the relay operating characteristic.

Fig. 2 Siemens 7SD operating characteristic

This caused me some confusion until I coined ITOT for use on the horizontal axes with
the restraint characteristic as in Fig. 4.2.4 and I REST which has its rightful place on the
horizontal axis of the operating characteristic as in Fig. 2 above.
Taking things even further, the display of the diff hi-set characteristic as in Fig. 3 is
not correct:
WRONG
I-diff
Region of unrestrained operation
Iu
Operating
Region
slope 1

slope 2

Restraint
Region

I-min
Itp1

ITOT
Fig. 3 restraint characteristic with hi-set

V - Why is the above incorrect? Because it says that beyond a certain loading of the
trfr the amount of differential current required for operation is independent of the trfr
loading.
The hi-set characteristic more properly should reside in Fig. 2. See Fig. 4 below.

I-diff

Region of
unrestrained operation

Iu

Operate
Area
No-mans
land

45slope

Restrain
Area

I-biasmin

I-bias

Fig. 4 operating characteristic with hi-set


V - Note that these figures are all from different texts Ive written on diff relays so
forgive the interchanging of terms. But I think you get the message?
Slavag
Hi Veritas.
Good job!!!!!
Star to you.
Yes, possible add additional few ways, as Ohmly wrote.
But it more or less same way for slope calculation.
You can add other logic for restraint, base on the angle principle
Ir=sqrt(I1*I2*cos(I1;I2))
It's other way, really other way.
Best Regards.
Slava

V - Thanks for this Slava. Could you perhaps provide an example of a relay that
works on this principle please?
Nightfox 1925
Looking into your document. In calculating the full load currents references, are
we going to consider the FLA at OA rating or at FA rating (which gives the
maximum possible current at no-fault operating conditions)?
V - Firstly, you need to consider the worst case possible loading for CT sizing
purposes. Even temporary overloads of a few hours need to be factored in here.

However, for the differential slope calculations it does not matter. Why? Because a
given rating is chosen is usually with nominal tap in mind. Thus the changing of
the tap position leads to unbalance but the % slope setting remains constant
irrespective with which FLA you started with.
Veritas (or anyone), for example if I use SR 745 relays, the equation for the three
winding transformer is Ir = Imax(I1, I2, I3) therefore, the Ir will chosen to be the
maximum among I1, I2 and I3 inclusive of CTR correction. Do you agree with me?

V - Yes, always remember that the diff and restraint calculations are always made
AFTER compensation has been done for CTR correction, zero phase sequence
elimination and phase shift.
Veritas, do you have similar instruction for three winding transformers? I would
appreciate if you can also share them (for free).
V - No, the job I did this for only had two winding trfrs. Maybe one day Ill add
three winding but the principles are all the same and its not that much more
complicated to apply them to 3 winding trfrs.
From your equations, for three winding transformers:
I(second. winding) = IFLA(second wndg)*(CTRcflv/CTRlv)
I(tertiary winding) = IFLA(tertiary wndg)*(CTRcfter/CTRter)
V - Note these are my generic equations to effect CTR compensation. Every
manufacturer has its own way of achieving this. However, the end result should
be the same.
Note: 'ter' denotes tertiary
Equation 4.2.10 modified to take reference to each secondary and tertiary
winding.
During a fault within the differential zone:
1. IHV is flowing opposite to the SUM (Ilv + Itertiary)
2. Currents in above 1. calculated to apply CT correction factors.
Is this a valid approach?

V
1. Here one has to be very careful for it is a matter of convention and you need to be
specific as to whether youre talking about primary or secondary currents. The
convention I use is that under balanced load or throughfault conditions the secondary
currents IHV and ILV are equal in magn but opposite in phase thus their vector sum
is zero. The primary currents are in the same direction as they are flowing thru the trfr
diff zone.
With an inzone fault, assuming HV is source and LV and TER are possible sources,
then HV current does not change direction but LV and TER currents will both primary
and secondary. So I would rephrase to IHV and SUM(ILV, ITER) add to zero when
considering primary currents whilst secondary currents are now all in phase and diff is
now not equal to zero.

2. As I said before the diff and restraint calcs are classically only done after
compensation.
Hi Veritas,
I made a trial calculation using my understanding on the reference you provided
for a three winding transformer.
I am calculating a slope 1 of 80% at an Idiff =0.883A. I doubt this is quite high so I
think I might have made some errors.
Here is a sample of the trial I made:
[IMG]http://i532.photobucket.com/albums/ee327/primo_beltran/Trialcalc.j
pg[/IMG]
I am intending to use SR 745 with following relay algorithm:
Idiff = Abs[ I1+I2+I3]
Slope = Idiff/Irestraint where I restraint=Max[I1; I2; I3]
I am looking the I diff as +I1 + (-I2) + (-I3 V - it is obvious that the SR745 is
based on the same convention I use. Thus I would recommend that you stick to I
diff according to
Idiff = Abs[ I1+I2+I3]
Thus for balanced load conditions Idiff = 0, i.e. I1 = -(I2+I3) ONLY if IHV is the
only source.
I am hoping for comments and reactions. Thank you very much.
V - Youre welcome mate. Ive made a printout of your calcs and will give it a
squizz some time.

You might also like