Morning

For those of you who use Amtech, please note that the K2 value used for the armouring for a thermosetting swa cable, operating at 70 degrees is incorrect

Following a recent design project, I noticed that the figures did not stack up and following a conversation with their technical department, a patch will be sent out soon to everyone

If anyone wants the patch now, PM me

Regards

Jobbo

Have you also tried that particular cable in parallel Jobbo - the cable analysis report suggests some very strange goings on - ditto if you start using soft starters at load end in the load/motor dialogue - plays havoc with diversity assesments

regards

OMS

-------------------------
Failure is always an option

OMS

I also tried to establish, how the automatic minimum sizing of the cpc is calculated as the 'Min. Section' earth fault adiabatic requirement is being calculated from the manufacturer's I2t curve, at the fault level, on far end of the circuit. So a cpc size would already have to be selected, otherwise how would the fault level be known? Using near end data, would obviously result in an oversized conductor, as the fault level would drop considerably, as discussed previously

The adiabatic equation, only confirms the selected size, it doesn't select a size. Or am I wrong? I feel chapter 54 isn't clear enough on this, especially with reference to using manufacturer's data, without referring back to section 434

Regards

Jobbo

Does it depend on the kind of protective device? Traditionally, using fuses, worst case energy let-though was with lowest fault current, hence you use max Zs. A lot of circuit breakers let through more energy with higher fault currents though, so worst-case is at the start of the circuit rather than the end - in that case the c.p.c. would be correctly sized, not oversized.
- Andy.

I don't know how Amtech does it, but essentially if you have I2t then all you need is K2S2 - or rearranged to find S. From the results it actually calculates S.

The presumption is that you have I2t at the source end (as Andy highlighted) so the fault level is known. From there I guess it then iterates for say every metre along the stated circuit length.



Well if you exchange < for = you can select a size from that, then iterate, with knowledge of commercial cable sizes - Amtech won't allow you to install a 1.2mm2 CPC for example.

Best though, to consider it an adiabatic expression, not an equation

Which bit of chapter 54 don't you feel is clear enough ?

Regards

OMS

-------------------------
Failure is always an option

Ok so I'm on a tangent, after a conversation with a colleague and my chats with Amtech technical support

Yes, I should of mentioned, this is with regards to circuit breakers

Now from Amtech, they calculate the size from the fault current at the far end of the final circuit, then apply the fault level to the manufactures I2t data. Using sq root of I2t/115 for example. I have this confirmed in an email

They only use K2S2 on phase fault calculations

Now I was always taught to check near and far end fault levels, hence my questioning. You have to question things, when industry leaders in design software do something right?

Now BS7671 states in chapter 54

I = is the value in amperes (rms for a.c.) of fault current for a fault of negligible impedance, which can flow through the associated protective device, due account being taken of the current limiting effect of the circuit impedances and the limiting capability (I2t) of that protective device

How is this being interpreted? Why doesn't chaper 54 put more emphasis on using manufacturer data? Using anything else will require you to have already selected your conductor to establish a fault level and a disconnection time

I'm probably starting to confuse myself, as its been a long day at work, but fault levels can change in an installation quite considerably, as more parallel paths are introduced

Excuse any errors, typing this on my iPhone

Regards

Jobbo

True, but then parallel paths will usually only be in parallel with the c.p.c. - while they'll increase the overall fault current (i.e. in the line conductors), they should reduce the current the c.p.c. carries itself.


I didn't think there was any difference in principle between using manufacturer's data rather than the curves/tables in the back of BS 7671 - either way you're just comparing I2t with k2S2 (i.e. the same as S = sqrt(i2t)/k just rearranged). The only slight procedural difference is that for the BS 7671 charts you'd lookup t for I and work out I2t yourself, whereas much manufacturer's data is arranged to give you I2t for I directly - all the same result in the end.

- Andy.

True, but then parallel paths will usually only be in parallel with the c.p.c. - while they'll increase the overall fault current (i.e. in the line conductors), they should reduce the current the c.p.c. carries itself.

Yes I guess that is true and we have no way of knowing how that fault current is shared

I have a habit of extracting the cable report from Amtech and checking it freehand and when it doesn't stack up, I'm looking for answers

I initially used pro design, then switch to single cable to ensure the programme wasn't relying on a device upstream. Tomorrow I will question them further, as to why it uses far end figures for circuit breakers. Essentially at higher fault levels the disconnection times of circuit breakers are not further reduced and the I2t increases

Now looking at a 100A BS88 fuse (from an old cable report)
Ze=0.08179
Zs=0.08601
Earth Fault Current =2.68
Disconnection time=0.06
Min section=2.13mm2
Armour used (25mm 4c) so 70mm. So equivalent is 70x51/115=31mm2

Again the I2t details from the manufacturer are used

Regards

Jobbo

Of course - t is less than 0.1 seconds so you have to have I2t data in this region ?

Near end and far end faults (earth fault):

as the position of the fault approaches the breaker, EFLI decreases and the disconnection time will probably be at the definite minimum time of 0.01 seconds

You'll be familiar with the maximum value of Zs at load end being derived from :

U/xIn

Where x is the multiple of normal current for the MCB type to give 0.1 second disconection (ie x = 5 for a type B, 10 for a type C and 20 for a Type D)

At the near end condition, you need a minimum value of EFLI (so called Zdb). Assuming (dangerous) DMT is actually 0.01 seconds then the minimum Zdb that can be tolerated is:

0.1 x U/kS

anywhere between these two conditions is effectively thermally protected

To the best of my knowledge, this is how amtech deals with earth fault conditions - sketch an adiabatic line on an MCB curve (preferably an old curve with the tail below 0.1 seconds shown) - you'll soon see what I'm getting at

Regards

OMS

-------------------------
Failure is always an option

Update from Amtech

The earth fault adiabatic check in our software uses the earth fault current at the load end of the circuit. The earth fault adiabatic equation checks whether the CPC will be damaged by a through earth fault occurring at the load end of the circuit. The magnitude of the current used in this check therefore takes into account the current limiting affect of the circuit impedances.

This value of current is used to calculate I2t where t is in excess of 100ms; for a disconnection time below 100ms the manufacturer's energy let-through characteristics are used.

Any increase in the source earth fault level will result in a considerably smaller increase at the far end of the circuit. In your example, earth fault disconnection occurs at less than 0.1s. Consequently, the earth fault current at the load end of the circuit is used to obtain the energy let-through value from the manufacturer's I2t characteristic. As a result the I2t value returned from the manufacturer's data might smaller than anticipated.

ProDesign does not currently report the energy let-through value used in the earth fault adiabatic calculation; only the I2t value used in the phase fault adiabatic calculation is reported. This can be found near the foot of the Phase Fault area of the Cable Calculation report where disconnection occurs in less than 0.1s. I will log your request for the earth fault I2t value to be included in the report.

So the value at the source end are not considered it seems

Regards

Jobbo

Clearly the software does not consider a near end cable fault in accordance with 434.5.2 ( "fault occurring at any point in the circuit"). Although most of the time a far end fault may be more onerous but a near end fault could be in some circumstances?

-------------------------
John Peckham

http://www.astutetechnicalservices.co.uk/