Motor Locked Rotor Current and Short Circuit Studies

[darock]

When calculating the short circuit ratings of LV switchgear , does IEC require a tolerance of 20% to be added to motor manufacturers locked rotor current values. I am dealing with a piece of 400 V LV switchgear whose short circuit rating is marginal compared to the equipment type test certificates. The application or non application of 20% tolerance makes a significant difference to the acceptability of the equipment.

 

[mykh]

I have a feeling that even expensive commercial software for short circuit analysis do not necessarily take equipment tolerances into account. It's not a big deal when most accurate information is available for all the system components, so input data error is equal to none.

Problem comes when typical values are substituted instead of exact values (say, typical transformer, motor impedances, estimated cable run length etc). Any substitution like this introduces error into calculations which can very well impact final result. I witnessed brand name software producing results with far more significant digits and accuracy comparing to input data. It means final results have errors of order sometimes of 10e-5 percent and are more accurate than input data with tolerances say of 2 to 5%! I believe anyone who heard about things like capacity calculations and is familiar with concept of precision should be very cautious about using such software tools.

Neglecting input data tolerances is very shaky business and can impact results that arise from measurements and/or calculations in surprising ways. You should also be cautious of disclaimers saying that company doesn't take responsibility, expressed or implied warranty for accuracy of produced results, disclaimers stating that actual measures could be different from the given results, software assuming certain amount of load, power factor, operating on typical input values, and nevertheless producing results with three digits after decimal point.

From my point of view, any result is useless and potentially misleading if there is no estimate for it's tolerance.

 

[dpc]

I do not think IEC calculation methods require addition of 20% tolerance. The standard motor impedance values used are generally already conservative.

Motor contributions are always an approximation, like everything else in your model. The IEC and ANSI methods give different results, but both have been shown to be reasonably conservative when compared with actual fault data over the years.

 

[mykh]

"Reasonably conservative" does it mean that calculated short circuit currents are always above actual fault current values and therefore safe?

It that always sufficient to use conservative values not even knowing how much different they are comparing to actual experienced values?

Wouldn't it be more desirable to know approximate fault current values with accurate tollerance, so one can consider both minimum, maximum and any possible prospective fault current value within the limits?

I have more trust in numbers like 1.9kA +/- 5% rather than 1.924557kA. First value is obtained by rounding off three phase fault on load side of 100KVA, 600V, 5.0% impedance transformer with infinite short circuit current on primary available, second value is obtained by pluging same data into software developed by major marker player.

Selective co-ordination and arc flash analysis are the examples where rather more information other than "conservative" short circuit current is preferable.

 

[dpc]

More accurate is always better than less accurate, I suppose. I was just trying to explain that IEC and ANSI have specified methods for calculating short circuit currents for use in determining the adequacy of breaker interrupting ratings.

If I am calculating short circuits for some other purpose, I might use a different method. Most protection engineers don't even bother with motor contributions in their calculations at all.

Short-circuit calculations are full of many other approximations other than motor contributions. The entire notion of sub-transient reactance in motors is borrowed from Parks transformation for synchronous machines and is itself only an approximation.

As far as significant figures, I agree that carrying a fault calculation out to 5 decimal places doesn't make any sense. But that certainly doesn't make the computer software less useful or more untrustworthy than a manual calculation. All you have to do is apply your experience and common sense and use however many significant figures you would like. The program I use allows me to choose the number of decimal places displayed. I do generally perform a sanity check calculation to make sure that the computer results make sense. The computer certainly can't replace engineering experience and knowledge. It's just a tool.

 

[mykh]

Yes, I agree that more accurate is always better than less accurate. But you can't make your result more accurate by choosing the number of decimal places displayed in the program or using however many significant figures you like. Significant figures are not place holders, they are SIGNIFICANT because they are there for purpose. You don't have to rely only on your experience and common sense by doing the calculations.

Technically, you can calculate resulting error based on input data, respective tollerances (error of measurement) and model used (error of calculation). Mathematical procedures for this are well known. In fact, not much would be known about subatomic interactions if experimental physicists didn't use same measurement and calculation techniques in there work. We are talking about heavy duty equipment accompanied with readily available manufacturer specifications and tollerances. I guess the problem is that even expensice 10K+ software dont have the procedures for error analysis incorporated in it.

Also, I agree that conservative results are safe for use in determining the adequacy of breaker interrupting ratings. 50kA rated breaker can be used in circuits with prospective fault current of say 15kA and you'll be safe. In the example with circuit breaker, no short circuit analysis would be required at all if you had whole wack of money to spend on big expensive oversized breakers.

I don't want to judge anybody but I think protection engineers should have a pride in the work they perform. It includes collecting most accurate information available, specifying tollerances, considering different modes of power system operations inlcuding motor contributions etc. when performing short circuit analysis. Software is as good as people who developed it, and even the best computer used to run the program or software high price won't make it any better.

 

[dpc]

mykh,

You seem to have a big concern about significant figures. It doesn't bother me that much. Even when doing a complicated calcuation by hand, I still have to rely on experience and common sense to verify the results and to know what an acceptable level of accuracy is.

So whether I use a calculator, slide rule, Excel or dedicated electrical analysis software, there is no substitute for the training, education and experience to be able to interpret the results.

Cheers.

Dave

 

[mykh]

dpc,

Not considering tolerances of input data, using however many figures you would like and adjusting tolerances by fooling around with the number of decimal places displayed, one can quite easily "justify" poor calculations.

I would personally prefer to rely on reputable procedures for capacity calculations and proper math rather than on experience and common sense to verify the results.

Cheers, mykh

 

[Electic]

1) Some analytical software does allow you to specify equipment tolerance.

2) I have yet to witness a computer operator, with even the most robust programs, to come close to proper fault current calculations, without some experienced engineer qualifying the results as being appropriate real world answers. In the final analysis, usually a few digits of accuracy are sufficient. I.E.: switchgear will be rated at 28kA withstand, not 27.99kA (although extra digits are helpful with other ratings such as crest factor etc.)

Specifying two digit AIC values does not 'justify' poor calculations, however; does accurately convey that this is only a fault current calculation that should by nature be liberal. Anyone who thinks that calculating to one amp accuracy is fooling themselves.

3) Rather than 6 decimal fault current values, arc flash calculations will be most affected by accurate device clearing times, for the given fault current levels.

 

[mykh]

Electric,

Could you please list software that allow you to specify equipment tolerance. I understand that an engineer is the one to review and approve the short circuit study. I think he would have sweeter dreams if he knew software takes best care of both the data and their tollerances so he doesn't have to guess how's accurate is the result: is that within 0.001%, 0.01%, 0.1%, 1%, 10% tollerance?

Different pieces of equipment have different impact on short circuit current. Some small motors or short circuit sources far away from point of fault will probably introduce small error into final result even if the source data is not accurate. On the other side, large short circuit sources in the vicinity of the point of fault may introduce significant error into the resulting value even if quite accurate data is available for the equipment. I can envision how's common sence and experience can be applied to system consisting just of few components, but I don't think this approach is any helpfull in estimating the accuracy of final result for systems consisting of tens and hundreds components.

Fuse melting and clearing time current characteristics tollerance is typically +/-5%, so incident energy and flash protection boundaries will have significant error even if most accurate fault currents and equipment data is available. I think procedures for arc flash analysis are safe as they generate conservative results or require analysis performed at reduced fault current values which again might produce arc flash energies well above than experienced in field. Being overly protected is still better than being unprotected and ignorant about arc flash hazards.

Thanks, Michael

 

[mykh]

As a follow up to the above discussion, please check

http://www.arcadvisor.com/sitemap.html

for error propagation analysis in action. Since recently, ArcAd's short circuit online calculator features second to none input data analysis and hard coded error propagation rules ensuring that the resulting fault current values are not more precise than justified by the accuracy of input data. I find it very important but widely ignored issue. Please check the website F.A.Q. section for more info.