EE Power Handbook Loadflow fails in Mathcad 14

[Kernie]

Electrical Engineering Handbook for Power Transmission and Distribution Section 1.3b
provides incorrect load flow answers in Mathcad 14. The same handbook provides correct answers in Mathcad 13. The handbook uses Newton-Raphson interative solution that apparently fails to converge in Mathcad 14. The power mismatch from specified to results shows to be 9.051 in Mathcad 14 (which is way too much)and 1.41x10^-6 in Mathcad 13 (which is very acceptable). This makes Mathcad 14 unusable for me since it give me the wrong answers. Just upgraded. Upgrade is worthless to me with this bug. Any suggestions?

[TomGutman]

On the information provided, the only advice is to return MC14 and stick with MC13. Was there anything in particular that you want from MC14?

I do hope you have at least the basic service pack (M010 version). Without that it's completely hopeless.

For anything else one would need the actual failing worksheet. No way to know what the cause of the failure, hence how it might be avoided, otherwise. A programmed Newtonian iteration is not something I would normally expect to fail in MC14.
__________________
� � � � Tom Gutman

[RichardJ]

I doubt anyone wants to download and install an e-book just to look at this, so could you post the relevant page please.

Richard

[alnstevens]

On 5/18/2009 11:20:12 AM, rijackson wrote:
>I doubt anyone wants to
>download and install an e-book
>just to look at this, so could
>you post the relevant page
>please.
>
>Richard

Attached is the appropriate section from the e-book (saved in MC13 format). I thought at first that the problem was with the multiple assignment - the logic selected in 'Format/Worksheet options/Compatibility' is MC12, which had a different assignment order. However, changing this to the MC11 option makes only a small difference.

I can't see anything else at a quick glance.

stv

[ELSID]

Posted as MathCAD 11 file for greater Collab.

Look at your origin. Many Collabs will only assist if origin is set to 0

[TomGutman]

The MC12 option for multiple assignments is the correct one. MC11's implementation is a bug.

The problem is that the iteration is set up with two range variables. Mathcad does not define the order in which such an iteration takes place, and MC13 and MC14 implement a different order. Rewriting the iteration using explicit iteration, thus fixing the order, results in identical results in MC13 and MC14.
__________________
� � � � Tom Gutman

[Kernie]

Wow! I did not expect such a fast and helpful response. This was the first time ever using any forum like this.

That fixed my load flow problem with the EE Power Handbook. It now works for the actual what-if solutions I was studying.

I had received the EE Handbook as part of my upgrade from version 13 to 14 and had incorrectly assumed the handbook was compatible with the Mathcad version it was bundled with. I had written Newton-Raphson power flow solutions in Fortran and C++ before and was amazed at how compact the Mathcad version was. Your explanation and fix also helps me understand it better.

Perhaps the handbook will be made compatible in future.

Thank you so much for the help.

[TomGutman]

The sheet carefully avoids program structures, and was probably developed for an early version of MC, perhaps 5 or 6, before programming structures were standard (they were first introduced in MC6, but only in the pro version). I doubt that the e-book has been reviewed or revised since then.

If this is an e-book that PTC has for sale and included as a bennie in the MC14 upgrade, then you could, if you have a maintenance contract, file a bug report against that e-book. If this is one of the many e-books from the library, freely available for download, and PTC merely pacakaged a copy with the MC14 disk then nobody has any responsibility for the e-book -- library files are strictly on an as-is basis.

Interesting difference in viewpoints. You looked at the sheet and was impressed by how compact it was. I looked at it and thought it was remarkably convoluted. I find the repeated use of the sin/cos of a modified arg (encapsulated in an if) to be quite strange, as some simple complex arithmetic would do the job. I also suspect that it is unduly complicated by being done by components (magnitude and phase) rather than just using complex values, but don't know enough about the application to know how to do that (no idea what the different types of busses mean).
__________________
� � � � Tom Gutman