Power system tools?

[Binraider666]

Hi all,

I have dabbled in programming for years, dodging between C++ and endless dialects of Basic. In the last 18 months or my job has had me heavily involved in analysis of the thermal performance of high voltage electrical systems (e.g. buried cables, overhead lines, transformers etc.)

There is a whole range of code in use, of various vintages (the oldest is Fortran IV, the more recent stuff is in Matlab). The methods themselves are in many cases well ahead of what the much of the world does, yet there is clearly an opportunity here to build something much more capable than the mathematical modelling done in 1977 (and it's inevitable restrictions on computing power). I would love to be able to bring these ideas into the 21st century, adding new functionality to certain aspects of the problem, and maybe a GUI. Certain types of problem are horrifically CPU intensive hence parallel operation would also be a significant benefit. The end benefits of such a project are basically a better way of running electricity grids for everyone - hopefully resulting in reduced costs, minimised land use, and so forth - there are all sorts of social benefits to be able to do this.

RPi probably isn't the ideal platform from a performance perspective for a system to analyse this type of problem, however I think it could be very useful for prototyping it. I'd also like to avoid any tangles with Java or Microsoft (as in either case, problems with pre-requisites inevitably arise!)

Most software of this nature is proprietary in origin and thus expensive to develop (and in most cases, very poor in terms of capability compared to our legacy code!)

On the other hand, research in this field is usually the subject of international standards and/or funded by public monies therefore the algorithms, etc. are readily available for use... To someone motivated enough to implement them.

I'd like to start fleshing out the bones of a design for a fully open sourced solution; but thought I'd also ask here if anyone has any experience of building large, complex packages like this in terms of getting them off the ground. Community contributions would, especially from the research community, be a really powerful way to get this thing to fly.

Look forward to hearing your advice.

[ghans]

Is this of any use to you ?
http://www.openelectrical.org/wiki/inde ... l_Software

ghans

[Binraider666]

Thanks for the suggestions, however the tools in that list are mainly to do with electrical properties rather than thermal ones.

Essentially any electrical network is limited by how hot you can allow it's components to get to, and having determined those limits you would then move onto a study of the electrical properties in one of those tools in that list. (or some propietary software - bleuch!)

To give a bit more background to the problem, many existing methods e.g. IEC 60287, Electra 143 and others work on a steady-state basis and therefore result in under-estimating capacity because you have to use pessimistic assumptions about ambient temperatures and so forth (e.g. summer ambient air temp is often assumed to be 30 deg C, 24 hours a day). This is clearly excessive. If you could load in a more representative load profile, you could safely get more performance out of the same equipment. Even better, with suitable condition monitoring, you could dynamically re-calculate performance (provided an appropriate safety margin and forecast method is built in!)

What I'd like to do is go one step further with these ideas and build a toolbox to study thermal properties in a transient manner; permitting the use of load profiles etc. There is software out there already to do this to a certain extent (but it is of highly dubious quality and often limited to specific installation environments). There is also a good deal of research into the topic available in the public domain. I'm not aware of an open source project that aims to implement these ideas, but I can see it being the kind of thing the power industry would be willing to invest in if enough of a "start" could be made to get it off the ground.

I think what I need to do is write up a brief on the basic problem and exactly what I'd want to achieve with such a project.

[ghans]

Perhaps have a look at this guide:
http://producingoss.com/en/

ghans