open source concentrator design software

[...]

Greetings all:

As far as I can tell, there is no open source ray tracing program suitable for solar concentrator design.  Most optical ray tracing programs are concerned with figuring aberation, not collection efficiency, and they assume optical components are sequential (like a series of lenses and mirrors).  Also, I found nothing that included the geometry of non-imaging or trough components.

So, I found the best open source, non-sequential ray tracing program in my language of choice (Python!) and began to make something suitable for solar concentrator design.  Its taken alot of time so far, and the project is still very alpha, but its looking good.  Fortunately, I started with a 3d rendered GUI and ray tracing math already working.  Almost all of the relevant trough geometry is now solved (for instance, parabolic troughs and CPCs, as well as semi-tubes and their involutes). More on the way (including a light source that mimics the motion of the sun and collector efficency statistics).

Check out a screen shot of a simple fresenl trough (similar to the OSE design) along with code for building that model at http://www.permafacture.org/tracer.html

Good thing we didn't build that!  A program like this to point out missed rays and blocked mirrors is an indespensible friend for a second opinion.

I'm at the point that having a couple systems to model would be helpful in figuring out how to calcuate statitics and how to let users interface with them.  Then it could be animated as the sun moves through the day/seasons and collection efficency kept track of.

Marcin or other practicing engineers out there- have you a specific design I can model?

The maths are ready to model that half cylinder secondary reflector that was proposed instead of a CPC.  I haven't modeled it yet, but I'll send out the results when I get to it!

later,

elliot

[Marcin Jakubowski]

Can we do a test on the design at:

http://openfarmtech.org/index.php?title=Solar_Concentrator_Technical_Drawings

with varying number of slats, varying the spacing, and height of receiver? The slats are to be 6" wide.

How do we interpret the ray tracing results?

Marcin

--
--------------------------------------------------
Marcin Jakubowski, Ph.D.
Open Source Ecology
http://openfarmtech.org
opensourceecology at gmail dot com
Skype: marcin_ose
--------------------------------------------------

Nobody said that building the world's first open source village would be easy.

-- Anonymous, 2009

A human being should be able to change a diaper, plan an invasion, butcher a hog, conn a ship, design a building, write a sonnet, balance accounts, build a wall, set a bone, comfort the dying, take orders, give orders, cooperate, act alone, solve equations, analyze a new problem, pitch manure, program a computer, cook a tasty meal, fight efficiently, die gallantly. Specialization is for insects.

-- Robert A. Heinlein

[Josef Davies-Coates]

I'm not really technical enough to take advantage of this, but its sounds wonderful and really important work.

Thank you!

Keep up the great work,

Warm regards,

Josef.

2009/9/14 ... <offonoff...@gmail.com>

--
Josef Davies-Coates
07974 88 88 95
http://uniteddiversity.com
Together We Have Everything

[...]

Vinay,

Your experiments confirm the pen and paper analysis I did when you mentioned light funnels in this group a long time ago.  The light cone ends up having lots of internal reflections and turns away lots of light.  The CPC is the ideal version of what you are thinking.  looking into a CPC from inside the collection angle, the walls are all black (or whatever color the target is).  From outside that angle the walls look shiney.

Looking into your light cone from where you expect light to be coming from (ie, your eye is the sun), you could see how much of the reflector surface is blacked by the reflection of the target.  That shows you what areas of the reflector are actively engaged when your eye is replaced by the sun.  That will tell you better than a simulation would because this design is very inefficent and not very sensitive to varations.

A simulation will tell you that the angle of a ray becomes rapidly more shallow with secondary and tertiary reflections, and with multiple reflections the ray is turned back and shot out of the cone.  (I dont have those original notes with me right now).  with sides at 45 degrees, all light is rejected from the cone.  at 60 degrees, light coming straight in can bounce once, but a second bounce and its on its way out again.

Thats with light coming straight in.  With light coming in at some angle (due to the sun having some size, and much more due to imperfect tracking), the performance is much worse.

If you make the cone too long, then the extra length is not doing anything because any light that hits the sides thee will bounce several times and escape.

So, you make the sides steeper, and rays can sustain more bounces before they are rejected, but the entrance becomes narrower.  adding length doesn't fix this.

http://books.google.com/books?id=MliJHWwTnVQC&lpg=PP1&dq=nonimaging%20optics&pg=PA49#v=onepage&q=&f=false

A 2d version would be easy to model, the 3d would require writing new geometry, which I wont do.  The 2d CPC is an ideal concentrator, but its 3d version is actually less than ideal. I bet the 3d light cone rejects more light than the 2d also.

I will model a CPC using two or three flat sections rather than an continuous curve and see how that goes.

[...]

Marcin,

I was wondering if you had a specific design.  Like, isn't there some sister group that has a design ready to build that OSE was going to replicate?  testing different design rationale will be more effective than spending days moving individual mirrors around in the simulator and settling upon what seems best (which may just be a local maximum in a muddled design).

But this software isn't really at the stage to be evaluating different designs yet.  In developing this capacity, I thought it would be helpful to have a purposefully designed system, and compare it to a haphazardly designed one.

Marcin asked:

How do we interpret the ray tracing results?

Thats what we are trying to figure out.  I'm thinking that defining an "entrance aperture" (in this case a rectangle defining the square area of the colector array placed a few inches above the mirrors) is the way to go.  Then the collector efficency is the rays that hit the absorber divided by the # of rays that entered the apperature.

concentration ratio would be the area of entrance aperature divided by the area of absorber.

Then concentration ration and collection efficency are the values we have to evaluate different designs.  We then run trough all the angles we expect the sun to be at and get a spreadsheet of performace throughout the days and seasons.

Are there other values we should be gleaning from the simulator?  Are there alternative definitions of these values we should look into implementing?

entrance aperature and efficency code should be done in a couple weeks.

later,

elliot