I am interested in implementing a heliostat field module in the "Tracer", a python based ray-tracing program I am currently developing and using for receiver optimisation problems. There are different methods used in commercial softwares and in the scientific litterature and it is quite uneasy to reference them all or get enough details.
The interest here is to get a good radiative flux distribution match at the receiver location, whatever the field model.
Salut,
Je crois que tu es français. Pour la modélisation de flux concentré j'utilise la environnement de développement EDStaR. Est-ce que tu connais? Tu trouveras des informations sur le profil de Jean-Jacques Bézian.
I have used finite element methods to solve radiative heat transfer problem in an enclosure which has multiple surfaces. I used MATLAB. I do not have much idea about your configuration, however we can write the program with flexibility to address many configurations. In the program I developed, main factor decided the computational time was the number of elements.
Can you be a more precise when you mean a reasonable price!
I am focusing on multi-physics design optimisation. As a consequence I need to be simulating a large number of different configurations and cannot afford to spend the time simulating a full heliostat field for each configuration.
My question might have been a bit confusing. I am looking for the methods that can be used to simplify and approximate heliostat fields. As we already have a rather complex optimisation work going on on our Python program, I would rather stick to it and implement a new module to it.
I Dont know if it will be useful or not, but if computation cost is your concern,
You might want to look into convolution techniques, normally it is used to optimize heliostat field, it will give less accurate but reasonably good result.(refer Collado's papers and Helios documentation(available online) (Helios is based on convolution))
this technique is more complex to implement, but once implemented it drastically reduce computation load.
disclaimer: this is what i inferred from papers i haven't implemented this technique my self yet, although it seems logical to think this way.
I suggest you can try sequential rays trace mode. Make best use of prioir information about sunshape. So you can decrease the complexity of the algorithm.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Software