Context¶

The author¶

My name is José Antonio Fernández-Fernández and I am from Motril, a coastal city in the South of Spain, in Andalucía. I was born in 1992 and I spent my first eighteen years in my hometown before move to Granada to start my universitary studies.

In Granada, I studied a Bachelor’s Degree on Civil Engineering from 2010 until 2015. There I fell in love with computer simulations applied to complex and real problems in engineering. In last years of my degree I learnt all I could in modelling dynamical problems in a computer as well as computational geometry and artificial intelligence.

In the last year, I have had the opportunity to participate in national (Spain) and international competitions and congresses on the application of computational techniques to the Renewable Energies. In this time, I have published pyny3d, and this web, and I am currently working in the transcription, and improvement, of the whole Solarized Project from proprietry code to Python. As soon as possible, all the tools I have developed for now will be accessible for the open source community.

If you are interested in my CV, you can check my LinkedIn. Please, feel free to contact for me for any doubt or suggestion you have or even if you are interested in count on my services.

Solarized Project¶

The Solarized Project is about generating a robust methodology to address the optimization process on Microsystems with Renewable Energies. The project has been published under the name of “PVT-Biomass-Battery systems optimized with evolutionary algorithms for small and highly shaded environments” in the University of Granada (Spain), and its authors are:

José Antonio Fernández-Fernández. Department of Civil Engineering

Enrique Alameda-Hernández. Department of Civil Engineering

Pedro González-Rodelas. Department of Applied Mathematics

The publication is accessible through this link.

Abstract¶

When the space where we have to build a PV Solar Plant is not enough to satisfy appropriately the user’s demand, a geometric problem arises in its design. To find the optimal location of the panels, the distances between them, their tilt, or even the number of panels itself, may result a complex work; much more in small spaces with obstacles near the modules. In these cases, the economically optimal solution(s) is strongly related to these geometric constraints, in addition to the more common variables considered in this type of problems: like electrical loads, or local temperature and cloudiness.

This work collects two proposals to mitigate the problem: a Photovoltaic thermal hybrid solar collectors (PVT)-Biomass-Battery system which significantly increases the generated energy per square meter; and a computational methodology, based on evolutionary algorithms, to calculate it. Additional power sources are welcomed, but only when the available resources (space) are fully exhausted

State of the art¶

When it comes to design microsystems based on renewable energies, there is no software that addresses, in the same calculation process, the following issues:

They will expect for you to detail every single part of the design, regardless whether you know that it is a good design or not.

Devices, connections and geometric parameters have to be defined before the simulation starts.

All of these are related to a previous combinatorial analysis to figure out the bests plausible combinations. This project use Artificial Intelligence techniques to supply autonomy and creativity to the process.

Solarized methodology¶

The methodology we are working on provides a change of approach to the design process because it has been conceived to address the problem as a searching for the bests in a complex field of possible solutions; while the majority of commercial software are focused on trying to be extremely precise solving a single design which, by definition, will not behave so predictably.

It is important to note that this tool has been created to orientate, to guide, the predesigning phase through combinatorial analysis. Never to substitute specialised tools like PVsyst, for example.

Finally, I want to mention a special part of the project that is an Evolutionary Algorithm to find the best connections (series-parallel) among all the pannels in a highly dense PV Plant with obstacles (a lot of shades).

Some figures from the work¶