Artificial Intelligence for the characterization and analysis of photovoltaic systems

The operating conditions of a photovoltaic module determine the amount of electrical energy it generates. In order to characterize a photovoltaic module it is necessary to know its electrical behaviour (short-circuit current, open circuit voltage and the voltage and current at the maximum power point) under certain environmental variables (radiation and temperature).  

DaSCI researchers work collaboratively with the research group “Research and Development in Solar Energy” (TEP 101) of the University of Jaén, in the development of methods based on Artificial Intelligence for the characterization and optimization of photovoltaic systems with different technologies (flat panel, thin film and concentration) in different scenarios. Information from data acquisition systems from installations of different types and locations has been merged with data from geographic and meteorological information systems for the design of methods based on neural networks, fuzzy systems and descriptive rules. Work is currently underway on the combination of prediction and description methods for the configuration of domestic photovoltaic systems adapted to the curve and consumption characteristics in order to reduce costs and increase the use of clean energies.

This collaboration has made it possible to transfer research results in the field of intelligent construction. Proof of this is that the house Patio 2.12, with which the University of Jaen participated in the international competition Solar Decathlon Europe in 2012, incorporated knowledge extracted by some of the models developed achieving the first position in energy management and second globally.

Some significant results:

  • Models for the characterisation of concentration photovoltaic modules by means of competitive cooperative radial basis function network.
  • Differential evolution models to estimate the maximum power supplied by concentration photovoltaic modules in real outdoor conditions.
  • Evolutionary models for the detection of exceptions in concentration photovoltaic modules.


2008 – Present


Cristóbal Carmona, María José del Jesus, Pedro González, María Dolores Pérez, Antonio Rivera


Participation in the international competition Solar Decathlon Europe in 2012, with the Patio’s House 2.12