This present paper deals with the parameter determination of solar cell under different values of irradiance and temperature by using an artificial neural network. This latter is trained by an optimization algorithm based on gradient descent. In this work we used two distinguished algorithms from different order of gradient descent: Levenberg-Marquardt and conjugate gradient. The use of these two algorithms is to conduct a comparative study on their performances. The results revealed that the Levenberg-Marquardt algorithm presents the best potential in providing accurate electrical parameters values compared to conjugate gradient algorithm. Moreover, the trends of electrical parameters according to irradiance and temperature show the effect of each of these two meteorological factors on the values of the intrinsic parameters of solar cell.
Show abstract
Full Text
