سال انتشار: ۱۳۸۲

محل انتشار: هجدهمین کنفرانس بین المللی برق

تعداد صفحات: ۱۱

نویسنده(ها):

Sarvi – Department of Electrical Engineering Iran University of Science & Technology Tehran, Iran, 16844.
Masoum – Department of Electrical Engineering Iran University of Science & Technology Tehran, Iran, 16844.
Amani – Department of Electrical Engineering Iran University of Science & Technology Tehran, Iran, 16844.

چکیده:

Solar arrays have nonlinear insolation and temperature dependent characteristics. This paper proposes a neural network based method for
simulation and modeling of solar arrays considering impacts of insolation and temperature variations. Simulations and measurements are performed for MLP and RBF neural networks and the advantages and limitations of each technique are presented. Inputs of these neural network models are solar array voltage ( sa V ), short circuit current ( sc I ) and temperature (T) while solar array current ( sa I ) is selected as the output. Inorder to validate the proposed neural network model and investigate it’s accuracy, an experimental setup consisting of a personal computer,an interface board, a thermal sensor and one OFFC silicon solar panel is used. Theoretical and experimental results are compared andnalyzed.