A new offline mobile device for extracting I-V characteristic curve for photovoltaic modules

No Thumbnail Available
Nabulsi, Lina
Khushashi, Heba
Journal Title
Journal ISSN
Volume Title
This project introduces an improved device using hybrid learning machine system for I-V curve prediction. Two cascaded forward neural network were the basis for the proposed device. Factor X is the predicted output from the first cascaded neural network which was used as input for the second cascaded neural network in order to eliminate correlation between variables. The cas caded neural network was used to predict the actual current for a solar cell. In the learning process of the proposed device, an experimental dataset that involves records of ambient temperature, solar radiation, voltage and current for different photovoltaic modules is used. The device built to be general for all solar cell modules where the device input’s are ambient temperature, solar radiation, and datasheet specifications of photovoltaic module (short circuit current and open circuit voltage). Matlab is used to train, test and validate the proposed model, then the weights were used in equations in the python program, in order to download it on the Raspberry Pi to propose an easy and general device with a friendly GUI. The proposed device is very helpful and useful in predicting I-V curves for various photovoltaic modules.