Jump to Main Content
An enhanced dynamic modeling of PV module using Levenberg-Marquardt algorithm
- Blaifi, Sid-ali, Moulahoum, Samir, Taghezouit, Bilal, Saim, Abdelhakim
- Renewable energy 2019 v.135 pp. 745-760
- algorithms, diodes, dynamic models, empirical models, equations, irradiation, renewable energy sources, solar collectors, temperature, weather
- An improved dynamic modeling of PV cell/modules based on automatic parameters extraction is proposed in this paper. For the sake of clarity, three models are compared in this study including, Single Diode (SDM), Double Diode (DDM) and the empirical model developed by Sandia National Laboratory (SANDIA). The use of nominal parameters or the values given by manufacturer in both SDM and DDM diode saturation current I0 and photo-generation current Iph equations can engender a significant error depending on the operating conditions and the consumed lifetime. Hence, these values can be handled as model parameters, and can be adjusted using automatic parameters extraction algorithms. Moreover, parameters based on static extraction methods (with fixed irradiation and temperature) namely, Rs, Rsh and n do not give satisfactory results under variable irradiation and temperature, which involve the use of a dynamic adjustment method to improve these parameters. In this way, static parameters extraction using genetic algorithm (GA) is proposed as a first stage for both SDM and DDM. After that, a dynamic parameters extraction based on the Levenberg-Marquardt algorithm (LMA) has been employed in the purpose to adjust some nominal parameters provided by the literature and the manufacturer, and those given by the static method. The idea consists of considering the PV module and the MPPT as a single system with dynamic inputs (irradiation and temperature) and output (Impp, Vmpp and Pmpp) to minimize the error between the measured and the simulated outputs. The validity of the proposed approach is compared with dynamic LMA models, nominal parameters based models, and the models based on static GA extracted parameters under of different weather conditions and out-door measurements. The improved models show promising results in terms of agreement with real data.