Skip to Main content Skip to Navigation

Modélisation des panneaux photovoltaïques et adaptation de la cyclostationnarité pour le diagnostic

Abstract : Photovoltaic (PV) systems can be operated in different locations. The exhibition (Wind, rain, snow, heat, lightning, shading, etc.) which cause their degradation over time and reduc their efficiency. Diagnosis is one of the interesting solutions to make PV panels work at their optimum power and in order to maximize the efficiency of PV conversion in order to reduce maintenance costs. In this thesis work, we are interested only in the diagnosis of PV generators. The aim of this thesis is to propose signal processing tools to detect and locate faults leading to a drop in efficiency. To carry out this work, we first make a stateof the art on the panels. Photovoltaic cells of microscopic appearance (cell) with macroscopic appearance (fields). To begin, we present the principle of operation of a photovoltaic cell as well as the various parameters affecting its performance. The combination of cells to create a photovoltaic panel, panels to create fields, are the studied. It is then shown how to connect these elements to a Load and network. At the same time, we describe the different types of defects and present an overview of the methods of their detection. A third part dealing with the modeling of defects shows how to build a database of electrical signals by simulation. Many electrical models are used in the literature to understand the functioning of PV panels. The Bishop model has been chosen in this study because it represents the current voltage characteristic (I-V) of the functioning of the PV cells in direct regime as well as in the inverse regime in the case where a cell is occulted. We explain how the different types of defects affect the I-V characteristic of solar panels. The electrical signals of the indicators (maximum power, short circuit current and open circuit voltage) are then calculated from the characteristic [V of the PV panel obtained for specific conditions (irradiance, temperature, mismatch defect, bypass diode fault). .). The originality of our work is to simulate the signals using real sunlight characteristics obtained by satellite measurements. We introduce the notion of seasonality into the characteristic I-V which then depends on time. We then analyze the first signals obtained by simulation. The time evolution of these indicators shows a CS aspect and the ATSA method is applied for these signals to have a good separation of the cyclic pattern and the random pattern of the time signals. The separation of these two components. To work on different CS commands. The cyclic autocorrelation function is applied to the random parts of the signals to work on the CS to order2 (CS2). In the fourth part, we show how to combine tools such as ATSA to diagnose the signals we simulated. We first present our choice of types of defects and severity used to build our database. Next, we describe and illustrate the various indicator in detail for a shading defect. A larger study is then carried out on all the simulated defects. In this study, the CS2 of the signals gave good results to make the diagnosis by comparing by the time and frequency analysis
Complete list of metadatas

Cited literature [103 references]  Display  Hide  Download
Contributor : Abes Star :  Contact
Submitted on : Wednesday, April 24, 2019 - 12:57:40 PM
Last modification on : Thursday, November 21, 2019 - 2:21:08 AM


Version validated by the jury (STAR)


  • HAL Id : tel-02108716, version 1



Mohammed Telidjane. Modélisation des panneaux photovoltaïques et adaptation de la cyclostationnarité pour le diagnostic. Traitement du signal et de l'image [eess.SP]. Université de Lyon, 2017. Français. ⟨NNT : 2017LYSES023⟩. ⟨tel-02108716⟩



Record views


Files downloads