Conception of optimal MPPT controller based artificial intelligence of a Photovoltaic system
Description
The grid connected the photovoltaic system performance is strongly affected by the environmental conditions that undergoes, such as random atmospheric variations. This thesis work aims to improve the DC / DC converter and the PV inverter controllers' performance against brutal climatic fluctuations. Therefore, the first part of this thesis is devoted to the comparative study between the following maximum power point tracking algorithms (MPPT): (i) the algorithm of the Incremental of Conductance (IC), (ii) Fuzzy Logic (FL) and (iii) Particle Swarm Optimization algorithm (PSO). These algorithms are tested under various atmospheric conditions such as partial shading and evaluated in terms of efficiency, stability, speed and robustness. According to the simulation results, PSO is superior to IC and FL, especially during partial shading. The second part of this thesis deals with improving the efficiency of the DC / AC control system which includes an internal DC link voltage control loop (VDC) and an external control loop for direct current regulation and in quadrature (Id, Iq) provided by the PLL. Each of these two loops includes a PI controller whose gains are optimized using meta-heuristic techniques to improve the dynamic performance of the three-phase PV system connected to the network. Therefore, a comparative study is carried out for proposed meta-heuristics techniques such as: (i) whale optimization algorithm (WOA), (ii) gray wolf optimization algorithm (GWO) (iii) the Ant-Lion Optimization algorithm (ALO) and (iv) of the Moth-Flame Optimization algorithm (MFO). The results obtained, via MatlabTM-Simulink, reveal that the proposed WOA technique performance is relevant than the other studied techniques in terms of efficiency, robustness and stability which optimizes the PI controllers gains in order to obtain the best power factor and THD values. (author)
Abstract (French)
La performance du systeme photovoltaique connecte au reseau est fortement affectee par les conditions environnementales auxquelles est soumis tels que les variations atmospheriques aleatoires. Le travail de cette these vise a ameliorer les performances des controleurs du hacheur DC/DC et l'onduleur PV face aux changements climatiques brutaux. A cet effet, la premiere partie de cette these est consacree a l'etude comparative entre les algorithmes de de recherche de point de puissance maximale (MPPT) suivants: (i) l'algorithme de l'incrementale de conductance (IC), la logique floue (FL) et l'algorithme d'optimisation d'essaim de particules (PSO). Ces algorithmes sont testes sous diverses conditions atmospheriques telles que l'ombrage partiel et evaluees en termes d'efficacite, de stabilite, de rapidite et de robustesse. D'apres les resultats de la simulation, la PSO est meilleure par rapport a IC et FL, particulierement durant l'ombrage partiel. La seconde partie de cette these a pour but l'amelioration de l'efficacite du systeme de controle DC/AC qui comprend une boucle interne de controle de tension de liaison DC (VDC) et une boucle de controle externe pour la regulation des courants directs et en quadrature (Id, Iq) fournis par la PLL. Chacune de ces deux boucles comprend un controleur PI dont les gains sont optimises en utilisant des techniques meta-heuristiques afin d'ameliorer les performances dynamiques du systeme PV triphase connecte au reseau. Par consequent, une etude comparative est effectuee pour les techniques meta-heuristiques proposees telles que: (i) l'algorithme d'optimisation des baleines a bosse (WOA), (ii) l'algorithme d'optimisation des loups gris (GWO), (iii) l'algorithme d'optimisation des fourmilions (ALO) et (iv) de l'algorithme d'optimisation Heterocere-Flamme (MFO). Les resultats obtenus, via MatlabTM-Simulink, revelent que la technique WOA proposee est plus performante que les autres techniques etudiees en termes d'efficacite et de stabilite et qui permet d'optimiser les gains des controleurs PI afin d'obtenir les meilleures valeurs de facteur de puissance et de THD. (auteur)
Files
Restricted
The record is publicly accessible, but files are restricted to users with access.
Additional details
Additional titles
- Original title (French)
- Conception d'une commande MPPT optimale a base d'intelligence artificielle d'un systeme photovoltaique
Publishing Information
- Imprint Pagination
- 108 p.
- Report number
- FRNC-TH--13978
INIS
- Country of Publication
- France
- Country of Input or Organization
- France
- INIS RN
- 54032266
- Subject category
- S14: SOLAR ENERGY; S42: ENGINEERING;
- Resource subtype / Literary indicator
- Thesis
- Descriptors DEI
- ALGORITHMS; ARTIFICIAL INTELLIGENCE; COMPARATIVE EVALUATIONS; COMPUTERIZED SIMULATION; DC TO DC CONVERTERS; FUZZY LOGIC; INVERTERS; LOAD MANAGEMENT; OPTIMIZATION; PHOTOVOLTAIC POWER PLANTS; POWER DISTRIBUTION SYSTEMS; TRANSFORMERS
- Descriptors DEC
- ELECTRICAL EQUIPMENT; EQUIPMENT; EVALUATION; MANAGEMENT; MATHEMATICAL LOGIC; POWER PLANTS; SIMULATION; SOLAR POWER PLANTS
Optional Information
- Notes
- 120 refs.; Available from the INIS Liaison Officer for France, see the INIS website for current contact and E-mail addresses