Smart Algorithm Based on the Optimization of SVR Technique by k-NNR Method for the Prognosis of the Open-Circuit and the Reversed Polarity Faults in a PV Generator - Archive ouverte HAL Accéder directement au contenu
Article Dans Une Revue International Review on Modelling and Simulations (IREMOS) Année : 2015

Smart Algorithm Based on the Optimization of SVR Technique by k-NNR Method for the Prognosis of the Open-Circuit and the Reversed Polarity Faults in a PV Generator

Résumé

– This paper deals with a new smart algorithm allowing open-circuit and reversed polarity faults prognosis in photovoltaic generators. Its contribution lies on the optimization of support vector regression (SVR) technique by a k-NN regression tool (k-NNR) for undetermined outputs. To testing the performance of the proposed algorithm, we used a significant data base containing the generator functioning history, and as indicators we selected variance, standard deviation, Confidence interval, absolute and relative errors. Nomenclature PV Photovoltaic SVM Support Vector Machines SVR Support Vector Regression k-NNR k-Nearest Neighbor Regression X SVR input vector Y SVR output vector f Linear function Ф Nonlinear mapping function w Weight vector e Squared loss function x Problem variable x * New problem variable α Lagrange multipliers N Number of classes m Number of index of minimum distances I / V Current / Voltage IPH Photocurrent
Fichier principal
Vignette du fichier
PWP IREMOS 2015 REZGUI.pdf (436.87 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-01154195 , version 1 (21-05-2015)

Identifiants

Citer

Wail Rezgui, Nadia Kinza Mouss, Leïla-Hayet Mouss, Mohamed Djamel Mouss, Yassine Amirat, et al.. Smart Algorithm Based on the Optimization of SVR Technique by k-NNR Method for the Prognosis of the Open-Circuit and the Reversed Polarity Faults in a PV Generator. International Review on Modelling and Simulations (IREMOS), 2015, 8 (1), pp.18-25. ⟨10.15866/iremos.v8i1.5197⟩. ⟨hal-01154195⟩
249 Consultations
355 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More