Contribution to the optimization of a sliding mode maximum power point tracking for fast variations of solar irradiation

Résumé

ABSTRACT This thesis focuses on the contribution to the optimization of a sliding mode Maximum Power Point Tracking (MPPT) for the control of a photovoltaic system evolving in rapidly changing atmospheric conditions, by exploiting the stochastic particle swarm optimization algorithm to establish the stability and global convergence conditions of the control law. To do this, an optimal sliding surface selection study for photovoltaic (PV) systems is conducted. We used the stochastic particle swarm optimization algorithm to guarantee the global stability and fast convergence of the control; the reticence phenomenon inherent to sliding mode control is thus suppressed by using the smooth switching function sat. The results show that the control developed in this way is not affected by unstable and even severe atmospheric conditions. The effectiveness of the proposed system is tested and validated by simulation under Matlab/Simulink software under the standard test condition, extreme irradiance variation conditions, and one-second intervals. These numerical simulation results confirm the analytical predictions. Indeed, we observe a fast tracking of the maximum power point with a convergence speed that reaches 0.05 seconds in standard condition and 0.03 seconds in extreme irradiance variation conditions, without overshooting. The control also demonstrates its robustness and ability to track the maximum power point in the case of variations occurring within one second. In the second part of this thesis, two sliding surfaces are established. The reticence phenomenon is suppressed by means of the tanh switching function. This time, two more controllers are added: a fuzzy controller to generate voltage and current references optimally for low irradiance, and a PI controller to optimize the trajectory tracking. The stability of the entire system is guaranteed by the stochastic particle swarm optimization algorithm. The convergence speed obtained reaches 0.026 seconds with energy optimization achieved for low irradiations. The power gain reaches 70 W compared to the P&O at low irradiation. From these results, we observe that the stochastic particle swarm optimization algorithm combined with the sliding mode control law produces a very efficient hybrid control law under rapidly changing operating conditions, especially for low irradiation.

Citer ce document

Accès au document

Ce document est hébergé sur son dépôt institutionnel d'origine.

Auteur(s)

Statistiques

Consultations : 2

Téléchargements : 0