Data-driven optimal adaptive MPPT techniques for grid-connected photovoltaic systems

• Published in: Ain Shams Engineering Journal
• Main Authors: Ahmed H. EL-Ebiary, Mostafa I. Marei, Mohamed Mokhtar
• Format: Article
• Language: English
• Published: Elsevier 2025-03-01
• Subjects: MPPT; Adaptive PI controllers; Machine learning; Optimization techniques
• Online Access: http://www.sciencedirect.com/science/article/pii/S2090447925000590

The constant fluctuations in the maximum power obtained from Photovoltaic (PV) systems are due to variations of temperature and irradiance. Maximum Power Point Tracking (MPPT) techniques are used to guarantee the best possible efficiency and performance for the PV systems. In this paper, an Incremental Conductance (IC) MPPT technique based on adaptive controllers is proposed. This paper presents two different types of adaptive PI controllers, including optimized Fractional Order Adaptive PI (FOAPI), and Single Perceptron Adaptive PI (SP-API). The IC technique along with the adaptive controllers ensure accurate extraction of maximum power under sudden changes and different weather conditions. Moreover, machine learning is utilized to initialize the duty cycle of PV system converter, where different regression models are compared and the model with the least Root Mean square error (RMSE) is exploited. Three case studies are carried out to compare and validate the performance of the suggested adaptive MPPT controllers.