Processor in the loop implementation for nonlinear control based on artificial intelligence MPPT used in photovoltaic systems

Abstract

Photovoltaic (PV) systems are increasingly significant contributors to global energy supply, currently accounting for 4% of total needs. However, extracting maximum power from these systems presents a complex challenge due to the nonlinear nature of their electrical characteristics. Factors like temperature fluctuations and varying irradiance levels can significantly impact power output. This work contributes a novel design, testing, and validation methodology for robust Maximum Power Point Tracking (MPPT) control in PV systems. The methodology leverages two newly developed artificial intelligence (AI)-based nonlinear control strategies, Adaptive Neuro-Fuzzy Inference System based on Terminal Sliding Mode Control (ANFIS-TSMC), and Backstepping combined with ANFIS (ANFIS-BS). A validated Model-Based Design (MBD) approach and a two-stage testing sequence provide a comprehensive framework for developing and verifying embedded software for MPPT algorithms. Evaluation results demonstrate the proposed system’s ability to control reference power precisely under diverse atmospheric conditions. Successful integration with a 32-bit ARM microcontroller also paves the way for real-world implementation. The embedded software consistently exhibited high compliance and performance in meeting MPPT requirements across various test scenarios.