Data-driven green energy extraction: Machine learning-based MPPT control with efficient fault detection method for the hybrid PV-TEG system

Abstract

The hybrid photovoltaic-thermoelectric generation system (PVTEG) gives two-fold benefits; firstly, it efficiently utilizes the available solar energy as it converts both solar irradiance and solar thermal energy into electricity, secondly, it enhances PV efficiency by reducing the PV module surface temperature. However, the efficiency of the hybrid PVTEG system is usually low because both PV and TEG are not highly efficient devices. Under changing environmental conditions, a well-designed maximum power point tracking (MPPT) controller can enhance the generation efficiency by 10%–15%. Therefore, this research explores the evolutionary Neural Network based MPPT control technique for the hybrid PVTEG systems. The snake optimizer optimally tuned weight and biases of the Multilayer perceptron neural network (MLPNN), which provides fast real-time global maxima (GM) tracking. Furthermore, to enhance the robustness of MPPT control, PID gains are tuned using the SO algorithm. Use of the snake optimizer based PID (SOPID) controller with the snake optimizer based neural network (SOANN) results in stable, accurate and fast MPPT under varying environmental conditions. The effectiveness of the proposed SOANN MPPT controller is validated by comparing it with PSOANN, RSANN and GWOANN. SOANN based MPPT controller provides very fast real-time global maxima (GM) tracking with negligible power oscillations. The SOANN controller extract optimal power with an average efficiency of 99.928% and tracking time of less than 5ms. Furthermore, an intelligent data driven based fault detection algorithm is proposed, which do not require any temperature or irradiance sensors reducing cost of the system. Comparative, simulation, quantitative, and statistical results second superior performance of SOANN controller in terms of efficiency, tracking time, stability and fault detection capability under various practical condition.