Development of Maximum Power Extraction Algorithms for PV system With Non-Uniform Solar Irradiances

Abstract

This thesis addresses the problem of extraction of maximum power from PV arrays subjected to non-uniform solar irradiances e.g partial shading. In the past, a number of maximum power point tracking algorithms (MPPTs) such as Perturb & Observe, Hill climbing, Incremental Conductance, etc. have been proposed. These are extensively used for obtaining maximum power from a PV module to maximize power yield from PV systems under uniform solar irradiance. However, these techniques have not considered partial shading conditions and the stochastic nature of solar insolation. In the event of non-uniform solar insolation, a number multiple maximum power points (MPPs) appear in the power-voltage characteristic of the PV module. In the present thesis, the stochastic nature of the solar insolation is considered to obtain the global MPP of a PV module with a focus on developing global optimization techniques for MPPT that would handle the multiple MPPs. Thus, the thesis will address the above problem by developing a number of global MPPT algorithms.
In this thesis, an extensive review on MPPT algorithms for both uniform and non-uniform insolation levels is presented. Subsequently, an analysis with respect to their merits, demerits and applications have been provided in order to design new MPPTs to achieve higher MPPT efficiency under non-uniform solar irradiances.
Firstly, PV modules are modelled with and without bypass diodes for handling Partial shading conditions (PSCs). Then, a new Ring pattern (RP) configuration has been proposed which is compared with different existing configurations such as Series parallel (SP), Total cross tied(TCT) and Bridge linked(BL) configurations on the basis of maximum power and fill factor.
As described earlier, under non-uniform irradiances the MPPT problem boil down to determining the global MPP. Thus, the MPPT problem can be cast as a global optimization problem. It may be noted that evolutionary computing approaches are extensively used for obtaining global optimum solutions. One of the most recent evolutionary optimization techniques called grey wolf optimization technique has gained enormous popularity as an efficient global optimization approach. In view of this, Grey wolf optimization is employed to design a global MPPT such that maximum power from PV modules can be extracted which will work under partial shading conditions. Its performance has been compared with two existing MPPTs namely P&O and IPSO based MPPT methods. From the obtained simulation and experimental results, it was found that the GWO based MPPT exhibits superior MPPT performance as compared to both P&O and IPSO MPPTs on the basis of dynamic response, faster convergence to GP and higher tracking efficiency.
Further, in order to scale down the search space of GWO which helps to speed up for achieving convergence towards the GP, a fusion of GWO-MPPT with P&O MPPT for obtaining maximum power from a PV system with different possible patterns is developed. An experimental setup of 600W solar simulator is used in the laboratory having characteristics of generating partial shading situation. Firstly, the developed algorithms were implemented for a PV system using MATLAB/SIMULINK. Subsequently, the aforesaid experimental setup is used to implement the proposed global MPPT algorithms. From the obtained simulation and experimental results it is observed that the Hybrid-MPPT converges to the GP with least time enabling highest possible maximum power from the solar PV system.
In this thesis, analytical modeling of PV modules for handling non-uniform irradiances is pursued as well as a new RP configuration of PV modules is developed to achieve maximum power and fill factor. In order to extract maximum power from PV panels subjected to non-uniform solar irradiances, two new MPPT algorithms are developed namely Grey wolf optimization based MPPT (GWO-MPPT) and GWO assisted PO (GWO-PO).