Optimized Parameter Estimation, Array Configuration and MPPT Control of Standalone Photovoltaic System

World energy requirements are increasing at a faster pace, and at the same time fossil fuels are depleting at an alarming rate. As a result, for energy sustainability, renewable energy is playing a crucial role in today‟s life. Among all the renewable energy sources, like wind, tidal, biofuel, hydrogen energy, and fuel cell, etc., photovoltaic (PV) based solar energy appears to be the most promising alternative source of energy, because it is abundantly available and produces noise-less, maintenance-free, clean energy. Although PV energy is a popular source of energy, it faces lots of challenges because of its high dependency on environmental condition and poor conversion efficiency. Further, the characteristics of PV module are nonlinear in nature, and has single maximum power point (MPP) which changes with the environmental condition. To increase the efficiency of the PV system, it must be operated at its MPP which is realizable with the help of a suitable MPP tracking (MPPT) control scheme. To design an appropriate MPPT control structure, a reliable and an accurate PV model is necessary for analysis and simulation. Accurate modelling of PV module is only possible if its model parameters are known apriori. However, all model parameters are not available in the manufacturer data sheet. Therefore for accurate modelling of PV module, precise estimation of the module parameters is needed. This thesis proposes different parameter estimation techniques such as Nonlinear Least Square (NLS) method, Nelder-Mead (NM) optimization method and Grey Wolf Optimization (GWO) method to estimate PV module parameters accurately with less statistical errors. The estimated model parameters are validated considering both current-voltage (I-V) and power-voltage (P-V) characteristics for various PV modules at different environmental conditions and found matching accurately with the experimental characteristics. Maximum power point tracking is an essential part of any PV system. Therefore, the thesis is subsequently focused on development of different MPPT techniques. Since dc-dc converter is a fundamental unit of any MPPT design, selection of an appropriate converter for different load condition is also discussed in this thesis. Among different MPPT methods, perturb and observe (P&O) method is the most popular method for its easy implementation and simple structure. But the conventional P&O method is sluggish in nature due to its fixed perturb amplitude. To overcome this, an adaptive P&O algorithm is proposed in this thesis, which has faster dynamics with less steady state oscillations compared to the conventional P&O method. The effectiveness of the proposed adaptive P&O method is verified through Matlab simulation and experimentation. As the conventional MPPT algorithms fails to track maximum power from the PV panel under rapidly changing weather conditions, an artificial neural network (ANN) based MPPT is implemented under very fast changing weather condition to track the maximum power. The result shows that ANN MPPT can track the maximum power more accurately with much reduced steady-state oscillations with a faster rate of convergence as compared to P&O method. For large power applications, PV array is used instead of PV modules. There is a great possibility of occurrence of partial shading condition (PSC) on PV array which causes module characteristic mismatch, affecting the performance of PV system. Mathematical equations have been derived to calculate the mismatch loss for series connected PV string. To reduce the severity of mismatch loss, two hybrid array configurations named as total cross tied with bridge link (TCT+BL) and total cross tied with honey-comb (TCT+HC) are proposed. It is observed these hybrid array configurations are better options to mitigate the mismatch loss and to minimize the occurrence of local peaks. Under PSC, occurrence of multiple peaks in P-V characteristics makes the MPPT problem more complex to detect and track global maximum power point (GMPP) from all local and global power peaks. This thesis proposes a new meta-heuristic swarm based optimization algorithm named as whale optimization algorithm (WOA) to track GMPP of PV system under PSC. The performance of WOA is compared with other two popular MPPT methods: perturb and observe (P&O) and particle swarm optimization (PSO) in terms of their transient and steady-state behaviour. It is observed that the proposed WOA can track GMPP with faster rate with less steady state oscillations as compared to P&O and PSO algorithm