Investigations on optimization algorithms for estimating the equivalent circuit parameters in solar photovoltaic modules

Abstract

Over the past several decades, photovoltaic (PV) systems for solar newlinepower harvesting often see accelerated progress. This increasing trend of PV newlinesystems offers new opportunities for PV cells/module manufacturers, who are newlinecharged with making their products more robust and of excellent quality and newlinepose challenges.The precise PV model identification technique is becoming newlinenecessary to significantly reduce PV energy loss due to the simulation newlinenecessities and the latest implementations of model-based controllers. The newlinesimulation, assessment, and harvesting of maximum energy of the PV system newlinerequire accurate and fast parameter estimation for solar cell/module models. newlineThe parameters must be accurately calculated from the characteristic of solar newlinecells and/or a collection of data set for simulations and PV system control. newlineNevertheless, it might not be easy to access experimental data or newlinereliable characteristic data of a PV panel. The PV system manufacturer newlinetypically provides vital data on short-circuit, open-circuit, and maximum newlineoutput levels. The mathematical model of the PV module is built on its newlinecurrent (I)-voltage (V) curve, which is a non-linear relationship. Therefore, no newlinecomplete information on the PV module parameters is provided by the newlinemanufacturers. This leads to a non-linear PV model with numerous newlineunidentified parameters. The problem of defining the model parameters of the newlinePV cells is considered as the problem of estimating the PV cell parameter, newlineand finally, it is considered as an optimization problem. newlineIn different engineering fields, such as modeling, detection, newlineoptimization, forecasting, predicting, and managing complicated processes, newlineoptimization methods or algorithms are commonly utilized. Recently, many newlineoptimization algorithms are reported by different researchers for various newlineengineering applications. However, none of the algorithms is suitable for all newlinethe engineering problems as per the No-Free-Lunch theory principle. newlineTherefore, in this thesis, two improved versions of single-objective algorithms newlineare introd