Modeling of photovoltaic array based on support vector machines under partial shaded conditions

Due to shadow caused by neighboring buildings and trees, photovoltaic (PV) array often works under shaded conditions. PV array is the key part of a PV generation system, and its electrical characteristics are susceptible to temperature, solar radiation intensity and load impedance and other factors. In order to extracting maximum power of the PV array, the electrical model of PV array is required. Firstly, a method to quickly draw the voltage-current characteristic curve of partial shaded PV array is presented based on its physical mechanism. Next, a model of PV array under partial shaded conditions is established by applying support vector machines (SVM). To obtain the optimum parameters in SVM, the method of cross validation and genetic algorithm (GA) are separately used. Experimental results demonstrate that the model based on SVM is in accordance with real PV array and has high ability of generation, so it can be an effective method to set up the model of PV array under shaded conditions.