With the increasingly serious energy crisis and environmental issues worldwide, the research and application of new energy has aroused widespread concern because of its wide distribution, huge storage and environment-friendly. The photovoltaic array is the source of the whole PV system. In order to enable the system work in high efficiency, the fist task is to be familiar with the electrical characteristics of the photovoltaic array. PV array are often in partial shade in application. And the electrical characteristic of PV module can be acquired only under standard test condition which is from the makers.It is very important to build the model of PV array in partial shadow, with respect to realize the maximum power point tracking and ensure the inverter to work efficiently.Traditionally, the configuration of PV array is fixed. When shadow occurs, the output power is reduced greatly, because of the influence between modules.In order to reduce the impact of the shadow, Reconfiguration of PV array is established automatically in real time according to their own work status, which is an important method to improve the output power.In this paper, the model and configuration optimization of PV array under partial shade are studied. The main works are as follows:(1) Analysis of the working and hot spots principles of the PV array are elaborated in detail. External characteristics of the PV array and the effect of the number of bypass diodes on the maximum output power in a PV module are simulated through the use of PV system software. The models of PV array under partial shade conditions are obtained by MATLAB simulation. The changes in curves of IV and PV are observed and analyzed according to the following factors that are irradiancy intensity, cell temperature, partial shadows and the numbers of bypass diode. Finally, methods to improve the maximum output power are summarized.(2) Detailed comparative analysis of existing PV array modeling is executed. Based on the experimental platform, sample data and parameters optimization using cross validation and genetic algorithm separately, the model of PV array based on SVM under partial shadow is established. Through the comparison between the method of SVM and RBF neural network, it is found that higher accuracy but the longer computation time of the model based on Support Vector Machines.(3) Control strategies for optimizing the configuration of photovoltaic array are detailed analyzed and compared. Finally, suggestions during the optimization of configuration are presented. Based on the principle of irradiation equalization, the MATLAB simulation in order to realize the optimization of configuration is implemented. The relationship between the sum of irradiation of each PV module in the same line and the maximum output power of PV array is proposed, when the array is in the TCT configuration.(4) Hardware circuit which includes switch matrix, sampling circuit, detection circuit and driver circuit and software program which is based on CCS integrated development environment for the control system optimizing configuration of PV array are both designed. Difficulties and related solutions are summarized during the process of programming. The feasibility of the configuration optimization algorithm is verified based on the hardware and software platforms. Finally, the idea of the optimization of PV configuration based on model predictive control is proposed. |