Research And Design Of Photovoltaic Array Efficiency Based On Fault Diagnosis

With the increasingly serious of energy shortage and environmental pollution,photovoltaic power generation has attracted people's attention due to its unique advantages.However,compared with other renewable energy sources,photovoltaic power generation is more susceptible to failures owing to external conditions,resulting in a significant drop in power generation efficiency.In this paper,the fault detection and diagnosis of photovoltaic arrays are studied,a new type of SI-SP fault detection structure is proposed and the influence of faults on the output efficiency of photovoltaic arrays is analyzed.Besides,the diagnosis of photovoltaic array shading degree is realized based on neural networks and a configuration scheme is carried out according to the diagnostic results to optimize the efficiency of photovoltaic array.In the paper,first,an engineering application model of a solar cell under the basic structure and working principle is established in order to analyze the factors that affect its output characteristics.Besides,the output characteristics of different photovoltaic array structures are analyzed and SP photovoltaic array structure is determined as the subject of this study.In addition,a fault diagnosis structure of photovoltaic array based on neural network is proposed,which lays the foundation for the following research on fault detection and diagnosis.Second,a new type of SI-SP detection structure is proposed on the basis of the theory of sensor-crossed detection.When compared with traditional detection methods,the number of voltage sensors is greatly reduced.Next,the fault location process of SI-SP structure detection is analyzed in detail and a 6*4 array model is established by Matlab/Simulink to verify the effectiveness of SI-SP structure detection.Furthermore,the influence of different faults on the output power of the array is analyzed.Third,the weight ratio of the component parameters is analyzed by SPSS Modeler,thus the input parameters of the neural network model for fault diagnosis are determined.In the process of BP neural network modeling,the LM algorithm is used to optimize the neural network training process,which improves the accuracy of diagnosis.In order to shorten the diagnosis time,the extreme learning machine(ELM)algorithm is included.Additionally,the KELM algorithm is designed by adding kernel function to improve the training speed and diagnostic accuracy of the model.Based on the research of array efficiency under different shadow distributions,a configuration scheme based on KELM algorithm is proposed in order to optimize the efficiency of the array.Finally,a photovoltaic array test system is set up and various hardware modules of the test system are designed,on which experiments are performed.The result indicates that the KELM algorithm with high accuracy and good real-time performance can be used in the field of shade degree diagnosis.Besides,the efficiency of the photovoltaic array is improved according to the configuration scheme.As a result,the effectiveness and feasibility of the optimized configuration is verified.