Research On Photovoltaic Fault Diagnosis Based On Feature Extraction

As an important foundation for social progress and economic development,the development and utilization of Energy play a significant role in the progress of civilization.Face to the lack of traditional fossil energy and the increasing demands of environment,photovoltaic power must become the strategic energy.However,solar energy is the intermittent energy and the working environment of photovoltaic system is complex.In order to protect the normal operation of photovoltaic system and reduce the life loss of PV array and power loss,the study of accurate and fast fault diagnosis is needed.It makes real-time targeted maintenance and state monitoring possible and greatly expanding the photovoltaic system intelligence and information technology.PV output characteristics in normal or fult condition is the basic for the study of PV fault diagnosis.The output characteristics of PV system are affected by environmental factors such as solar radiation and temperature,its non-stationarity makes the diagnosis of PV fault more difficult.This paper analysis the influence of weather,light,temperature and humidity,air quality on the output of the system,The classification of the operational faults of the components and their causes,as well as the external characteristics under different types,to establish the mapping relationship between the output characteristics and the fault types.So that it has the capability of simulating the non-linear transfer characteristic of the external behavior of the PV module.Which can ensure the simulation ability under the shading,and reduce the modeling difficulty and the running time on the basis of ensuring the precision.The model is established by electronic models of Matl ab,So the model itself can be scalable,and is conducive to large-scale photovoltaic power station analysis and research work.In this paper an on-line parameter identification method based on output characteristics is proposed and Simplifies parameter identification,improves the accuracy and reduces the fitting time,and greatly reduces the model parameter error caused by the deviation of operating conditions.The method can just need maximum power point current voltage,short circuit current,open circuit Voltage can be carried out,greatly reducing the cost of other testing equipment.The study of this paper is based on a large number of actual data,and has good generalization,which can provide reference for the researchers in the same field.