Fault Analysis Of Photovoltaic Grid Connected Inverter Based On Dummy Variable

With the proposal of sustainable development strategy,the harmonious development of human and nature is becoming more and more important.The application of new energy has become an important strategic plan of the country,in which solar energy as a typical clean energy is widely used,and the most common use method is photovoltaic grid connected power generation.For such widely used grid connected power generation system,its fault will cause serious consequences,so fault diagnosis plays an important role in its reliability research.In addition,for such a complex system,if the output parameters are analyzed considering only one kind of fault,it may lead to misdiagnosis and increase the difficulty of fault isolation or maintenance.In order to simplify the difficulty of fault diagnosis and reduce the demand for the number of maintenance personnel and experience,make the power generation system operate reliably and isolate faults quickly,this paper proposes a fault analysis strategy of photovoltaic grid connected system based on coordinate transformation and cluster analysis,so as to improve the comprehensiveness of the diagnosis system.In this paper,the traditional three-phase voltage source photovoltaic inverter is studied(1)The topology and control method of grid connected inverter system are analyzed,and the faults are classified according to their location,which are DC side fault,AC side fault and inverter main circuit fault.For several typical faults,the causes and evolution mechanism of faults are analyzed,and verified by MATLAB simulation model and rt-box hardware in the loop simulation model.(2)Through the mechanism analysis,it is pointed out that the characteristic changes of the output parameters of the system are the same when the inverter has different faults.And because the threshold judgment method is often used to detect faults in practice,this phenomenon will make fault classification more difficult.In order to solve the problem that the causes of the fault are different,but the parameters are the same,this paper proposes a virtual feature extraction method based on coordinate transformation,which takes DQ axis current as the basis of fault classification.(3)According to the corresponding relationship between the fault type and DQ axis current,a fault diagnosis method based on cluster analysis model is proposed.In view of the sparsity and nearest neighbor of the high-dimensional space of the clustering model,this paper normalizes and rearranges the data set to obtain the dimensionless clustering center.By calculating the Euclidean distance between the fault data set and the cluster center,the fault classification of inverter system can be realized.Finally,a complete model of inverter fault diagnosis is designed,and the practicability and accuracy of the proposed method are verified by experiments.