A New Photovoltaic Array Fault Diagnosis Method

With the vigorous development of the photovoltaic industry,the service life and safety issues of photovoltaic arrays are getting more and more attention.If the photovoltaic arrays which under fault cannot be diagnosed and processed in time,it may cause serious consequences such as fire and so on.Therefore,the research on the detection and diagnosis technology of photovoltaic array fault has a great significance.Based on the traditional method of fault diagnosis of photovoltaic array,a fault diagnosis method for photovoltaic array based on threshold and BP neural network is proposed in this paper.The main contents of the thesis are summarized as follows:Firstly,in view of the photovoltaic array's own structure,the photovoltaic arrays are connected in series,parallel,SP and TCT structures are studied.The advantages and disadvantages of each structure in the actual photovoltaic power generation system are analyzed,and the difficulty degree in the fault diagnosis of each structure is studied.For the major failures of photovoltaic array,the cause and harm of the short-circuit faults,shadow faults and the problem of lobule are studied,and the hot spot failures with the greatest damage to the photovoltaic arrays have been analyzed.Secondly,in view of the output characteristics of photovoltaic arrays under different environment and fault conditions,the photovoltaic arrays are modeled under the conditions of complex illumination,uniform illumination,short circuit fault and shadow fault.By studying the I-V curve and the P-V curve of the photovoltaic array,the output characteristics of the photovoltaic array under different environment and fault condition are analyzed.The output characteristics of the photovoltaic array under different occlusion conditions and the short-circuit fault are analyzed,the fault characteristics of the photovoltaic array are defined.Then,in view of the problems of the photovoltaic array fault detection and diagnosis,a method of PV array fault detection and diagnosis based on threshold is proposed.This method first classifies 9 common faults of photovoltaic array,by comparing the simulation curves under the faults and the normal conditions,the characteristic quantities of the nine kinds of faults are defined.Then according to the uncertainties in the production process of the PV modules and the measurement errors,calculating the threshold of photovoltaic array which need to be diagnosed under the standard condition of short circuit current and open circuit voltage and maximum power.Finally,the simulation of the PV array to be diagnosed is performed through the simulation software(the external conditions of the simulation are the actual PV plant environmental conditions),and the obtained simulation value is compared with the actual value of the collected PV power plant,the difference in value by comparing the maximum power and threshold,detecting the fault of photovoltaic array.If there is a fault,based on nine kinds of fault characteristic,the PV array is fault diagnosed by comparing the maximum number of power points and the thresholds and differences between the short-circuit current and the open-circuit voltage.Finally,in view of the photovoltaic array with the same fault characteristic quantity,according to the characteristics of the faults in the PV array whose output characteristics are not the same,a fault diagnosis algorithm based on BP neural network is proposed.The algorithm can be diagnosed based on the threshold method which faults with the same characteristic quantity.The steps are as follows: Firstly,establish the photovoltaic array fault identification model,and propose a fault diagnosis algorithm for photovoltaic array based on BP neural network.Then,the neural network is selected and trained by simulation software,and finally the fault with the same characteristic quantity is diagnosed.The two fault diagnosis algorithms mentioned in this paper can detect and diagnose nine kinds of faults.