Font Size: a A A

Research On Fault Detection Of Photovoltaic Array Based On Probabilistic Neural Networ

Posted on:2021-04-02Degree:MasterType:Thesis
Country:ChinaCandidate:W H SongFull Text:PDF
GTID:2392330605467810Subject:Engineering
Abstract/Summary:PDF Full Text Request
Photovoltaic power generation systems mainly include photovoltaic arrays,controllers,DC/AC converters,batteries and so on.The photovoltaic array is the source of the entire photovoltaic power generation system,and its role is extremely important.The photovoltaic array is located in a complex environment with a certain scale,which is difficult to detect when faults occur.Therefore,the effective detection of photovoltaic array fault is a prerequisite to ensure the normal operation of photovoltaic power generation systems.In this paper,the method of measuring the photovoltaic array's fault and the techniques for fault diagnosis are studied as follows:(1)The generation principle of photovoltaic cell is analyzed,and the mathematical model of photovoltaic cell is established.Using Matlab/Simulink to simulate the photovoltaic cell,analyze the output characteristics of the photovoltaic cell,simulate the open circuit,short circuit,abnormal aging,shadow four kinds of faults,select the open circuit voltage,short circuit current,voltage and current at the maximum power point as the fault characteristic signal.(2)The structure and fault type analysis of photovoltaic array are studied.The fault causes of the photovoltaic array during operation were analyzed.According to the selection of fault characteristic signals,the four faults of open circuit,short circuit,abnormal aging and shadow were simulated in the photovoltaic array simulation environment.Then collecting the fault sample data of this paper.(3)Investigating the fault location methods of SP structure and TCT structure of photovoltaic array.Aiming at the shortcoming of using too many sensors in fault location of TCT structured photovoltaic array,a new sensor embedding method is proposed.In this method,the TCT structured photovoltaic array is first equivalent to a series branch,embedded with a voltage sensor to detect the fault row position,and then embedded with a current sensor to detect the fault row position according to the number and resolution of parallel cells in each row,so as to determine the fault range.On the premise of ensuring the same accuracy and detection accuracy,this fault location method is more suitable for sensor embedding,reducing the number of sensors used.(4)Studying the fault mode diagnosis of photovoltaic arrays.Taking the SP structure photovoltaic array as the research object,the fault diagnosis of the photovoltaic array isperformed by using the probabilistic neural network.In this method,after the data of fault samples are normalized,the fault diagnosis model of probabilistic neural network is used to predict and identify the fault mode by Bayesian decision theory.The simulation results show that the accuracy of fault pattern recognition is more than 90%,which can complete the fault diagnosis of photovoltaic array.(5)An improved probabilistic neural network fault detection method based on whale optimization algorithm is proposed.In order to further improve the correct recognition rate of fault modes of the probabilistic neural network fault diagnosis model,this paper proposed to use the whale optimization algorithm to improve the probabilistic neural network fault diagnosis model.The smoothing factor of probabilistic neural network is optimized by whale optimization algorithm,and the most appropriate smoothing factor value is determined.And the improved fault diagnosis model is used to identify the fault pattern.The simulation results show that,compared with the original probabilistic neural network fault diagnosis model,the whale optimization algorithm improves the probabilistic neural network fault diagnosis model to recognize the fault mode with the accuracy of 100%,which can completely identify the fault mode of photovoltaic array,and this method can achieve effective detection of the fault mode.
Keywords/Search Tags:photovoltaic array, fault detection, whale optimization algorithm, probabilistic neural network, smooth factor
PDF Full Text Request
Related items