Font Size: a A A

Research On Fault Diagnosis Of Inverter Power Supply Based On Gaussian Mixture Model

Posted on:2022-07-11Degree:MasterType:Thesis
Country:ChinaCandidate:W X ZhangFull Text:PDF
GTID:2512306494992139Subject:Electrical engineering
Abstract/Summary:PDF Full Text Request
Inverter power supplies have now been widely used in industrial production,such as new energy vehicles,UPS power supplies and so on.The high-frequency switching characteristics of the inverter power supply make device damage and performance degradation inevitable.Failure of the inverter power supply will result in poor performance of system equipment.If the fault cannot be located and responded in time,it will pose a serious threat to the safety of electrical equipment and production.Therefore,accurate and effective fault diagnosis of the inverter power supply is necessary.This paper proposes a fault diagnosis scheme based on Gaussian mixture model for the possible faults in the parallel resonant inverter power system.This scheme mainly takes the parallel resonant inverter power supply as the research object,and takes the voltage at the output terminal of the inverter power supply as the information source.The scheme is based on the crest factor of the voltage and the wavelet energy entropy to form a two-dimensional feature vector,and establishes and trains the system Gaussian mixture model to Realize system fault identification.Firstly,the working mode of the parallel resonant inverter power supply in the whole switching cycle is studied and a mathematical model of the parallel resonant inverter power supply system based on the circuit structure is established,and the fault types of the inverter power supply are classified and studied.It focuses on the fault characteristics in five failure modes: normal state,VT1 open-circuit state,VT2open-circuit state,VT2 short-circuit state and resonant resistance R open-circuit state.Secondly,it analyzes the fault feature extraction methods of the output voltage under different fault modes,focusing on the time domain characteristic crest factor of the voltage and the frequency domain characteristic wavelet energy entropy,and establishes the Gaussian model parameters combined with the EM algorithm.Finally,a system simulation model of the parallel resonant inverter power supply is built through the Simulink platform,and the Gaussian mixture model is used to diagnose the above five types of inverter power failure types.The experimental results show that the proposed model is based on the Gaussian mixture model.The parallel resonant inverter power supply fault diagnosis method has high diagnostic accuracy.
Keywords/Search Tags:Parallel resonance inverter, Crest factor, Wavelet energy entropy, Fault diagnosis, Gaussian mixing model
PDF Full Text Request
Related items