Research On Condition Monitoring Method Of IGBT And Capacitor In Power Converter

Power converters are widely used in power generation,substations,electric vehicles,and other fields with the development of new energy technology.As the key components of power converters,Insulated Gate Bipolar Transistors(IGBTs)and capacitors play the role of energy conversion and storage,which failure will inevitably affect the normal operation of the converter.Hence,the condition monitoring of IGBTs and capacitors can detect faults in time and make corresponding operation and maintenance plans,which is very important to ensure the safe and reliable operation of converters.This paper takes the condition monitoring of IGBTs and capacitors in power converters as the research object.The research documents in this thesis are addressed based on methods of bond wires failure detection,chip failure detection of multichip module,state detection of multichip IGBT module under high power conditions,and capacity monitoring.The main studies and achievements are as follows.(1)Bond wire failure is the direct factor that induces open circuit failure of IGBT module.Aiming at the problem of bond wire failure monitoring in IGBT module,a bond wire failure monitoring method based on measuring thermal resistance reduction is proposed,which overcomes the limitations that traditional methods using zerotemperature point measurement and cannot be applied under normal operating conditions.Firstly,based on the principle of heat flow during IGBT operation,the influence of bond wire failure on heat flow conduction is analyzed in depth.Secondly,the reason for the reduction of measured thermal resistance is analyzed according to the mechanism of heat flow change caused by bond wires failure.Finally,a monitoring method based on the thermal resistance reduction of heat sink is proposed to detect the failure of bond wire,which can monitor the status of wires in normal working conditions.The effectiveness of proposed method is verified by experiments.(2)Chip failure is the main cause of multichip module failure.Aiming at the problem of chip failure monitoring in multichip module,a chip failure detection method based on the change of turn-off delay is proposed,which reduces the requirement of complete gate signal measurement and reduces the difficulty of monitoring.Firstly,the characteristics of turn-off process in multichip module are analyzed,and a theoretical model of turn-off process duration considering the number of chips is constructed.Secondly,by studying the influence of chip failure on turn-off delay,a chip failure monitoring method based on the reduction of turn-off delay is proposed.Finally,the proposed method is verified under different working conditions.(3)Aiming at the problem that it is difficult for measuring equipment to meet the requirements of high sampling bandwidth and high precision required by electrical parameters measurement under high power conditions,which leads to the difficulty in realizing the condition monitoring of the IGBT module,a neural network prediction model based on case temperature changes is proposed.Firstly,the influence of high power conditions on the traditional detection methods is analyzed,and the case temperature is indicated as the characteristic parameter of IGBT aging monitoring based on the influence of IGBT aging on heat flow change.Secondly,a neural network model is constructed to reflect IGBT aging in multichip module through the case temperature based on the mapping relationship between case temperature and aging of each chip.To solve the problem of identifying the continuously changing state in the aging process of IGBT module,a state monitoring result processing method based on demand preference is proposed.Finally,the effectiveness of the proposed method is verified on a test platform based on a mission profile emulating platform for high power Modular multilevel converter(MMC)system,and the state monitoring of IGBT module under high-power operating conditions is realized.(4)The capacitance is an important indicator to characterize the health status of capacitors.Aiming at the problem of capacitance monitoring,a non-invasive monitoring method based on the Haar wavelet transformation extraction switching sequence is proposed,which eliminates the need for switching sequence and reduces the complexity of monitoring system.Firstly,the operation principle of capacitor in MMC system is analyzed,and the relationship between switching sequence and capacitor voltage is analyzed.Secondly,a calculation method of extracting switching sequence using Haar wavelet transform is proposed based on the relationship between switching sequence and capacitor voltage,and the capacitance value is calculated through the extracted switching sequence.Finally,the accuracy and effectiveness of the proposed method are verified on the MMC simulation model and experimental platform,and the influence of high power conditions on the results of capacitance monitoring is analyzed.(5)Based on the key components condition monitoring methods proposed in this paper,a set of methods for designing and developing a state monitoring system for key components of power converters are formed.The main functional structure of power converter condition detection system is analyzed.The design and implementation methods of hardware and software systems are proposed.Then,a condition monitoring system under real working conditions is demonstrated through an engineering example.The research results of this thesis can be used for the design of power converter monitoring system and provide theoretical support and design reference to guarantee the stable operation of converter.It is of great academic value and practical significance to improve the operation reliability of power converters and reduce the maintenance cost.