Fault Diagnosis And Parameter Identification Of Three-phase Inverter In Multi-energy Storage System

The 13th Five-Year Plan(2016-2020)first incorporates energy storage into national planning.Power electronics technology is the use of power electronic devices for efficient power conversion and control,and is the interface technology between energy storage and the clients.For long,safe,reliable and efficient disaster prevention power demand,wind power,photovoltaic and other new energy demand,multi-energy storage systems need to be highly reliable and with the ability of fault diagnosis and fault prediction for uninterrupted work and reducing repair fees.Fault diagnosis and parameter identification technology as a more effective way to improve multi-energy storage system reliability has become a hot focus of the research.Studies have shown that semiconductor devices and passive components are prone to failure in power converters especially inverters in a multi-energy storage systems.Therefore,the in-depth study of fault diagnosis and parameter identification of the three-phase inverter in multi-energy storage system is of important theoretical significance and application value.In this paper,the specific contents include the following four aspects:(1)Analytical model is proposed applicable to voltage fault diagnosis of three-phase inverter circuit.And two fault diagnosis methods based on bridge arm pole voltage and bridge arm line voltage are proposed to improve the speed of diagnosis and diagnostic reliability for all load situations.Fault diagnosis method based on voltage is performed quickly with the use of additional sensors and hardware circuits.But diagnostic time is more important for multi-energy storage systems application.This paper reports investigation conducted on two diagnostic methods based on switching voltage pattern of IGBTs open-circuit faults in voltage-source inverters.With an additional simple circuit,these two diagnostic methods detected and identified single and multiple open-circuit faults of the inverter IGBTs effectively.The comparison of time for diagnosis and anti-interference features between these two methods are presented.The diagnostic time for the method based on bridge arm pole voltage was less than that of the method based on bridge arm line voltage and was 1/2 of the fundamental period in worst case.(2)A novel diagnosis method based on error voltage for IGBTs open-circuit faults in dual-loop controlled voltage-source inverters is proposed.The phase voltage changes when open-circuit faults occur,and then there are error voltages of each phase which can be calculated directly by control algorithm with phase voltages and their corresponding flexible references both available from the main control system.Therefore,the proposed diagnosis method requires no additional sensors or electrical circuits to detect open-circuit faults,further with the features of fast diagnostic time,being suitable for various load conditions,strong robustness against false alarms and being easily inserted to the existing control algorithms as a subroutine.(3)Ant colony system is a novel simulated evolutionary algorithm,which provides a new method for complicated combinatorial optimization problems such as traveling salesman problem,sequential ordering problem,generalized assignment problem,etc.In this paper,an evolutionary approach based on ant colony system is applied to system-level fault diagnosis.It only needs a small amount of communication data with the host controller,while avoids circuit-level fault diagnosis which usually takes up hardware and software resources.Firstly,characteristics of each power module in the multi-modular system are collected up to the host controller and a circular graph is formed.Then ant colony algorithm(ACA)is applied to the circular graph combining with symmetric comparison model.Finally,experimental results from simulation study and theoretical analysis demonstrate the effectiveness of the ACA-based diagnosis approach for multi-modular system,making it a viable addition to system-level fault diagnosis.(4)A class of magnetic powder core material inductors in high power applications has been studied.Based on this,the inductor model and the failure criteria of inductor are proposed.The failure prediction of multi-energy storage system has been achieved by the online least squares method.Actual inductor can be modeled as an ideal inductor in series with equivalent series resistance.Then,the inductor inductance value of equivalent circuit model can be selected as fault feature parameter.When the inductance value decreases more than 20%,the inductor is invalid.Then a fault diagnostic parameter identification method can be effectively applied to inductor based on the three-phase inverter circuit model.Meanwhile,for the three-phase inverter circuit inductance parameter changing with the load current,an improved parameter identification algorithm is proposed for online identification of inductor parameters.