Noninvasive Health Condition Monitoring And Fault Warning For Key Devices Of Buck Converter

Switching power supply plays an important role in various industrial fields.The performance of its internal key DC-DC converter circuit directly affects its output quality,thereby affecting the subsequent circuit of switching power supply.When the internal device performance of the converter degrades,the output quality will decrease,which affects the normal operation of the later stage equipment.When the internal device fails,it will lead to abnormal operation of the rear equipment,and serious failure occurs.Therefore,it is of great significance to evaluate the health of DC-DC converter,grasp its health status in real time,and make early warning and related countermeasures in time.In this thesis,the buck DC-DC converter is taken as the research object.Starting from the design of non-invasive system,the health status monitoring and fault warning of the electrolytic capacitor and switch tube with the highest failure rate in the converter are carried out.The main contents of this thesis are as follows :Firstly,the basic principle of Buck converter is introduced.Then,the electrolytic capacitor and switch tube with the highest internal failure rate are studied,and their basic structure,failure mechanism and degradation index are analyzed.Finally,the capacitance C of the electrolytic capacitor and the equivalent series resistance ESR,and the abnormal working phenomenon of short-term continuous conduction or switching of the switch tube are determined as the degradation or fault indexes of the two,which provides the basis for the following analysis.For the health condition monitoring of electrolytic capacitor,according to the ripple voltage of the output capacitor,the calculation expressions of C and ESR under different working modes are deduced by the parameter relationship in the circuit.In the formula,the duty cycle D value,the ripple voltage value at 0 time and DT time are obtained.The above parameters are obtained by using the corresponding relationship between the output ripple voltage and the driving signal.The corresponding calculation formulas under different modes are substituted to obtain the C and ESR values,and the simulation model is built to verify the above method.Two methods are used in the study of switch tube fault early warning.The first method uses an output voltage differential discriminant method to discriminate abnormal faults by analyzing the corresponding relationship between abnormal faults and output voltage differential signals,and then gives a warning.The second method decomposes the output voltage under different states by wavelet packet to obtain the voltage characteristics,and then trains the neural network to distinguish different states to achieve the purpose of abnormal fault identification.Finally,the non-invasive health condition monitoring and fault warning system for the key components of Buck converter is designed and implemented.An experimental platform is built to simulate the degradation and abnormal faults of the two key components,and the feasibility of the proposed method is verified by experiments.The test results show that the monitoring and warning system has good performance.