Search

CN-120800492-B - IGBT module performance test method based on multi-parameter collaborative analysis

CN120800492BCN 120800492 BCN120800492 BCN 120800492BCN-120800492-B

Abstract

The invention discloses a performance test method of an IGBT module based on multi-parameter collaborative analysis, which comprises the steps of S1, collecting test data of the IGBT module, wherein the test data comprise an electrical parameter set, a thermal parameter set and a vibration parameter set, S2, executing collaborative feature extraction operation to generate a joint feature vector based on parameter association relation among the electrical parameter set, the thermal parameter set and the vibration parameter set, S3, processing the joint feature vector by utilizing a feature mapping algorithm, outputting a performance analysis result comprising a switching loss rate, a junction temperature rising slope and vibration power spectral density, and S4, comparing the performance analysis result with a parameter threshold value in a preset performance standard library, judging the performance grade of the IGBT module and generating a test report. The comprehensive performance evaluation and objective grade judgment of the IGBT module under the complex working condition are realized, and the testing precision and reliability prediction capability are remarkably improved.

Inventors

  • ZHAO MENGFAN

Assignees

  • 深圳市芯威能半导体有限公司

Dates

Publication Date
20260508
Application Date
20250813

Claims (8)

  1. 1. The IGBT module performance test method based on the multi-parameter collaborative analysis is characterized by comprising the following steps of: S1, collecting test data of an IGBT module, wherein the test data comprises an electrical parameter set, a thermal parameter set and a vibration parameter set; S2, based on the parameter association relation among the electrical parameter set, the thermal parameter set and the vibration parameter set, performing collaborative feature extraction operation to generate a joint feature vector, wherein the method comprises the following steps: S21, analyzing the electrical parameter set to obtain an electrical feature subset, analyzing the thermal parameter set to obtain a thermal feature subset, and analyzing the vibration parameter set to obtain a vibration feature subset; S22, calculating correlation coefficients of the electrical feature subset and the thermal feature subset to construct a first correlation matrix, calculating mutual information entropy of the electrical feature subset and the vibration feature subset to construct a second correlation matrix, and calculating covariance of the thermal feature subset and the vibration feature subset to construct a third correlation matrix; s23, combining the first association matrix, the second association matrix and the third association matrix to generate a joint feature vector, wherein the S23 comprises the following steps: s231, respectively executing eigenvalue decomposition on the first incidence matrix, the second incidence matrix and the third incidence matrix; S232, selecting a feature vector corresponding to the maximum feature value from the feature value decomposition result of each matrix as a dominant feature component; S233, merging the dominant feature components according to preset weights to generate a joint feature vector; s3, processing the joint feature vector by utilizing a feature mapping algorithm, and outputting a performance analysis result comprising a switching loss rate, a junction temperature rising slope and a vibration power spectral density; And S4, comparing the performance analysis result with a parameter threshold value in a preset performance standard library, judging the performance grade of the IGBT module and generating a test report.
  2. 2. The IGBT module performance test method based on multi-parameter collaborative analysis according to claim 1, wherein step S1 includes: s11, starting an IGBT module performance test system, applying an electrothermal load to the IGBT module based on preset voltage and current parameters, and synchronously collecting an original data stream; s12, performing signal separation processing on the original data stream, and extracting electric original data, thermal original data and vibration original data; And S13, linearly compensating the electrical original data by calling a pre-stored electrical calibration coefficient matrix, non-linearly correcting the thermal original data by calling a thermal calibration coefficient matrix, and performing frequency domain calibration on the vibration original data by calling a vibration calibration coefficient matrix to generate a calibrated parameter set.
  3. 3. The IGBT module performance test method based on multi-parameter collaborative analysis according to claim 1, wherein step S4 includes: s41, extracting a switching loss rate, a junction temperature rising slope and a vibration power spectral density from a performance analysis result; S42, comparing the switching loss rate, the junction temperature rising slope and the vibration power spectral density with corresponding parameter threshold intervals in a preset performance standard library; s43, marking the switching loss rate, the junction temperature rising slope and the vibration power spectral density as qualified grades when all the switching loss rate, the junction temperature rising slope and the vibration power spectral density are in the corresponding parameter threshold value intervals, and otherwise, executing a fault diagnosis program.
  4. 4. The IGBT module performance test method based on multi-parameter collaborative analysis according to claim 2, wherein step S13 includes: s131, performing matrix operation on the electric calibration coefficient matrix and electric original data to realize linear compensation; S132, performing nonlinear function operation on the thermal calibration coefficient matrix and the thermal original data to realize temperature drift correction; and S133, carrying out convolution operation on the vibration calibration coefficient matrix and the frequency spectrum of the vibration original data to realize frequency domain amplitude calibration.
  5. 5. The IGBT module performance test method based on multi-parameter collaborative analysis according to claim 1, wherein step S21 includes: S211, extracting rising time, turn-off overshoot voltage and turn-on voltage drop from an electrical parameter set; s212, extracting junction temperature gradient distribution and thermal resistance curve slope from the thermal parameter set; s213, extracting the resonance frequency, the root mean square value of the vibration acceleration and the frequency spectrum energy distribution from the vibration parameter set.
  6. 6. The IGBT module performance test method based on multi-parameter collaborative analysis according to claim 2, wherein step S11 further comprises: s111, collecting vibration original data; S112, calculating a vibration acceleration root mean square value of vibration original data; s113, judging whether the root mean square value of the vibration acceleration exceeds a preset threshold value; and S114, when the root mean square value of the vibration acceleration exceeds a preset threshold value, the sampling rate of the electrical parameter is improved.
  7. 7. The IGBT module performance test method based on multi-parameter collaborative analysis according to claim 4, wherein the fault diagnosis procedure in step S43 includes: S431, analyzing high-frequency components of the second incidence matrix, and marking the high-frequency components as lead bonding falling faults when the mutation quantity of the high-frequency components exceeds a preset critical value; s432, analyzing the low-frequency energy duty ratio of the third incidence matrix, and marking the third incidence matrix as a layered fault of the heat dissipation substrate when the low-frequency energy duty ratio exceeds a preset threshold value; s433, generating a fault report comprising fault type codes and positioning coordinates.
  8. 8. The IGBT module performance test method based on multi-parameter collaborative analysis according to claim 1, wherein step S22 further comprises: s221, analyzing frequency domain energy distribution characteristics of the vibration characteristic subset; S222, determining an optimized sampling frequency according to the frequency domain energy distribution characteristics; s223, applying the optimized sampling frequency to the subsequent collection of the electrical feature subset; and S224, updating the electrical characteristic subset according to the optimized sampling frequency.

Description

IGBT module performance test method based on multi-parameter collaborative analysis Technical Field The invention relates to the technical field of performance test, in particular to an IGBT module performance test method based on multi-parameter collaborative analysis. Background Currently, in the industrial field, performance testing of IGBT (insulated gate bipolar transistor) modules generally adopts a single parameter evaluation method, that is, isolated testing is performed on electrical characteristics, thermal characteristics or mechanical characteristics, respectively. Although the test method is simple and convenient to operate, the interaction effect among parameters cannot be reflected, so that obvious deviation exists between a test result and an actual application scene. For example, conventional methods typically measure switching characteristics in a constant temperature environment, neglecting the effect of temperature fluctuations on electrical performance in actual operation, or just static parameters and neglecting performance in dynamic conditions. In addition, the prior art mostly relies on manual experience to judge test results, and lacks a standardized evaluation system, so that IGBT modules of different batches and different manufacturers are difficult to objectively compare, and the difficulty and risk of designing and selecting a power electronic system are increased. Limitations of the traditional testing method are particularly prominent in high-reliability application scenes, such as a new energy automobile driving system, a high-speed railway traction system and the like, and comprehensive performance requirements of the fields on IGBT modules are extremely high. Due to the lack of multi-parameter collaborative analysis capability, the performance boundary of the IGBT under the complex working condition is difficult to accurately evaluate in the prior art, and the reliability level of the IGBT in practical application cannot be effectively predicted. Meanwhile, the parameter extraction and analysis in the traditional test process are mostly manually operated, the efficiency is low, subjective deviation exists, and the quick screening and quality control on the production line are not facilitated. The technical defects seriously restrict the development of the IGBT module in the high-end application field, and a comprehensive performance test method capable of realizing multi-parameter collaborative analysis and objective evaluation is needed. Disclosure of Invention The invention provides an IGBT module performance test method based on multi-parameter collaborative analysis, which aims to solve the problems in the prior art. In order to achieve the above purpose, the present invention provides the following technical solutions: A multi-parameter collaborative analysis-based IGBT module performance test method comprises the following steps: S1, collecting test data of an IGBT module, wherein the test data comprises an electrical parameter set, a thermal parameter set and a vibration parameter set; S2, based on the parameter association relation of the electrical parameter set, the thermal parameter set and the vibration parameter set, performing cooperative feature extraction operation to generate a joint feature vector; s3, processing the joint feature vector by utilizing a feature mapping algorithm, and outputting a performance analysis result comprising a switching loss rate, a junction temperature rising slope and a vibration power spectral density; And S4, comparing the performance analysis result with a parameter threshold value in a preset performance standard library, judging the performance grade of the IGBT module and generating a test report. Wherein, the step S1 comprises the following steps: s11, starting an IGBT module performance test system, applying an electrothermal load to the IGBT module based on preset voltage and current parameters, and synchronously collecting an original data stream; s12, performing signal separation processing on the original data stream, and extracting electric original data, thermal original data and vibration original data; And S13, linearly compensating the electrical original data by calling a pre-stored electrical calibration coefficient matrix, non-linearly correcting the thermal original data by calling a thermal calibration coefficient matrix, and performing frequency domain calibration on the vibration original data by calling a vibration calibration coefficient matrix to generate a calibrated parameter set. Wherein, the step S2 comprises: S21, analyzing the electrical parameter set to obtain an electrical feature subset, analyzing the thermal parameter set to obtain a thermal feature subset, and analyzing the vibration parameter set to obtain a vibration feature subset; S22, calculating correlation coefficients of the electrical feature subset and the thermal feature subset to construct a first correlation matrix, calculating