CN113987900A - IGBT service life prediction method based on extended Kalman particle filter - Google Patents
IGBT service life prediction method based on extended Kalman particle filter Download PDFInfo
- Publication number
- CN113987900A CN113987900A CN202111258761.3A CN202111258761A CN113987900A CN 113987900 A CN113987900 A CN 113987900A CN 202111258761 A CN202111258761 A CN 202111258761A CN 113987900 A CN113987900 A CN 113987900A
- Authority
- CN
- China
- Prior art keywords
- igbt
- particle
- extended kalman
- data
- life
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 239000002245 particle Substances 0.000 title claims abstract description 49
- 238000000034 method Methods 0.000 title claims abstract description 37
- 238000007781 pre-processing Methods 0.000 claims abstract description 4
- 239000011159 matrix material Substances 0.000 claims description 12
- 230000032683 aging Effects 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 9
- 238000005259 measurement Methods 0.000 claims description 6
- 238000002474 experimental method Methods 0.000 claims description 4
- 238000012545 processing Methods 0.000 claims description 4
- 238000013507 mapping Methods 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 3
- 230000007704 transition Effects 0.000 claims description 3
- 238000001914 filtration Methods 0.000 abstract description 5
- 238000012549 training Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/25—Design optimisation, verification or simulation using particle-based methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2111/00—Details relating to CAD techniques
- G06F2111/08—Probabilistic or stochastic CAD
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/04—Ageing analysis or optimisation against ageing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
本发明公开了一种基于扩展卡尔曼粒子滤波的IGBT寿命预测方法，其包括以下步骤：S1、获取IGBT器件的历史数据；S2、数据预处理并建立IGBT的寿命模型；S3、利用扩展卡尔曼粒子滤波算法建立预测方程；S4、预测IGBT的剩余使用寿命。本发明对粒子滤波算法进行了改进，并且用来预测IGBT的剩余使用寿命，这种方法提高了寿命预测的准确性。
The invention discloses an IGBT life prediction method based on extended Kalman particle filtering, which comprises the following steps: S1, acquiring historical data of IGBT devices; S2, data preprocessing and establishing a life model of IGBT; S3, using extended Kalman The particle filter algorithm establishes a prediction equation; S4, predicts the remaining service life of the IGBT. The invention improves the particle filter algorithm and is used to predict the remaining service life of the IGBT, and the method improves the accuracy of the life prediction.
Description
Technical Field
The invention relates to the field of semiconductors, in particular to an IGBT service life prediction method based on extended Kalman particle filtering.
Background
In recent years, power electronic equipment is rapidly developed and applied to many fields, but due to the complexity of the working environment, the core device, namely an insulated gate bipolar transistor (IGBT for short), of the power electronic equipment often fails. Therefore, the reliability assessment of the IGBT operation condition is of great significance, if an effective state assessment and early warning scheme can be provided, the occurrence of catastrophic accidents can be avoided, and therefore the research on the service life prediction direction of the IGBT device is of great practical significance.
Disclosure of Invention
The invention provides an IGBT service life prediction method based on extended Kalman particle filtering. The method utilizes the improved particle filter algorithm to predict the residual service life of the IGBT, and compared with the prior art, the method improves the accuracy of prediction.
In order to achieve the purpose of the invention, the invention adopts the technical scheme that:
the IGBT service life prediction method based on the extended Kalman particle filter comprises the following steps:
s1, acquiring historical data of the IGBT device;
s2, preprocessing data and establishing a life model of the IGBT;
s3, establishing a prediction equation by using an extended Kalman particle filter algorithm;
and S4, predicting the residual service life of the IGBT.
Further, the specific method of step S1 is:
and carrying out a power cycle accelerated aging experiment on the IGBT device, acquiring the conduction voltage drop between a collector and an emitter of the IGBT device in the aging process, and establishing a historical data set.
Further, the specific method of step S2 includes the following steps:
s2-1, eliminating error points generated by measurement by using a Rhein criterion;
s2-2, taking the average value of the data in the single cycle as the characteristic value of the cycle;
s2-3, mapping the data samples into the same range according to a uniform standard through scale transformation;
s2-4, performing curve fitting on the processed data, establishing a primary life model, and determining the initial values of the model parameters.
Further, the specific method of step S3 includes the following steps:
s3-1, establishing a state equation f and an observation equation h according to the life model;
s3-2, setting the number of particles as N, and initializing a particle set, namely extracting an initial state from the prior distribution P;
s3-3, calculating the Jacobian of the state transition matrix for each particleComparable matrix F(i)Noise driving matrix G(i)And observe the noise drive matrix U(i)(ii) a Wherein i corresponds to the ith particle;
s3-4, updating the particle set by using an extended Kalman filter algorithm in the importance sampling stage, and specifically calculating as follows:
wherein k corresponds to the kth data in the data set, i.e. represents the kth cycle; pre represents a corresponding predicted value; k represents a Kalman gain; r is the noise variance; z is a system observation value, namely data in a historical data set;
obtaining the mean value of the sample by the above calculation
Sum covariance
And then updating the particle set by using the particle set, wherein the specific calculation is as follows:
s3-5, calculating the weight value of each particle
The specific calculation is as follows:
wherein p represents prior probability and q represents posterior probability;
s3-6, carrying out normalization processing on the weight values, and then calculating the weighted average value of the particle set as output, thereby carrying out state updating, namely updating the parameters in the life model and determining the prediction equation.
Further, the specific method of step S4 is:
and setting an early warning threshold value, substituting the early warning threshold value into a prediction equation, and calculating to obtain the remaining service life of the IGBT device.
The invention has the beneficial effects that: according to the invention, the residual service life under the current working state can be obtained only by acquiring historical data of the IGBT device in the aging process and then processing and calculating by using the proposed method, and the more the data volume is, the more accurate the prediction result is. Compared with the prior art, the method greatly improves the accuracy of prediction, and is simple and high in practicability.
Drawings
FIG. 1 is a schematic flow diagram of the process;
fig. 2 is a diagram comparing the method with the basic particle filter.
Detailed Description
The following description of the embodiments of the present invention is provided to facilitate the understanding of the present invention by those skilled in the art, but it should be understood that the present invention is not limited to the scope of the embodiments, and it will be apparent to those skilled in the art that various changes may be made without departing from the spirit and scope of the invention as defined and defined in the appended claims, and all matters produced by the invention using the inventive concept are protected.
As shown in fig. 1, the IGBT life prediction method based on the extended kalman particle filter includes the following steps:
s1, acquiring historical data of the IGBT device;
s2, preprocessing data and establishing a life model of the IGBT;
s3, establishing a prediction equation by using an extended Kalman particle filter algorithm;
and S4, predicting the residual service life of the IGBT.
The specific method of step S1 is:
and carrying out a power cycle accelerated aging experiment on the IGBT device, acquiring the conduction voltage drop between a collector and an emitter of the IGBT device in the aging process, and establishing a historical data set.
The specific method of step S2 includes the steps of:
s2-1, eliminating error points generated by measurement by using a Rhein criterion;
s2-2, taking the average value of the data in the single cycle as the characteristic value of the cycle;
s2-3, mapping the data samples into the same range according to a uniform standard through scale transformation;
s2-4, performing curve fitting on the processed data, establishing a primary life model, and determining the initial values of the model parameters.
The specific method of step S3 includes the steps of:
s3-1, establishing a state equation f and an observation equation h according to the life model;
s3-2, setting the number of particles as N, and initializing a particle set, namely extracting an initial state from the prior distribution P;
s3-3, calculating Jacobian matrix F of state transition matrix of each particle(i)Noise driving matrix G(i)And observe the noise drive matrix U(i)(ii) a Wherein i corresponds to the ith particle;
s3-4, updating the particle set by using an extended Kalman filter algorithm in the importance sampling stage, and specifically calculating as follows:
wherein k corresponds to the kth data in the data set, i.e. represents the kth cycle; pre represents a corresponding predicted value; k represents a Kalman gain; r is the noise variance; z is a system observation value, namely data in a historical data set;
obtaining the mean value of the sample by the above calculation
Sum covariance
And then updating the particle set by using the particle set, wherein the specific calculation is as follows:
s3-5, calculating the weight value of each particle
The specific calculation is as follows:
wherein p represents prior probability and q represents posterior probability;
s3-6, carrying out normalization processing on the weight values, and then calculating the weighted average value of the particle set as output, thereby carrying out state updating, namely updating the parameters in the life model and determining the prediction equation.
The specific method of step S4 is:
and setting an early warning threshold value, substituting the early warning threshold value into a prediction equation, and calculating to obtain the remaining service life of the IGBT device.
In the specific implementation process, when the service life of a certain IGBT device is predicted, a power cycle accelerated aging experiment needs to be performed on the IGBT device first to obtain the conduction voltage drop between a collector and an emitter of the IGBT device in the aging process as a data set of experimental measurement, as shown in fig. 2, then a part of data is selected as training data, and another part of data is selected as verification data, in this embodiment, 10 data points are predicted, so that the last 10 points of the experimental measurement data are used as verification data, and the rest of data are used as training data, and a prediction equation is obtained by training and calculating the method provided by the present invention, in the embodiment, the conduction voltage drop of the next cycle is only predicted each time, as shown in fig. 2, it can be seen from the experimental results that the prediction result of the method is closer to the real experimental measurement data, compared with the basic particle filtering algorithm, the method provided by the invention has better accuracy, so that the method can predict the residual service life more accurately, and the accuracy of the service life prediction is improved.
In conclusion, the invention provides an IGBT service life prediction method based on extended Kalman particle filtering. The method improves the particle filter algorithm, and improves the accuracy of the particle filter algorithm for service life prediction. Compared with the prior art, the method greatly improves the accuracy of prediction, is simple, has strong practicability, and improves the reliability of the device in practical application.
Claims (5)
1. The IGBT service life prediction method based on the extended Kalman particle filter is characterized by comprising the following steps:
s1, acquiring historical data of the IGBT device;
s2, preprocessing data and establishing a life model of the IGBT;
s3, establishing a prediction equation by using an extended Kalman particle filter algorithm;
and S4, predicting the residual service life of the IGBT.
2. The IGBT life prediction method based on extended Kalman particle filter according to claim 1, characterized in that the specific method of step S1 is:
and carrying out a power cycle accelerated aging experiment on the IGBT device, acquiring the conduction voltage drop between a collector and an emitter of the IGBT device in the aging process, and establishing a historical data set.
3. The IGBT life prediction method based on extended Kalman particle filter according to claim 1, characterized in that the specific steps of step S2 are:
s2-1, eliminating error points generated by measurement by using a Rhein criterion;
s2-2, taking the average value of the data in the single cycle as the characteristic value of the cycle;
s2-3, mapping the data samples into the same range according to a uniform standard through scale transformation;
s2-4, performing curve fitting on the processed data, establishing a primary life model, and determining the initial values of the model parameters.
4. The IGBT life prediction method based on extended Kalman particle filter according to claim 1, characterized in that the specific steps of step S3 are:
s3-1, establishing a state equation f and an observation equation h according to the life model;
s3-2, setting the number of particles as N, and initializing a particle set, namely extracting an initial state from the prior distribution P;
s3-3, calculating Jacobian matrix F of state transition matrix of each particle(i)Noise driving matrix G(i)And observe the noise drive matrix U(i)(ii) a Wherein i corresponds to the ith particle;
s3-4, updating the particle set by using an extended Kalman filter algorithm in the importance sampling stage, and specifically calculating as follows:
wherein k corresponds to the kth data in the data set, i.e. represents the kth cycle; pre represents a corresponding predicted value; k represents a Kalman gain; r is the noise variance; z is a system observation value, namely data in a historical data set;
obtaining the mean value of the sample by the above calculation
Sum covariance
And then updating the particle set by using the particle set, wherein the specific calculation is as follows:
s3-5, calculating the weight value of each particle
The specific calculation is as follows:
wherein p represents prior probability and q represents posterior probability;
s3-6, carrying out normalization processing on the weight values, and then calculating the weighted average value of the particle set as output, thereby carrying out state updating, namely updating the parameters in the life model and determining the prediction equation.
5. The IGBT life prediction method based on extended Kalman particle filter according to claim 1, characterized in that the specific method of step S4 is:
and setting an early warning threshold value, substituting the early warning threshold value into a prediction equation, and calculating to obtain the remaining service life of the IGBT device.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111258761.3A CN113987900A (en) | 2021-10-26 | 2021-10-26 | IGBT service life prediction method based on extended Kalman particle filter |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111258761.3A CN113987900A (en) | 2021-10-26 | 2021-10-26 | IGBT service life prediction method based on extended Kalman particle filter |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113987900A true CN113987900A (en) | 2022-01-28 |
Family
ID=79742903
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111258761.3A Pending CN113987900A (en) | 2021-10-26 | 2021-10-26 | IGBT service life prediction method based on extended Kalman particle filter |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113987900A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117764271A (en) * | 2023-11-24 | 2024-03-26 | 华南理工大学 | Power grid dynamic state estimation method and system based on extended Kalman particle filtering |
| CN118625087A (en) * | 2024-06-17 | 2024-09-10 | 广东电网有限责任公司 | IGBT device remaining life prediction method and system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106845866A (en) * | 2017-02-27 | 2017-06-13 | 四川大学 | Equipment method for predicting residual useful life based on improved particle filter algorithm |
| CN111783243A (en) * | 2020-06-18 | 2020-10-16 | 东南大学 | A filter algorithm-based fatigue crack growth life prediction method for metal structures |
| CN111832125A (en) * | 2020-06-18 | 2020-10-27 | 东南大学 | A method for predicting fatigue crack growth life of metal structures of metallurgical cranes |
| CN113393064A (en) * | 2021-08-17 | 2021-09-14 | 中南大学 | Method for predicting service life of cadmium-nickel storage battery of motor train unit and terminal equipment |
-
2021
- 2021-10-26 CN CN202111258761.3A patent/CN113987900A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106845866A (en) * | 2017-02-27 | 2017-06-13 | 四川大学 | Equipment method for predicting residual useful life based on improved particle filter algorithm |
| CN111783243A (en) * | 2020-06-18 | 2020-10-16 | 东南大学 | A filter algorithm-based fatigue crack growth life prediction method for metal structures |
| CN111832125A (en) * | 2020-06-18 | 2020-10-27 | 东南大学 | A method for predicting fatigue crack growth life of metal structures of metallurgical cranes |
| CN113393064A (en) * | 2021-08-17 | 2021-09-14 | 中南大学 | Method for predicting service life of cadmium-nickel storage battery of motor train unit and terminal equipment |
Non-Patent Citations (1)
| Title |
|---|
| 郭稳: "功率MOSFET剩余使用寿命预测方法及热疲劳建模研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》, no. 135, 15 March 2021 (2021-03-15), pages 135 - 102 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117764271A (en) * | 2023-11-24 | 2024-03-26 | 华南理工大学 | Power grid dynamic state estimation method and system based on extended Kalman particle filtering |
| CN118625087A (en) * | 2024-06-17 | 2024-09-10 | 广东电网有限责任公司 | IGBT device remaining life prediction method and system |
| CN118625087B (en) * | 2024-06-17 | 2025-06-27 | 广东电网有限责任公司 | IGBT device remaining life prediction method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2016107246A1 (en) | Wavelet noise reduction and relevance vector machine-based method for predicting remaining life of lithium battery | |
| CN111190088B (en) | Method for extracting characteristic parameters of IGBT (insulated Gate Bipolar transistor) performance degradation | |
| CN112415414A (en) | Method for predicting remaining service life of lithium ion battery | |
| CN113987900A (en) | IGBT service life prediction method based on extended Kalman particle filter | |
| CN113761751A (en) | Lithium ion battery residual life prediction method and system based on temperature acceleration factor | |
| US12050455B2 (en) | State prediction apparatus and semiconductor manufacturing apparatus | |
| CN113759225B (en) | IGBT residual life prediction and state evaluation realization method | |
| CN113987932A (en) | A MOSFET lifetime prediction method based on time series model | |
| CN112288147B (en) | A Method of BP-Adaboost Strong Predictor to Predict the Insulation State of Generator Stator | |
| CN113391244A (en) | VMD-based transformer switching-on vibration signal characteristic frequency calculation method | |
| CN116740900B (en) | SVM-based power construction early warning method and system | |
| CN112000081A (en) | Fault monitoring method and system based on multi-block information extraction and Mahalanobis distance | |
| CN117129872A (en) | An adaptive estimation method for lithium-ion battery health status | |
| CN113379252A (en) | Bayesian network high-voltage switch cabinet health system evaluation method based on multi-scale arrangement entropy | |
| CN115659768A (en) | Method and device for predicting remaining service life of IGBT | |
| CN105868544B (en) | The method of three parameter Weibull distributions processing flashover voltage analysis insulating material properties | |
| CN115270410A (en) | Prediction method and system of GIS particle weak discharge development trend based on ARMA model | |
| CN115728672A (en) | Circuit system health prediction method based on low-frequency noise and deep learning | |
| CN110532629A (en) | A kind of method for predicting residual useful life of multi-modal degenerative process | |
| CN111079270A (en) | Bearing residual life prediction method based on binary mixed random process | |
| CN104499001B (en) | Feature based subspace optimizes the aluminium cell condition diagnostic method of relative matrix | |
| CN113987899A (en) | An IGBT life prediction method based on unscented Kalman particle filter | |
| CN118246599A (en) | A risk assessment method, system, medium and equipment for gas-insulated fully enclosed combined electrical equipment | |
| Sedaghat et al. | Modeling and sensitivity analysis of coronavirus disease (COVID-19) outbreak prediction | |
| CN117151012A (en) | A method for predicting the reliability of Schottky barrier diodes |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20220128 |