IGBT life prediction method, device, storage medium and program product

Abstract

The present invention belongs to the field of electronic power technology, and specifically relates to an IGBT life prediction method, device, storage medium and program product. The present invention obtains the Miller platform duration and collector current of the IGBT, and obtains the junction temperature corresponding to the obtained Miller platform duration according to the relationship between the Miller platform duration and the junction temperature of the IGBT device under the collector current; wherein, the relationship between the Miller platform duration and the junction temperature of the IGBT device is derived based on the relationship between the Miller platform duration and the collector-emitter voltage, and the fitting relationship between the collector-emitter voltage and the junction temperature of the IGBT device under different collector currents; and then the difference obtained by subtracting the shell temperature from the junction temperature corresponding to the obtained Miller platform duration is used as the junction temperature difference, and the junction temperature difference is used to predict the IGBT life. The problem of low accuracy of IGBT life prediction by the existing IGBT life prediction method is solved.

Description

Technical Field

[0001] The present invention belongs to the field of electronic power technology, and specifically relates to an IGBT life prediction method, device, storage medium and program product. Background Art

[0002] IGBT (Insulated Gate Bipolar Transistor) is a key power semiconductor device used in power electronic conversion circuits, and is widely used in inverters, frequency converters, power amplifiers, etc. In the actual service process of IGBT, IGBT is subjected to harsh environments such as high voltage, high current and high temperature, and is easily affected by thermal stress and voltage stress, resulting in device performance degradation and shortened life. It is of great significance to predict the life of IGBT under different working conditions by monitoring the relevant state parameters of IGBT.

[0003] The existing IGBT life prediction methods mainly include: machine learning methods, physical model methods, experimental data methods and accelerated life prediction models. Among them, the machine learning method uses a large amount of historical data to train the prediction model. By establishing a complex machine learning model (such as a neural network), it can capture the relationship between the IGBT life and multiple factors to a certain extent; however, this method has a complex tuning process and high requirements for data and computing resources; and the model training and verification process is often time-consuming. The physical model method is based on a detailed understanding of the internal structure and material properties of the IGBT device to establish a mathematical model to predict the life, which usually involves complex mathematical calculations and simulations, such as thermal simulation, stress analysis, etc., involving complex mathematical calculations and simulations, and requires full consideration of the internal structure and material properties of the device; in a complex working environment, the parameter values ​​under different conditions may vary greatly, resulting in large deviations in the life prediction results and low accuracy. Although the experimental data method has high prediction accuracy, this method belongs to offline testing and cannot achieve online monitoring. The commonly used accelerated life prediction model can predict the life in real time without the need to establish a complex prediction model. It mainly predicts the life by obtaining the junction temperature difference and average junction temperature of the IGBT; however, since the device junction temperature is difficult to obtain accurately, the accuracy of the life prediction result is low. Summary of the invention

[0004] The object of the present invention is to provide an IGBT life prediction method, device, storage medium and program product to solve the problem of low accuracy of IGBT life prediction in existing IGBT life prediction methods.

[0005] In order to solve the above technical problems, the present invention provides an IGBT life prediction method, which comprises:

[0006] The Miller platform duration and collector current of the IGBT are obtained, and according to the relationship between the Miller platform duration and the junction temperature of the IGBT device under the collector current, the junction temperature corresponding to the obtained Miller platform duration is obtained; wherein the relationship between the Miller platform duration and the junction temperature of the IGBT device is derived according to the relationship between the Miller platform duration and the collector-emitter voltage, and the fitting relationship between the collector-emitter voltage and the junction temperature of the IGBT device under different collector currents; wherein the relationship between the Miller platform duration and the collector-emitter voltage is obtained based on the actual circuit connection condition of the IGBT;

[0007] The difference between the junction temperature corresponding to the obtained Miller platform duration and the shell temperature is taken as the junction temperature difference, and the IGBT life prediction is performed using the junction temperature difference.

[0008] Furthermore, the calculation formula for predicting the life of the IGBT using the junction temperature difference is:

[0009]

[0010] Among them, N f is the number of device power cycles, i.e., device life, ΔT j is the junction temperature difference, T m is the mean junction temperature, f is the temperature cycle frequency, E a is the activation energy, and α, k, n1, and n2 are correlation coefficients.

[0011] Furthermore, the fitting relationship between the collector-emitter voltage and the junction temperature of the IGBT device is obtained by fitting using a least squares method.

[0012] Furthermore, the process of obtaining the Miller platform duration includes: converting the collector-emitter voltage into two pulse signals according to the two moments when the collector-emitter voltage starts to drop during the IGBT conduction process and when the Miller platform ends, and calculating the time difference between the rising edges of the two pulse signals as the Miller platform duration.

[0013] Furthermore, the collector-emitter voltage of the IGBT when it is turned on is a voltage after filtering, and the filtering includes filtering out high-frequency noise and / or common-mode noise.

[0014] In order to solve the above technical problem, the present invention further provides a computer device, including a processor, wherein the processor executes a computer program to implement the steps of the above-mentioned IGBT life prediction method.

[0015] In order to solve the above technical problem, the present invention further provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the IGBT life prediction method as described above are implemented.

[0016] In order to solve the above technical problem, the present invention further provides a computer program product, including a computer program, which implements the steps of the above-mentioned IGBT life prediction method when executed by a processor.

[0017] The beneficial effects are as follows: the present invention is an improved invention creation. The present invention no longer directly uses the collected junction temperature to calculate the junction temperature difference and then predicts the life of the IGBT as in the prior art, but derives the relationship between the Miller platform duration and the junction temperature of the IGBT device according to the relationship between the Miller platform duration and the collector-emitter voltage, and the fitting relationship between the collector-emitter voltage and the junction temperature of the IGBT device under different collector currents. Thus, by obtaining the Miller platform duration and collector current of the IGBT, according to the above-derived relationship, the junction temperature corresponding to the obtained Miller platform duration can be obtained, and then the difference obtained by subtracting the shell temperature from the junction temperature corresponding to the obtained Miller platform duration is used as the junction temperature difference, and the IGBT life prediction is performed using the junction temperature difference. The junction temperature calculated by the Miller platform duration in the present invention is more accurate than the directly collected junction temperature, and the Miller platform duration is closely related to the actual working state of the IGBT device, and can more directly reflect the working state of the IGBT, and the Miller platform duration monitoring is not difficult, which improves the accuracy of life prediction and solves the problem of low accuracy of IGBT life prediction by the existing IGBT life prediction method. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 is a flow chart of a method for predicting the remaining life of an IGBT according to an embodiment of the present invention;

[0019] Figure 2 is T of the embodiment of the present invention m =60 life prediction result diagram;

[0020] Figure 3 is T of the embodiment of the present invention m =80 life prediction result diagram;

[0021] Figure 4 is T of the embodiment of the present invention m =100 life prediction result graph. DETAILED DESCRIPTION

[0022] The present invention obtains the Miller platform duration and collector current of the IGBT, and obtains the junction temperature corresponding to the obtained Miller platform duration according to the relationship between the Miller platform duration and the junction temperature of the IGBT device under the collector current; the difference obtained by subtracting the shell temperature from the junction temperature corresponding to the obtained Miller platform duration is used as the junction temperature difference, and the IGBT life prediction is performed using the junction temperature difference. The relationship between the Miller platform duration and the junction temperature of the IGBT device is derived based on the relationship between the Miller platform duration and the collector-emitter voltage, and the fitting relationship between the collector-emitter voltage and the junction temperature of the IGBT device under different collector currents; the problem of low accuracy of IGBT life prediction by the existing IGBT life prediction method is solved.

[0023] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention is further described in detail below with reference to the accompanying drawings and embodiments.

[0024] Method Example:

[0025] In this embodiment, a method for predicting the life of an IGBT is used to obtain the collector-emitter voltage V by fitting. ce The junction temperature T j At different collector currents I c The functional relationship under the Miller platform duration t GP and junction temperature T j The IGBT life model is established based on the relationship between the IGBT life model and the remaining service life of the IGBT can be predicted. Figure 1 As shown, the specific steps include:

[0026] Step 1: Connect the IGBT to the actual working circuit, and collect and store the gate-source voltage waveform when the IGBT is turned on.

[0027] Among them, the actual working circuit is a Boost DC / DC boost circuit.

[0028] Step 2: Extract the Miller platform duration from the collected gate-source voltage waveform and store it.

[0029] The method for extracting the duration of the Miller platform is to collect the collector-emitter voltage V ce After preprocessing such as filtering, the GP The signal processing circuit will turn on when V ce The voltage is converted into two pulses and connected to the gate driver without affecting the normal operation. The time difference between the rising edges of the two pulses is measured by a time-to-digital converter to obtain the Miller platform duration.

[0030] When IGBT is turned on, V ceThe voltage will experience a significant drop process. During the Miller plateau, the voltage drops relatively slowly or remains unchanged. GP The signal processing circuit detects V ce The collector-emitter voltage at the two moments when the collector-emitter voltage starts to drop and the Miller platform ends during the IGBT conduction process, and the voltage changes corresponding to these two moments are converted into two pulse signals; the rising edges of these two pulse signals correspond to V ce By calculating the time difference between the rising edges of the two pulses at the moment when the decline begins and the Miller platform ends, the duration of the Miller platform t can be accurately determined. GP .

[0031] Among them, this embodiment uses an LC filter to filter out high-frequency noise, a differential amplifier to filter out common-mode noise, and Laplace transform to perform signal processing from the perspectives of time domain and frequency domain. As other implementation methods, other means can be selected for preprocessing according to actual needs.

[0032] Step 3: Construct the IGBT remaining life model. The construction process of the IGBT remaining life model includes:

[0033] Fitting the collector-emitter voltage V by the least squares algorithm ce and junction temperature T j At different collector currents I c The functional relationship below. Among them, the fitting method used in this embodiment is the least squares method, and other implementation methods can be selected according to actual needs. The least squares algorithm fitting process is as follows:

[0034] Y=β0+β1x1+…+β k x k +ε (1)

[0035] ε~N(0,σ 2 )

[0036] Among them, β i (i=0,1,....k) is the regression coefficient, σ 2 , β i are all unknown quantities. The dependent variable Y represents the actual observed value between the measured collector-emitter voltage and junction temperature, that is, Y is the collector-emitter voltage V obtained by experiment or measurement. ce A collection of data; the independent variable x represents the factor that affects the dependent variable Y and is used to predict the dependent variable Y; ε represents the measurement error or residual, which reflects the deviation between the model predicted value and the actual observed value.

[0037] In order to facilitate calculation, the present embodiment rewrites formula (1) into a matrix form:

[0038] Y=Xβ+ε (2)

[0039] Where Y=(Y1,Y2,...,Y i ) T , i is the amount of sample data; β=(β0,β1,...,β k ) T ,ε=(ε1,ε2,...,ε i ) T .

[0040] Then the sum of squares of the measurement deviations is:

[0041] s E 2 =||Y-Xβ|| (3)

[0042] Since measurement errors are inevitable, the least squares solution is:

[0043]

[0044] in, is an estimator of β, based on different I c The functional relationship between the lower junction temperature and the collector-emitter voltage. When the collector current is I c When the junction temperature of the device is j and the collector-emitter voltage V ce The functional relationship is:

[0045] T j =βV ce +ε (5)

[0046] Right now:

[0047] T j =F(V ce ) (6)

[0048] Among them, different collector current I c There are corresponding parameters β and ε respectively, so different collector currents I c There are corresponding methods for calculating the junction temperature respectively.

[0049] The duration of the Miller plateau can be characterized by formula (8):

[0050]

[0051] Among them, t GP is the duration of the Miller platform, R G is the gate resistance, C GC is the junction capacitance, V DCis the DC bus voltage, V G is the gate voltage, V GP is the Miller plateau voltage.

[0052] During the switching process of the IGBT, the Miller platform reflects the process in which the collector-emitter voltage decreases when the gate voltage rises. The gate-emitter voltage at this stage is maintained at a relatively stable level, while the collector current of the IGBT remains unchanged. G , junction capacitance C GC , DC bus voltage V DC , Collector-emitter voltage V ce , gate voltage V G and Miller platform voltage V GP And other factors, combined with the actual circuit connection of IGBT, the relationship between the Miller platform duration and the collector-emitter voltage is obtained.

[0053] Then from formula (7) we can get:

[0054]

[0055] Then from formulas (5) to (8), we can get:

[0056]

[0057] The conventional IGBT life prediction model is shown in formula (10):

[0058]

[0059] Among them, N f is the number of device power cycles, i.e., device life, ΔT j is the junction temperature difference, T m is the average junction temperature, and f is the temperature cycle frequency. a is the activation energy, α, k, n1, and n2 are correlation coefficients, which can be measured experimentally.

[0060] Since the junction temperature of the device is difficult to obtain accurately, the junction temperature difference ΔT in formula (10) j It is difficult to obtain accurately, which affects the accuracy of the life prediction of the IGBT. Therefore, the present invention characterizes the junction temperature difference as the junction temperature T of the device. j With shell temperature T c , using the Miller platform duration t GP and junction temperature T j The relationship between the Miller platform duration t GP Establish the IGBT remaining life model, specifically:

[0061] ΔT j =Tj -T c (11)

[0062] According to formulas (9) to (11), the remaining life model of IGBT is:

[0063]

[0064] Among them, T m is the junction temperature T calculated according to formula (9) j Calculated. m It is reflected by the change of periodic junction temperature, which can be calculated by formula (13):

[0065]

[0066] Among them, T j,i is the junction temperature at each time point, and N is the number of time points recorded.

[0067] The life prediction model proposed in the present invention no longer relies on the actual junction temperature data which is difficult to obtain accurately, but obtains the mean junction temperature T by calculating the relevant electrical parameters. m , thus realizing the prediction of IGBT life. This improvement effectively solves the problem of difficulty in collecting junction temperature in existing life prediction methods.

[0068] Step 4: Use the Miller platform duration extracted in step 2 and the IGBT remaining life model constructed in step 3 to predict the life of the IGBT. Specifically:

[0069] 1) Obtain the collector current I of the IGBT to be tested c , Gate resistance R G , junction capacitance C GC , DC voltage V DC , gate voltage V G and Miller platform voltage V GP The collector current I is calculated by combining the following parameters with formula (9): c The junction temperature under the condition of

[0070] 2) Get the device case temperature T c , using the junction temperature calculated in step 2) and the device case temperature T c , the IGBT life is predicted through the IGBT remaining life model constructed in step three.

[0071] After the IGBT remaining life model is established, different device junction temperature averages are selected to study the remaining life of the IGBT. m = 60 when the life prediction curve is as follows Figure 2 As shown, when T m At 60°C: At a lower average junction temperature, the Miller platform duration t of the IGBT is GP (Right now Figure 2 , 3 , the horizontal coordinate t in 4 gp ) is longer. This is because when the temperature is low, the carrier lifetime of the device is prolonged, resulting in a longer release time of the gate charge in the Miller platform stage. GP The longer the device is, the greater the switching loss. However, due to the lower overall temperature, the device ages relatively slowly, so the lifetime N f Relatively high.

[0072] T m =80 when the life prediction curve is as follows Figure 3 As shown, when T m At 80℃: As T m The increase of t GP This is because at higher temperatures, the mobility of carriers increases, resulting in a shorter duration of the Miller platform. However, although t GP The shorter the N, the lower the switching loss. However, the higher the temperature, the greater the thermal stress and the faster the aging of the device. f Compared with 60℃, it has decreased.

[0073] T m =100 when the life prediction curve is as follows Figure 4 As shown, when T m At 100℃: At this temperature level, t GP The high temperature accelerates the carrier transfer process in the Miller platform phase, resulting in t GP However, high temperature also causes the device to be subjected to greater thermal stress, which accelerates the aging of the material and leads to a shorter life span. f Significantly reduced. GP The shortening may reduce the switching losses, but its contribution to the lifetime is not enough to offset the degradation effects caused by high temperature, which ultimately leads to a significant shortening of the IGBT's lifetime at this temperature.

[0074] The IGBT life prediction method of the present invention utilizes the Miller platform duration t in the IGBT conduction waveform. GP Construct a remaining life model to predict the remaining life of IGBT; the main parameters involved in the model are t GPIt is closely related to the actual working state of the IGBT. The established model can more directly reflect the working state of the IGBT, and the parameter monitoring is not difficult. The IGBT life prediction method of the present invention can also perform real-time online monitoring of the IGBT without offline testing, which can help users understand the remaining life of the device in real time to reduce unnecessary failure losses.

[0075] Computer equipment example:

[0076] A computer device of the present invention includes a memory, a processor and an internal bus, and the processor and the memory communicate and exchange data with each other through the internal bus. The memory includes at least one software function module stored in the memory, and the processor executes various functional applications and data processing by running the computer program and module stored in the memory to implement an IGBT life prediction method introduced in the method embodiment of the present invention. The principle, implementation process and achievable effect of the method have been fully introduced in the method embodiment and will not be repeated here.

[0077] Computer readable storage medium embodiment:

[0078] A computer-readable storage medium of the present invention stores a computer program thereon, and when the computer program is executed by a processor, an IGBT life prediction method described in the method embodiment of the present invention is implemented. The principle, implementation process, and effect that can be achieved by the method have been fully described in the method embodiment and will not be repeated here.

[0079] Computer program product embodiment:

[0080] A computer program product of the present invention includes a computer program, and when the computer program is executed by a processor, an IGBT life prediction method described in the method embodiment of the present invention is implemented. The principle, implementation process, and effect that can be achieved by the method have been fully described in the method embodiment and will not be repeated here.

Claims

1. A method for predicting IGBT life, characterized in that: The method includes: The Miller platform duration and collector current of the IGBT are obtained, and according to the relationship between the Miller platform duration and the junction temperature of the IGBT device under the collector current, the junction temperature corresponding to the obtained Miller platform duration is obtained; wherein the relationship between the Miller platform duration and the junction temperature of the IGBT device is derived according to the relationship between the Miller platform duration and the collector-emitter voltage, and the fitting relationship between the collector-emitter voltage and the junction temperature of the IGBT device under different collector currents; wherein the relationship between the Miller platform duration and the collector-emitter voltage is obtained based on the actual circuit connection condition of the IGBT; The difference between the junction temperature corresponding to the obtained Miller platform duration and the shell temperature is taken as the junction temperature difference, and the IGBT life prediction is performed using the junction temperature difference.

2. The IGBT life prediction method according to claim 1, characterized in that: The calculation formula for predicting the life of the IGBT using the junction temperature difference is: Among them, N f is the number of device power cycles, i.e., device life, ΔT j is the junction temperature difference, T m is the mean junction temperature, f is the temperature cycle frequency, E a is the activation energy, and α, k, n1, and n2 are correlation coefficients.

3. The IGBT life prediction method according to claim 1, characterized in that: The fitting relationship between the collector-emitter voltage and the junction temperature of the IGBT device is obtained by fitting using the least square method.

4. The IGBT life prediction method according to claim 1, characterized in that: The process of obtaining the Miller platform duration includes: converting the collector-emitter voltage into two pulse signals according to the two moments when the collector-emitter voltage starts to drop and the Miller platform ends during the IGBT conduction process, and calculating the time difference between the rising edges of the two pulse signals as the Miller platform duration.

5. The IGBT life prediction method according to claim 4, characterized in that: The collector-emitter voltage when the IGBT is turned on is a voltage after filtering, and the filtering includes filtering out high-frequency noise and / or common-mode noise.

6. A computer device comprising a processor, characterized in that: The processor executes a computer program to implement the steps of the IGBT lifetime prediction method according to any one of claims 1 to 5.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the IGBT life prediction method according to any one of claims 1 to 5 are implemented.

8. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the IGBT life prediction method according to any one of claims 1 to 5 are implemented.

Images