Power device junction temperature monitoring method based on multivariate thermosensitive electrical parameters

Abstract

The invention discloses a power device junction temperature monitoring method based on multivariate thermo-sensitive electrical parameters, which comprises the following steps of: performing integrated modeling on dynamic characteristics of a power device and a power circuit, and considering the influence of factors such as parasitic parameters, driving parameters, operating parameters and temperature; thermosensitive parameter change rules such as switching time and voltage and current change rates of the device are analyzed; according to the dynamic model, thermosensitive parameters with low operation parameter dependence are obtained, and the junction temperature of the power device is jointly estimated through multiple sets of thermosensitive parameters. In addition, a thermosensitive parameter weight self-correction coefficient based on a fitting error is designed, so that higher estimation precision can be obtained while the robustness of the algorithm is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of power device junction temperature monitoring, and in particular relates to a power device junction temperature monitoring method based on multi-element thermal-sensitive electrical parameters. Background Art

[0002] Due to the inherent flaws of power device junction temperature estimation methods based on thermal models, junction temperature monitoring methods based on thermosensitive electrical parameters have gradually become the mainstream junction temperature monitoring method. Since the intrinsic carrier concentration in power devices is approximately exponentially related to junction temperature, most of the device's electrical parameters (such as turn-on time, turn-off time, turn-on delay, and turn-off delay) are related to junction temperature and are referred to as thermosensitive electrical parameters (TSEP). Offline calibration of the TSEP-junction temperature curve transforms the junction temperature monitoring problem into an online measurement problem of thermosensitive electrical parameters. The analytical expressions for the thermosensitive electrical parameters versus junction temperature are complex and coupled, often presented in tabular form. Approximate discretized results are then approximated to the true value through mathematical methods such as interpolation and fitting. The accuracy of the extreme junction temperature monitoring method using the thermosensitive electrical parameter method depends on the sensitivity of the selected thermosensitive electrical parameter, which is the change in the thermosensitive electrical parameter per unit change in junction temperature.

[0003] At present, junction temperature monitoring methods based on thermosensitive electrical parameters (such as [GRIFFO A, WANG J, COLOMBAGE K, et al. Real-time measurement of temperature sensitive electrical parameters in SiC power MOSFETs [J]. IEEE Transactions on Industrial Electronics, 2018, 65 (3): 2663-2671] and [YANG F, PU S, XU C, et al. Turn-on Delay Based Real-Time Junction Temperature Measurement for SiC MOSFETs With Aging Compensation [J]. IEEE Transactions on Power Electronics, 2021, 36 (2): 1280-1294]) still face certain limitations, which are summarized as follows:

[0004] 1. The change mechanism of thermistor electrical parameters is complex and difficult to model. The accuracy of existing models still needs to be improved, and the theoretical system is incomplete, which affects the reliability of junction temperature monitoring.

[0005] 2. Affected by device characteristics, the sensitivity of thermal electrical parameters is low, which restricts the accuracy of junction temperature monitoring;

[0006] 3. Thermistor parameters change rapidly, especially in SiC devices, where switching times vary in the nanosecond range. Measurement errors have a significant impact on junction temperature monitoring accuracy.

[0007] Therefore, it is urgent to propose a robust monitoring method for the junction temperature of power devices with high accuracy. Summary of the Invention

[0008] In view of the above, the present invention provides a power device junction temperature monitoring method based on multi-element thermal-sensitive electrical parameters, which utilizes multiple sets of thermal-sensitive parameters to jointly estimate the device junction temperature, and introduces a weight self-correction coefficient so that the thermal parameter weight ratio can be adaptively corrected, thereby improving the accuracy of the estimated value while obtaining a highly robust estimation result.

[0009] A method for monitoring junction temperature of a power device based on multi-element thermal electrical parameters comprises the following steps:

[0010] (1) Select multiple thermal electrical parameters used to characterize the junction temperature of the power device, and obtain mathematical expressions of these thermal electrical parameters and the junction temperature of the power device through theoretical analysis;

[0011] (2) Build a test circuit containing power devices, collect the values ​​of various thermosensitive electrical parameters at different temperatures during the normal operation of the circuit, and then fit the mathematical expressions of the thermosensitive electrical parameters and the junction temperature of the power devices based on these values;

[0012] (3) Obtaining the fitted value of the thermistor electrical parameter according to the fitted mathematical expression, and calculating the residual between the fitted value and the sampled value and its square root mean;

[0013] (4) Obtain the junction temperature expression of each thermistor parameter by finding the inverse function of the fitted mathematical expression, and then calculate the weight of each junction temperature expression according to the square root mean of the residual;

[0014] (5) Collect and obtain the values ​​of each thermistor electrical parameter at the current temperature, and substitute these values ​​into each junction temperature expression to obtain the corresponding junction temperature estimation value, and then perform the weighted summation of these junction temperature estimation values ​​to obtain the current junction temperature of the power device.

[0015] Furthermore, in step (1), the number of dependent variables of the thermosensitive electrical parameters needs to be controlled, and N thermosensitive electrical parameters with low correlation with the operating parameters are selected from a large number of thermosensitive electrical parameters, where N is a natural number greater than 2.

[0016] Furthermore, in step (2), for any thermosensitive electrical parameter, the mathematical expression of the thermosensitive electrical parameter and the junction temperature of the power device is fitted by the least squares method according to the values ​​of the thermosensitive electrical parameter collected at different temperatures.

[0017] Furthermore, in step (3), the residual between the thermosensitive electrical parameter fitting value and the sampling value and its square root mean are calculated by the following formula;

[0018] Δr k (T j )=r k_t (T j )-r k (T j )

[0019]

[0020] Where: r k_t (T j ) is the temperature T j Thermistor electrical parameter r k The fitted value, r k (T j ) is the temperature T j Thermistor electrical parameter r k The sampling value of Δr k (T j ) is the fitted value r k_t (T j ) and the sampling value r k (T j ), subscript j = 1, 2, ..., J, J is the number of temperature discrete points, subscript k = 1, 2, ..., N, N is the number of selected thermal electrical parameters, H k is the thermal electrical parameter r k The root mean square of the residual errors between the fitted and sampled values.

[0021] Furthermore, in step (4), the weight of each junction temperature expression is calculated by the following formula:

[0022]

[0023] Where: α k is the thermal electrical parameter r k The junction temperature expression weight, H k is the thermal electrical parameter r k The root mean square of the residuals between the fitted values ​​and the sampled values. Subscript k = 1, 2, ..., N, where N is the number of selected thermal electrical parameters.

[0024] Furthermore, in step (5), the current junction temperature of the power device is obtained by weighted summing the estimated junction temperature values ​​using the following expression;

[0025] T=α1·T(r1)+α2·T(r2)+…+α N ·T(r N )

[0026] Where: T is the current junction temperature of the power device, α k is the thermal electrical parameter r k The junction temperature expression weight, T(r k ) is the thermal electrical parameter r k The corresponding junction temperature estimation value, subscript k = 1, 2, ..., N, N is the number of selected thermal electrical parameters.

[0027] A computer device includes a memory and a processor, wherein the memory stores a computer program, and the processor is used to execute the computer program to implement the above-mentioned power device junction temperature monitoring method based on multi-element thermal sensitive electrical parameters.

[0028] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the above-mentioned power device junction temperature monitoring method based on multi-element thermal-sensitive electrical parameters.

[0029] The power device junction temperature monitoring method based on multi-element thermistor parameters of the present invention has obvious advantages in terms of junction temperature monitoring accuracy and robustness in the face of thermistor parameter measurement errors. The selected thermistor parameters need to have a clear mathematical relationship with the junction temperature of the power device. The dual-pulse test circuit used is consistent with the actual application system in terms of stray parameters. The thermistor parameters used for expression fitting are tested for consistency through no less than 20 sets of data. The present invention reduces parameter sensitivity by combining multi-element thermistor parameters to improve the stability and reliability of the junction temperature monitoring results. The residual root mean square is used to characterize the reliability of the fitting expression, and the residual root mean square inverse function is used to calculate the weights of different junction temperature characterization expressions. Therefore, the present invention has the following beneficial technical effects:

[0030] 1. The present invention takes into account the influence of system parasitic parameters and temperature on the dynamic performance of power devices, establishes a dynamic mathematical model of power devices, and has a high degree of matching with actual working conditions.

[0031] 2. The present invention determines the thermistor parameters used to characterize the junction temperature of the power device based on the dynamic mathematical model of the power device, and fits the mathematical expressions of the thermistor parameters and the junction temperature of the power device based on the measured data. It has the advantages of being simple and easy to operate, and the algorithm is highly efficient and suitable for online estimation of the device junction temperature.

[0032] 3. The present invention uses multivariate thermal electrical parameters to estimate the junction temperature of power devices, thereby improving the accuracy and robustness of power device junction temperature estimation. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 This is a flow chart of the device junction temperature monitoring method based on multi-element thermal sensitive electrical parameters of the present invention.

[0034] Figure 2This is a schematic diagram of the structure of a two-level double-pulse test circuit constructed in an embodiment of the present invention.

[0035] Figure 3 Schematic diagram of the transient waveform of the power device switch in an embodiment of the present invention, (a) corresponds to the turn-on process, and (b) corresponds to the turn-off process.

[0036] Figure 4 Schematic diagram of the device junction temperature result estimated based on a single thermal electrical parameter in an embodiment of the present invention, in which (a) corresponds to the turn-on delay t don , (b) corresponds to the off time t off , (c) corresponds to the opening time t on , (d) corresponds to the turn-off delay t doff .

[0037] Figure 5 Schematic diagram of device junction temperature estimation based on multi-element thermosensitive electrical parameters in an embodiment of the present invention. DETAILED DESCRIPTION

[0038] In order to describe the present invention more specifically, the technical solution of the present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0039] like Figure 1 As shown, the power device junction temperature monitoring method based on multi-element thermal sensitive electrical parameters of the present invention includes the following steps:

[0040] Step 1: Obtain the mathematical expressions of thermistor electrical parameters and junction temperature of power devices through theoretical analysis.

[0041] Step 2: Select N groups of different thermal electrical parameter physical quantities r1, r2, ..., r that can characterize the junction temperature of the power device N , N≥3; it is necessary to control the number of dependent variables of thermistor parameters and use thermistor parameters with low correlation with operating parameters as much as possible.

[0042] Step 3: Build Figure 2 The double pulse test circuit shown in the figure, where L load is the load inductance, and the collected temperature T j (j=1,2,...,J, J is the number of temperature discrete points) under r1(T j ),r2(T j ),...,r N (T j ), when measuring the upper-arm power devices, it is necessary to ensure that the lower-arm power devices are reliably shut down.

[0043] Step 4: Fit r using the least squares method k (k=1,2,...,N) and different junction temperatures T jThe mathematical expression of r k_t (T j ) is the temperature T j Fitting values ​​of the thermal electrical parameters below.

[0044] Step 5: Calculate r k_t (T j ) and the sampling value r k (T j )’s residual:

[0045] Δr k (T j )=r k_t (T j )-r k (T j )

[0046] Step 6: Calculate the lower r at different junction temperatures k_t (T j ) and the sampling value r k (T j ) is the root mean square of the residuals H k :

[0047]

[0048] Among them: H k The smaller it is, the better the representation effect is.

[0049] Step 7: By solving r k_t (T j ) is the inverse function of the junction temperature under different thermistor parameters, and the mathematical expression T j (r k_t ), and T j (r k_t )≈T j (r k ).

[0050] Step 8: Take the weighted sum of the junction temperature expressions under different thermal parameters to get the estimated junction temperature of the power device:

[0051] T=α1·T(r1)+α2·T(r2)+…+α N ·T(r N )

[0052] Where: α k is the weight of the kth group of junction temperature expressions.

[0053] Step 9: Solve the weight α of the N-group junction temperature expression k ,make:

[0054]

[0055] Step 10: Alpha k Substitute the estimated junction temperature of the power device into the expression to obtain the estimated junction temperature.

[0056] Step 11: Collect the thermal electrical parameters and substitute them into the estimated value of the power device junction temperature to estimate the power device junction temperature.

[0057] Example

[0058] This embodiment takes an IGBT power device as an example to verify the power device junction temperature estimation method of the present invention under the working conditions of 600V and 225A.

[0059] The gate voltage v during the power device switching transient process gs , drain-source voltage v ds and drain current i d Waveform Figure 3 As shown in the figure, V P is the driving voltage, V m is the Miller voltage, V th is the threshold voltage, I L is the load current, V dc is the bus voltage, I pk is the peak current, t don is the opening delay, t ri1 is the current i d The time from zero to load current rise, t ri2 The current is I L Time from rising to peak current, t fv is the voltage drop time, t on is the opening time, t doff is the turn-off delay, t rv is the voltage rise time, t fi is the current fall time, t off is the off time, V pk The transient process of device turn-on and turn-off is studied in stages, where the turn-on process is divided into 4 stages and the turn-off process is divided into 3 stages.

[0060] At time t0, the gate voltage begins to increase and can be expressed as:

[0061]

[0062] Where: C gs is the gate-source capacitance, C gd is the gate-drain capacitance, R g is the gate resistance.

[0063] Until time t1, v gs_on Rising to V th, the device starts to conduct, so the turn-on delay t don for:

[0064]

[0065] After the device is turned on, the drain current increases until it reaches I L , at this time v gs_on Rising to V m , so the current rise time t ri1 for:

[0066]

[0067] Due to the reverse recovery process of the diode current, there will be an overshoot after the current reaches the steady-state value, i d From t2 to t3, it rises to I pk , the time required is:

[0068]

[0069] Where: C F is the parasitic capacitance of the anti-parallel diode.

[0070] L loop =L bus +L d +L s

[0071] Where: L bus is the busbar inductance, L d is the leakage inductance, L s is the source inductance.

[0072] After t3, the voltage enters the drop phase and ends at t4. The voltage drop time t fv for:

[0073]

[0074] At t5, the gate voltage begins to decrease, which can be expressed as:

[0075]

[0076] Until time t6, v gs_off Drop to V m , the voltage starts to rise, so the turn-off delay t doff for:

[0077]

[0078] Until time t7, v ds Rising to V dc , voltage rise time trv for:

[0079]

[0080] After t7, v gs_off Continue to decline; at the same time, enter the current decline stage, until t8 moment, v gs_off Down to V th , the current drops to zero, so the current drop time t fi for:

[0081]

[0082] Figure 3 In the device turn-on time t on and device off time t off They can be expressed as:

[0083] t on =t don +t ri1 +t ri2 +t fv

[0084] t off =t doff +t rv +t fi

[0085] Since temperature affects threshold voltage, transconductance coefficient, parasitic capacitance, etc., according to the above analysis, it can be found that t don , t on , t doff , t off , t ri1 , t ri2 , t fv , t rv , t fi All are affected by temperature and can be used as thermosensitive electrical parameters; in addition, threshold voltage, gate current, Miller platform, device terminal voltage change rate and drain current change rate are also thermosensitive electrical parameters. It should be noted that t fv and t rv Affected by the DC voltage, in the actual operating conditions of the system, the DC voltage fluctuation is small, and the junction temperature and t fv , t rv The mathematical relationship between the junction temperature and the junction temperature is improved.

[0086] In this embodiment, we determine t don , t on , t doff , t off As four groups of thermal electrical parameters to characterize the junction temperature of power devices, build Figure 2 The double pulse test circuit shown in the figure uses 600V and 225A as the system operating conditions, and collects t values ​​at 6 different temperatures: 25℃, 50℃, 75℃, 100℃, 125℃, 150℃, and 175℃. don , t on , t doff , t off , as shown in Table 1:

[0087] Table 1: Thermistor parameter sampling values ​​at different temperatures

[0088]

[0089] The least squares fitting method is used to obtain t don , t off , t on , t doff and junction temperature T j The mathematical expression is:

[0090] t don =-1.188571428571427×10 -2 ×T j +2.203428571428570×10

[0091] t off =6.034285714285709×10 -2 ×T j +6.029714285714284×10

[0092] t on =-2.8799999999999997×10 -2 ×T j +4.370285714285713×10

[0093] t doff =4.114285714285709×10 -2 ×T j +3.355428571428571×10

[0094] Calculate the fitting value of the thermosensitive electrical parameter at different temperatures by fitting the thermosensitive parameter expression t don_t , t on_t , t doff_t , t off_t , and then obtain t don_t , t on_t , t doff_t , t off_t With the sampling value t don , t on, t doff , t off The residuals of thermistor are obtained, and the fitting values ​​of thermistor parameters at different junction temperatures are obtained. don_t , t on_t , t doff_t , t off_t With the sampling value t don , t on , t doff , t off The root mean square of the residuals H k :

[0095] H1=0.4441; H2=0.1567; H3=0.1404; H4=0.4049

[0096] By solving the fitting expression of the thermal parameters, 4 sets of estimated junction temperature and thermal electrical parameters t are obtained. don , t off , t on , t doff The mathematical expression is:

[0097] T j (t don )=-(t don -2.203428571428570×10) / 1.188571428571427×10 -2

[0098] T j (t off )=(t off -6.029714285714284e×10) / 6.034285714285709×10 -2

[0099] T j (t on )=-(t on -4.370285714285713e×10) / 2.8799999999999997×10 -2

[0100] T j (t doff )=(t doff -3.355428571428571×10) / 4.114285714285709×10 -2

[0101] Four sets of junction temperature expressions based on single thermal parameters are used to jointly estimate the junction temperature of the power device. The junction temperature expression of the power device based on multiple thermal parameters is:

[0102] Tj =α1·T j (t off )+α2·T j (t don )+α3·T j (t on )+α4·T j (t doff )

[0103] Where: α k is the weight of the kth group of junction temperature expressions.

[0104] Solving for weight α k :

[0105] α1=0.1235; α2=0.35; α3=0.39; α4=0.1365

[0106] Then we get:

[0107] T j =0.1235·T j (t off )+0.35·T j (t don )+0.39·T j (t on )+0.1365·T j (t doff )

[0108] Finally, the thermal electrical parameters are collected and substituted into the above formula to calculate the junction temperature of the power device.

[0109] Based on a single thermal electrical parameter t don , t off , t on , t doff The estimated device junction temperatures are as follows: Figure 4 As shown in (a) to (d) in FIG, the junction temperature of the device estimated based on the multivariate thermal electrical parameters is as follows Figure 5 As shown. Figure 4 and Figure 5 It can be seen that the device junction temperature estimated based on multivariate thermal electrical parameters is highly consistent with the actual junction temperature, and its consistency is second only to the t-based method with the highest weight coefficient. don The estimation formula of .

[0110] Next, we use the test data under a 100°C environment as an example to analyze the advantages of the power device junction temperature estimation method based on multi-element thermal electrical parameters. The calculation results are shown in Table 2:

[0111] Table 2: Junction temperature estimation results based on thermal parameters at 100°C

[0112]

[0113]

[0114] It can be seen from Table 2 that the reliability of the device junction temperature estimated based on multi-element thermal electrical parameters is better than that estimated based on a single parameter.

[0115] Table 3: Impact of measurement error on junction temperature estimation results (t don (large)

[0116]

[0117] Table 4: Impact of measurement error on junction temperature estimation results (t don Small)

[0118]

[0119] Tables 3 and 4 are artificially adjusted t don Sampling results, t in Table 3 don The sampling value is 0.02ns larger than the sampling value. don It is 0.02ns smaller than the sampling value. Based on the thermal parameter t don The maximum deviation of the estimated device junction temperature reaches 7.2°C, while the maximum deviation of the junction temperature estimated based on multi-element thermistor electrical parameters is less than 3°C. In addition, the junction temperature estimated based on multi-element thermistor electrical parameters is far less affected by measurement deviation fluctuations than the estimation method based on a single parameter. Anti-interference performance is also a key advantage of the present invention.

[0120] The above description of the embodiments is intended to facilitate understanding and application of the present invention by those skilled in the art. It is apparent that those skilled in the art can readily make various modifications to the above embodiments and apply the general principles described herein to other embodiments without requiring creative effort. Therefore, the present invention is not limited to the above embodiments. Any improvements or modifications made by those skilled in the art based on the disclosure of the present invention should fall within the scope of protection of the present invention.

Claims

1. A method for monitoring junction temperature of a power device based on multi-element thermal electrical parameters, characterized in that: The steps include: (1) Select multiple thermal electrical parameters used to characterize the junction temperature of the power device, and obtain mathematical expressions of these thermal electrical parameters and the junction temperature of the power device through theoretical analysis; (2) Build a test circuit containing power devices, collect the values ​​of various thermosensitive electrical parameters at different temperatures during the normal operation of the circuit, and then fit the mathematical expressions of the thermosensitive electrical parameters and the junction temperature of the power devices based on these values; (3) Obtaining the fitted value of the thermistor electrical parameter according to the fitted mathematical expression, and calculating the residual between the fitted value and the sampled value and its square root mean; (4) Obtain the junction temperature expression of each thermistor parameter by finding the inverse function of the fitted mathematical expression, and then calculate the weight of each junction temperature expression according to the square root mean of the residual; (5) Collect and obtain the values ​​of each thermistor electrical parameter at the current temperature, and substitute these values ​​into each junction temperature expression to obtain the corresponding junction temperature estimation value, and then perform the weighted summation of these junction temperature estimation values ​​to obtain the current junction temperature of the power device.

2. The method for monitoring junction temperature of a power device based on multi-element thermal electrical parameters according to claim 1, characterized in that: In the step (1), the number of dependent variables of the thermoelectric parameters needs to be controlled, and N thermoelectric parameters with low correlation with the operating parameters are selected from a large number of thermoelectric parameters, where N is a natural number greater than 2.

3. The method for monitoring junction temperature of a power device based on multi-element thermal electrical parameters according to claim 1, characterized in that: In the step (2), for any thermosensitive electrical parameter, the mathematical expression of the thermosensitive electrical parameter and the junction temperature of the power device is fitted by the least squares method according to the values ​​of the thermosensitive electrical parameter collected at different temperatures.

4. The method for monitoring junction temperature of a power device based on multi-element thermal electrical parameters according to claim 1, characterized in that: In the step (3), the residual between the thermoelectric parameter fitting value and the sampled value and its square root mean are calculated by the following formula; Δr k (T j )=r k_t (T j )-r k (T j ) Where: r k_t (T j ) is the temperature T j Thermistor electrical parameter r k The fitted value, r k (T j ) is the temperature T j Thermistor electrical parameter r k The sampling value of Δr k (T j ) is the fitted value r k_t (T j ) and the sampling value r k (T j ), subscript j = 1, 2, ..., J, J is the number of temperature discrete points, subscript k = 1, 2, ..., N, N is the number of selected thermal electrical parameters, H k is the thermal electrical parameter r k The root mean square of the residual errors between the fitted and sampled values.

5. The method for monitoring junction temperature of a power device based on multi-element thermal electrical parameters according to claim 1, characterized in that: In step (4), the weight of each junction temperature expression is calculated by the following formula: Where: α k is the thermal electrical parameter r k The junction temperature expression weight, H k is the thermal electrical parameter r k The root mean square of the residuals between the fitted values ​​and the sampled values. Subscript k = 1, 2, ..., N, where N is the number of selected thermal electrical parameters.

6. The method for monitoring junction temperature of a power device based on multi-element thermal electrical parameters according to claim 1, characterized in that: In the step (5), the current junction temperature of the power device is obtained by weighted summing the estimated junction temperature values ​​using the following expression; T=α1·T(r1)+α2·T(r2)+…+α N ·T(r N ) Where: T is the current junction temperature of the power device, α k is the thermal electrical parameter r k The junction temperature expression weight, T(r k ) is the thermal electrical parameter r k The corresponding junction temperature estimation value, subscript k = 1, 2, ..., N, N is the number of selected thermal electrical parameters.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: The processor is configured to execute the computer program to implement the power device junction temperature monitoring method based on multi-element thermosensitive electrical parameters as claimed in any one of claims 1 to 6.

8. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method for monitoring junction temperature of a power device based on multi-element thermosensitive electrical parameters according to any one of claims 1 to 6 is implemented.

Images