Method for predicting loss junction temperature of power module

Abstract

The invention provides a loss junction temperature prediction method of a power module, which relates to the technical field of power electronics, and comprises the following steps: reading a working condition file containing voltage, current, temperature ratio and resistance, and importing characteristic parameters of a power module data manual; based on the characteristic parameters, using an interpolation algorithm to construct a loss and thermal resistance model dynamically changing along with temperature, current and resistance; calculating a pulse sequence in a current half cycle, and calculating conduction loss and switching loss corresponding to each pulse through an accumulation method to obtain total loss; and calculating an initial junction temperature based on the loss and thermal resistance model, and feeding back the initial junction temperature to a loss calculation process for loop iteration until the junction temperature difference value of two successive calculations is smaller than a preset threshold value, and taking the last calculation result as the highest junction temperature for life prediction. The method has the beneficial effects that the pulse loss is accurately calculated through an accumulation method, a multi-dimensional dynamic model is constructed, thermoelectric coupling feedback iterative logic is introduced, and efficient and accurate junction temperature calculation under mass working conditions is realized.

Description

Technical Field

[0001] This invention relates to the field of power electronics technology, and in particular to a method for predicting the junction temperature of power modules. Background Technology

[0002] Loss calculation and junction temperature prediction for power modules (such as IGBTs and diodes) are core aspects of power electronic system design and reliability assessment. In existing technologies, loss and junction temperature calculation typically face the following technical challenges:

[0003] First, the calculation accuracy is limited. Traditional loss calculations often use integration methods or simple averaging methods, which cannot accurately capture the loss characteristics in dynamic pulses when dealing with nonlinear waveforms under complex modulation methods (such as SPWM and SVPWM), resulting in a large deviation between the calculated results and the actual measured values.

[0004] Secondly, the models have poor adaptability. Existing calculation models are often based on a single temperature, voltage, or resistance reference in the datasheet. However, in actual operating conditions, current, temperature, and gate resistance fluctuate in real time. Static models cannot dynamically adjust loss and thermal resistance parameters according to operating conditions, making it difficult to adapt to complex and changing operating environments.

[0005] Finally, the calculation process lacks feedback logic. Due to the strong thermo-electric coupling between the power consumption and junction temperature of the power module (i.e., losses change with increasing temperature, and temperature is also affected by losses), traditional open-loop calculation methods ignore the real-time correction process between losses and temperature, often leading to inaccurate predictions of the maximum junction temperature. Furthermore, when faced with massive operating condition files (such as hundreds of thousands of rows of data), manual or semi-automated processing methods are extremely inefficient and cannot meet the needs of large-scale lifetime prediction. Summary of the Invention

[0006] To address the problems existing in the prior art, this invention provides a method for predicting the junction temperature of a power module, comprising the following steps:

[0007] Step S1: Read the operating condition file containing voltage, current, temperature ratio and resistance, and import the characteristic parameters from the power module datasheet;

[0008] Step S2: Based on the characteristic parameters and the operating condition file, construct a loss and thermal resistance model that dynamically changes with temperature, current and resistance using an interpolation algorithm;

[0009] Step S3: Calculate the pulse sequence within the current half-cycle. Based on the loss and thermal resistance model, calculate the conduction loss and switching loss corresponding to each pulse using the accumulation method, and then obtain the total loss.

[0010] Step S4: Calculate the preliminary junction temperature based on the loss and thermal resistance model, and feed the preliminary junction temperature back to the loss calculation process for iterative iteration until the difference between the two calculated junction temperatures is less than a preset threshold. Take the last calculation result as the highest junction temperature for staff to use in predicting the lifespan of the power module.

[0011] Preferably, the loss and thermal resistance model includes an output characteristic model, then the process of constructing the output characteristic model in step S2 includes:

[0012] First, the output characteristic curve data at different temperatures are extracted from the datasheet. Then, linear interpolation is used to interpolate between the normal temperature curve and the high temperature curve based on the real-time operating temperature, generating a function curve of saturated on-state voltage drop versus current that conforms to the operating temperature as the output characteristic model.

[0013] Preferably, the loss and thermal resistance model includes a loss characteristic model, then the process of constructing the loss characteristic model in step S2 includes:

[0014] Step A1: Obtain the current loss curve data and resistance loss curve data under normal temperature and high temperature conditions from the data manual, and use the linear interpolation algorithm to perform temperature correction interpolation to fit the initial current loss curve and initial resistance loss curve at the corresponding operating temperature.

[0015] Step A2: If the operating current is within the range of the initial current loss curve, then determine the two sampling points on the curve that are closest to the operating current and perform linear interpolation. If the operating current is less than the minimum current in the initial current loss curve, then select the two sampling points at the front of the initial current loss curve for processing to obtain the loss value. If the operating current is greater than the maximum current in the initial current loss curve, then select the two sampling points at the end of the initial current loss curve for processing to obtain the loss value. Multiply the processed loss value by the ratio of the operating voltage to the specification voltage to obtain the intermediate loss value.

[0016] Step A3: Using the same interpolation method as in step A2, interpolate the resistance dimension of the initial resistance loss curve to obtain the first resistance loss component under the operating resistance and the second resistance loss component under the specification resistance.

[0017] Step A4: Divide the intermediate loss value by the second resistance loss component, and multiply the result by the first resistance loss component, so as to finally synthesize the target loss model that simultaneously covers the four dimensions of operating voltage, operating current, operating resistance and operating temperature.

[0018] Preferably, the loss and thermal resistance model includes a thermal characteristic model, then the process of constructing the thermal characteristic model in step S2 includes:

[0019] Step B1: Obtain the transient thermal resistance curve data from the power module datasheet, and calculate the time parameter corresponding to half of the positive half-cycle of the current sine wave based on the output frequency in the operating condition file.

[0020] Step B2: Find the sampling point corresponding to the time parameter in the transient thermal resistance curve data. If the time parameter is not a direct sampling point, select two adjacent sampling points before and after the time parameter and perform linear interpolation to fit the equivalent thermal resistance coefficient corresponding to the time parameter, thereby completing the construction of the thermal characteristic model for the current working condition.

[0021] Preferably, step S3 includes:

[0022] Step S31: Divide the switching frequency in the operating condition file by twice the output frequency to determine the number of voltage pulses in the positive half-cycle of the current sine wave, and calculate the instantaneous current point corresponding to each pulse based on the position information of each voltage pulse in the cycle.

[0023] Step S32: Substitute the instantaneous current point into the output characteristic model for interpolation to obtain the corresponding instantaneous on-state voltage drop, and combine it with the duty cycle corresponding to the pulse under the current operating condition to calculate the on-state loss component of a single pulse.

[0024] Step S33: Use the instantaneous current point to perform multidimensional interpolation processing in the loss characteristic model to obtain the switching energy consumption value corresponding to a single pulse, as the switching loss component.

[0025] Step S34: Sum the conduction loss components and switching loss components of all pulses in the positive half-cycle of the current to obtain the total loss of the power module.

[0026] Step S35: The total loss is used as a thermophysical input, and combined with the equivalent thermal resistance coefficient obtained by fitting the thermal characteristic model, it is used to calculate the junction temperature rise of the power module.

[0027] Preferably, step S4 includes:

[0028] Step S41: Multiply the total loss by the equivalent thermal resistance coefficient, and add the product to the case temperature read in real time from the operating condition file to obtain the preliminary junction temperature;

[0029] Step S42: The initial junction temperature is used as a feedback temperature parameter and fed back into the loss calculation process. The updated initial current loss curve and initial resistance loss curve are obtained by refitting through an interpolation algorithm.

[0030] Step S43: Calculate the updated total loss based on the refitted initial current loss curve and initial resistance loss curve, and recalculate the current junction temperature using the updated total loss, and calculate the absolute value of the difference between the current junction temperature and the junction temperature calculated last time.

[0031] Step S44: Determine whether the absolute value is less than a preset threshold.

[0032] If so, stop the iteration process and determine the current junction temperature obtained from the last calculation as the highest junction temperature of the power module;

[0033] If not, update the current junction temperature to the feedback temperature parameter and return to step S42 to continue execution until the convergence condition is met.

[0034] Preferably, the preset threshold in step S44 is 2°C.

[0035] Preferably, after performing step S4, the method further includes:

[0036] Step S5: Perform rainflow counting processing on the junction temperature curves formed by the multiple junction temperatures obtained from the calculation, and extract the temperature rise fluctuation characteristics for power module lifetime prediction.

[0037] Preferably, the rainflow counting process in step S5 includes: extracting all peak and trough values ​​in the junction temperature curve and calculating the difference between adjacent peak and trough values. When the difference is greater than a preset minimum peak-to-trough value range, the junction temperature curve is retained for lifetime prediction.

[0038] Preferably, the calculation process of the duty cycle corresponding to the pulse under the current operating condition in step S32 includes: selecting the corresponding calculation formula according to the modulation method selected by the user on the main interface, wherein the modulation method includes SPWM modulation or SVPWM modulation.

[0039] The above technical solution has the following advantages or beneficial effects:

[0040] 1. By calculating the pulse sequence within a half-cycle of the current and using the accumulation method to calculate the conduction and switching losses corresponding to each pulse, compared with the traditional integration method, this invention can more accurately capture the energy loss of each switching action during the modulation process, greatly improving the accuracy of the total loss calculation.

[0041] 2. By using interpolation algorithms to construct loss and thermal resistance models that dynamically change with temperature, current, and resistance, the computing system can overcome the limitations of discrete reference points in the datasheet and dynamically adjust the model characteristics according to real-time operating parameters, ensuring the reliability and universality of the calculation results under complex and variable operating conditions.

[0042] 3. A feedback loop mechanism for junction temperature and loss calculation was established, feeding the calculated preliminary junction temperature back to the loss model in real time for iterative correction until the results converge. This process fully considers the thermal-electric coupling effect of the power module, thus obtaining a maximum junction temperature that more closely matches the physically measured value.

[0043] 4. By automating the reading of operating condition files and module characteristic parameters, batch processing of massive amounts of data is achieved. When dealing with the large-scale data input required for subsequent life calculations, it not only improves calculation efficiency but also reduces the difficulty of manual maintenance for users through standardized summary reports. Attached Figure Description

[0044] Figure 1 A flowchart illustrating a method for predicting the junction temperature of a power module loss in a preferred embodiment of the present invention is shown below.

[0045] Figure 2 In a preferred embodiment of the present invention, a schematic diagram of the process of constructing the loss characteristic model in step S2 is shown.

[0046] Figure 3 In a preferred embodiment of the present invention, a schematic diagram of the construction process of the thermal characteristic model in step S2 is shown.

[0047] Figure 4 This is a schematic diagram of a sub-process of step S3 in a preferred embodiment of the present invention;

[0048] Figure 5 This is a schematic diagram of the sub-process of step S4 in a preferred embodiment of the present invention. Detailed Implementation

[0049] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. The present invention is not limited to this embodiment; other embodiments that conform to the spirit of the present invention may also fall within the scope of the present invention.

[0050] In a preferred embodiment of the present invention, based on the above-mentioned problems existing in the prior art, a method for predicting the junction temperature of a power module loss is provided, such as... Figure 1 As shown, it includes the following steps:

[0051] Step S1: Read the operating condition file containing voltage, current, temperature, frequency, modulation ratio and gate resistance, and import the characteristic parameters from the power module datasheet.

[0052] Step S2: Based on the characteristic parameters, construct a loss and thermal resistance model that dynamically changes with temperature, current and resistance using an interpolation algorithm.

[0053] Step S3: Calculate the pulse sequence within the current half-cycle. Based on the loss and thermal resistance model, calculate the conduction loss and switching loss corresponding to each pulse using the accumulation method, and then obtain the total loss.

[0054] Step S4: Calculate the preliminary junction temperature based on the loss and thermal resistance model, and feed this preliminary junction temperature back into the loss calculation process for iterative iteration until the difference between the two calculated junction temperatures is less than a preset threshold. Take the last calculation result as the highest junction temperature for staff to use in predicting the lifespan of the power module.

[0055] Specifically, this embodiment provides a method for predicting the junction temperature and losses of a power module, utilizing the Python programming language to automate the batch processing of massive operating condition files. First, the operating condition files are read, which require the following parameters: temperature, voltage, current, switching frequency, output frequency, modulation ratio, turn-on resistance, and turn-off resistance. Simultaneously, the characteristic parameter module from the power module's datasheet is imported. The datasheet must include four data tables: output characteristic curves for IGBTs and diodes, current loss curves, resistance loss curves, and thermal resistance curves. Next, based on these characteristic parameters, an interpolation algorithm is used to construct a dynamic loss and thermal resistance model that varies with temperature, current, and resistance. In the core calculation part, the program analyzes the pulse sequence within the current half-cycle and uses an accumulation method to accurately calculate the conduction and switching losses corresponding to each pulse, accumulating them to obtain the total loss, thus avoiding the errors of traditional integral methods under nonlinear modulation. Finally, a preliminary junction temperature is calculated based on the loss and thermal resistance model, and a feedback mechanism between junction temperature and loss is established for iterative calculation until the difference between two consecutive calculated junction temperatures is less than a preset threshold, thereby locking in the module's true maximum junction temperature under actual operating conditions.

[0056] The loss junction temperature prediction method in this embodiment has the following beneficial effects:

[0057] Improved calculation accuracy and precision: By calculating the pulse sequence within a half-cycle of the current and using the accumulation method to calculate the conduction and switching losses corresponding to each pulse, compared with the traditional integration method, this invention can more accurately capture the energy loss of each switching action during the modulation process, greatly improving the accuracy of the total loss calculation.

[0058] Multidimensional dynamic mapping of the model was achieved: the loss and thermal resistance model that dynamically changes with temperature, current and resistance was constructed using interpolation algorithms, which enabled the computing system to break through the limitations of discrete reference points in the datasheet and dynamically adjust the model characteristics according to real-time operating parameters, thus ensuring the reliability and universality of the calculation results under complex and variable operating conditions.

[0059] The accuracy of junction temperature prediction is significantly improved through closed-loop iteration: This invention establishes a feedback loop mechanism for junction temperature calculation and loss calculation, feeding back the calculated preliminary junction temperature to the loss model in real time for iterative correction until the result converges. This process fully considers the thermal-electric coupling effect of the power module, thereby obtaining a maximum junction temperature that more closely matches the physically measured value.

[0060] Significantly reducing data processing time costs: By automating the reading of operating condition files and module characteristic parameters, this invention enables batch processing of massive amounts of data. When dealing with the large-scale data input required for subsequent lifespan calculations, it not only improves computational efficiency but also reduces the difficulty of manual maintenance for users through standardized summary reports.

[0061] In a preferred embodiment of the present invention, the loss and thermal resistance model includes an output characteristic model, and the process of constructing the output characteristic model in step S2 includes:

[0062] First, the output characteristic curve data at different temperatures are extracted from the datasheet. Then, the linear interpolation method is used to interpolate between the normal temperature curve and the high temperature curve based on the real-time operating temperature, generating a function curve of saturated on-state voltage drop versus current that conforms to the operating temperature as the output characteristic model.

[0063] Specifically, in this embodiment, the program first extracts static output characteristic curve data at different temperatures (e.g., 25℃ and 150℃) from the datasheet. Then, using linear interpolation, it dynamically interpolates between the room temperature curve and the high temperature curve based on the real-time operating temperature. In this way, the program can generate a function curve of saturated on-state voltage drop versus current that perfectly matches the current real-time operating temperature, and defines it as the output characteristic model. This model provides a high-precision instantaneous voltage reference for subsequent pulse-level loss calculations.

[0064] In a preferred embodiment of the present invention, the loss and thermal resistance model includes a loss characteristic model. Therefore, the process of constructing the loss characteristic model in step S2 is as follows: Figure 2 The following are included:

[0065] Step A1: Obtain the current loss curve data and resistance loss curve data under normal temperature and high temperature conditions from the original data manual, and use the linear interpolation algorithm to perform temperature correction interpolation to fit the initial current loss curve and initial resistance loss curve at the corresponding operating temperature.

[0066] Step A2: If the operating current is within the range of the initial current loss curve, then determine the two sampling points on the curve that are closest to the operating current and perform linear interpolation. If the operating current is less than the minimum current in the initial current loss curve, then select the two sampling points at the front of the initial current loss curve for processing to obtain the loss value. If the operating current is greater than the maximum current in the initial current loss curve, then select the two sampling points at the end of the initial current loss curve for processing to obtain the loss value. Multiply the processed loss value by the ratio of the operating voltage to the specification voltage to obtain the intermediate loss value.

[0067] Step A3: Using the same interpolation method as in step A2, interpolate the resistance dimension of the initial resistance loss curve to obtain the first resistance loss component under the operating resistance and the second resistance loss component under the specification resistance.

[0068] Step A4: Divide the intermediate loss value by the second resistance loss component, and multiply the result by the first resistance loss component to finally synthesize the target loss model that simultaneously covers the four dimensions of operating voltage, operating current, operating resistance, and operating temperature.

[0069] Specifically, in this embodiment, the loss characteristic model is constructed by performing multi-dimensional processing on the original loss data. The core of this process is to perform bilinear interpolation fitting on the original data according to the actual operating conditions. The specific processing procedure is as follows:

[0070] First, temperature-dimensional fitting is performed. The program extracts current loss curve data (such as E) under normal and high temperature conditions from the original datasheet. on E off E rec The system also obtains the current loss curve data and the resistance loss curve data. A linear interpolation algorithm is used to perform temperature correction interpolation between normal temperature and high temperature references based on the real-time operating temperature, thereby fitting the initial current loss curve and the initial resistance loss curve at the corresponding operating temperature.

[0071] Secondly, current-dimensional interpolation and voltage correction are performed. The program performs piecewise linear interpolation on the obtained initial current loss curve:

[0072] Interpolation within the interval: If the operating current is within the numerical range of the initial current loss curve, then determine the two sampling points on the curve that are closest to the operating current and perform linear interpolation.

[0073] Boundary extrapolation processing: If the operating current is less than the minimum current value in the initial current loss curve, then the two sampling points at the very beginning of the curve (minimum value end) are selected for processing; if the operating current is greater than the maximum current value in the initial current loss curve, then the two sampling points at the end of the curve (maximum value end) are selected for processing. This selection method ensures the rationality and robustness of the interpolation calculation at the data range boundaries.

[0074] The loss value obtained by the above interpolation is further multiplied by the operating voltage and divided by the standard voltage in the datasheet to complete the correction of the voltage dimension, and finally the intermediate loss value (i.e. the loss ei1 under the operating current, voltage and resistance conditions) is obtained.

[0075] Subsequently, resistance dimension interpolation is performed. Using the same piecewise linear interpolation method and boundary processing logic as for the current dimension, the initial resistance loss curve at the operating temperature is interpolated. Through this step, the first resistance loss component (i.e., loss er1 under the operating resistance) and the second resistance loss component (i.e., loss er2 under the datasheet reference resistance) are obtained respectively.

[0076] Finally, multi-dimensional parameter fusion processing is performed. The program divides the intermediate loss value by the second resistance loss component and multiplies the result by the first resistance loss component (ei1 divided by er2 multiplied by er1). Through this ratio compensation logic, a target loss model that simultaneously covers four dimensions—operating voltage, operating current, operating resistance, and operating temperature—is finally synthesized. This model can accurately respond to the instantaneous energy consumption changes of power modules under complex power electronic operating environments.

[0077] In a preferred embodiment of the present invention, the loss and thermal resistance model includes a thermal characteristic model, and the process of constructing the thermal characteristic model in step S2 is as follows: Figure 3 As shown, it includes:

[0078] Step B1: Obtain the transient thermal resistance curve data from the power module datasheet, and calculate the time parameter corresponding to half of the positive half-cycle of the current sine wave based on the output frequency in the operating condition file.

[0079] Step B2: Find the sampling point corresponding to the time parameter in the transient thermal resistance curve data. If the time parameter is not a direct sampling point, select two adjacent sampling points before and after the time parameter and perform linear interpolation to fit the equivalent thermal resistance coefficient corresponding to the time parameter, thereby completing the construction of the thermal characteristic model for the current working condition.

[0080] Specifically, in this embodiment, the construction of the thermal characteristic model aims to provide accurate thermal resistance parameters for junction temperature calculation. The program reads the transient thermal resistance curve and fourth-order thermal resistance parameters from the power module datasheet. The steady-state thermal resistance is defined as the sum of the fourth-order thermal resistances. For the calculation of junction temperature fluctuations, the program needs to determine half of the positive half-cycle of the current sine wave (i.e., T / 2, where T is the output period, determined by the output frequency f). out The program retrieves the first thermal resistance sampling point with a time greater than T / 2 from the transient thermal resistance curve dataset. If this point is the nth point on the curve, it performs linear interpolation between this nth point and the previous n-1 point to fit the equivalent thermal resistance coefficient at a time precise to T / 2. This equivalent thermal resistance coefficient serves as the core output of the thermal characteristic model and is used for subsequent quantitative calculations of junction temperature fluctuations.

[0081] In a preferred embodiment of the present invention, such as Figure 4 As shown, step S3 includes:

[0082] Step S31: Divide the switching frequency in the operating condition file by twice the output frequency to determine the number of voltage pulses in the positive half-cycle of the current sine wave, and calculate the instantaneous current point corresponding to each pulse based on the position information of each voltage pulse in the cycle.

[0083] Step S32: Substitute the instantaneous current point into the output characteristic model for interpolation to obtain the corresponding instantaneous on-state voltage drop, and combine it with the duty cycle corresponding to the pulse under the current operating condition to calculate the conduction loss component of a single pulse.

[0084] Step S33: Use the instantaneous current point to perform multidimensional interpolation processing in the loss characteristic model to obtain the switching energy consumption value corresponding to a single pulse, which is used as the switching loss component.

[0085] Step S34: Sum the conduction loss components and switching loss components of all pulses in the positive half-cycle of the current to obtain the total loss of the power module.

[0086] Step S35: The total loss is used as a thermophysical input and combined with the equivalent thermal resistance coefficient obtained from the fitting in the thermal characteristic model to calculate the junction temperature rise of the power module.

[0087] In a preferred embodiment of the present invention, the calculation process of the duty cycle corresponding to the pulse under the current operating condition in step S32 includes: automatically selecting the corresponding calculation formula according to the modulation method selected by the user on the main interface, wherein the modulation method includes SPWM modulation or SVPWM modulation.

[0088] In this embodiment, the total loss P avIt is derived by adding the conduction loss and the switching loss. When constructing the output characteristic model, the program not only extracts the curve but also performs further parameterization: obtaining the intercept v of the straight line connecting the current value at the 1 / 3 and 2 / 3 points on the output characteristic curve. ce0 and slope r ce1 This serves as a characteristic parameter for calculating conduction losses. During the loss accumulation calculation process, the program reads key variables such as voltage, current, modulation ratio m, power factor cosφ, and output frequency in real time from the operating condition file. Switching loss (E on and E off The loss value is obtained through interpolation using the aforementioned loss characteristic model. It should be noted that the loss result obtained through interpolation at this time only corresponds to the initial case temperature under operating conditions and does not yet include the reaction effect of temperature rise caused by power consumption on the loss. Therefore, this loss value will be used as the initial input in the subsequent iteration process.

[0089] Furthermore, dividing the switching frequency by the output frequency and then by two yields the number of voltage pulses within the positive half-cycle of the current sine wave. Based on the voltage pulse position, the current point corresponding to each pulse can be calculated. Substituting this current point into the output characteristic curve yields the corresponding on-state voltage drop. The duty cycle of each pulse can be obtained based on the operating conditions. For example, when the modulation method is SPWM, the duty cycle is d = (1 + m * sin(2 * π * fout * n / fout + arccos(cosf)), where m is the modulation ratio, fout is the output frequency, n is the current point position, fsw is the switching frequency, and cosφ is the power factor. When the modulation method is SVPWM, the angular displacement a = ω * t = 2 * π * fout * n / fsw. When 0 < a <= π / 3, ... When π / 3<a<=2*π / 3, d=(1 / 2)*(1+3 / 2*mcos(a)), when 2*π / 3<=a<=π, When -π < a <= -2π / 3 When -2*π / 3<a<=-π / 3, d=(1 / 2)*(1+3 / 2*m*cos(a)), when -π / 3<=a<=0, The program will automatically determine whether the operation is stalled. In stalled operation, the duty cycle is 0.5. Then, the duty cycle is substituted into the formula to calculate the conduction loss using the cumulative method. The formula is: pic = vce0 * i * d / (fsw * / fout), where v... ce0Let i be the intercept of the straight line connecting the 1 / 3 and 2 / 3 points of the output characteristic curve, and let i be the current corresponding to each current point in the positive half-cycle. Sum the results to obtain the final conduction loss. Substitute the current points into the loss curve fitted according to the datasheet under the operating conditions to obtain the corresponding turn-on and turn-off losses. Sum the corresponding losses obtained from the interpolation of all current points and substitute them into the switching loss formula Psw=(eon+eoff)*fout to obtain the total switching loss. Substitute the calculated losses back into the junction temperature formula to obtain the final junction temperature.

[0090] Two modulation methods are available: SPWM and SVPWM. If SPWM is selected, the number of current points *n* within the positive half-cycle can be obtained by dividing the switching frequency *fsw* by the output frequency *fout* and then dividing by two. The accumulation range for this calculation is 0 to *n*, and the formula for the current at each point is as follows:

[0091] Where io is the output current, and ni

[0092] This represents the position of the current point within the accumulation range. For example, when calculating the first current point, ni is 0, and the second current point, ni is 1. If the modulation method is SVPWM, the starting point of the accumulation range is n1 = (-1)*((fsw / fout) / 4, and the ending point is n2 = (fsw / fout) / 4. The formula for the current at each point is... The position ni corresponding to the first current point is n1 minus 1, and the position corresponding to the last current point is n2 minus 1.

[0093] Simultaneously, the switching loss component is obtained in the loss characteristic model using this current point. Finally, the losses of all pulses in the sequence are summed to obtain the total loss. This process realizes the function mapping of the position-determined value, ensuring that the loss of each switching action is taken into account.

[0094] In a preferred embodiment of the present invention, such as Figure 5 As shown, step S4 includes:

[0095] Step S41: Multiply the total loss by the equivalent thermal resistance coefficient, and add the product to the shell temperature value read in real time from the operating condition file to obtain the preliminary junction temperature.

[0096] Step S42: The initial junction temperature is used as a feedback temperature parameter and fed back into the loss calculation process. The updated initial current loss curve and initial resistance loss curve are obtained by refitting through an interpolation algorithm.

[0097] Step S43: Calculate the updated total loss based on the refitted loss curve, and recalculate the current junction temperature using the updated total loss. Calculate the absolute value of the difference between the current junction temperature and the junction temperature calculated previously.

[0098] Step S44: Determine whether the absolute value is less than a preset threshold.

[0099] If so, stop the iteration process and determine the current junction temperature obtained from the last calculation as the highest junction temperature of the power module;

[0100] If not, update the current junction temperature to the feedback temperature parameter and return to step S42 to continue execution until the convergence condition is met.

[0101] In a preferred embodiment of the present invention, the preset threshold in step S44 is 2°C.

[0102] Specifically, in this embodiment, a junction temperature closed-loop iterative correction based on the thermo-electric coupling effect is implemented. First, substituting the obtained P... av intercept v ce0 Slope r ce1 Substitute parameters such as modulation ratio and equivalent thermal resistance coefficient into the preset formula. Calculate the highest junction temperature of the module, where P av For conduction loss P ic and switching loss P sw The total loss obtained by adding them together, of which Psw = (eon + eoff) * fout, where v ce0 r is the intercept of the straight line connecting the 1 / 3 point and the 2 / 3 point of the output characteristic curve current. ce1 The slope of the line is given by , m is the modulation ratio, cosφ is the power factor, and io is the output frequency. These three values ​​can be read from the operating condition file, combined with the case temperature T read from the operating condition file. c The preliminary junction temperature (E) is then obtained. on and E off The result obtained by interpolating the above losses represents the losses corresponding to the case temperature under operating conditions. It does not include the heat generated by power consumption, so subsequent junction temperature iteration is required to obtain the true temperature and the corresponding true losses. Rth(t=∞) is the steady-state thermal resistance, which is the sum of the fourth-order thermal resistances. To account for junction temperature fluctuations, find the first thermal resistance point in the thermal resistance curve where the time is greater than T / 2, and the value of T is the output frequency f. out Divide by 2. If this point is the nth point on the curve, use linear interpolation between the nth and n-1th points to obtain the thermal resistance at time T / 2. This thermal resistance is given by the formula above. Tc is the temperature read during operation, i.e., the case temperature.

[0103] Because the power consumption of the power module is strongly correlated with temperature, in order to obtain the true temperature and corresponding true loss, the program initiates an iterative mechanism: the calculated preliminary junction temperature is used as a feedback temperature parameter, and re-substituted into the interpolation process of the loss characteristic model to obtain the updated total loss and recalculate the junction temperature. The program continues to repeat this operation until the absolute value of the difference between the newly calculated junction temperature and the previous calculation result is less than 2℃ (a preset threshold). At this point, the result of the last iteration is taken as the highest junction temperature of the module, and it is re-substituted into the loss interpolation process for the final calculation to output high-precision loss and junction temperature data in the final summary report.

[0104] In a preferred embodiment of the present invention, after performing step S4, the method further includes:

[0105] Step S5: Perform rainflow counting processing on the junction temperature curves formed by the multiple calculated junction temperatures to extract the temperature rise fluctuation characteristics used for lifetime prediction.

[0106] In a preferred embodiment of the present invention, the rainflow counting process in step S5 includes: extracting all peak and trough values ​​in the junction temperature curve and calculating the difference between adjacent peak and trough values. When the difference is greater than the preset minimum peak-trough value range difference, the fluctuation data is retained for lifetime prediction.

[0107] Specifically, in this embodiment, after the program completes the junction temperature iterative calculation, it will generate the final calculation result set. The calculation results specifically include key parameters such as the switching loss, conduction loss, total loss, maximum junction temperature, and junction temperature fluctuation of the IGBT and diode.

[0108] To meet the demands of subsequent lifespan prediction (which may involve up to a million input operating conditions), this embodiment features highly efficient data management and processing capabilities. The program maps all input operating condition files to the calculated results and writes them uniformly into a data summary file. This method supports processing hundreds of thousands to millions of operating conditions in a single run, significantly improving processing efficiency.

[0109] After the data aggregation is completed, this embodiment automatically executes the lifespan preprocessing procedure:

[0110] The program automatically extracts the highest junction temperature data for IGBTs and diodes from the summary file and plots a dynamic curve of junction temperature change over time.

[0111] The program uses the rainflow counting method to discretize the junction temperature curve. The processing logic is as follows: starting from the beginning of the curve, the search proceeds, and moving the position forward simulates the start of rainflow.

[0112] Rainwater streams that begin flowing from a wave crest continue to flow toward subsequent wave troughs. If they encounter a wave crest that is higher than the initial wave crest, the rainwater stream stops flowing.

[0113] Rainwater streams originating from a trough continue flowing toward subsequent troughs. If they encounter a trough that is lower than the initial trough, the rainwater stream stops flowing.

[0114] The program records all peak and trough values ​​during the process and calculates the difference between adjacent peak and trough values. Then, it compares the difference with the minimum peak-to-trough difference preset by the user on the main interface. The program only retains data points where the difference is greater than the minimum peak-to-trough difference, thus eliminating the interference of minor fluctuations on the lifetime calculation.

[0115] The processed time-temperature characteristic data will be automatically saved in a new folder generated by the system. This data file, through simplification and feature extraction, directly provides high-value input data for the subsequent lifetime calculation module.

[0116] The above are merely preferred embodiments of the present invention and are not intended to limit the implementation methods and protection scope of the present invention. Those skilled in the art should recognize that any equivalent substitutions and obvious changes made using the content of this specification and illustrations should be included within the protection scope of the present invention.

Claims

1. A method for predicting the junction temperature of a power module, characterized in that, Includes the following steps: Step S1: Read the operating condition file containing voltage, current, temperature ratio and resistance, and import the characteristic parameters from the power module datasheet; Step S2: Based on the characteristic parameters and the operating condition file, construct a loss and thermal resistance model that dynamically changes with temperature, current and resistance using an interpolation algorithm; Step S3: Calculate the pulse sequence within the current half-cycle. Based on the loss and thermal resistance model, calculate the conduction loss and switching loss corresponding to each pulse using the accumulation method, and then obtain the total loss. Step S4: Calculate the preliminary junction temperature based on the loss and thermal resistance model, and feed the preliminary junction temperature back to the loss calculation process for iterative iteration until the difference between the two calculated junction temperatures is less than a preset threshold. Take the last calculation result as the highest junction temperature for staff to use in predicting the lifespan of the power module.

2. The method for predicting junction temperature loss according to claim 1, characterized in that, The loss and thermal resistance model includes an output characteristic model, so the process of constructing the output characteristic model in step S2 includes: First, the output characteristic curve data at different temperatures are extracted from the datasheet. Then, linear interpolation is used to interpolate between the normal temperature curve and the high temperature curve based on the real-time operating temperature, generating a function curve of saturated on-state voltage drop versus current that conforms to the operating temperature as the output characteristic model.

3. The method for predicting junction temperature loss according to claim 1, characterized in that, The loss and thermal resistance model includes a loss characteristic model, so the process of constructing the loss characteristic model in step S2 includes: Step A1: Obtain the current loss curve data and resistance loss curve data under normal temperature and high temperature conditions from the data manual, and use the linear interpolation algorithm to perform temperature correction interpolation to fit the initial current loss curve and initial resistance loss curve at the corresponding operating temperature. Step A2: If the operating current is within the range of the initial current loss curve, then determine the two sampling points on the curve that are closest to the operating current and perform linear interpolation. If the operating current is less than the minimum current in the initial current loss curve, then select the two sampling points at the front of the initial current loss curve for processing to obtain the loss value. If the operating current is greater than the maximum current in the initial current loss curve, then select the two sampling points at the end of the initial current loss curve for processing to obtain the loss value. Multiply the processed loss value by the ratio of the operating voltage to the specification voltage to obtain the intermediate loss value. Step A3: Using the same interpolation method as in step A2, interpolate the resistance dimension of the initial resistance loss curve to obtain the first resistance loss component under the operating resistance and the second resistance loss component under the specification resistance. Step A4: Divide the intermediate loss value by the second resistance loss component, and multiply the result by the first resistance loss component, so as to finally synthesize the target loss model that simultaneously covers the four dimensions of operating voltage, operating current, operating resistance and operating temperature.

4. The method for predicting junction temperature loss according to claim 1, characterized in that, The loss and thermal resistance model includes a thermal characteristic model, so the process of constructing the thermal characteristic model in step S2 includes: Step B1: Obtain the transient thermal resistance curve data from the power module datasheet, and calculate the time parameter corresponding to half of the positive half-cycle of the current sine wave based on the output frequency in the operating condition file. Step B2: Find the sampling point corresponding to the time parameter in the transient thermal resistance curve data. If the time parameter is not a direct sampling point, select two adjacent sampling points before and after the time parameter and perform linear interpolation to fit the equivalent thermal resistance coefficient corresponding to the time parameter, thereby completing the construction of the thermal characteristic model for the current working condition.

5. The method for predicting junction temperature loss according to claim 4, characterized in that, Step S3 includes: Step S31: Divide the switching frequency in the operating condition file by twice the output frequency to determine the number of voltage pulses in the positive half-cycle of the current sine wave, and calculate the instantaneous current point corresponding to each pulse based on the position information of each voltage pulse in the cycle. Step S32: Substitute the instantaneous current point into the output characteristic model for interpolation to obtain the corresponding instantaneous on-state voltage drop, and combine it with the duty cycle corresponding to the pulse under the current operating condition to calculate the on-state loss component of a single pulse. Step S33: Use the instantaneous current point to perform multidimensional interpolation processing in the loss characteristic model to obtain the switching energy consumption value corresponding to a single pulse, as the switching loss component. Step S34: Sum the conduction loss components and switching loss components of all pulses in the positive half-cycle of the current to obtain the total loss of the power module. Step S35: The total loss is used as a thermophysical input, and combined with the equivalent thermal resistance coefficient obtained by fitting the thermal characteristic model, it is used to calculate the junction temperature rise of the power module.

6. The method for predicting junction temperature loss according to claim 5, characterized in that, Step S4 includes: Step S41: Multiply the total loss by the equivalent thermal resistance coefficient, and add the product to the case temperature read in real time from the operating condition file to obtain the preliminary junction temperature; Step S42: The initial junction temperature is used as a feedback temperature parameter and fed back into the loss calculation process. The updated initial current loss curve and initial resistance loss curve are obtained by refitting through an interpolation algorithm. Step S43: Calculate the updated total loss based on the refitted initial current loss curve and initial resistance loss curve, and recalculate the current junction temperature using the updated total loss, and calculate the absolute value of the difference between the current junction temperature and the junction temperature calculated last time. Step S44: Determine whether the absolute value is less than a preset threshold. If so, stop the iteration process and determine the current junction temperature obtained from the last calculation as the highest junction temperature of the power module; If not, update the current junction temperature to the feedback temperature parameter and return to step S42 to continue execution until the convergence condition is met.

7. The method for predicting junction temperature loss according to claim 6, characterized in that, The preset threshold in step S44 is 2℃.

8. The method for predicting junction temperature loss according to claim 1, characterized in that, After performing step S4, the method further includes: Step S5: Perform rainflow counting processing on the junction temperature curves formed by the multiple junction temperatures obtained from the calculation, and extract the temperature rise fluctuation characteristics for power module lifetime prediction.

9. The method for predicting junction temperature loss according to claim 8, characterized in that, The rainflow counting process in step S5 includes: extracting all peak and trough values ​​in the junction temperature curve and calculating the difference between adjacent peak and trough values. When the difference is greater than the preset minimum peak-trough value range difference, the junction temperature curve is retained for lifetime prediction.

10. The method for predicting junction temperature loss according to claim 5, characterized in that, The calculation process of the duty cycle corresponding to the pulse under the current operating condition in step S32 includes: selecting the corresponding calculation formula according to the modulation method selected by the user on the main interface, wherein the modulation method includes SPWM modulation or SVPWM modulation.

Images