Processing method and device based on IGBT module, electronic equipment and medium

Abstract

The embodiment of the invention provides a processing method and device based on an IGBT module, electronic equipment and a medium, and the method comprises the steps: obtaining the current operation data of the IGBT module, inputting the current operation data into a pre-trained neural network model, and obtaining the first junction temperature information of the IGBT module; determining second junction temperature information of the IGBT module according to the power consumption change of the IGBT module; and optimizing the neural network model according to the first junction temperature information and the second junction temperature information so as to perform junction temperature prediction on the IGBT module by adopting the optimized neural network model. According to the embodiment of the invention, the method achieves the prediction of the junction temperature of the IGBT module through employing the neural network model, improves the prediction accuracy and real-time performance of the junction temperature of the IGBT module, can achieve the optimization of the neural network model according to the power consumption change of the IGBT module, and further improves the prediction accuracy.

Description

Technical Field

[0001] The present invention relates to the field of power electronics technology, and in particular to a processing method, device, electronic equipment and medium based on an IGBT module. Background Art

[0002] Insulated Gate Bipolar Transistor (IGBT) is a high-power semiconductor device that is widely used in electric vehicles, wind power generation, photovoltaic systems, high-speed railways, power grid systems and other fields, especially for high-voltage and high-current circuits. However, IGBT modules may face severe environments such as temperature rise, electrical shock, power cycling, etc. during long-term operation, which may cause the performance of IGBT modules to degrade or even fail.

[0003] During the operation of the IGBT module, the junction temperature is a key parameter, which directly reflects the safety margin, health status and operating performance of the device. In the prior art, the detection methods of the IGBT module junction temperature are mainly physical contact method, optical method, thermal model prediction method and temperature sensitive electrical parameter (TSEP) method.

[0004] However, the existing IGBT module junction temperature detection methods have certain limitations. For example, the physical contact method requires additional sensors and may interfere with equipment operation, the optical method is not suitable for online monitoring, and the accuracy of the temperature-sensitive electrical parameter method is limited by electrical characteristic parameters. These problems cannot meet the needs of accurate and real-time IGBT module junction temperature prediction in engineering practice. Summary of the invention

[0005] In view of the above problems, a processing method, device, electronic device and medium based on an IGBT module are proposed to overcome the above problems or at least partially solve the above problems, including:

[0006] A processing method based on an IGBT module, the method comprising:

[0007] Acquire current operating data of the IGBT module, and input the current operating data into a pre-trained neural network model to obtain first junction temperature information of the IGBT module;

[0008] Determining second junction temperature information of the IGBT module according to a change in power consumption of the IGBT module;

[0009] The neural network model is optimized according to the first junction temperature information and the second junction temperature information, so as to use the optimized neural network model to predict the junction temperature of the IGBT module.

[0010] Optionally, before determining the second junction temperature information of the IGBT module according to the power consumption change of the IGBT module, the method further includes:

[0011] The power consumption of the IGBT module is determined according to the conduction loss and the switching loss of the IGBT module, and the power consumption change of the IGBT module is determined according to the power consumption of the IGBT module at different times.

[0012] Optionally, determining the power consumption of the IGBT module according to the conduction loss and switching loss of the IGBT module includes:

[0013] The power consumption of the IGBT module is determined according to the conduction loss and switching loss of the IGBT module and the power consumption of a fast recovery diode connected in parallel with the IGBT module.

[0014] Optionally, before determining the power consumption of the IGBT module according to the conduction loss and switching loss of the IGBT module, the method further includes:

[0015] The conduction loss of the IGBT module is determined according to the forward voltage, equivalent resistance, load current, power factor, and phase angle of the power factor of the IGBT module.

[0016] Optionally, before determining the power consumption of the IGBT module according to the conduction loss and switching loss of the IGBT module, the method further includes:

[0017] The switching loss of the IGBT module is determined according to the switching frequency of the IGBT module, the energy loss when the switch is turned off, the energy loss when the switch is turned on, the DC link voltage, the nominal voltage, the load current, and the nominal current.

[0018] Optionally, determining the second junction temperature information of the IGBT module according to the power consumption change of the IGBT module includes:

[0019] The second junction temperature information of the IGBT module is determined according to the power consumption change of the IGBT module, the thermal resistance of the IGBT module, and the temperature of the heat sink.

[0020] Optionally, optimizing the neural network model according to the first junction temperature information and the second junction temperature information includes:

[0021] A prediction error is determined according to the first junction temperature information and the second junction temperature information, and the neural network model is optimized according to the prediction error.

[0022] Optionally, before acquiring the current operating data of the IGBT module and inputting the current operating data into a pre-trained neural network model to obtain the first junction temperature information of the IGBT module, the method further includes:

[0023] Obtain test operation data of IGBT modules under different working conditions;

[0024] The test run data is used to train the preset neural network model.

[0025] Optionally, before the test operation data is used to train the preset neural network model, the method further includes:

[0026] The test operation data is adjusted to simulate the aging condition of the IGBT module.

[0027] A processing device based on an IGBT module, the device comprising:

[0028] A first junction temperature information obtaining module, used to obtain current operating data of the IGBT module, and input the current operating data into a pre-trained neural network model to obtain first junction temperature information of the IGBT module;

[0029] A second junction temperature information determining module, used to determine second junction temperature information of the IGBT module according to a change in power consumption of the IGBT module;

[0030] The model optimization and junction temperature prediction module is used to optimize the neural network model according to the first junction temperature information and the second junction temperature information, so as to use the optimized neural network model to predict the junction temperature of the IGBT module.

[0031] An electronic device comprises a processor, a memory and a computer program stored in the memory and capable of running on the processor, wherein the computer program implements the method described above when executed by the processor.

[0032] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method described above is implemented.

[0033] The embodiments of the present invention have the following advantages:

[0034] In an embodiment of the present invention, the first junction temperature information of the IGBT module is obtained by acquiring the current operating data of the IGBT module and inputting the current operating data into a pre-trained neural network model; the second junction temperature information of the IGBT module is determined according to the power consumption change of the IGBT module; the neural network model is optimized according to the first junction temperature information and the second junction temperature information, so as to use the optimized neural network model to predict the junction temperature of the IGBT module, thereby realizing the use of the neural network model to predict the junction temperature of the IGBT module, improving the accuracy and real-time performance of the junction temperature prediction of the IGBT module, and being able to optimize the neural network model according to the power consumption change of the IGBT module, further improving the accuracy of the prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solution of the present invention, the accompanying drawings required for use in the description of the present invention will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative labor.

[0036] Figure 1 is a flowchart of a processing method based on an IGBT module provided by some embodiments of the present invention;

[0037] Figure 2 A neural network processing architecture diagram based on IGBT junction temperature prediction provided by some embodiments of the present invention;

[0038] Figure 3 is a schematic diagram of an IGBT junction temperature test architecture provided by some embodiments of the present invention;

[0039] Figure 4 A flowchart of another processing method based on an IGBT module provided by some embodiments of the present invention;

[0040] Figure 5 It is a structural block diagram of a processing device based on an IGBT module provided in some embodiments of the present invention. DETAILED DESCRIPTION

[0041] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0042] In the related technologies, the traditional IGBT junction temperature detection method has problems such as physical contact interference, difficulty in achieving online prediction through optical monitoring, inability of thermal models to consider the impact of device aging, and limitation of temperature-sensitive electrical parameters by electrical characteristic parameters. These problems cannot meet the demand for accurate and real-time IGBT junction temperature prediction in engineering practice. Moreover, since the IGBT junction temperature prediction is affected by multiple factors, such as direct current (DC), switching frequency, load current, and negative temperature coefficient (NTC) temperature, the prediction results are unstable and easily affected by factors such as device aging.

[0043] In practical applications, neural networks have powerful nonlinear modeling capabilities and can learn from multiple input parameters and predict the output of complex systems.

[0044] Therefore, the present invention proposes a high-precision, real-time IGBT junction temperature prediction method based on a neural network. By reasonably designing the network structure and training algorithm, and taking into account factors such as device aging and environmental changes, the high-precision prediction of the IGBT junction temperature can be achieved according to the neural network, and it can be applied in practical engineering to improve the reliability and performance of power electronic equipment.

[0045] Specifically, the present invention can have the following advantages:

[0046] 1. High-precision prediction: The present invention adopts an IGBT junction temperature prediction method based on a neural network, which utilizes the powerful nonlinear modeling ability of the neural network to learn and predict the IGBT junction temperature from multiple input parameters, thereby achieving high-precision junction temperature prediction.

[0047] 2. Real-time application: The neural network model has fast response speed and real-time performance, and can realize real-time IGBT junction temperature monitoring and prediction in engineering practice, providing timely and effective protection for the safe operation of power electronic equipment.

[0048] 3. Consideration of aging effects: The present invention can effectively take into account factors such as device aging effects and environmental changes. Through sufficient learning and training of training data, it can achieve accurate prediction of IGBT junction temperature, thereby improving the reliability and performance of the equipment.

[0049] The embodiments of the present invention are further described below with reference to the accompanying drawings:

[0050] Reference Figure 1 , shows a flowchart of a processing method based on an IGBT module provided by some embodiments of the present invention, which may specifically include the following steps:

[0051] Step 101, obtaining current operating data of the IGBT module, and inputting the current operating data into a pre-trained neural network model to obtain first junction temperature information of the IGBT module.

[0052] As some examples, the neural network model may be a Foster network model, which is pre-trained by acquiring operating data of the IGBT module; the first junction temperature information is chip junction temperature prediction information of the IGBT module.

[0053] As some examples, sensors can be pre-installed in the power system, which can be used to measure the collector-emitter on-state voltage (V ce,on ), chip junction temperature (T j ) and load current (I C ) devices; such as voltage probes, thermocouples, infrared cameras, and current sensors.

[0054] In practical applications, the trained neural network model can be embedded in the power system to collect the DC link voltage (V dc ), switching frequency (f sw ), load current (Ic) and temperature (T NTC ), input it into the pre-trained neural network model, and calculate the chip junction temperature prediction information of the IGBT module.

[0055] like Figure 2 , is a neural network processing architecture diagram based on IGBT junction temperature prediction provided by some embodiments of the present invention, wherein the design of the neural network model will adopt a BP neural network with two hidden layers; the input layer will include 4 nodes, corresponding to the DC link voltage (V dc ), switching frequency (f sw ), load current (I c ) and temperature (T NTC ) parameters, the output layer will include 1 node for predicting the IGBT junction temperature information (T j ).

[0056] In some embodiments of the present invention, before obtaining the current operating data of the IGBT module and inputting the current operating data into a pre-trained neural network model to obtain the first junction temperature information of the IGBT module, the method further includes:

[0057] Step 11, obtaining test operation data of the IGBT module under different working conditions;

[0058] In practical applications, an experimental test bench can be built in advance to collect test operation data of the IGBT module under different working conditions.

[0059] like Figure 3 , is a schematic diagram of an IGBT junction temperature test architecture provided by some embodiments of the present invention. An experimental test bench is pre-built, and the test bench includes: a control terminal, an infrared camera, an IGBT test box, a DC power supply, and an inductive load.

[0060] As some examples, the operating conditions include different ambient temperatures (35°C to 110°C) and different pulse currents (30A to 70A); these conditions are simulated by controlling the temperature of the IGBT module in the IGBT test box and applying pulse currents of different sizes; the test run data includes the collector current (I C ), collector-emitter conduction voltage (V ce,on ) and chip junction temperature (T j ), on the test bench, a sensor is installed to monitor the IGBT’s DC link voltage (V dc ), switching frequency (f sw ), load current (I c ) and the temperature of the negative temperature coefficient thermistor (NTC) (T NTC ) to obtain the test operation data of the IGBT module under different working conditions.

[0061] Step 12: Use the test operation data to train the preset neural network model.

[0062] As an example, the training of the neural network model is to divide the test run data into training set and test set data, and calculate the error of the output layer through the back propagation algorithm, back propagate the error to the front layer of the network, calculate the contribution of the weight and bias of each layer to the error, adjust the weight and bias according to the contribution value, and iterate this process until the network's prediction error reaches an acceptable level or reaches a predetermined training cycle, so that the neural network model can accurately predict the IGBT junction temperature.

[0063] As some examples, before training the preset neural network model, the test run data can also be preprocessed, such as data cleaning, noise removal, missing value filling, outlier detection and processing, and data normalization to ensure the quality and consistency of the data so that the range of input data is suitable for neural network processing; the test run data is then divided into a training set and a test set for use in training and verification of the neural network model.

[0064] As some examples, after the test run data is pre-processed, it can also be stored in a computer system in a digital format, and the stored test run data can be read through the written software code and sent as an input vector to the neural network model for training.

[0065] In some embodiments of the present invention, before using the test operation data to train the preset neural network model, the method further includes:

[0066] The test operation data is adjusted to simulate the aging condition of the IGBT module.

[0067] In practical applications, IGBT aging will affect the junction temperature prediction results during the training and verification process of the neural network model; therefore, aging parameters can be used or the training data set can be adjusted. For example, based on the experimental results and model performance feedback, a data subset that better represents the aging characteristics of the IGBT module can be selected, and different aging conditions can be simulated through data enhancement techniques (such as adding noise and performing data transformation). The preset neural network model is trained to improve the generalization and robustness of the model and the model's adaptability to aging effects, thereby improving the accuracy and reliability of the model's predictions.

[0068] Step 102: Determine second junction temperature information of the IGBT module according to the power consumption change of the IGBT module.

[0069] As some examples, the second junction temperature information of the IGBT module is determined by calculating the power consumption change of the IGBT module through a formula according to the current operation data of the IGBT module; wherein the second junction temperature information is theoretical IGBT module chip junction temperature information.

[0070] In practical applications, since the increase in chip junction temperature is related to the change in power consumption, the power consumption change of the IGBT module can be measured by the sensor in the power system. ce,on ) and load current (I C ) to estimate, that is, the power consumption change P = V ce,on ×I c .

[0071] In some embodiments of the present invention, determining the second junction temperature information of the IGBT module according to the power consumption change of the IGBT module includes:

[0072] The second junction temperature information of the IGBT module is determined according to the power consumption change of the IGBT module, the thermal resistance of the IGBT module, and the temperature of the heat sink.

[0073] In practical applications, after determining the power consumption change of the IGBT module, the thermal resistance of the IGBT module and the temperature of the heat sink can be obtained through sensors, and the second junction temperature information of the IGBT module can be calculated through a formula.

[0074] As some examples, the calculation expression of the second junction temperature information is:

[0075] T j =ΔP loss ·Z th +T c

[0076] Among them, ΔP loss is the power consumption change of the IGBT module, Z th is the temperature of the thermal resistance, T c is the temperature of the heat sink, T j It is the second junction temperature information.

[0077] In some embodiments of the present invention, before determining the second junction temperature information of the IGBT module according to the power consumption change of the IGBT module, the method further includes:

[0078] The power consumption of the IGBT module is determined according to the conduction loss and the switching loss of the IGBT module, and the power consumption change of the IGBT module is determined according to the power consumption of the IGBT module at different times.

[0079] As some examples, the conduction loss of the IGBT module can be calculated by the equivalent resistance, forward voltage, power factor, load current, and phase angle of the power factor of the IGBT module; the switching loss of the IGBT module can be accurately measured by a double pulse test on the voltage and current waveforms of the IGBT module during the switching process, and then the energy loss is calculated by integrating these waveforms as the switching loss of the IGBT module.

[0080] As some examples, after obtaining the conduction loss and switching loss of the IGBT module, the forward voltage drop (V f ) and the current flowing through the diode (I FRD ) to calculate the power consumption of the fast recovery diode, that is, the power consumption of the fast recovery diode FRD=V f ×I FRD , then determine the power consumption of the IGBT module, and record the power consumption of the IGBT module at different times, and determine the power consumption change of the IGBT module according to the power consumption of the IGBT module at different times.

[0081] In some embodiments of the present invention, determining the power consumption of the IGBT module according to the conduction loss and switching loss of the IGBT module includes:

[0082] The power consumption of the IGBT module is determined according to the conduction loss and switching loss of the IGBT module and the power consumption of a fast recovery diode connected in parallel with the IGBT module.

[0083] As some examples, the power dissipation of an IGBT module is calculated as:

[0084] P loss =PT loss +PD loss

[0085] Among them, PT loss is the power consumption of IGBT, PT loss and PD loss They are the power consumption of IGBT and fast recovery diode (FRD) respectively.

[0086] The power consumption of FRD refers to the power consumed by the fast recovery diode connected in parallel with the IGBT module when it is turned on. The power consumption of FRD can be measured by measuring the forward voltage drop of the diode (V f ) and the current flowing through the diode (I FRD ) to calculate, that is, FRD=V f ×I FRD .

[0087] In some embodiments of the present invention, before determining the power consumption of the IGBT module according to the conduction loss and switching loss of the IGBT module, the method further includes:

[0088] The conduction loss of the IGBT module is determined according to the forward voltage, equivalent resistance, load current, power factor, and phase angle of the power factor of the IGBT module.

[0089] As some examples, the conduction loss calculation expression for an IGBT is:

[0090] PT cond =V ce 2π(1+π4mcosφ)Ic+R T 8(1+83πmcosφ)I c

[0091] Among them, R T is the equivalent resistance of the IGBT module, V ce is the forward voltage of the IGBT, m is the power factor, I c is the load current, φ is the phase angle of the power factor, PT cond is the conduction loss of the IGBT; the power loss caused by the internal resistance of the module in the on state is calculated through the conduction loss expression of the IGBT, which is important for evaluating the thermal performance of the module and designing the heat dissipation system.

[0092] In some embodiments of the present invention, before determining the power consumption of the IGBT module according to the conduction loss and switching loss of the IGBT module, the method further includes:

[0093] The switching loss of the IGBT module is determined according to the switching frequency of the IGBT module, the energy loss when the switch is turned off, the energy loss when the switch is turned on, the DC link voltage, the nominal voltage, the load current, and the nominal current.

[0094] In some examples, the switching loss calculation expression of IGBT is:

[0095] PT sw =1πf sw (E on +E off )V dc V nom I c I nom

[0096] Among them, f sw is the switching frequency, E on and E off are the energy loss when turning on and off, V dc is the DC link voltage, V nom and I nom They are the nominal voltage and nominal current, PT sw is the switching loss of the IGBT. on and E off The energy loss is calculated by accurately measuring the voltage and current waveforms of the IGBT module during the switching process (double pulse test) and then integrating these waveforms to calculate the energy loss.

[0097] In practical applications, the voltage and current waveforms of the IGBT module during the switching process can be measured through a double pulse test, and the energy loss can be calculated by integrating these waveforms. The switching loss of the IGBT module can be determined based on the switching frequency of the IGBT module, the energy loss when the switch is turned off, the energy loss when the switch is turned on, the DC link voltage, the nominal voltage, the load current, and the nominal current.

[0098] Step 103: Optimize the neural network model according to the first junction temperature information and the second junction temperature information, so as to use the optimized neural network model to predict the junction temperature of the IGBT module.

[0099] In practical applications, the first junction temperature information predicted by the neural network model through the current operating data of the IGBT module and the calculated theoretical junction temperature information of the IGBT module chip can be used to make a judgment, so as to optimize the neural network model according to the judgment result, and use the optimized neural network model to predict the junction temperature of the IGBT module when predicting the junction temperature of the IGBT module next time, so as to provide effective monitoring and guarantee for the safe and stable operation of the power system.

[0100] In some embodiments of the present invention, optimizing the neural network model according to the first junction temperature information and the second junction temperature information includes:

[0101] A prediction error is determined according to the first junction temperature information and the second junction temperature information, and the neural network model is optimized according to the prediction error.

[0102] In practical applications, a judgment is made based on the first junction temperature information predicted by the neural network model through the current operating data of the IGBT module and the calculated second junction temperature information, and a prediction error value between the first junction temperature information and the second junction temperature information is predicted. The neural network model is optimized based on the prediction error value, so that the optimized neural network model can be used to predict the junction temperature of the IGBT module when the junction temperature of the IGBT module is predicted next time, thereby providing effective monitoring and protection for the safe and stable operation of the power system.

[0103] As some examples, in order to optimize the performance of a neural network model, the structure and parameters of the neural network model can be continuously adjusted, and methods such as cross-validation and parameter adjustment can be used to optimize the performance.

[0104] In some examples, it is necessary (and may be necessary) to try different numbers of hidden layers, the number of neurons in each layer, and adjust the connection mode of the neural network model, and then fine-tune the parameters, such as choosing an appropriate learning rate and optimization algorithm; it is also necessary to try different learning rate strategies, such as learning rate decay or using an adaptive learning rate optimizer, and choose different optimization algorithms such as stochastic gradient descent (SGD), adaptive motion estimation algorithm (Adam) or gradient descent algorithm (Root mean square propagation, RMSprop).

[0105] Next, cross-validation is implemented to evaluate the generalization ability of the model. The running data is divided into a data set, and the data set is divided into k subsets. One of them is used as a test set in turn, and the rest are used as training sets. Finally, the average of the k test results is taken as an evaluation of the performance of the neural network model.

[0106] In addition, it is necessary to refine the selection of performance indicators, such as the root mean square error (RMSE), mean absolute error (MAE), or R square value (R 2 ) and other indicators to comprehensively evaluate the predictive performance of the model.

[0107] Next, we conduct in-depth feature engineering analysis to analyze the collector-emitter on-state voltage (V ce,on ), chip junction temperature (T j ) and collector current (I C ) and other key parameters for feature selection, feature extraction, and feature transformation to enhance the model's sensitivity to input data.

[0108] During model training, apply regularization techniques such as L1 or L2 regularization, batch normalization to improve the generalization ability of the model and prevent overfitting. Use systematic methods for hyperparameter optimization, such as grid search, random search, or Bayesian optimization, to find the optimal hyperparameter combination.

[0109] Develop a model fusion strategy to combine the prediction results of multiple models through ensemble learning techniques such as bootstrap aggregating (Bagging), boosting method (Boosting) or stacked generalization (Stacking) to improve the overall performance.

[0110] Establish a continuous monitoring and model update mechanism, and regularly use newly collected data to retrain or fine-tune the model to ensure that the model can adapt to long-term changes in IGBT module performance.

[0111] In the experimental verification phase, the effectiveness of the optimization measures is verified through experiments, and the optimization strategy is adjusted according to the experimental results to form a feedback loop from experiment to model optimization; multi-scale modeling and feature fusion are considered to capture short-term and long-term change trends and improve the model's adaptability to data at different time scales.

[0112] Integrating these steps into an iterative optimization process ensures that the model can continue to provide accurate prediction results when faced with new data and conditions; through this coherent approach, the performance and reliability of the neural network model in predicting IGBT junction temperature can be systematically improved to adapt to the needs of IGBT junction temperature prediction under different operating conditions.

[0113] In some examples, a pulse current of known magnitude can be applied to the IGBT module at a specific ambient temperature, the collector-emitter on-state voltage can be measured using a specially designed circuit, and the chip junction temperature (T j ), and then use these measured values ​​as input data to predict the junction temperature through the neural network model. If the junction temperature predicted by the model is close to the actual measured junction temperature, it means that the model can accurately reflect the thermal behavior of the IGBT module under this working condition, that is, the junction temperature prediction.

[0114] In some embodiments of the present invention, the first junction temperature information of the IGBT module is obtained by acquiring the current operating data of the IGBT module and inputting the current operating data into a pre-trained neural network model; the second junction temperature information of the IGBT module is determined according to the power consumption change of the IGBT module; the neural network model is optimized according to the first junction temperature information and the second junction temperature information, so as to use the optimized neural network model to predict the junction temperature of the IGBT module, thereby realizing the use of the neural network model to predict the junction temperature of the IGBT module, improving the accuracy and real-time performance of the IGBT module junction temperature prediction, and being able to optimize the neural network model according to the power consumption change of the IGBT module, further improving the accuracy of the prediction.

[0115] Reference Figure 4 , shows a flowchart of another processing method based on an IGBT module provided by some embodiments of the present invention, which may specifically include the following steps:

[0116] Step 401, obtaining test operation data of the IGBT module under different working conditions.

[0117] In practical applications, an experimental test bench can be built in advance to collect test operation data of the IGBT module under different working conditions.

[0118] like Figure 3 , is a schematic diagram of an IGBT junction temperature test architecture provided by some embodiments of the present invention. An experimental test bench is pre-built, and the test bench includes: a control terminal, an infrared camera, an IGBT test box, a DC power supply, and an inductive load.

[0119] As some examples, the operating conditions include different ambient temperatures (35°C to 110°C) and different pulse currents (30A to 70A); these conditions are simulated by controlling the temperature of the IGBT module in the IGBT test box and applying pulse currents of different sizes; the test run data includes the collector current (IC), the collector-emitter on-state voltage (V ce,on ) and chip junction temperature (T j ), on the test bench, a sensor is installed to monitor the IGBT’s DC link voltage (V dc ), switching frequency (f sw ), load current (Ic), and temperature of the negative temperature coefficient thermistor (NTC) (T NTC ) to obtain the test operation data of the IGBT module under different working conditions.

[0120] Step 402: Adjust the test operation data to simulate the aging condition of the IGBT module.

[0121] In practical applications, IGBT aging will affect the junction temperature prediction results during the training and verification process of the neural network model; therefore, aging parameters can be used or the training data set can be adjusted. For example, based on the experimental results and model performance feedback, a data subset that better represents the aging characteristics of the IGBT module can be selected, and different aging conditions can be simulated through data enhancement techniques (such as adding noise and performing data transformation). The preset neural network model is trained to improve the generalization and robustness of the model and the model's adaptability to aging effects, thereby improving the accuracy and reliability of the model's predictions.

[0122] As some examples, the test run data can also be preprocessed, such as data cleaning, noise removal, missing value filling, outlier detection and processing, and data normalization to ensure data quality and consistency so that the range of input data is suitable for neural network processing; the test run data is then divided into a training set and a test set for use in training and verification of the neural network model.

[0123] Step 403: Use the test operation data to train the preset neural network model.

[0124] As an example, the training of the neural network model is to divide the test run data into training set and test set data, and calculate the error of the output layer through the back propagation algorithm, back propagate the error to the front layer of the network, calculate the contribution of the weight and bias of each layer to the error, adjust the weight and bias according to the contribution value, and iterate this process until the network's prediction error reaches an acceptable level or reaches a predetermined training cycle, so that the neural network model can accurately predict the IGBT junction temperature.

[0125] Step 404 , obtaining current operating data of the IGBT module, and inputting the current operating data into a pre-trained neural network model to obtain first junction temperature information of the IGBT module.

[0126] As some examples, the neural network model may be a Foster network model, which is pre-trained by acquiring operating data of the IGBT module; the first junction temperature information is chip junction temperature prediction information of the IGBT module.

[0127] As some examples, sensors can be pre-installed in the power system, which can be used to measure the collector-emitter on-state voltage (V ce,on ), chip junction temperature (T j ) and load current (I C ) devices; such as voltage probes, thermocouples, infrared cameras, and current sensors.

[0128] In practical applications, the trained neural network model can be embedded in the power system to collect the DC link voltage (V dc ), switching frequency (f sw ), load current (Ic) and temperature (T NTC ), input it into the pre-trained neural network model, and calculate the chip junction temperature prediction information of the IGBT module.

[0129] like Figure 2 , is a neural network processing architecture diagram based on IGBT junction temperature prediction provided by some embodiments of the present invention, wherein the design of the neural network model will adopt a BP neural network with two hidden layers; the input layer will include 4 nodes, corresponding to the DC link voltage (V dc ), switching frequency (f sw ), load current (I c ) and temperature (T NTC ) parameters, the output layer will include 1 node for predicting the IGBT junction temperature information (T j ).

[0130] Step 405 , determining the power consumption of the IGBT module according to the conduction loss and switching loss of the IGBT module, and determining the power consumption change of the IGBT module according to the power consumption of the IGBT module at different times.

[0131] As some examples, the conduction loss of the IGBT module can be calculated by the equivalent resistance, forward voltage, power factor, load current, and phase angle of the power factor of the IGBT module; the switching loss of the IGBT module can be accurately measured by a double pulse test on the voltage and current waveforms of the IGBT module during the switching process, and then the energy loss is calculated by integrating these waveforms as the switching loss of the IGBT module.

[0132] As some examples, after obtaining the conduction loss and switching loss of the IGBT module, the forward voltage drop (V f ) and the current flowing through the diode (I FRD ) to calculate the power consumption of the fast recovery diode, that is, the power consumption of the fast recovery diode FRD=V f ×I FRD , then determine the power consumption of the IGBT module, and record the power consumption of the IGBT module at different times, and determine the power consumption change of the IGBT module according to the power consumption of the IGBT module at different times.

[0133] In some embodiments of the present invention, before determining the power consumption of the IGBT module according to the conduction loss and switching loss of the IGBT module, the method further includes:

[0134] The conduction loss of the IGBT module is determined according to the forward voltage, equivalent resistance, load current, power factor, and phase angle of the power factor of the IGBT module.

[0135] As some examples, the conduction loss calculation expression for an IGBT is:

[0136] PT cond =V ce 2π(1+π4mcosφ)Ic+R T 8(1+83πmcosφ)I c

[0137] Among them, R T is the equivalent resistance of the IGBT module, V ce is the forward voltage of the IGBT, m is the power factor, I c is the load current, φ is the phase angle of the power factor, PT cond is the conduction loss of the IGBT; the power loss caused by the internal resistance of the module in the on state is calculated through the conduction loss expression of the IGBT, which is important for evaluating the thermal performance of the module and designing the heat dissipation system.

[0138] In some embodiments of the present invention, before determining the power consumption of the IGBT module according to the conduction loss and switching loss of the IGBT module, the method further includes:

[0139] The switching loss of the IGBT module is determined according to the switching frequency of the IGBT module, the energy loss when the switch is turned off, the energy loss when the switch is turned on, the DC link voltage, the nominal voltage, the load current, and the nominal current.

[0140] In some examples, the switching loss calculation expression of IGBT is:

[0141] PT sw =1πf sw (E on +E off )V dc V nom I c I nom

[0142] Among them, f sw is the switching frequency, E on and E off are the energy loss when turning on and off, V dc is the DC link voltage, V nom and I nom They are the nominal voltage and nominal current, PT swis the switching loss of the IGBT. on and E off The energy loss is calculated by accurately measuring the voltage and current waveforms of the IGBT module during the switching process (double pulse test) and then integrating these waveforms to calculate the energy loss.

[0143] In practical applications, the voltage and current waveforms of the IGBT module during the switching process can be measured through a double pulse test, and the energy loss can be calculated by integrating these waveforms. The switching loss of the IGBT module can be determined based on the switching frequency of the IGBT module, the energy loss when the switch is turned off, the energy loss when the switch is turned on, the DC link voltage, the nominal voltage, the load current, and the nominal current.

[0144] In some embodiments of the present invention, determining the power consumption of the IGBT module according to the conduction loss and switching loss of the IGBT module includes:

[0145] The power consumption of the IGBT module is determined according to the conduction loss and switching loss of the IGBT module and the power consumption of a fast recovery diode connected in parallel with the IGBT module.

[0146] As some examples, the power dissipation of an IGBT module is calculated as:

[0147] P loss =PT loss +PD loss

[0148] Among them, PT loss is the power consumption of IGBT, PT loss and PD loss They are the power consumption of IGBT and fast recovery diode (FRD) respectively.

[0149] The power consumption of FRD refers to the power consumed by the fast recovery diode connected in parallel with the IGBT module when it is turned on. The power consumption of FRD can be measured by measuring the forward voltage drop of the diode (V f ) and the current flowing through the diode (I FRD ) to calculate, that is, FRD=V f ×I FRD .

[0150] Step 406: Determine second junction temperature information of the IGBT module according to the power consumption change of the IGBT module.

[0151] As some examples, the second junction temperature information of the IGBT module is determined by calculating the power consumption change of the IGBT module through a formula according to the current operation data of the IGBT module; wherein the second junction temperature information is theoretical IGBT module chip junction temperature information.

[0152] In practical applications, since the increase in chip junction temperature is related to the change in power consumption, the power consumption change of the IGBT module can be measured by the sensor in the power system. ce,on ) and load current (I C ) to estimate, that is, the power consumption change P = V ce,on ×I c .

[0153] In some embodiments of the present invention, determining the second junction temperature information of the IGBT module according to the power consumption change of the IGBT module includes:

[0154] The second junction temperature information of the IGBT module is determined according to the power consumption change of the IGBT module, the thermal resistance of the IGBT module, and the temperature of the heat sink.

[0155] In practical applications, after determining the power consumption change of the IGBT module, the thermal resistance of the IGBT module and the temperature of the heat sink can be obtained through sensors, and the second junction temperature information of the IGBT module can be calculated through a formula.

[0156] As some examples, the calculation expression of the second junction temperature information is:

[0157] T j =ΔP loss ·Z th +T c

[0158] Among them, ΔP loss is the power consumption change of the IGBT module, Z th is the temperature of the thermal resistance, T c is the temperature of the heat sink, T j It is the second junction temperature information.

[0159] Step 407: Optimize the neural network model according to the first junction temperature information and the second junction temperature information, so as to use the optimized neural network model to predict the junction temperature of the IGBT module.

[0160] In practical applications, the first junction temperature information predicted by the neural network model through the current operating data of the IGBT module and the calculated theoretical junction temperature information of the IGBT module chip can be used to make a judgment, so as to optimize the neural network model according to the judgment result, and use the optimized neural network model to predict the junction temperature of the IGBT module when predicting the junction temperature of the IGBT module next time, so as to provide effective monitoring and guarantee for the safe and stable operation of the power system.

[0161] In some embodiments of the present invention, optimizing the neural network model according to the first junction temperature information and the second junction temperature information includes:

[0162] A prediction error is determined according to the first junction temperature information and the second junction temperature information, and the neural network model is optimized according to the prediction error.

[0163] In practical applications, a judgment is made based on the first junction temperature information predicted by the neural network model through the current operating data of the IGBT module and the calculated second junction temperature information, and a prediction error value between the first junction temperature information and the second junction temperature information is predicted. The neural network model is optimized based on the prediction error value, so that the optimized neural network model can be used to predict the junction temperature of the IGBT module when the junction temperature of the IGBT module is predicted next time, thereby providing effective monitoring and protection for the safe and stable operation of the power system.

[0164] As some examples, in order to optimize the performance of a neural network model, the structure and parameters of the neural network model can be continuously adjusted, and methods such as cross-validation and parameter adjustment can be used to optimize the performance.

[0165] In some examples, it is necessary (and may be necessary) to try different numbers of hidden layers, the number of neurons in each layer, and adjust the connection mode of the neural network model, and then fine-tune the parameters, such as choosing an appropriate learning rate and optimization algorithm; it is also necessary to try different learning rate strategies, such as learning rate decay or using an adaptive learning rate optimizer, and choose different optimization algorithms such as stochastic gradient descent (SGD), adaptive motion estimation algorithm (Adam) or gradient descent algorithm (Root mean square propagation, RMSprop).

[0166] Next, cross-validation is implemented to evaluate the generalization ability of the model. The running data is divided into a data set, and the data set is divided into k subsets. One of them is used as a test set in turn, and the rest are used as training sets. Finally, the average of the k test results is taken as an evaluation of the performance of the neural network model.

[0167] In addition, it is necessary to refine the selection of performance indicators, such as the root mean square error (RMSE), mean absolute error (MAE), or R square value (R 2 ) and other indicators to comprehensively evaluate the predictive performance of the model.

[0168] Next, we conduct in-depth feature engineering analysis to analyze the collector-emitter on-state voltage (V ce,on ), chip junction temperature (T j ) and collector current (I C ) and other key parameters for feature selection, feature extraction, and feature transformation to enhance the model's sensitivity to input data.

[0169] During model training, apply regularization techniques such as L1 or L2 regularization, batch normalization to improve the generalization ability of the model and prevent overfitting. Use systematic methods for hyperparameter optimization, such as grid search, random search, or Bayesian optimization, to find the optimal hyperparameter combination.

[0170] Develop a model fusion strategy to combine the prediction results of multiple models through ensemble learning techniques such as bootstrap aggregating (Bagging), boosting method (Boosting) or stacked generalization (Stacking) to improve the overall performance.

[0171] Establish a continuous monitoring and model update mechanism, and regularly use newly collected data to retrain or fine-tune the model to ensure that the model can adapt to long-term changes in IGBT module performance.

[0172] In the experimental verification phase, the effectiveness of the optimization measures is verified through experiments, and the optimization strategy is adjusted according to the experimental results to form a feedback loop from experiment to model optimization; multi-scale modeling and feature fusion are considered to capture short-term and long-term change trends and improve the model's adaptability to data at different time scales.

[0173] Integrating these steps into an iterative optimization process ensures that the model can continue to provide accurate prediction results when faced with new data and conditions; through this coherent approach, the performance and reliability of the neural network model in predicting IGBT junction temperature can be systematically improved to adapt to the needs of IGBT junction temperature prediction under different operating conditions.

[0174] In some examples, a pulse current of known magnitude can be applied to the IGBT module at a specific ambient temperature, the collector-emitter on-state voltage can be measured using a specially designed circuit, and the chip junction temperature (T j ), and then use these measured values ​​as input data to predict the junction temperature through the neural network model. If the junction temperature predicted by the model is close to the actual measured junction temperature, it means that the model can accurately reflect the thermal behavior of the IGBT module under this working condition, that is, the junction temperature prediction.

[0175] In some embodiments of the present invention, the first junction temperature information of the IGBT module is obtained by acquiring the current operating data of the IGBT module and inputting the current operating data into a pre-trained neural network model; the second junction temperature information of the IGBT module is determined according to the power consumption change of the IGBT module; the neural network model is optimized according to the first junction temperature information and the second junction temperature information, so as to use the optimized neural network model to predict the junction temperature of the IGBT module, thereby realizing the use of the neural network model to predict the junction temperature of the IGBT module, improving the accuracy and real-time performance of the IGBT module junction temperature prediction, and being able to optimize the neural network model according to the power consumption change of the IGBT module, further improving the accuracy of the prediction.

[0176] It should be noted that, for the sake of simplicity, the method embodiments are described as a series of action combinations, but those skilled in the art should be aware that the embodiments of the present invention are not limited by the order of the actions described, because according to the embodiments of the present invention, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions involved are not necessarily required by the embodiments of the present invention.

[0177] Reference Figure 5 , shows a schematic diagram of the structure of a processing device based on an IGBT module provided by some embodiments of the present invention, which may specifically include the following modules:

[0178] A first junction temperature information obtaining module 501 is used to obtain current operating data of the IGBT module and input the current operating data into a pre-trained neural network model to obtain first junction temperature information of the IGBT module;

[0179] A second junction temperature information determining module 502, configured to determine second junction temperature information of the IGBT module according to a change in power consumption of the IGBT module;

[0180] The model optimization and junction temperature prediction module 503 is used to optimize the neural network model according to the first junction temperature information and the second junction temperature information, so as to use the optimized neural network model to predict the junction temperature of the IGBT module.

[0181] In some embodiments of the present invention, the device further comprises:

[0182] The power consumption change determination module is used to determine the power consumption of the IGBT module according to the conduction loss and switching loss of the IGBT module, and to determine the power consumption change of the IGBT module according to the power consumption of the IGBT module at different times.

[0183] In some embodiments of the present invention, the power consumption change determination module includes:

[0184] The power consumption determination submodule is used to determine the power consumption of the IGBT module according to the conduction loss and switching loss of the IGBT module and the power consumption of the fast recovery diode connected in parallel with the IGBT module.

[0185] In some embodiments of the present invention, the power consumption change determination module further includes:

[0186] The conduction loss determination submodule is used to determine the conduction loss of the IGBT module according to the forward voltage, equivalent resistance, load current, power factor, and phase angle of the power factor of the IGBT module.

[0187] In some embodiments of the present invention, the power consumption change determination module further includes:

[0188] The switching loss determination submodule is used to determine the switching loss of the IGBT module according to the switching frequency of the IGBT module, the energy loss when the switch is turned off, the energy loss when the switch is turned on, the DC link voltage, the nominal voltage, the load current, and the nominal current.

[0189] In some embodiments of the present invention, the second junction temperature information determining module 302 includes:

[0190] The second junction temperature information determination submodule is used to determine the second junction temperature information of the IGBT module according to the power consumption change of the IGBT module, the thermal resistance of the IGBT module, and the temperature of the heat sink.

[0191] In some embodiments of the present invention, the model optimization and junction temperature prediction module 303 includes:

[0192] A prediction error determination submodule is used to determine a prediction error according to the first junction temperature information and the second junction temperature information, and optimize the neural network model according to the prediction error.

[0193] In some embodiments of the present invention, the device further comprises:

[0194] A test operation data acquisition module is used to acquire the test operation data of the IGBT module under different working conditions;

[0195] The neural network model training module is used to train the preset neural network model using the test operation data.

[0196] In some embodiments of the present invention, the device further comprises:

[0197] The data adjustment module is used to adjust the test operation data to simulate the aging condition of the IGBT module.

[0198] In some embodiments of the present invention, the first junction temperature information of the IGBT module is obtained by acquiring the current operating data of the IGBT module and inputting the current operating data into a pre-trained neural network model; the second junction temperature information of the IGBT module is determined according to the power consumption change of the IGBT module; the neural network model is optimized according to the first junction temperature information and the second junction temperature information, so as to use the optimized neural network model to predict the junction temperature of the IGBT module, thereby realizing the use of the neural network model to predict the junction temperature of the IGBT module, improving the accuracy and real-time performance of the IGBT module junction temperature prediction, and being able to optimize the neural network model according to the power consumption change of the IGBT module, further improving the accuracy of the prediction.

[0199] Some embodiments of the present invention further provide an electronic device, comprising a processor, a memory, and a computer program stored in the memory and capable of running on the processor, wherein the above method is implemented when the computer program is executed by the processor.

[0200] Some embodiments of the present invention further provide a computer-readable storage medium, on which a computer program is stored, and the computer program implements the above method when executed by a processor.

[0201] Some embodiments of the present invention further provide a computer program product, including a computer program, which implements the above method when executed by a processor.

[0202] As for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0203] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0204] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0205] Those skilled in the art will appreciate that the embodiments of the present invention may be provided as methods, devices, or computer program products. Therefore, the embodiments of the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the embodiments of the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0206] The embodiments of the present invention are described with reference to the flowcharts and / or block diagrams of the methods, terminal devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing terminal device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0207] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing terminal device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0208] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device so that a series of operating steps are executed on the computer or other programmable terminal device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable terminal device to implement the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0209] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the embodiments of the present invention.

[0210] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or terminal device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or terminal device. In the absence of further restrictions, the elements defined by the sentence "including one..." do not exclude the existence of other identical elements in the process, method, article or terminal device including the above elements.

[0211] The above is a detailed introduction to a processing method, device, electronic device and medium based on an IGBT module. Specific examples are used in this article to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A processing method based on an IGBT module, characterized in that: The method comprises: Acquire current operating data of the IGBT module, and input the current operating data into a pre-trained neural network model to obtain first junction temperature information of the IGBT module; Determining second junction temperature information of the IGBT module according to a change in power consumption of the IGBT module; The neural network model is optimized according to the first junction temperature information and the second junction temperature information, so as to use the optimized neural network model to predict the junction temperature of the IGBT module.

2. The method according to claim 1, characterized in that Before determining the second junction temperature information of the IGBT module according to the power consumption change of the IGBT module, the method further includes: The power consumption of the IGBT module is determined according to the conduction loss and the switching loss of the IGBT module, and the power consumption change of the IGBT module is determined according to the power consumption of the IGBT module at different times.

3. The method according to claim 2, characterized in that The step of determining the power consumption of the IGBT module according to the conduction loss and the switching loss of the IGBT module comprises: The power consumption of the IGBT module is determined according to the conduction loss and switching loss of the IGBT module and the power consumption of a fast recovery diode connected in parallel with the IGBT module.

4. The method according to claim 2, characterized in that: Before determining the power consumption of the IGBT module according to the conduction loss and switching loss of the IGBT module, the method further includes: The conduction loss of the IGBT module is determined according to the forward voltage, equivalent resistance, load current, power factor, and phase angle of the power factor of the IGBT module.

5. The method according to claim 2, characterized in that: Before determining the power consumption of the IGBT module according to the conduction loss and switching loss of the IGBT module, the method further includes: The switching loss of the IGBT module is determined according to the switching frequency of the IGBT module, the energy loss when the switch is turned off, the energy loss when the switch is turned on, the DC link voltage, the nominal voltage, the load current, and the nominal current.

6. The method according to any one of claims 2 to 5, characterized in that: The determining the second junction temperature information of the IGBT module according to the power consumption change of the IGBT module includes: The second junction temperature information of the IGBT module is determined according to the power consumption change of the IGBT module, the thermal resistance of the IGBT module, and the temperature of the heat sink.

7. The method according to claim 1, characterized in that The optimizing the neural network model according to the first junction temperature information and the second junction temperature information includes: A prediction error is determined according to the first junction temperature information and the second junction temperature information, and the neural network model is optimized according to the prediction error.

8. The method according to claim 1, characterized in that Before obtaining the current operation data of the IGBT module and inputting the current operation data into a pre-trained neural network model to obtain the first junction temperature information of the IGBT module, the method further includes: Obtain test operation data of IGBT modules under different working conditions; The test run data is used to train the preset neural network model.

9. The method according to claim 8, characterized in that Before the test operation data is used to train the preset neural network model, the method further includes: The test operation data is adjusted to simulate the aging condition of the IGBT module.

10. A processing device based on an IGBT module, characterized in that: The device comprises: A first junction temperature information obtaining module, used to obtain current operating data of the IGBT module, and input the current operating data into a pre-trained neural network model to obtain first junction temperature information of the IGBT module; A second junction temperature information determining module, used to determine second junction temperature information of the IGBT module according to a change in power consumption of the IGBT module; The model optimization and junction temperature prediction module is used to optimize the neural network model according to the first junction temperature information and the second junction temperature information, so as to use the optimized neural network model to predict the junction temperature of the IGBT module.

11. An electronic device, characterized in that: The method comprises a processor, a memory, and a computer program stored in the memory and capable of running on the processor, wherein the computer program implements the method according to any one of claims 1 to 9 when executed by the processor.

12. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 9 is implemented.

Images