An IGBT module junction temperature prediction method and system based on an LSTM hot neural network

Abstract

This invention discloses a method and system for predicting IGBT module junction temperature based on an LSTM thermal neural network, belonging to the field of IGBT application technology. It includes: analyzing the unsteady-state heat transfer characteristics of the multilayer materials inside the IGBT module and constructing a Discrete Heat Transfer (DHT) equation; generating a training dataset containing temperature response data under different power loss conditions through finite element (FEM) simulation; constructing an LSTM thermal neural network junction temperature prediction model with an embedded physical layer, the physical layer containing sub-neural networks to extract the mapping relationship between input features and thermal resistance and thermal capacity; training the model using the training dataset, transforming the DHT equation as a physical constraint into dynamic constraints for model parameters, thereby achieving the prediction of IGBT module junction temperature under different operating conditions. This invention overcomes the limitations of existing methods by integrating the physical heat transfer process with the LSTM network, achieving accurate and rapid prediction of IGBT module junction temperature, effectively improving the reliability and economy of IGBT module thermal management.

Description

Technical Field

[0001] This invention belongs to the field of IGBT application technology, specifically relating to a method and system for predicting the junction temperature of an IGBT module based on an LSTM thermal neural network. Background Technology

[0002] Currently, there are four common methods for predicting IGBT junction temperature: contact sensor method, infrared thermal imaging method, thermal network method, and thermosensitive electrical parameter method. The contact sensor method uses a temperature sensor to directly contact the monitoring points inside the IGBT module. The infrared thermal imaging method is a non-contact optical measurement method that relies on infrared equipment to measure the monitoring points inside the specially treated IGBT module. The thermal network method constructs a thermal impedance network based on the equivalent heat dissipation method of the IGBT module's multi-layer material structure to estimate the junction temperature. The thermosensitive electrical parameter method derives the junction temperature by working backward from the relationship between the device's junction temperature and external electrical parameters. However, in practical applications, IGBT module manufacturers prefer to calculate the junction temperature of their IGBT products using an offline system.

[0003] To address this challenge, a method for predicting the junction temperature of IGBT modules based on an LSTM thermal neural network was devised. Summary of the Invention

[0004] The technical problem to be solved by this invention is to address the shortcomings of the prior art by providing an IGBT module junction temperature prediction method and system based on LSTM thermal neural network. By combining the LSTM thermal neural network junction temperature prediction model with the physical heat transfer process, the limitations of existing methods are overcome, and accurate and rapid prediction of IGBT module junction temperature is achieved. This solves the technical problems of existing IGBT junction temperature prediction methods, such as intrusiveness, complex operation, high cost, difficulty in dynamically reflecting nonlinear heat transfer characteristics, weak generalization ability, and inaccurate offline prediction.

[0005] The present invention adopts the following technical solution:

[0006] A method for predicting the junction temperature of an IGBT module based on an LSTM thermal neural network includes the following steps:

[0007] S1. Analyze the unsteady heat transfer characteristics of the multilayer materials inside the IGBT module, and construct the DHT equation based on the structure and material properties;

[0008] S2. Generate a training dataset containing temperature response data under different power loss conditions through FEM simulation;

[0009] S3. Construct an LSTM thermal neural network junction temperature prediction model, embedding a physical layer into the LSTM thermal neural network junction temperature prediction model. The physical layer contains two sub-neural networks.R net and C net These are used to extract the mapping relationship between input features and thermal resistance and thermal capacity, respectively;

[0010] S4. Train the constructed LSTM thermal neural network junction temperature prediction model using the training dataset, and embed the DHT equation as a physical constraint of the heat transfer process into the network, transforming it into a dynamic constraint condition for the parameters of the LSTM thermal neural network junction temperature prediction model. Use the trained thermal neural network junction temperature prediction model to predict the junction temperature of the IGBT module under different operating conditions.

[0011] Preferably, in step S1, the DHT equation is:

[0012]

[0013] in, T i ( t ) is a node i Temperature, node i and j There is thermal resistance between them. R ij ( t ) and heat capacity C ij ( t This represents the heat transfer relationship between two nodes. T ij ( t () represents the temperature difference between two nodes. T a This refers to ambient temperature or coolant temperature. P i ( t ) is the node through which the flow passes. i Power loss, using matrix M To describe the connectivity of nodes.

[0014] Preferably, step S2 specifically includes:

[0015] A FEM simulation model including the geometry, material properties, and boundary conditions of the IGBT module is constructed. Temperature monitoring points are set at the center of the IGBT module chip, the center of the chip solder layer, the center of the substrate solder layer, and the center of the copper substrate. Different power losses are applied to simulate different operating conditions in actual operation. The temperature response curves of each monitoring point from the initial moment to the steady state are recorded with high time resolution and integrated to form a training dataset.

[0016] Preferably, in step S3, the sub-neural network R net and Cnet Output dynamic thermal resistance R ij ( t ) and dynamic heat capacity C ij ( t The dynamic thermal resistance and thermal capacity are concatenated with the input features to form an enhanced feature vector. z ( t This will enhance the feature vector. z ( t The input is fed into the LSTM-DHT layer for time series modeling.

[0017] Preferably, sub-neural network R net and C net Output for:

[0018]

[0019] Preferably, sub-neural network R net and C net Output dynamic thermal resistance R ij ( t ) and dynamic heat capacity C ij ( t They are respectively:

[0020]

[0021]

[0022] in, P ( t ) are the input features.

[0023] Preferably, in step S4, the output of the LSTM layer is mapped to the target temperature through a fully connected layer, and the prediction result of the sub-neural network is substituted into the DHT equation to calculate the node temperature at the next moment, thereby realizing real-time prediction of the IGBT module temperature.

[0024] Preferably, the output of the LSTM layer is mapped to the target temperature through a fully connected layer. Specifically:

[0025]

[0026] in, It is a fully connected layer. for t The output of the LSTM network at that moment.

[0027] Preferably, in step S4, the LSTM thermal neural network junction temperature prediction model constructed using the training dataset is specifically trained as follows:

[0028] S401. Perform feature extraction and normalization on the dataset generated by FEM simulation;

[0029] S402. Define a loss function, including mean square error loss and first derivative difference loss, to minimize the difference between the actual temperature value and the predicted value.

[0030] S403. Hyperparameters are tuned using time-series cross-validation and Bayesian optimization methods.

[0031] Secondly, embodiments of the present invention provide an IGBT module junction temperature prediction system based on an LSTM thermal neural network, comprising:

[0032] The equation module analyzes the unsteady heat transfer characteristics of the multilayer materials inside the IGBT module and constructs the DHT equation based on the structural and material properties.

[0033] The data module generates a training dataset containing temperature response data under different power loss conditions through FEM simulation;

[0034] The module constructs an LSTM thermal neural network junction temperature prediction model and embeds a physical layer within it. This physical layer contains two sub-neural networks. R net and C net These are used to extract the mapping relationship between input features and thermal resistance and thermal capacity, respectively;

[0035] The prediction module uses an LSTM thermal neural network junction temperature prediction model trained on a training dataset. The DHT equation is embedded into the network as a physical constraint of the heat transfer process, and transformed into a dynamic constraint condition for the parameters of the LSTM thermal neural network junction temperature prediction model. The trained thermal neural network junction temperature prediction model is used to predict the junction temperature of the IGBT module under different operating conditions.

[0036] Thirdly, a computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described method for predicting the junction temperature of an IGBT module based on an LSTM thermal neural network.

[0037] Fourthly, embodiments of the present invention provide a computer-readable storage medium including a computer program, which, when executed by a processor, implements the steps of the above-described method for predicting the junction temperature of an IGBT module based on an LSTM thermal neural network.

[0038] Fifthly, a chip includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described IGBT module junction temperature prediction method based on an LSTM thermal neural network.

[0039] In a sixth aspect, embodiments of the present invention provide an electronic device, including a computer program, which, when executed by the electronic device, implements the steps of the above-described method for predicting the junction temperature of an IGBT module based on an LSTM thermal neural network.

[0040] Compared with the prior art, the present invention has at least the following beneficial effects:

[0041] A method for predicting IGBT module junction temperature based on an LSTM thermal neural network is proposed. This method constructs a DHT equation to characterize the unsteady-state heat transfer properties of multilayer materials, overcoming the problem of oversimplification of nonlinear heat transfer processes in traditional thermal network methods. This improves the model's adaptability to complex operating conditions. The DHT equation is transformed into dynamic constraints on LSTM network parameters, allowing the neural network to follow the physical laws of heat transfer during training. This avoids the overfitting risk of purely data-driven models. From physical modeling and data generation to physical-guided neural network training, a closed loop of physical law-guided and data-driven optimization is formed, significantly improving prediction accuracy. The dynamic extraction of thermal resistance / capacity subnets allows for adaptive adaptation to changes in material properties, making it applicable to different IGBT module models. This method addresses the poor generalization problem of traditional methods by utilizing the DHT equation to capture the physical laws of heat transfer and leveraging LSTM to handle temperature time-series dependence, overcoming the shortcomings of fixed parameters in traditional thermal network methods and the lack of physical constraints in deep learning. The advantages include achieving high-precision prediction of junction temperature under different operating conditions, balancing accuracy and generalization ability, meeting offline evaluation needs, and reducing reliance on invasive measurements.

[0042] Furthermore, the thermal resistance and heat capacity network structure of the DHT equation is clearly defined. Node connectivity is described using matrices, dynamically reflecting the changes in thermal resistance and heat capacity with material properties and boundary conditions. This overcomes the limitations of static parameters in traditional thermal network methods, accurately characterizing the nonlinear properties of unsteady-state heat transfer. The model can adapt to different material combinations and operating conditions, improving the modeling accuracy for complex heat transfer processes and providing a reliable physical basis for subsequent temperature calculations.

[0043] Furthermore, the process of generating datasets for FEM simulations is standardized, covering data from multiple monitoring points, multiple operating conditions, and high temporal resolution. By simulating complex scenarios in actual operation, comprehensive and realistic samples are provided for model training. The dataset includes complete temperature responses at key locations such as the chip and solder layer, preserving temporal details at high resolution to ensure that the LSTM can learn subtle temperature dynamics, thereby improving model training effectiveness and generalization ability.

[0044] Furthermore, through sub-neural networks R net and C net Dynamically extracting thermal resistance and heat capacity to form an enhanced feature vector input to the LSTM solves the problem of dynamic changes in thermal resistance and heat capacity under operating conditions. By fusing physical parameters with input features, the model's ability to capture heat transfer characteristics is enhanced. This avoids errors from manually setting thermal parameters, making the extraction of thermal resistance and heat capacity more consistent with actual operating conditions, and improving the relevance and accuracy of temperature prediction.

[0045] Furthermore, an LSTM layer output formula is defined, utilizing the temporal modeling capabilities of LSTM to process enhanced feature vectors. The LSTM gating mechanism is used to capture the short- and long-term dependencies of temperature changes, adapting to the time-cumulative effect of the heat transfer process. Compared to traditional time-series models, this more accurately depicts the dynamic evolution of temperature over time, making it particularly suitable for full-cycle temperature prediction of IGBT modules from startup to steady state.

[0046] Furthermore, the input features are directly correlated with thermal resistance and heat capacity, establishing a mapping relationship between input parameters and core thermal parameters. This gives the extraction of thermal resistance and heat capacity a clear physical orientation, improves the interpretability of thermal parameter extraction, avoids the black box problem of deep learning, ensures that the extracted thermal resistance and heat capacity conform to heat transfer logic, and provides reliable parameters for temperature calculation.

[0047] Furthermore, by mapping the target temperature through a fully connected layer and substituting it into the DHT equation to calculate the temperature at the next moment, the predictive power of the neural network and the constraints of the physical equation are combined to achieve dual verification of data-driven and physical laws. This leverages the rapid prediction advantage of LSTM while ensuring that the results conform to the heat transfer laws through the DHT equation, achieving high-precision real-time prediction and meeting the rapid response requirements in engineering.

[0048] Furthermore, the details of mapping the target temperature to the fully connected layer are supplemented, clarifying the adaptation relationship between the output and physical parameters. The high-dimensional features of the LSTM are transformed into directly computable temperature values ​​through the fully connected layer, ensuring that the output dimension matches the DHT equation. This eliminates the dimensionality difference between the features and the target temperature, improves the accuracy of temperature mapping, makes the model output more closely match the actual temperature range, and reduces conversion errors.

[0049] Furthermore, standardize the model training process, including data normalization, custom loss functions, and hyperparameter optimization. Improve input quality through data preprocessing, ensure the loss function considers both error and trend, and enhance model performance through hyperparameter optimization. Avoid the impact of data distribution differences on training, reduce prediction bias, quickly find the optimal parameter configuration, and ensure the model's stability and accuracy under different operating conditions.

[0050] It is understood that the beneficial effects of the second to sixth aspects mentioned above can be found in the relevant descriptions in the first aspect mentioned above, and will not be repeated here.

[0051] In summary, the method of this invention integrates a physical heat transfer model with an LSTM network, overcoming the limitations of traditional methods such as intrusiveness and fixed parameters. It not only improves prediction accuracy but also reduces computational costs, thus possessing high practical value.

[0052] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0053] Figure 1 This represents the temperature monitoring points in the IGBT module and the thermal resistance and thermal capacity network represented by the DHT equation.

[0054] Figure 2 Material layers and boundary conditions for FEM simulation of IGBT modules;

[0055] Figure 3 This invention provides a junction temperature prediction model for IGBT modules based on an LSTM thermal neural network.

[0056] Figure 4 To compare the junction temperature of IGBT modules under different operating conditions using FEM simulation, experiments, and LSTM thermal neural networks, where (a) converter power is 200 kVA, (b) converter power is 300 kVA, (c) converter power is 400 kVA, and (d) converter power is 500 kVA;

[0057] Figure 5 A schematic diagram of a computer device provided in an embodiment of the present invention;

[0058] Figure 6 This is a block diagram of a chip provided according to an embodiment of the present invention.

[0059] Among them, 60. Computer equipment; 61. Processor; 62. Memory; 63. Computer program; 600. Electronic device; 610. Processing unit; 620. Storage unit; 6201. Random access memory unit; 6202. Cache memory unit; 6203. Read-only memory unit; 6204. Program / utility; 6205. Program module; 630. Bus; 640. Display unit; 650. Input / output interface; 660. Network adapter; 700. External device. Detailed Implementation

[0060] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0061] In the description of this invention, it should be understood that the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0062] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0063] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes such combinations. For example, A and / or B can represent three cases: A alone, A and B simultaneously, and B alone. Additionally, the character " / " in this invention generally indicates that the preceding and following objects have an "or" relationship.

[0064] It should be understood that although terms such as first, second, third, etc., may be used in the embodiments of the present invention to describe the preset range, these preset ranges should not be limited to these terms. These terms are only used to distinguish the preset ranges from one another. For example, without departing from the scope of the embodiments of the present invention, the first preset range may also be referred to as the second preset range, and similarly, the second preset range may also be referred to as the first preset range.

[0065] Depending on the context, the word "if" as used here can be interpreted as "when," "when," "in response to determination," or "in response to detection." Similarly, depending on the context, the phrase "if determination" or "if detection (of the stated condition or event)" can be interpreted as "when determination," "in response to determination," "when detection (of the stated condition or event)," or "in response to detection (of the stated condition or event)."

[0066] The accompanying drawings illustrate various structural schematic diagrams according to embodiments disclosed in this invention. These drawings are not to scale, and some details have been enlarged for clarity, and some details may have been omitted. The shapes of the various regions and layers shown in the drawings, as well as their relative sizes and positional relationships, are merely exemplary and may deviate from reality due to manufacturing tolerances or technical limitations. Furthermore, those skilled in the art can design regions / layers with different shapes, sizes, and relative positions as needed.

[0067] This invention provides a method for predicting IGBT module junction temperature based on an LSTM thermal neural network. Based on the structure and material properties of the IGBT module, the unsteady-state heat transfer characteristics of its internal multilayer materials are analyzed, and a Discrete Heat Transfer (DHT) equation is constructed. A training dataset is generated through finite element method (FEM) simulation. An LSTM thermal neural network junction temperature prediction model architecture is constructed, embedding the physical constraints of the heat transfer process into the network and transforming the DHT equation into dynamic constraints on the network parameters. A physical layer is embedded in the LSTM thermal neural network junction temperature prediction model, and thermal resistance and heat capacity are dynamically extracted from the DHT equation through a sub-neural network. The node temperature at the next moment is calculated by combining the input power loss to satisfy the boundary dependence of thermal resistance and heat capacity. The trained LSTM thermal neural network is used for junction temperature prediction, and the accuracy of the prediction results is verified by comparison with FEM data and experimental data. This invention achieves the prediction of IGBT module junction temperature and critical layer temperature under different operating conditions, effectively improving the reliability and economy of IGBT module thermal management.

[0068] Please see Figure 3 This invention discloses a method for predicting the junction temperature of an IGBT module based on an LSTM thermal neural network, comprising the following steps:

[0069] S1. Analyze the unsteady-state heat transfer characteristics within the IGBT module and construct the DHT equation;

[0070] The DHT equations are used to represent the heat transfer relationships between nodes, and the node connections are described in matrix form to derive the formula for the node temperature at the next moment. The thermal resistance and heat capacity in the DHT equations can vary with the material properties and boundary conditions of the IGBT module. Therefore, the constructed DHT equations can dynamically reflect the nonlinear characteristics of the heat transfer process in the IGBT module.

[0071] Please see Figure 1 Based on the structure and material properties of the IGBT module, the DHT equation is constructed and represented as a thermal resistance-thermal capacity network. This network includes heat sources (power losses) P ( t )), thermal resistance ( R ij ( t )) and heat capacity (C ij ( t The system's state space is discretized, and the connection relationships between nodes are described in matrix form. Formulas for the node temperatures at the next time step are derived. This includes temperature monitoring points for each material layer, with each pair of thermal resistance and thermal capacity representing several material layers along the heat transfer path.

[0072] The heat transfer relationships between nodes in an IGBT module can be represented in matrix form, where the temperature change of each node can be described by the DHT equation:

[0073] (1)

[0074] Wherein, the DHT equation represents m+n The heat transfer relationship between nodes can be represented by a thermal resistance-thermal capacity network. T i ( t ) is a node i Temperature. Node i and j There is thermal resistance between them. R ij ( t ) and heat capacity C ij ( t ), which represents the heat transfer relationship between two nodes. T ij ( t () represents the temperature difference between two nodes. T a This refers to ambient temperature or coolant temperature. P i ( t ) is the node through which the flow passes. i Power loss, using matrix M To describe the connectivity of nodes.

[0075] S2. Generate a complete dataset based on FEM simulation to train the LSTM thermal neural network junction temperature prediction model;

[0076] A junction temperature prediction model based on an LSTM thermal neural network is constructed, comprising an input layer, an LSTM layer, and a fully connected layer. The input is the power loss of the IGBT module. P ( t The output is the junction temperature. T j ( tThis method gradually captures temporal dependency patterns in a sequence through memory units and gating mechanisms (input gate, forget gate, output gate). It obtains high-level features of long-range temporal correlations through information transfer and state updates between time steps. Fully connected layers concatenate features and output data. Time series features are gradually extracted through LSTM layers, and global high-level features are obtained by utilizing cell states and hidden states. Fully connected layers concatenate features and output data. Considering that the output feature information of thermal resistance curves is relatively small, three LSTM layers and one fully connected layer are used for training, thereby constructing an LSTM neural network thermal resistance prediction model.

[0077] Please see Figure 2 In ANSYS Fluent, an IGBT module FEM simulation was established, temperature monitoring points were set, and temperature response data under asynchronous power loss conditions were extracted.

[0078] In ANSYS Fluent, the material properties of the IGBT module were set, and the inlet temperature of the water-cooled heatsink was configured to 20°C. Temperature monitoring points were set at the center of the chip and the critical material layer to record temperature changes under different power loss conditions. The data sampling frequency was 1000 Hz, and the recording duration was 10 seconds, generating a training dataset containing 10,000 samples.

[0079] Please see Figure 3 The power loss of the junction temperature prediction model of the LSTM thermal neural network is calculated, with the connection relationship between nodes as the input signal and the IGBT junction temperature as the output.

[0080] The data is normalized, and the weighted average and standard deviation of each input variable at each time point are calculated using the exponentially weighted moving average (EWMA) and exponentially weighted moving standard deviation (EWMS) methods. These values ​​are then used as input features of the model. Thermal resistance and thermal capacity from the DHT equations are embedded into the physical layer of the LSTM thermal neural network junction temperature prediction model, forming an LSTM-DHT unit. This unit is trained in parallel through sub-neural networks, extracting thermal parameters from the input power loss and combining them with the connectivity of the thermal network to predict the temperature at the next time step. The forget gate, input gate, and output gate mechanisms of LSTM are used to accurately capture the temporal characteristics of the heat transfer process.

[0081] The process of generating a complete dataset based on FEM simulation and using it to train an LSTM thermal neural network junction temperature prediction model is as follows:

[0082] S201. Construct an FEM simulation that includes the geometry, material properties, and boundary conditions of the IGBT module to ensure that the trained thermal neural network can accurately reflect the heat transfer characteristics of the IGBT module under different operating conditions.

[0083] S202. Set up multiple temperature monitoring points in the FEM simulation. These monitoring points are distributed in key locations of the IGBT module, including the chip center, chip solder layer center, substrate solder layer center, and copper substrate center, etc., to obtain temperature response data at each key location.

[0084] S203. Apply different power losses to simulate different operating conditions that IGBT modules may encounter in practical applications.

[0085] S204. Acquire the temperature response curves of each temperature monitoring point. These curves are recorded with high time resolution (e.g., 1000Hz) and cover the complete temperature change process from the initial moment to the steady state.

[0086] S205. Integrate the temperature response data obtained under all operating conditions into a complete dataset. This dataset contains rich temperature information and can provide comprehensive and diverse data support for training the LSTM thermal neural network junction temperature prediction model.

[0087] S3. A physical layer is constructed within the neural network framework, dynamically extracting thermal resistance and heat capacity through sub-neural networks. Specifically, a physical layer is embedded in the LSTM thermal neural network junction temperature prediction model, containing two sub-neural networks that respectively map the relationship between input power loss and thermal resistance and heat capacity. After training, the extracted thermal resistance and heat capacity are substituted into the DHT equation to calculate the node temperature at the next time step. This combination utilizes prior physical knowledge and the nonlinear fitting capability of deep learning to improve the accuracy and generalization ability of temperature prediction.

[0088] The PyTorch deep learning framework was chosen, combining momentum optimization and the Adam optimizer with an adaptive learning rate. A loss function was defined, including mean squared error loss and first-order derivative difference loss, to minimize the difference between actual and predicted temperature values. The dataset was divided into training (70%), validation (15%), and test (15%) sets, and time-series cross-validation was employed to prevent time leakage bias. Bayesian optimization was used to optimize the hyperparameters (learning rate, batch size, number of hidden layer nodes, span) of the LSTM thermal neural network junction temperature prediction model, achieving fast and accurate optimization of the objective function.

[0089] Construct two sub-neural networks R net and C net These are used to extract the mapping relationship between input features and thermal resistance and thermal capacity, respectively. R net and C net The outputs are dynamic thermal resistance. R ij ( t ) and dynamic heat capacityC ij ( t Its mathematical expression is:

[0090] (2)

[0091] (3)

[0092] in, P ( t ) is the input characteristic, namely power loss.

[0093] Dynamic thermal resistance and thermal capacity are concatenated with the input features to form an enhanced feature vector. z ( t ):

[0094] (4)

[0095] Enhance feature vectors z ( t The input is fed into an LSTM-DHT layer for time series modeling, and the sub-neural network... R net and C net The output is:

[0096] (5)

[0097] S4. Use the trained thermal neural network junction temperature prediction model to predict the junction temperature of the IGBT module under different operating conditions; combine the thermal parameters predicted by the LSTM thermal neural network junction temperature prediction model with the DHT equation, and calculate the junction temperature at the next moment based on the current junction temperature, thereby realizing the rapid prediction of the junction temperature of the IGBT module.

[0098] Feature extraction from feature data is performed using an LSTM sub-network, and neural network training is conducted based on the extraction results, including:

[0099] S401. Perform feature extraction and normalization on the dataset generated by FEM simulation;

[0100] The specific process is as follows: feature extraction is performed on the temperature response data generated by FEM simulation to extract key features related to temperature changes, and these features are normalized to facilitate the training of the LSTM thermal neural network junction temperature prediction model.

[0101] S402. Define a loss function, including mean squared error loss and first derivative difference loss, to minimize the difference between the actual temperature value and the predicted value; by optimizing the loss function, the LSTM thermal neural network junction temperature prediction model can more accurately predict the junction temperature change of the IGBT module.

[0102] S403. Hyperparameters are tuned using time-series cross-validation and Bayesian optimization methods.

[0103] Time series cross-validation can effectively avoid time leakage problems in time series data, while Bayesian optimization methods can efficiently search for the optimal combination of hyperparameters, thereby improving the prediction performance of LSTM hot neural networks.

[0104] Using a trained LSTM thermal neural network, based on the input power loss... P ( t Predicting the junction temperature of IGBT modules under different operating conditions. T j ( t The accuracy of the prediction results was verified by comparing them with FEM data and experimental data.

[0105] By substituting the thermal parameters predicted by the LSTM thermal neural network junction temperature prediction model into the DHT equation, the node temperature at the next time step is calculated, achieving high-precision real-time prediction of IGBT module temperature. Model training is optimized using custom loss functions (including MSE loss and differential loss) to ensure that the LSTM thermal neural network can accurately capture the dynamic characteristics of temperature changes.

[0106] The mathematical expression for mapping the output of the LSTM layer to the target temperature using a fully connected layer is as follows:

[0107] (6)

[0108] Finally, the prediction results of the sub-neural network are substituted into the DHT equation, as shown in equation (1), to calculate the node temperature at the next time step.

[0109] In another embodiment of the present invention, an IGBT module junction temperature prediction system based on LSTM thermal neural network is provided. This system can be used to implement the above-mentioned IGBT module junction temperature prediction method based on LSTM thermal neural network. Specifically, the IGBT module junction temperature prediction system based on LSTM thermal neural network includes an equation module, a data module, a construction module, and a prediction module.

[0110] Among them, the equation module analyzes the unsteady heat transfer characteristics of the multilayer materials inside the IGBT module and constructs the DHT equation based on the structural and material properties.

[0111] The data module generates a training dataset containing temperature response data under different power loss conditions through FEM simulation;

[0112] The module constructs an LSTM thermal neural network junction temperature prediction model and embeds a physical layer within it. This physical layer contains two sub-neural networks. R net and C net These are used to extract the mapping relationship between input features and thermal resistance and thermal capacity, respectively;

[0113] The prediction module uses an LSTM thermal neural network junction temperature prediction model trained on a training dataset. The DHT equation is embedded into the network as a physical constraint of the heat transfer process, and transformed into a dynamic constraint condition for the parameters of the LSTM thermal neural network junction temperature prediction model. The trained thermal neural network junction temperature prediction model is used to predict the junction temperature of the IGBT module under different operating conditions.

[0114] This invention provides a terminal device comprising a processor and a memory. The memory stores a computer program, which includes program instructions. The processor executes the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, graphics processing units (GPUs), tensor processing units (TPUs), digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions to achieve a corresponding method flow or function. The processor described in this embodiment can be used for the operation of an IGBT module junction temperature prediction method based on an LSTM thermal neural network, including:

[0115] The unsteady-state heat transfer characteristics of the multilayer materials inside the IGBT module were analyzed, and the DHT equation was constructed based on the structural and material properties. A training dataset containing temperature response data under different power loss conditions was generated through FEM simulation. An LSTM thermal neural network junction temperature prediction model was constructed, and a physical layer containing two sub-neural networks was embedded in the LSTM thermal neural network junction temperature prediction model. R net and Cnet These are used to extract the mapping relationship between input features and thermal resistance and thermal capacity, respectively; an LSTM thermal neural network junction temperature prediction model is trained using the training dataset, and the DHT equation is embedded into the network as a physical constraint of the heat transfer process, which is transformed into a dynamic constraint condition for the parameters of the LSTM thermal neural network junction temperature prediction model. The trained thermal neural network junction temperature prediction model is used to predict the junction temperature of the IGBT module under different operating conditions.

[0116] Please see Figure 5 The terminal device is a computer device. In this embodiment, the computer device 60 includes a processor 61, a memory 62, and a computer program 63 stored in the memory 62 and executable on the processor 61. When executed by the processor 61, the computer program 63 implements the method for estimating the concentration of radioactive iodine species in the containment structure after an accident, as described in this embodiment. To avoid repetition, these details are not elaborated here. Alternatively, when executed by the processor 61, the computer program 63 implements the functions of each model / unit in the IGBT module junction temperature prediction system based on an LSTM thermal neural network, as described in this embodiment. To avoid repetition, these details are not elaborated here.

[0117] Computer device 60 can be a desktop computer, laptop, handheld computer, cloud server, or other computing device. Computer device 60 may include, but is not limited to, a processor 61 and a memory 62. Those skilled in the art will understand that... Figure 5 This is merely an example of computer device 60 and does not constitute a limitation on computer device 60. It may include more or fewer components than shown, or combine certain components, or different components. For example, computer device may also include input / output devices, network access devices, buses, etc.

[0118] The processor 61 may be a Central Processing Unit (CPU), or other general-purpose processors, graphics processing units (GPUs), tensor processing units (TPUs), digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0119] The memory 62 can be an internal storage unit of the computer device 60, such as a hard disk or RAM of the computer device 60. The memory 62 can also be an external storage device of the computer device 60, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the computer device 60.

[0120] Furthermore, memory 62 may include both internal storage units and external storage devices of the computer device 60. Memory 62 is used to store computer programs and other programs and data required by the computer device. Memory 62 can also be used to temporarily store data that has been output or will be output.

[0121] Please see Figure 6 The terminal device is an electronic device 600, which is manifested in the form of a general-purpose computing device. The components of the electronic device may include, but are not limited to: at least one processing unit 610, at least one storage unit 620, a bus 630 connecting different platform components (including storage unit 620 and processing unit 610), a display unit 640, etc.

[0122] The storage unit stores program code that can be executed by the processing unit 610, causing the processing unit 610 to perform the steps described in the above-described method section of this specification according to various exemplary embodiments of the present invention. For example, the processing unit 610 can perform the steps of the illustrated method.

[0123] Storage unit 620 may include a readable medium in the form of a volatile storage unit, such as random access memory (RAM) 6201 and / or cache memory 6202, and may further include a read-only memory (ROM) 6203.

[0124] Storage unit 620 may also include a program / utility 6204 having a set (at least one) of program modules 6205, such program modules 6205 including but not limited to: an operating system, one or more application programs, other program modules and program data, each or some combination of these examples may include an implementation of a network environment.

[0125] Bus 630 can represent one or more of several types of bus structures, including a memory cell bus or memory cell controller, a peripheral bus, a graphics acceleration port, a processing unit, or a local bus using any of the multiple bus structures.

[0126] Electronic device 600 can also communicate with one or more external devices 700 (e.g., keyboard, pointing device, Bluetooth device, etc.), and with one or more devices that enable a user to interact with electronic device 600, and / or with any device that enables electronic device 600 to communicate with one or more other computing devices (e.g., router, modem). This communication can be performed via input / output interface 650. Furthermore, electronic device 600 can also communicate with one or more networks (e.g., local area network, wide area network, and / or public network, such as the Internet) via network adapter 660. Network adapter 660 can communicate with other modules of electronic device 600 via bus 630. It should be understood that, although not shown in the figures, other hardware and / or software modules can be used in conjunction with electronic device 600, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage platforms.

[0127] Example 4

[0128] This invention also provides a storage medium, specifically a computer-readable storage medium, which is a memory device in a terminal device for storing programs and data. It is understood that the computer-readable storage medium here can include both built-in storage media in the terminal device and extended storage media supported by the terminal device; it can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. The computer-readable storage medium provides storage space that stores the terminal's operating system. Furthermore, the storage space also stores one or more instructions suitable for loading and execution by a processor, which can be one or more computer programs (including program code). More specific examples of the computer-readable storage medium include: an electrical connection with one or more wires, a portable disk, a hard disk, random access memory, read-only memory, erasable programmable read-only memory, optical fiber, portable compact disk read-only memory, optical storage device, magnetic storage device, or any suitable combination thereof.

[0129] Computer-readable storage media also include data signals propagated in baseband or as part of a carrier wave, carrying readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A readable storage medium can also be any readable medium other than a readable storage medium that can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the readable storage medium can be transmitted using any suitable medium, including but not limited to wireless, wired, optical fiber, radio frequency, etc., or any suitable combination thereof.

[0130] Program code for performing the operations of this invention can be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java and C++, and conventional procedural programming languages ​​such as C or similar languages. The program code can execute entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0131] One or more instructions stored in the computer-readable storage medium can be loaded and executed by the processor to implement the corresponding steps of the IGBT module junction temperature prediction method based on LSTM thermal neural network in the above embodiments; one or more instructions in the computer-readable storage medium are loaded and executed by the processor to perform the following steps:

[0132] The unsteady-state heat transfer characteristics of the multilayer materials inside the IGBT module were analyzed, and the DHT equation was constructed based on the structural and material properties. A training dataset containing temperature response data under different power loss conditions was generated through FEM simulation. An LSTM thermal neural network junction temperature prediction model was constructed, and a physical layer containing two sub-neural networks was embedded in the LSTM thermal neural network junction temperature prediction model. R net and C netThese are used to extract the mapping relationship between input features and thermal resistance and thermal capacity, respectively; an LSTM thermal neural network junction temperature prediction model is trained using the training dataset, and the DHT equation is embedded into the network as a physical constraint of the heat transfer process, which is transformed into a dynamic constraint condition for the parameters of the LSTM thermal neural network junction temperature prediction model. The trained thermal neural network junction temperature prediction model is used to predict the junction temperature of the IGBT module under different operating conditions.

[0133] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0134] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0135] This invention presents a comprehensive experimental evaluation of the IGBT module junction temperature prediction method based on an LSTM thermal neural network under different operating conditions (converter power of 200kVA, 300kVA, 400kVA, and 500kVA). Experimental results show that this method outperforms traditional finite element method (FEM) simulation in both prediction accuracy and computational efficiency.

[0136] like Figure 4 As shown, we compared the junction temperature prediction results of LSTM thermal neural network, FEM simulation, and experimental data under different load conditions. It can be seen that the prediction results of LSTM thermal neural network are highly consistent with the experimental data, while FEM simulation shows significant deviations under certain operating conditions. According to the error statistics, the mean absolute error (MAE) of the junction temperature prediction by LSTM thermal neural network compared to the experimentally measured junction temperature is 0.163~0.281, the root mean square error (RMSE) is 0.185~0.325, and the coefficient of determination (R²) is... 2The accuracy rates were 92.36%–99.71%. In comparison, the MAE and RMSE of FEM simulations compared to experiments were 0.464–0.552, and the R... 2 The range is 78.50% to 98.52%. To facilitate the use of simulation results for power semiconductor device lifetime assessment, the relative error of the maximum junction temperature is also considered. The accuracy of LSTM thermal neural network and FEM simulation is compared using the percentage (%). As the converter power increases from 200 kVA to 500 kVA, the relative error of the junction temperature predicted by LSTM thermal neural network compared to the experimentally measured maximum value decreased from 1.15% to 0.36%. In contrast, the relative error of FEM simulation compared to the experimental maximum value decreased from 2.13% to 1.08%. This demonstrates that LSTM thermal neural network has higher junction temperature prediction accuracy than FEM simulation.

[0137] The results show that the LSTM thermal neural network can more accurately capture the dynamic changes in the junction temperature of IGBT modules and has higher prediction reliability. This provides a more accurate and reliable method for junction temperature prediction.

[0138] In summary, this invention presents an IGBT module junction temperature prediction method and system based on an LSTM thermal neural network. By combining physical constraints with the time-series modeling capabilities of deep learning, it dynamically extracts thermal resistance and heat capacity, resulting in smaller prediction errors under different operating conditions. The MAE and RMSE metrics are superior to traditional methods, and the model shows high consistency with FEM simulation and experimental data. The multi-condition training dataset generated through FEM simulation allows the model to adapt to different power losses, overcoming the limitations of traditional methods in handling nonlinear heat transfer. The trained model offers fast prediction speeds, meeting real-time requirements. Furthermore, the embedded physical layer ensures that the prediction conforms to heat transfer laws, improving the reliability and economy of IGBT module thermal management. This method is particularly suitable for offline junction temperature assessment scenarios for manufacturers.

[0139] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0140] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0141] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed in this invention can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0142] In the embodiments provided by this invention, it should be understood that the disclosed devices / terminals and methods can be implemented in other ways. For example, the device / terminal embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0143] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0144] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0145] If the integrated module / unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random-access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0146] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus, and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0147] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0148] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0149] The above content is only for illustrating the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. Any modifications made to the technical solution based on the technical concept proposed in this invention shall fall within the scope of protection of the claims of this invention.

Claims

1. An IGBT module junction temperature prediction method based on an LSTM hot neural network, characterized in that, Includes the following steps: S1. Analyze the unsteady heat transfer characteristics of the multilayer materials inside the IGBT module, and construct the DHT equation based on the structure and material properties; S2. Generate a training dataset containing temperature response data under different power loss conditions through FEM simulation; S3, a LSTM thermal neural network junction temperature prediction model is constructed, a physical layer is embedded in the LSTM thermal neural network junction temperature prediction model, and the physical layer includes two sub neural networks R net and C net respectively used for extracting the mapping relationship between the input features and the thermal resistance and the thermal capacity. S4. The LSTM thermal neural network junction temperature prediction model is trained using the training dataset, and the DHT equation is embedded into the network as a physical constraint of the heat transfer process, which is transformed into a dynamic constraint condition for the parameters of the LSTM thermal neural network junction temperature prediction model. The trained thermal neural network junction temperature prediction model is used to predict the junction temperature of the IGBT module under different operating conditions.

2. The IGBT module junction temperature prediction method based on LSTM thermal neural network according to claim 1, characterized in that, In step S1, the DHT equation is: in, T i ( t ) is a node i Temperature, node i and j There is thermal resistance between them. R ij ( t ) and heat capacity C ij ( t This represents the heat transfer relationship between two nodes. T ij ( t () represents the temperature difference between two nodes. T a This refers to ambient temperature or coolant temperature. P i ( t ) is the node through which the flow passes. i Power loss, using matrix M To describe the connectivity of nodes.

3. The IGBT module junction temperature prediction method based on LSTM thermal neural network according to claim 1, characterized in that, Step S2 is as follows: A FEM simulation model including the geometry, material properties, and boundary conditions of the IGBT module was constructed; temperature monitoring points were set at the center of the IGBT module chip, the center of the chip solder layer, the center of the substrate solder layer, and the center of the copper substrate; different power losses were applied to simulate different operating conditions in actual operation. The temperature response curves of each monitoring point from the initial moment to the steady state are recorded with high temporal resolution and integrated to form a training dataset.

4. The IGBT module junction temperature prediction method based on LSTM thermal neural network according to claim 1, characterized in that, In step S3, the sub-neural network R net and C net Output dynamic thermal resistance R ij ( t and dynamic heat capacity C ij ( t The dynamic thermal resistance and thermal capacity are concatenated with the input features to form an enhanced feature vector. z ( t This will enhance the feature vector. z ( t The input is fed into the LSTM-DHT layer for time series modeling.

5. The IGBT module junction temperature prediction method based on LSTM thermal neural network according to claim 4, characterized in that, Sub-neural network R net and C net Output for:

6. The IGBT module junction temperature prediction method based on LSTM thermal neural network according to claim 4, characterized in that, Sub-neural network R net and C net Output dynamic thermal resistance R ij ( t and dynamic heat capacity C ij ( t They are respectively: in, P ( t ) are the input features.

7. The IGBT module junction temperature prediction method based on LSTM thermal neural network according to claim 1, characterized in that, In step S4, the output of the LSTM layer is mapped to the target temperature through a fully connected layer, and the prediction result of the sub-neural network is substituted into the DHT equation to calculate the node temperature at the next moment, thereby realizing real-time prediction of the IGBT module temperature.

8. The IGBT module junction temperature prediction method based on LSTM thermal neural network according to claim 7, characterized in that, The output of the LSTM layer is mapped to the target temperature using a fully connected layer. Specifically: in, It is a fully connected layer. for t The output of the LSTM network at that time.

9. The IGBT module junction temperature prediction method based on LSTM thermal neural network according to claim 1, characterized in that, In step S4, the LSTM thermal neural network junction temperature prediction model is trained using the training dataset, specifically as follows: S401. Perform feature extraction and normalization on the dataset generated by FEM simulation; S402. Define a loss function, including mean square error loss and first derivative difference loss, to minimize the difference between the actual temperature value and the predicted value. S403. Hyperparameters are tuned using time-series cross-validation and Bayesian optimization methods.

10. A junction temperature prediction system for IGBT modules based on an LSTM thermal neural network, characterized in that, include: The equation module analyzes the unsteady heat transfer characteristics of the multilayer materials inside the IGBT module and constructs the DHT equation based on the structural and material properties. The data module generates a training dataset containing temperature response data under different power loss conditions through FEM simulation; The module constructs an LSTM thermal neural network junction temperature prediction model and embeds a physical layer within it. This physical layer contains two sub-neural networks. R net and C net These are used to extract the mapping relationship between input features and thermal resistance and thermal capacity, respectively; The prediction module uses an LSTM thermal neural network junction temperature prediction model trained on a training dataset. The DHT equation is embedded into the network as a physical constraint of the heat transfer process, and transformed into a dynamic constraint condition for the parameters of the LSTM thermal neural network junction temperature prediction model. The trained thermal neural network junction temperature prediction model is used to predict the junction temperature of the IGBT module under different operating conditions.

Images