Real-time evaluation method for reduced-order junction temperature of thermal model of traction converter IGBT (Insulated Gate Bipolar Translator) module

Abstract

The invention discloses a traction converter IGBT module thermal model reduced-order junction temperature real-time evaluation method. The method comprises the following steps: establishing a circuit model of an IGBT to calculate loss; establishing an IGBT structure model based on the spatial positions and geometric dimensions of the IGBT and the water cooling plate; taking the temperature field of the water-cooled IGBT module under different boundary conditions as sample data, and extracting the modal of the temperature field of the water-cooled IGBT based on a singular value decomposition method; projecting a heat conduction high-order equation to the subspace by combining Galerkin projection based on the extracted main mode to obtain a reduced-order thermal model of the water-cooled IGBT module; and substituting the IGBT loss and the water-cooling convection coefficient into the water-cooling IGBT thermal reduced order model to calculate the IGBT junction temperature. According to the method, the two-dimensional temperature field of the IGBT module is predicted, only the convection coefficient and the loss need to be input, an external circuit or a sensor does not need to be additionally arranged, and compared with a traditional method, calculation is simplified, the junction temperature monitoring precision of the IGBT module is guaranteed, meanwhile, the generalization ability of the model is improved, and practical engineering application is facilitated.

Description

Technical Field

[0001] The present invention belongs to the technical field of reliability of power electronic devices, and in particular relates to a real-time evaluation method for reduced-order junction temperature of a thermal model of an IGBT module of a traction converter. Background Art

[0002] Insulated-Gate Bipolar Transistor (IGBT) integrates the advantages of insulated-gate field-effect transistors and power transistors, and is a key component of rail transit traction drive systems. According to industry surveys, IGBT modules are one of the most vulnerable devices in power electronic systems, and about 55% of device failures are caused by temperature. The average life of IGBT modules decreases with rising temperature. The existing service life is based on condition-based repair, and the life is not used to the full extent. Therefore, monitoring the working junction temperature of IGBT modules is of great significance to improving the reliability of devices.

[0003] At present, the commonly used IGBT module junction temperature monitoring methods can be divided into five categories: physical contact method, optical measurement method, temperature-sensitive electrical parameter method, thermal model method and finite element method. They have the following disadvantages: 1) IGBT module has packaging characteristics; 2) the cost of monitoring equipment is high; 3) the accuracy depends on the pre-calibration of the module; 4) the extraction of thermal resistance and thermal capacitance parameters is relatively complicated; 5) the modeling process is relatively complicated. In order to solve the problem of accurate and real-time monitoring of the junction temperature of the traction converter IGBT module, it is necessary to propose a real-time evaluation method of the junction temperature of the traction converter IGBT module thermal model reduction. Summary of the invention

[0004] In view of the defects in the prior art, the present invention aims to provide a real-time evaluation method for reduced-order junction temperature of a traction converter IGBT module thermal model. The method simplifies the calculation, ensures the monitoring accuracy of the IGBT module junction temperature, and improves the calculation speed of the model.

[0005] In order to achieve the above purpose, the technical solution adopted by the present invention is: In a first aspect, the present invention provides a real-time evaluation method for reduced-order junction temperature of a thermal model of an IGBT module of a traction converter, comprising the steps of: Establish the circuit model of IGBT to calculate the loss; Establish an IGBT structural model based on the spatial position and geometric dimensions of the IGBT and water cooling plate; The temperature field of the water-cooled IGBT module under different boundary conditions is used as sample data, and the modes of the water-cooled IGBT temperature field are extracted based on the singular value decomposition method. Based on the extracted main modes and combined with Galerkin projection, the high-order heat conduction equations are projected into the subspace to obtain the reduced-order thermal model of the water-cooled IGBT module. Substitute the IGBT loss and water cooling convection coefficient into the cold IGBT thermal reduction model to calculate the IGBT junction temperature.

[0006] Furthermore, the establishment of the IGBT circuit model to calculate the loss includes: Extract key parameters according to the IGBT module product data sheet, including initial saturation voltage drop, on-resistance, and thermal characteristics; The conduction loss of the IGBT module is a function of the initial saturation voltage drop and the on-resistance. It is characterized by interpolation calculation and a linear equation for the output characteristic curve. The switching loss is based on the relationship between the switching energy loss and the current in the data sheet, and the polynomial interpolation method is used to construct a quadratic equation for the switching energy loss. The correction coefficient under actual operating conditions is determined based on the slope of the curve where the switching energy loss changes with voltage and temperature.

[0007] Furthermore, the IGBT structural model is established based on the spatial position and geometric dimensions of the IGBT module and the water cooling plate, including: The precise dimensions and thickness parameters of each layer structure inside the IGBT module and the water-cooling plate are obtained, and a three-dimensional structural model of the water-cooled IGBT module is established.

[0008] Furthermore, the three-dimensional structure model drawing software is SolidWorks.

[0009] Furthermore, the method of using the temperature field of the water-cooled IGBT module under different boundary conditions as sample data and extracting the mode of the water-cooled IGBT temperature field based on the singular value decomposition method includes: The three-dimensional structural model of the water-cooled plate and IGBT module was imported into the ANSYS Icepak simulation software, meshed, and the boundary conditions were set; the water-cooled plate temperature field was obtained by fluid-thermal simulation, and then the equivalent forced convection heat transfer coefficient of the contact surface between the water-cooled plate and the IGBT was obtained; the water-cooled IGBT temperature field was obtained in ANSYS Workbench based on the convection heat transfer coefficient, and several modes of the water-cooled IGBT temperature field were extracted based on the singular value decomposition method.

[0010] Furthermore, the extraction of several modes of the water-cooled IGBT temperature field based on the singular value decomposition method includes: The heat conduction equation is: ; In the above formula, ρ is the IGBT chip material density; c is the specific heat capacity of the IGBT chip; T is temperature; is the divergence of heat flow, which indicates the heat flow rate per unit area in a certain direction; λ is the thermal conductivity of the IGBT chip;Q is the internal heating rate of the IGBT chip; The convection equation is: ; In the above formula, q ' is the heat flux density per unit area during convection; h is the convection coefficient; T c is the IGBT junction temperature; T w The temperature of the fluid in the environment that performs convection heat exchange with the IGBT module housing; The temperature field of the IGBT module at different time points is expressed by a finite number of basis functions and represented as a linear equation system; by solving the linear equation system, M coefficients are obtained, each coefficient corresponds to a basis function, so as to obtain an approximate solution of the temperature field at the current moment.

[0011] Furthermore, based on the extracted main modes, combined with Galerkin projection, the high-order heat conduction equation is projected into the subspace to obtain a reduced-order thermal model of the water-cooled IGBT module, including: The singular value decomposition method is used to extract the basis functions, and the original temperature field is expressed as a linear superposition of basis functions. The complex original temperature field is represented by a combination of a few basis functions. Based on the Galerkin projection, the inner product operation is performed on both sides of the temperature equation using the extracted basis functions to obtain the thermal reduced-order model of the IGBT module.

[0012] Furthermore, the step of substituting the IGBT loss and the water-cooling convection coefficient into the water-cooled IGBT thermal reduction model to calculate the IGBT junction temperature includes: The calculated IGBT loss and water-cooling convection coefficient are substituted into the reduced-order thermal model to solve the junction temperature of the IGBT module. Combined with real-time data, the junction temperature calculation is dynamically updated, so that the temperature change of the IGBT module can be evaluated in real time under different working conditions.

[0013] On the other hand, the present invention also provides a terminal device, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor, wherein the memory stores a computer program capable of running on the processor, and when the processor loads and executes the computer program, a real-time evaluation method for reduced-order junction temperature of a traction inverter IGBT module thermal model as described above is adopted.

[0014] The beneficial effects of the present invention are as follows: The present application provides a real-time evaluation method for reduced-order junction temperature of the thermal model of the IGBT module of a traction converter. First, the circuit model of the IGBT is established to calculate the loss. Then, the IGBT structure model is established based on the spatial position and geometric dimensions of the IGBT and the water-cooled plate. The temperature field of the water-cooled IGBT module under different boundary conditions is used as sample data. The mode of the water-cooled IGBT temperature field is extracted based on the singular value decomposition method. Finally, based on the extracted main mode, combined with the Galerkin projection, the high-order heat conduction equation is projected to the subspace to obtain the reduced-order thermal model of the water-cooled IGBT module. This method predicts the two-dimensional temperature field of the IGBT module. It only needs to input the convection coefficient and loss, and does not need to add external circuits or sensors. Compared with traditional methods, it simplifies the calculation, ensures the accuracy of the IGBT module junction temperature monitoring, and improves the generalization ability of the model, which is convenient for practical engineering applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The present invention has the following accompanying drawings: Figure 1 The finite element modeling flow chart of the IGBT module of the present invention; Figure 2 A process for establishing a sample space for the IGBT module temperature field snapshot of the present invention; Figure 3 Establishing a process for the sample space of convection coefficient of the water-cooled plate of the present invention; Figure 4 The POD modal extraction process of the finite element model of the water-cooled IGBT module of the present invention; Figure 5 The present invention provides a method flow for real-time evaluation of junction temperature of a water-cooled IGBT module. DETAILED DESCRIPTION

[0016] The preferred examples of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred examples described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0017] like Figure 1 As shown, the finite element modeling process of the IGBT module includes the following steps: Step 1: Based on the structure and material parameters of the IGBT module, use SolidWorks 3D graphics software to perform geometric modeling of the IGBT module; Step 2: Define the material properties of each component; Step 3: Extract key parameters including on-resistance, reverse recovery current and thermal characteristics according to the IGBT module product data sheet; The conduction loss of the IGBT module is a function of the initial saturation voltage drop and the on-resistance. The output characteristic curve is characterized by a linear equation through interpolation calculation. Switching loss Based on the relationship between switching energy loss and current in the manual, the polynomial interpolation method is used to construct the quadratic equation of switching energy loss. According to the slope of the curve of switching energy loss changing with voltage and temperature, the correction coefficient under actual working conditions is determined; The power load is applied based on the density, specific heat capacity, thermal conductivity, Young's modulus, etc. of the power device.

[0018] Step 4: Select the solver type. For fluid dynamics problems, including internal and external fluid flow, heat transfer, etc., select ANSYS Icepak as the fluid solver; Heat conduction problems, including steady-state and transient thermal analysis, use ANSYS Workbench as the heat conduction solver to simulate the heat distribution and temperature changes inside the object.

[0019] Step 5: Import the water cooling plate and IGBT three-dimensional structure model into ANSYS simulation software, select the appropriate mesh type and density, and divide the mesh; Step 6: Based on finite element analysis, the temperature field of the IGBT module is simulated in combination with the characteristics of the IGBT module and the heat generation mechanism, the heat generation of the IGBT module under different working conditions is calculated, and the simulation boundary conditions are obtained; The boundary conditions of the water-cooling plate are set to perform fluid-thermal simulation to obtain the temperature field of the water-cooling plate and then the forced convection heat transfer coefficient. Step 7: Determine convergence and verify the accuracy of the model; Step eight, perform data post-processing and design optimization to ensure that the performance and reliability of the IGBT module meet the requirements, and finally obtain the temperature field results.

[0020] like Figure 2 As shown in the figure, the process of establishing the IGBT module temperature field snapshot sample space includes the following steps: 1) Initialization process, set the initial value of the IGBT module loss value.

[0021] 2) Set the convection heat transfer coefficient of the cooling system to control heat loss.

[0022] 3) Collect data every 1 second and use it as a sample to record the current loss situation.

[0023] 4) Make a loss threshold judgment. If it is not reached, continue to collect samples; if it reaches the preset maximum loss value, the process ends.

[0024] like Figure 3 As shown in FIG. 1 , the process of establishing the sample space of the water-cooled plate convection coefficient includes the following steps: 1) Initialization process, set the IGBT module to static state; 2) Set the water cooling plate inlet flow rate and water temperature range; in the boundary condition setting, the fluid inlet is the water inlet, which is set to a fixed flow rate consistent with the actual water cooling machine; the outlet is the drain port, which is set to free outflow; 3) Obtain the convection heat transfer coefficient of the cooling system and verify the heat transfer capacity between the water cooling plate and the IGBT base plate.

[0025] The heat flux of the forced heat exchange process, that is, the heat transfer rate per unit area, is: ; In the above formula, q represents heat flux; Indicates heat; Indicates the area of ​​heat transfer; The equivalent commutation coefficient between the water cooling plate and the IGBT base plate is: ; In the above formula, h represents the equivalent commutation coefficient; q represents heat flux; Indicates the temperature difference between the water cooling plate and the IGBT base plate; 4) Collect data every 1 second, collect the flow field data under stable state as samples, and the process ends.

[0026] like Figure 4 As shown in the figure, the POD modal extraction process of the finite element model of the water-cooled IGBT module includes the following steps: Step 1: Establish an initial finite element model of the water-cooled IGBT module based on its physical structure and material properties; Step 2: In the finite element model, the model is divided into several grid units, and the temperature, position and other data of each grid node are obtained and stored in a data matrix, and the data matrix X is standardized and preprocessed; The standardized preprocessing is expressed as: ; In the above formula, is the mean of the data matrix, is the standard deviation of the data matrix.

[0027] Step 3: Based on the sample space and the preprocessed grid node data, a covariance matrix is ​​constructed to establish the thermal conduction correlation between nodes and capture the statistical characteristics of the thermal distribution of the IGBT module; The model input is represented as: ; In the above formula, h is the convection coefficient, P is the loss value.

[0028] The sample space is a matrix: ; In the above formula, U Represents the sample space, which reflects the state of the system at different points in time and space and includes all snapshot data.

[0029] Step 4: Singular value decomposition to construct the covariance matrix of the computational grid node data matrix C for: ; In the above formula, is the sample size; Perform SVD on the constructed covariance matrix to decompose the complex thermal characteristics into several modes; According to the principle of SVD, the decomposed mode of the temperature field of the IGBT module is obtained, and the mode decomposition is: ; In the above formula, is the eigenvector matrix; is the matrix of eigenvalue diagonal elements; The modal energy ratio is: ; In the above formula, E is the modal energy ratio, is the energy of the ith mode; is the sum of k modal energies; Step 5: According to the results of singular value decomposition, extract the main modal coefficients and screen out the modes that have a greater impact on temperature distribution; Step six, normalizing the extracted modal coefficients; Step seven, based on the normalized modal coefficients and key modal information, a simplified thermal model is obtained.

[0030] like Figure 5 As shown, the real-time evaluation method flow of the junction temperature of a water-cooled IGBT module includes the following steps: Step 1: Build a water-cooled IGBT thermal model; The partial differential equation of the temperature field of the IGBT module is: ; In the above formula, To simplify the evolution operator representing the change of temperature with time; The combined heat conduction equation and convection equation are: ; ; In the above formula, 0, indicating a convective heat transfer boundary; 1, indicating a non-convective heat transfer boundary; The integration of the heat conduction equation and the convection equation is: ; In the above formula, V is the volume of the chip; A is the convection surface contact area.

[0031] Step 2: After discretization according to the backward difference principle, the discretized heat conduction equation is further expanded and simplified to be: ; In the above formula, x Indicates the horizontal axis direction; The boundary conditions are x= The convection boundary condition is adopted at 0, namely: ; After discretizing the boundary condition, we get: ; Arranged: ; The internal nodes of the recursive equation are: ; The boundary nodes are: ; Step 3: Perform modal reduction on the temperature field based on the singular value decomposition method, and extract the main modes in the temperature field through eigenvalue decomposition; The temperature field is represented by a linear combination of a finite number of POD basis functions. The temperature field at this moment and the temperature field at the previous moment are respectively expressed as: ; In the above formula, It is The temperature field at time ; M is the number of selected modes; It is time The coefficients of the corresponding modes; ; In the above formula, It is Temperature field at each moment; It is time The coefficients of the corresponding modes; Furthermore, the equation is discretized as follows: ; To understand each The coefficient of And integrate over the entire space, using the properties of orthogonal basis functions, we get: ; In the above formula, for and The inner product of The inner product of each item is calculated to obtain the system dynamic equation after the order reduction based on the SVD method: ; Step 4: To approximate the finite element difference to zero, the projection target is: ; Project the physical equations onto the reduced-order modes, transform the original partial differential equations into a small number of ordinary differential equations, extract the main modes through step three, and express the original temperature field in the form of baseline linear superposition; based on the singular value decomposition results, use the Galerkin projection method to transform the POD reduced-order mode into a system of linear equations; Step 5, solving the linear equations to obtain the temperature field modal coefficients, which constitute the required parameters of the reduced-order model; The equation of unknown temperature is converted into an equation about basis function, and the temperature field at the current moment is indirectly obtained by solving the coefficients. Combined with Galerkin projection, the thermal reduction model of IGBT module is obtained. The specific steps are: The temperature field is a linear combination of a finite number of basis functions, expressed as: ; In the above formula, For spatial location x and time t The temperature field; M is the number of selected modes; is the time coefficient corresponding to the basis function; is the spatial basis function.

[0032] The heat conduction equation is discretized as: ; The singular value decomposition basis function is: ; In the above formula, is the coefficient of the kth mode at time step n; Perform the inner product on the discretized heat conduction equation, and the modal coefficient update formula is: ; In the above formula, is the coefficient matrix, is the projection coefficient related to the heat source term; The temperature field is reconstructed by coefficients and basis functions, forming a reduced-order model of the IGBT module: ; Step 6: Further simplify the linear equations into matrix form to simplify the calculation; Step seven, introduce the convection coefficient of the water cooling system into the reduced-order model, and build a reduced-order thermal model of the IGBT module by combining the reduced-order modal coefficients, linear equations and convection coefficients.

[0033] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A real-time evaluation method for reduced-order junction temperature of a traction converter IGBT module thermal model, characterized in that: Includes steps: Establish the circuit model of IGBT to calculate the loss; Establish an IGBT structural model based on the spatial position and geometric dimensions of the IGBT and water cooling plate; The temperature field of the water-cooled IGBT module under different boundary conditions is used as sample data, and the modes of the water-cooled IGBT temperature field are extracted based on the singular value decomposition method. Based on the extracted main modes and combined with Galerkin projection, the high-order heat conduction equations are projected into the subspace to obtain the reduced-order thermal model of the water-cooled IGBT module. Substitute the IGBT loss and water cooling convection coefficient into the cold IGBT thermal reduction model to calculate the IGBT junction temperature.

2. A real-time evaluation method for reduced-order junction temperature of a traction converter IGBT module thermal model as claimed in claim 1, characterized in that: The method of establishing a circuit model of the IGBT to calculate the loss includes: Extract key parameters according to the IGBT module product data sheet, including initial saturation voltage drop, on-resistance, and thermal characteristics; The conduction loss of the IGBT module is a function of the initial saturation voltage drop and the on-resistance. It is characterized by interpolation calculation and a linear equation for the output characteristic curve. The switching loss is based on the relationship between the switching energy loss and the current in the data sheet, and the polynomial interpolation method is used to construct a quadratic equation for the switching energy loss. The correction coefficient under actual operating conditions is determined based on the slope of the curve where the switching energy loss changes with voltage and temperature.

3. A real-time evaluation method for reduced-order junction temperature of a traction converter IGBT module thermal model as claimed in claim 2, characterized in that: The IGBT structural model is established based on the spatial position and geometric dimensions of the IGBT module and the water cooling plate, including: The precise dimensions and thickness parameters of each layer structure inside the IGBT module and the water-cooling plate are obtained, and a three-dimensional structural model of the water-cooled IGBT module is established.

4. A real-time evaluation method for reduced-order junction temperature of a traction converter IGBT module thermal model as claimed in claim 3, characterized in that: The three-dimensional structure model drawing software is SolidWorks.

5. A real-time evaluation method for reduced-order junction temperature of a traction converter IGBT module thermal model as claimed in claim 4, characterized in that: The method uses the temperature field of the water-cooled IGBT module under different boundary conditions as sample data and extracts the mode of the water-cooled IGBT temperature field based on the singular value decomposition method, including: The three-dimensional structural model of the water-cooled plate and IGBT module was imported into the ANSYS Icepak simulation software, meshed, and the boundary conditions were set; the water-cooled plate temperature field was obtained by fluid-thermal simulation, and then the equivalent forced convection heat transfer coefficient of the contact surface between the water-cooled plate and the IGBT was obtained; the water-cooled IGBT temperature field was obtained in ANSYS Workbench based on the convection heat transfer coefficient, and several modes of the water-cooled IGBT temperature field were extracted based on the singular value decomposition method.

6. A real-time evaluation method for reduced-order junction temperature of a traction converter IGBT module thermal model as claimed in claim 5, characterized in that: The method for extracting several modes of the water-cooled IGBT temperature field based on the singular value decomposition method includes: The heat conduction equation is: ; In the above formula, ρ is the IGBT chip material density; c is the specific heat capacity of the IGBT chip; T is temperature; is the divergence of heat flow, which indicates the heat flow rate per unit area in a certain direction; λ is the thermal conductivity of the IGBT chip; Q is the internal heating rate of the IGBT chip; The convection equation is: ; In the above formula, q ' is the heat flux density per unit area during convection; h is the convection coefficient; T c is the IGBT junction temperature; T w The temperature of the fluid in the environment that performs convection heat exchange with the IGBT module housing; The temperature field of the IGBT module at different time points is expressed by a finite number of basis functions and represented as a linear equation system; by solving the linear equation system, M coefficients are obtained, each coefficient corresponds to a basis function, so as to obtain an approximate solution of the temperature field at the current moment.

7. A real-time evaluation method for reduced-order junction temperature of a traction converter IGBT module thermal model as claimed in claim 6, characterized in that: Based on the extracted main modes, combined with Galerkin projection, the high-order heat conduction equation is projected into the subspace to obtain a reduced-order thermal model of the water-cooled IGBT module, including: The singular value decomposition method is used to extract the basis functions, and the original temperature field is expressed as a linear superposition of basis functions. The complex original temperature field is represented by a combination of a few basis functions. Based on the Galerkin projection, the inner product operation is performed on both sides of the temperature equation using the extracted basis functions to obtain the thermal reduced-order model of the IGBT module.

8. A real-time evaluation method for reduced-order junction temperature of a traction converter IGBT module thermal model as claimed in claim 7, characterized in that: Substituting the IGBT loss and the water-cooling convection coefficient into the water-cooled IGBT thermal reduction model to calculate the IGBT junction temperature includes: The calculated IGBT loss and water-cooling convection coefficient are substituted into the reduced-order thermal model to solve the junction temperature of the IGBT module. Combined with real-time data, the junction temperature calculation is dynamically updated, so that the temperature change of the IGBT module can be evaluated in real time under different working conditions.

9. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that: The memory stores a computer program that can be run on the processor, and when the processor loads and executes the computer program, a real-time evaluation method for reduced-order junction temperature of a traction converter IGBT module thermal model as described in any one of claims 1 to 8 is adopted.

Images