A semiconductor power module optimization method and system based on loss calculation

Abstract

The present invention discloses a semiconductor power module optimization method based on loss calculation, the method includes using finite element analysis to simulate the material characteristic data of the semiconductor power module to obtain a loss curve of the power loss data, using a particle swarm optimization algorithm to optimize the design parameters of the semiconductor power module, constructing a three-dimensional model of the semiconductor power module and performing meshing, performing multi-physical field coupling simulation analysis, evaluating the anti-loss performance of the three-dimensional model of the semiconductor power module at different operating temperatures according to the analysis results and a mapping relationship table, and optimizing and improving the semiconductor power module with the trained three-dimensional model of the semiconductor power module. The present invention simulates the change of loss data of the power module under different working environments, uses the relationship between the internal parameters of the semiconductor power module to optimize the design parameters of the power module using a particle optimization algorithm, and improves the reliability and stability of the semiconductor power module under different working environments.

Description

Technical Field

[0001] The present invention relates to the technical field of power module optimization, and in particular to a semiconductor power module optimization method and system based on loss calculation. Background Art

[0002] With the continuous advancement of semiconductor technology and the sustained development of the market, semiconductor power modules, as key components for power conversion and circuit control in electronic equipment, play an important role in various fields.

[0003] Semiconductor power modules face a changing working environment in actual applications. Different ambient temperatures will cause changes in the material properties inside the module, thus affecting its performance and reliability. As the core component of the power module, the microscopic defect evolution behavior of semiconductor devices shows obvious differences at different ambient temperatures. The influence of external environmental factors will accelerate the generation and diffusion of lattice defects, resulting in a decrease in carrier mobility, an increase in device on-resistance, and an increase in power loss. At the same time, the accumulation of defects inside the device will cause a local hot spot effect, exacerbating the uneven distribution of junction temperature and reducing heat dissipation efficiency.

[0004] It can be seen that how to optimize semiconductor power modules to reduce losses under environmental influences has become a technical problem that needs to be urgently solved by those skilled in the art. Summary of the invention

[0005] The present invention provides a semiconductor power module optimization method and system based on loss calculation to improve the anti-loss performance of the semiconductor power module.

[0006] In order to solve the above technical problems, an embodiment of the present invention provides a semiconductor power module optimization method based on loss calculation, comprising:

[0007] The intrinsic characteristic data of the target semiconductor power module is measured, and the material characteristic data of the target semiconductor power module is obtained according to the intrinsic characteristic data.

[0008] The finite element analysis method is used to perform loss simulation on the material characteristic data, and a mapping relationship table between the total power loss value of the semiconductor power module at each temperature and each of the material characteristic data is established according to the simulation results.

[0009] A particle swarm optimization algorithm is used to optimize the design parameters of the semiconductor power module at each operating temperature, and the optimized design parameters are used to construct a three-dimensional model of the semiconductor power module and evaluate the anti-loss performance of the three-dimensional model of the semiconductor power module. The anti-loss performance evaluation is designed to mesh the three-dimensional model of the semiconductor power module and perform multi-physical field coupling simulation analysis. The anti-loss performance of the three-dimensional model of the semiconductor power module at different operating temperatures is evaluated based on the analysis results and the mapping relationship table.

[0010] If the anti-loss performance of the semiconductor power module three-dimensional model does not meet the preset standards, adjust the target weight of the particle swarm optimization algorithm to re-optimize the design parameters, and return to the steps of using the optimized design parameters to construct the semiconductor power module three-dimensional model and evaluate the anti-loss performance of the semiconductor power module three-dimensional model until the corresponding anti-loss performance of the semiconductor power module three-dimensional model reaches the preset standards.

[0011] The semiconductor power module is optimized and improved by using the three-dimensional model of the semiconductor power module that meets the preset standard.

[0012] Further, the measuring of the intrinsic characteristic data of the target semiconductor power module and obtaining the material characteristic data of the target semiconductor power module according to the intrinsic characteristic data include at least one of the following:

[0013] Acquire lattice distortion data and dislocation density distribution map of the target semiconductor power module, and establish a defect feature associated data group based on the lattice distortion data and the dislocation density distribution map; obtain interface roughness parameters of the target semiconductor power module based on the defect feature associated data group, and obtain interface bonding state data of the target semiconductor power module using a synchrotron radiation photoelectron spectrometer based on the interface roughness parameters.

[0014] The carrier concentration data and resistivity data of the target semiconductor power module are obtained, and a carrier concentration gradient mapping association table is constructed according to the carrier concentration data and the resistivity data; the carrier lifetime data is calculated according to the carrier concentration gradient mapping association table using the photoconductivity attenuation method, and the carrier band structure data of the target semiconductor power data is obtained, and a scattering characteristic data group of the target semiconductor power module is constructed according to the carrier lifetime data and the carrier band structure data.

[0015] Furthermore, the finite element analysis method is used to perform loss simulation on the material characteristic data, and a mapping relationship table between the total power loss value of the semiconductor power module at each temperature and each of the material characteristic data is established according to the simulation results, including:

[0016] The finite element analysis method is used to perform loss simulation on the material characteristic data to obtain a loss curve reflecting the power loss data of the semiconductor power module at different operating temperatures.

[0017] A mapping relationship table between the total power loss value of the semiconductor power module at each temperature and each material characteristic data is established according to each of the loss curves at different temperatures.

[0018] Furthermore, the finite element analysis method is used to perform loss simulation on the material characteristic data to obtain a loss curve reflecting the power loss data of the semiconductor power module at different operating temperatures, including:

[0019] A finite element basic grid is constructed according to the material characteristic data, and the unit density distribution of the finite element basic grid is determined by hot spot prejudgment to obtain an initial grid of the temperature field.

[0020] Boundary temperature and external load current data are imported into the initial grid of the temperature field, and the temperature distribution value in the initial grid of the temperature field is calculated by a thermal field finite element solver. The temperature gradient value of the hot spot area in the initial grid of the temperature field is obtained through a heat source term boundary function according to the temperature distribution value.

[0021] The temperature stress initial value field in the temperature field initial grid is constructed according to the temperature gradient value, and the temperature stress initial value field is iteratively updated by using a multi-grid iteration method to obtain a stress distribution map.

[0022] The power loss density is calculated according to the failure area data in the stress distribution map and the current density distribution of the stress distribution map.

[0023] The power loss density obtained according to different time sequences is used to obtain the loss curve of the semiconductor power module changing with time at different operating temperatures.

[0024] Furthermore, constructing a finite element basic grid according to the material characteristic data, determining the unit density distribution of the finite element basic grid by hot spot pre-judgment, and obtaining an initial grid of the temperature field includes:

[0025] A finite element basic grid is constructed according to the material characteristic data, the cell density distribution of the finite element basic grid is determined by hot spot prejudgment, and an initial temperature field grid is obtained by using a tetrahedron to hexahedron grid converter.

[0026] Furthermore, the use of a particle swarm optimization algorithm to optimize the design parameters of the semiconductor power module at each operating temperature includes:

[0027] A multi-objective function is constructed with the goals of minimizing the temperature peak, minimizing the temperature gradient, minimizing the current density peak and minimizing the stress peak of the semiconductor power module at each operating temperature.

[0028] The design parameters of the semiconductor power module are optimized using a particle optimization algorithm according to the multi-objective function.

[0029] Furthermore, the use of the optimized design parameters to construct a three-dimensional model of a semiconductor power module and to evaluate the anti-loss performance of the three-dimensional model of the semiconductor power module includes:

[0030] A three-dimensional model of a semiconductor power module is constructed according to the optimized design parameters and the three-dimensional model of the semiconductor power module is meshed.

[0031] Obtain the temperature field distribution data and stress distribution data in each grid at each temperature.

[0032] According to the temperature field distribution data and the stress distribution data, a multi-physical field coupling simulation analysis method is used to perform loss simulation analysis on the semiconductor power module three-dimensional model, and the anti-loss performance of the semiconductor power module three-dimensional model at different operating temperatures is evaluated based on the simulation results and the mapping relationship table.

[0033] Furthermore, the preset standards for the anti-loss performance of the semiconductor power module three-dimensional model include that the hotspot temperature is higher than a preset hotspot temperature threshold, the current distribution uniformity is lower than the current distribution uniformity level, or the stress peak is higher than the material strength limit; if the anti-loss performance of the semiconductor power module three-dimensional model does not meet the preset standards, the target weight of the particle swarm optimization algorithm is adjusted to re-optimize the design parameters, and the steps of constructing the semiconductor power module three-dimensional model using the optimized design parameters and evaluating the anti-loss performance of the semiconductor power module three-dimensional model are returned until the corresponding anti-loss performance of the semiconductor power module three-dimensional model reaches the preset standards, including:

[0034] If the anti-loss performance of the three-dimensional model of the semiconductor power module does not meet any of the preset standards, the weight coefficient of each objective in the multi-objective function of the particle swarm optimization algorithm is adjusted.

[0035] The design parameters in the three-dimensional model of the semiconductor power module whose anti-loss performance does not meet any of the preset standards are re-optimized according to the adjusted particle swarm optimization algorithm, and the step of constructing the three-dimensional model of the semiconductor power module using the optimized design parameters and evaluating the anti-loss performance of the three-dimensional model of the semiconductor power module is returned to, until the corresponding anti-loss performance of the three-dimensional model of the semiconductor power module reaches the preset standard.

[0036] Furthermore, the method further comprises:

[0037] A semiconductor power module sample is designed according to each of the semiconductor power module three-dimensional models that meet preset standards, the semiconductor power module sample is put into use, and the reliability of the semiconductor power module three-dimensional model is evaluated based on the operating data of the semiconductor power module sample.

[0038] Another embodiment of the present invention provides a semiconductor power module optimization system based on loss calculation, comprising:

[0039] The data acquisition module is used to measure the intrinsic characteristic data of the target semiconductor power module and obtain the material characteristic data of the target semiconductor power module according to the intrinsic characteristic data.

[0040] The loss analysis module is used to perform loss simulation on the material characteristic data using the finite element analysis method, and to establish a mapping relationship table between the total power loss value of the semiconductor power module at each temperature and each of the material characteristic data according to the simulation results.

[0041] An optimization and evaluation module is used to use a particle swarm optimization algorithm to optimize the design parameters of the semiconductor power module at each operating temperature, use the optimized design parameters to construct a three-dimensional model of the semiconductor power module and evaluate the anti-loss performance of the three-dimensional model of the semiconductor power module, the anti-loss performance evaluation is designed to mesh the three-dimensional model of the semiconductor power module and perform multi-physical field coupling simulation analysis, and evaluate the anti-loss performance of the three-dimensional model of the semiconductor power module at different operating temperatures based on the analysis results and the mapping relationship table.

[0042] A weight adjustment module is used to adjust the target weight of the particle swarm optimization algorithm to re-optimize the design parameters if the anti-loss performance of the semiconductor power module three-dimensional model does not meet the preset standards, and return to the steps of constructing the semiconductor power module three-dimensional model using the optimized design parameters and evaluating the anti-loss performance of the semiconductor power module three-dimensional model until the corresponding anti-loss performance of the semiconductor power module three-dimensional model reaches the preset standards.

[0043] The optimization and improvement module is used to optimize and improve the semiconductor power module based on the three-dimensional model of the semiconductor power module that meets the preset standard.

[0044] Furthermore, the data acquisition module is specifically used for:

[0045] Acquire lattice distortion data and dislocation density distribution map of the target semiconductor power module, and establish a defect feature associated data group based on the lattice distortion data and the dislocation density distribution map; obtain interface roughness parameters of the target semiconductor power module based on the defect feature associated data group, and obtain interface bonding state data of the target semiconductor power module using a synchrotron radiation photoelectron spectrometer based on the interface roughness parameters.

[0046] The carrier concentration data and resistivity data of the target semiconductor power module are obtained, and a carrier concentration gradient mapping association table is constructed according to the carrier concentration data and the resistivity data; the carrier lifetime data is calculated according to the carrier concentration gradient mapping association table using the photoconductivity attenuation method, and the carrier band structure data of the target semiconductor power data is obtained, and a scattering characteristic data group of the target semiconductor power module is constructed according to the carrier lifetime data and the carrier band structure data.

[0047] Furthermore, the loss analysis module is specifically used for:

[0048] The finite element analysis method is used to perform loss simulation on the material characteristic data to obtain a loss curve reflecting the power loss data of the semiconductor power module at different operating temperatures.

[0049] A mapping relationship table between the total power loss value of the semiconductor power module at each temperature and each material characteristic data is established according to each of the loss curves at different temperatures.

[0050] Furthermore, the finite element analysis method is used to perform loss simulation on the material characteristic data to obtain a loss curve reflecting the power loss data of the semiconductor power module at different operating temperatures, including:

[0051] A finite element basic grid is constructed according to the material characteristic data, and the unit density distribution of the finite element basic grid is determined by hot spot prejudgment to obtain an initial grid of the temperature field.

[0052] Boundary temperature and external load current data are imported into the initial grid of the temperature field, and the temperature distribution value in the initial grid of the temperature field is calculated by a thermal field finite element solver. The temperature gradient value of the hot spot area in the initial grid of the temperature field is obtained through a heat source term boundary function according to the temperature distribution value.

[0053] The temperature stress initial value field in the temperature field initial grid is constructed according to the temperature gradient value, and the temperature stress initial value field is iteratively updated by using a multi-grid iteration method to obtain a stress distribution map.

[0054] The power loss density is calculated according to the failure area data in the stress distribution map and the current density distribution of the stress distribution map.

[0055] The power loss density obtained according to different time sequences is used to obtain the loss curve of the semiconductor power module changing with time at different operating temperatures.

[0056] Furthermore, constructing a finite element basic grid according to the material characteristic data, determining the unit density distribution of the finite element basic grid by hot spot pre-judgment, and obtaining an initial grid of the temperature field includes:

[0057] A finite element basic grid is constructed according to the material characteristic data, the cell density distribution of the finite element basic grid is determined by hot spot prejudgment, and an initial temperature field grid is obtained by using a tetrahedron to hexahedron grid converter.

[0058] Furthermore, the optimization evaluation module is specifically used to construct a multi-objective function with the goals of minimizing the temperature value peak, minimizing the temperature gradient, minimizing the current density peak and minimizing the stress peak of the semiconductor power module at each operating temperature.

[0059] The design parameters of the semiconductor power module are optimized using a particle optimization algorithm according to the multi-objective function.

[0060] Furthermore, the optimization evaluation module is also used to construct a three-dimensional model of a semiconductor power module according to the optimized design parameters and to mesh the three-dimensional model of the semiconductor power module.

[0061] Obtain the temperature field distribution data and stress distribution data in each grid at each temperature.

[0062] According to the temperature field distribution data and the stress distribution data, a multi-physical field coupling simulation analysis method is used to perform loss simulation analysis on the semiconductor power module three-dimensional model, and the anti-loss performance of the semiconductor power module three-dimensional model at different operating temperatures is evaluated based on the simulation results and the mapping relationship table.

[0063] Furthermore, the preset standards for the anti-loss performance of the semiconductor power module three-dimensional model include the hot spot temperature being higher than a preset hot spot temperature threshold, the current distribution uniformity being lower than the current distribution uniformity level, or the stress peak being higher than the material strength limit.

[0064] The weight adjustment module is specifically used for:

[0065] If the anti-loss performance of the three-dimensional model of the semiconductor power module does not meet any of the preset standards, the weight coefficient of each objective in the multi-objective function of the particle swarm optimization algorithm is adjusted.

[0066] The design parameters in the three-dimensional model of the semiconductor power module whose anti-loss performance does not meet any of the preset standards are re-optimized according to the adjusted particle swarm optimization algorithm, and the step of constructing the three-dimensional model of the semiconductor power module using the optimized design parameters and evaluating the anti-loss performance of the three-dimensional model of the semiconductor power module is returned to, until the corresponding anti-loss performance of the three-dimensional model of the semiconductor power module reaches the preset standard.

[0067] Furthermore, the system also includes a reliability assessment module, which is used to design semiconductor power module samples based on each of the semiconductor power module three-dimensional models that meet preset standards, put the semiconductor power module samples into use, and perform reliability assessment on the semiconductor power module three-dimensional models based on operating data of the semiconductor power module samples.

[0068] Another embodiment of the present invention provides a computer device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor implements the semiconductor power module optimization method based on loss calculation as described above when executing the computer program.

[0069] Yet another embodiment of the present invention provides a computer-readable storage medium storing a computer program, wherein when the device where the computer-readable storage medium is located executes the computer program, the semiconductor power module optimization method based on loss calculation as described above is implemented.

[0070] Compared with the prior art, the embodiments of the present invention have the following advantages:

[0071] (1) The finite element analysis method is used to simulate the change of power module loss data under different working environments based on parameters such as semiconductor device defects, interface characteristics, and carrier transport characteristics. The relationship between loss data and material properties is analyzed, and a loss mechanism model is constructed, revealing the intrinsic parameter connection in the loss process of semiconductor power modules.

[0072] (2) The relationship between the internal parameters of the semiconductor power module is used to optimize the design parameters of the power module using a particle optimization algorithm, thereby improving the reliability and stability of the semiconductor power module under different working environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] Figure 1 A flowchart of the steps of a semiconductor power module optimization method based on loss calculation provided by an embodiment of the present invention;

[0074] Figure 2 A structural block diagram of a semiconductor power module optimization system based on loss calculation provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0075] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. The purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0076] In the description of this application, the terms "first", "second", "third", etc. are used for descriptive purposes only and are not to be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first", "second", "third", etc. may explicitly or implicitly include one or more of the feature. In the description of this application, unless otherwise specified, "plurality" means two or more.

[0077] In the description of the present application, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be a connection between the two elements. The terms "vertical", "horizontal", "left", "right", "upper", "lower" and similar expressions used herein are only for illustrative purposes, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. The term "and / or" used herein includes any and all combinations of one or more related listed items. For those of ordinary skill in the art, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0078] In the description of this application, it should be noted that, unless otherwise defined, all technical and scientific terms used in the present invention have the same meaning as those commonly understood by those skilled in the art. The terms used in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. For those of ordinary skill in the art, the specific meanings of the above terms in this application can be understood by specific circumstances.

[0079] An embodiment of the present invention provides a semiconductor power module optimization method based on loss calculation. For details, see Figure 1 , Figure 1The flowchart is a step flow chart of a semiconductor power module optimization method based on loss calculation in one embodiment of the present invention, including steps S11 to S16:

[0080] Step S11 , measuring intrinsic characteristic data of a target semiconductor power module, and obtaining material characteristic data of the target semiconductor power module according to the intrinsic characteristic data.

[0081] The characteristic data of semiconductor power modules provides detailed information about the internal structure and performance parameters of the modules. By obtaining this data, the structural characteristics of the modules can be analyzed more accurately. When the module loses power, it can help to quickly locate the problem and reveal the relationship between the loss data and the material properties.

[0082] Generally speaking, the material property data of a semiconductor power module include parameters related to semiconductor device defects, interface characteristics, and carrier transport characteristics. Specifically, this embodiment measures and processes the intrinsic characteristic data of the target semiconductor power module to obtain material property data such as lattice distortion data, dislocation density distribution map, interface roughness parameters, interface bonding state data, carrier concentration data, resistivity data, carrier lifetime data, and carrier band structure data that are interrelated with each other.

[0083] Specifically, the process of obtaining the above data is as follows:

[0084] A high-temperature variable temperature Hall measurement device is used to obtain the lattice distortion data of the target semiconductor power module, and a transmission electron microscope is used to obtain the dislocation density distribution map of the target semiconductor power module. A defect feature correlation data group is established based on the lattice distortion data and the dislocation density distribution map.

[0085] The interface roughness parameters of the target semiconductor power module are measured according to the defect feature correlation data group, and then the interface bonding state data of the target semiconductor power module is obtained using a synchrotron radiation photoelectron spectrometer according to the interface roughness parameters.

[0086] If the carrier concentration gradient in the interface chemical bonding state data exceeds the concentration gradient threshold, the carrier concentration data is obtained through a Hall effect measurement device, and the carrier concentration data and the resistivity data measured by the four-probe method are used to construct a carrier concentration gradient mapping association table.

[0087] The photoconductivity attenuation method is used to obtain carrier lifetime data from the carrier concentration gradient mapping correlation table. The carrier lifetime data and the carrier band structure data measured by the ultraviolet photoelectron spectrometer are combined to generate scattering characteristic data.

[0088] The above material property data, such as lattice distortion data, dislocation density distribution map, interface roughness parameters, interface bonding state data, carrier concentration data, resistivity data, carrier lifetime data and carrier band structure data, are interrelated and influence each other, which helps to deeply understand the material performance and analyze the internal loss mechanism of semiconductor power modules.

[0089] Step S12, using finite element analysis to perform loss simulation on the material characteristic data to obtain a loss curve reflecting the power loss data of the semiconductor power module at different operating temperatures, and establishing a mapping relationship table between the total power loss value at each temperature and each material characteristic data according to the loss curves at different temperatures.

[0090] After obtaining the material property data of the semiconductor power module, it is necessary to analyze it to predict the relationship between the loss of the semiconductor power module and the material property data. The method used in this embodiment is to use the finite element analysis method to simulate the loss of the material property data to obtain a loss curve reflecting the power loss data of the semiconductor power module at different operating temperatures. The specific process is as follows:

[0091] The finite element basic mesh is constructed according to the material property data, the cell density distribution of the finite element basic mesh is determined by hot spot pre-judgment, and the initial temperature field mesh is obtained using a tetrahedron to hexahedron mesh converter.

[0092] Tetrahedral meshes are highly adaptable and can accommodate almost any complex structure. Therefore, using tetrahedral meshes in the initial stage makes it easier to handle complex geometries. With the tetrahedral to hexahedral mesh converter, tetrahedral meshes can be converted to more accurate hexahedral meshes in subsequent stages, thereby improving computational efficiency without sacrificing geometric accuracy.

[0093] The boundary temperature and external load current data are imported into the initial grid of the temperature field. The temperature distribution value in the initial grid of the temperature field is calculated by the thermal field finite element solver. The temperature gradient value of the hot spot area in the initial grid of the temperature field is obtained through the boundary function of the heat source term.

[0094] According to the temperature gradient value, the temperature stress initial value field in the initial grid of the temperature field is constructed, and the temperature stress initial value field is iteratively updated by using the multi-grid iteration method to obtain the stress distribution map; the power loss density is calculated according to the failure area data in the stress distribution map and the current density distribution of the stress distribution map;

[0095] The power loss density obtained at different time sequences is used to obtain the loss curve of the semiconductor power module that changes with time.

[0096] Specifically, the steps for obtaining the loss curve in this embodiment are:

[0097] The finite element basic grid is constructed according to the material property data. The grid unit size is divided into 0.1 mm for the geometric structure of the semiconductor power module. The grid density ratio is increased by 5 times through the hot spot pre-judgment area. The initial grid of the temperature field is generated from the basic grid data using the tetrahedron to hexahedron grid converter, and the current density distribution value of the unit node is calculated according to the electromagnetic field solver. The thermal field finite element solver is used to import the boundary temperature of 300 Kelvin and the external load current of 10 amperes. The implicit time integral solver is used to calculate the temperature distribution for the transient temperature field with a time step of 0.1 milliseconds. The temperature gradient value of the hot spot area is calculated by the boundary function of the heat source term, the thermal stress initial value field is constructed from the temperature distribution data, and the current concentration unit node position is calculated from the Ohm loss function. The stress calculation area is imported from the temperature stress initial value field. The stress concentration unit is calculated by multi-grid iteration with a residual less than 0.001 for the stress distribution. The maximum stress value is calculated by thermal stress coupling. The convolutional neural network is used to extract the characteristic vector from the stress distribution map, the failure area is calibrated according to the stress threshold function, and the initial loss calculation value is constructed from the failure area data. The power loss density distribution is calculated based on the current density distribution and stress failure area data. The power loss density distribution is calculated based on the material property data obtained at different time sequences to generate a loss curve of the semiconductor power module over time.

[0098] After the loss curve is obtained, a mapping relationship table between the total power loss value at each temperature and each material property data is established according to each loss curve calculated at different ambient temperatures. Through the mapping relationship table, the material property data corresponding to the total power loss value at different temperatures can be quickly obtained without recalculating or simulating, which improves the efficiency of the analysis, especially in scenarios that require multiple iterations or optimized designs.

[0099] Step S13, using a particle swarm optimization algorithm to optimize the design parameters of the semiconductor power module at each operating temperature.

[0100] In actual semiconductor power module applications, it is necessary to find a balance among multiple optimization objectives and comprehensively consider multiple performance indicators or constraints. These optimization objectives often conflict or restrict each other. Therefore, a single-objective optimization method is difficult to meet all requirements at the same time. This embodiment adopts a multi-objective optimization method to comprehensively consider multiple objectives, with the goals of minimizing the temperature numerical peak, minimizing the temperature gradient, minimizing the current density peak and minimizing the stress peak at each operating temperature, constructing a multi-objective function, and optimizing the design parameters of the power module using a particle optimization algorithm based on the multi-objective function.

[0101] Specifically, the material property data of the module is used to optimize the design parameters of the semiconductor power module under different temperature conditions according to the objective function, so as to obtain the most optimized design parameter combination of the semiconductor power module at each temperature.

[0102] Step S14, using the optimized design parameters to construct a three-dimensional model of a semiconductor power module, meshing the three-dimensional model of the semiconductor power module and performing multi-physical field coupling simulation analysis, and evaluating the anti-loss performance of the three-dimensional model of the semiconductor power module at different operating temperatures based on the analysis results and the mapping relationship table.

[0103] A three-dimensional model of a semiconductor power module is constructed according to the optimized design parameter combination. The three-dimensional model of the semiconductor power module is divided into grid units in the chip area to obtain a temperature field grid. The units of the temperature field grid are determined by the temperature gradient output by the convolutional neural network predictor.

[0104] Thermoelectric boundary conditions are introduced into the divided temperature field grid, the Joule heat distribution in the grid is calculated through the boundary conditions, and the temperature field distribution data is obtained from the Joule heat distribution.

[0105] The temperature field distribution data is used to construct the chip area material constitutive parameters of the semiconductor power module three-dimensional model, and the stress distribution data in the grid is calculated based on the material constitutive parameters.

[0106] According to the temperature field distribution data and stress distribution data, the multi-physical field coupling simulation analysis method is used to perform loss simulation analysis on the three-dimensional model of the semiconductor power module. According to the simulation results and the mapping relationship table, it is evaluated whether the anti-loss performance of the three-dimensional model of the semiconductor power module at different operating temperatures meets the pre-examination anti-loss standard.

[0107] Specifically, this embodiment constructs a three-dimensional solid geometry according to the optimized power module structural parameters, uses a 0.1 mm grid unit for the chip area in the ambient temperature range of 250 to 350 Kelvin, calibrates the initial value of the temperature field by the heat source power density of 0.5 watts per square millimeter, inputs the temperature gradient and stress peak data from the convolutional neural network predictor to generate a mesh encryption area, and uses a tetrahedron to hexahedron grid converter to construct the temperature field grid. The thermoelectric boundary conditions are imported from the multi-physics field solver, and the temperature coefficient is set to 0.01 per Kelvin for the material thermal conductivity in the range of 250 to 350 Kelvin. The Joule heat distribution is calculated from the conductivity temperature function, the enthalpy distribution is solved by the heat conduction equation, and the boundary node temperature is calculated according to the convection heat transfer coefficient of 25 watts per square meter Kelvin. The initial value of thermal stress is established from the multi-physics field coupler. The constitutive parameters of the chip area material are constructed from the initial value of thermal stress, the boundary displacement is set for the elastic modulus temperature function and the thermal expansion coefficient function, the plastic deformation is calculated by the Coulomb strength criterion, the residual stress value is calibrated according to the interface shear strength of 10 MPa, the stress distribution data is generated by the cubic spline interpolation function, and the stress level is calculated from the stress distribution spectrum. The key performance parameters are calculated based on the multi-physics field coupling data, the hot spot temperature distribution curve is output for the ambient temperature range of 250 to 350 Kelvin, the current distribution uniformity is calculated from the current density variance, the stress level change law is obtained through the stress peak curve, the radial basis function is used to construct the ambient temperature and performance index mapping relationship table, and the performance parameter group corresponding to different ambient temperatures is output from the mapping relationship table. The three-dimensional geometric mesh division of the power module plays a key role in the simulation accuracy. The chip area uses a basic mesh unit with a side length of 0.1 mm. The mesh density is increased by 5 times in the area where the temperature gradient is greater than 5 Kelvin per micron to ensure the accuracy of temperature field calculation. The convolutional neural network learns the characteristics of high temperature gradient and stress concentration from the existing simulation database, predicts the key heat dissipation area and stress concentration area, and realizes local mesh intelligent encryption. The setting of thermoelectric boundary conditions is based on the actual working environment. The thermal conductivity of the material changes significantly with temperature. At 250 Kelvin, the thermal conductivity of the silicon wafer is 150 watts per meter Kelvin, and it decreases to 90 watts per meter Kelvin when the temperature rises to 350 Kelvin. The Joule heat distribution calculation shows that the center temperature of the chip is the highest when the power is turned on, forming a heat source density of about 0.5 watts per square millimeter. The boundary layer convection heat transfer coefficient of 25 watts per square meter Kelvin reflects the forced air cooling condition, and the enthalpy distribution shows that the heat flux density from the chip to the substrate decays exponentially. The material constitutive relationship describes the temperature-stress coupling effect. The elastic modulus of the silicon wafer decreases from 170 GPa at room temperature to 150 GPa as the temperature increases, and the thermal expansion coefficient increases from 2.6 to 3.2 microns per meter Kelvin. The Coulomb strength criterion determines the starting point of plastic deformation. Plastic deformation occurs when the shear stress exceeds 10 MPa, and the residual stress is mainly concentrated at the interface between the chip and the substrate. Cubic spline interpolation reconstructs the continuous stress field, reflecting the spatial gradient characteristics of the stress distribution.The key performance indicator evaluation reflects the device reliability. In the ambient temperature range of 250 to 350 Kelvin, the hot spot temperature increases from 325 Kelvin to 425 Kelvin, and the temperature gradient increases from 3 to 7 Kelvin per micron. The current distribution uniformity is characterized by the standard deviation. The increase in the standard deviation from 5% to 12% indicates that the current distribution non-uniformity increases at high temperatures. The stress level increases approximately linearly with temperature, and the peak stress increases from 50 MPa to 80 MPa. The radial basis function fitting established the mapping relationship between the performance indicators and the ambient temperature. For every 10 Kelvin increase in temperature, the hot spot temperature increases by 8 Kelvin, the current non-uniformity increases by 0.8%, and the stress level increases by 3 MPa.

[0108] Step S15, if it is detected that the anti-loss performance of the three-dimensional model of the semiconductor power module does not meet the preset conditions, adjust the target weight of the particle swarm optimization algorithm to re-optimize the design parameters until the corresponding anti-loss performance of the three-dimensional model of the semiconductor power module reaches the preset standard.

[0109] The parameters involved in the anti-loss performance evaluation conditions preset in this embodiment are the hotspot temperature threshold, the current distribution uniformity level and the material strength limit. Therefore, the evaluation process is: if the hotspot temperature of the three-dimensional model of the semiconductor power module is higher than the preset hotspot temperature threshold, the current distribution uniformity is lower than the current distribution uniformity level or the stress peak is higher than the material strength limit, then return to step S13, adjust the weight coefficient of each objective in the multi-objective function of the particle swarm optimization algorithm, re-optimize the design parameters in the three-dimensional model of the semiconductor power module that does not meet the standards according to the adjusted particle swarm optimization algorithm, reconstruct the three-dimensional model of the semiconductor power module according to the optimized design parameters, and perform anti-loss performance evaluation and analysis again until the anti-loss performance of the corresponding three-dimensional model of the semiconductor power module reaches the preset standard.

[0110] Specifically, the hotspot temperature refers to the temperature of the point or area with the highest temperature in the semiconductor power module. When the semiconductor power module is working, heat is generated due to the passage of current and power loss. If the heat is unevenly distributed, some areas may become abnormally hot, and these areas are called hotspots. The hotspot temperature is an important indicator to measure the heat dissipation performance of the module. Excessive hotspot temperature may cause the module to perform poorly, reduce reliability, or even be damaged. The preset hotspot temperature threshold is determined based on factors such as the material, structure, and operating conditions of the module to ensure that the module does not fail due to overheating under normal operating conditions.

[0111] Current distribution uniformity is a parameter that describes the degree of uniformity of current distribution in a semiconductor power module. It reflects the uniformity of current flow inside the module. Ideally, the current distribution should be uniform so that each part can share the same current load, thereby avoiding local overheating and current concentration. However, in practical applications, due to limitations of module structure, materials, processes and other factors, the current distribution is often uneven. The level of current distribution uniformity is usually evaluated by measuring the current density at different locations inside the module and comparing it with the preset standard to determine whether it meets the requirements. Too low current distribution uniformity may cause excessive current in certain areas inside the module, thereby accelerating aging, damage or reducing efficiency.

[0112] Stress peak refers to the maximum stress value that a semiconductor power module is subjected to during operation. These stresses may come from thermal stress caused by temperature changes, electromagnetic stress caused by current flow, and mechanical stress. Excessively high stress peaks may cause failure modes such as deformation, cracking, or fracture of the internal materials of the module. The material strength limit is the maximum stress value that the material can withstand without being damaged. Therefore, the comparison between the stress peak and the material strength limit is an important method for evaluating the module's stress resistance. If the stress peak exceeds the material strength limit, the module may fail.

[0113] Specifically, this embodiment constructs a quantitative evaluation standard from the power module simulation data, performs a numerical comparison of the hot spot temperature with the 400 Kelvin temperature threshold, determines the uniformity index by the current distribution standard deviation being less than 5%, and uses a weighted sum function to calculate the performance score based on the stress peak and the 160 MPa strength limit, and determines the optimization scheme from the score data. If the performance score exceeds 90 points, a three-dimensional geometric structure drawing is generated from the design database, the solder ratio is calibrated for the silver content range of 75% to 85%, the reflow soldering temperature is calibrated by the temperature of 350 to 450 degrees Celsius, and the size specification table is generated from the chip thickness of 0.2 to 0.5 mm and the substrate thickness of 0.5 to 1.0 mm, and the manufacturing parameter data table is output according to the manufacturing process specification. If the performance score is less than 90 points, the temperature weight is modified from the particle swarm optimizer, and the new optimization objective function value is calculated using the deep neural network predictor. The power module parameters are reconstructed according to the optimization objective function value, and the convergence is determined to be terminated if the number of optimization iterations exceeds the preset number or the anti-loss performance score increment is less than the minimum optimization expectation, and the design scheme is output according to the convergence parameter group.

[0114] Step S16, optimizing and improving the semiconductor power module using the trained three-dimensional model of the semiconductor power module.

[0115] A semiconductor power module sample is designed based on the optimized three-dimensional model of the semiconductor power module at each temperature, the semiconductor power module sample is put into use, and the reliability of the optimization result is evaluated based on the operating data of the semiconductor power module sample.

[0116] The semiconductor power module optimization method based on loss calculation of the present invention uses finite element analysis to simulate the loss data changes of power modules under different working environments according to parameters such as semiconductor device defects, interface characteristics and carrier transport characteristics, and analyzes the correlation between loss data and material characteristics, constructs a loss mechanism model, and reveals the intrinsic parameter connection in the loss process of semiconductor power modules; the particle optimization algorithm is used to optimize the design parameters of the power module by using the relationship between the internal parameters of the semiconductor power module, thereby improving the reliability and stability of the semiconductor power module under different working environments. .

[0117] The embodiment of the present invention further provides a semiconductor power module optimization system based on loss calculation, which is used to execute the semiconductor power module optimization method based on loss calculation as described above. Figure 2 The structure block diagram of the semiconductor power module optimization system based on loss calculation according to an embodiment of the present invention is as follows:

[0118] The data acquisition module 21 is used to measure the intrinsic characteristic data of the target semiconductor power module and obtain the material characteristic data of the target semiconductor power module according to the intrinsic characteristic data.

[0119] The loss analysis module 22 is used to perform loss simulation on the material characteristic data using the finite element analysis method, and to establish a mapping relationship table between the total power loss value of the semiconductor power module at each temperature and each of the material characteristic data according to the simulation results.

[0120] The optimization and evaluation module 23 is used to use a particle swarm optimization algorithm to optimize the design parameters of the semiconductor power module at each operating temperature, use the optimized design parameters to construct a three-dimensional model of the semiconductor power module and evaluate the anti-loss performance of the three-dimensional model of the semiconductor power module. The anti-loss performance evaluation is designed to mesh the three-dimensional model of the semiconductor power module and perform multi-physical field coupling simulation analysis, and evaluate the anti-loss performance of the three-dimensional model of the semiconductor power module at different operating temperatures based on the analysis results and the mapping relationship table.

[0121] The weight adjustment module 24 is used to adjust the target weight of the particle swarm optimization algorithm to re-optimize the design parameters if the anti-loss performance of the semiconductor power module three-dimensional model does not meet the preset standards, and return to the steps of constructing the semiconductor power module three-dimensional model using the optimized design parameters and evaluating the anti-loss performance of the semiconductor power module three-dimensional model until the corresponding anti-loss performance of the semiconductor power module three-dimensional model reaches the preset standards.

[0122] The optimization and improvement module 25 is used to optimize and improve the semiconductor power module based on the three-dimensional model of the semiconductor power module that meets the preset standard.

[0123] The data acquisition module 21 is specifically used for:

[0124] Acquire lattice distortion data and dislocation density distribution map of the target semiconductor power module, and establish a defect feature associated data group based on the lattice distortion data and the dislocation density distribution map; obtain interface roughness parameters of the target semiconductor power module based on the defect feature associated data group, and obtain interface bonding state data of the target semiconductor power module using a synchrotron radiation photoelectron spectrometer based on the interface roughness parameters.

[0125] The carrier concentration data and resistivity data of the target semiconductor power module are obtained, and a carrier concentration gradient mapping association table is constructed according to the carrier concentration data and the resistivity data; the carrier lifetime data is calculated according to the carrier concentration gradient mapping association table using the photoconductivity attenuation method, and the carrier band structure data of the target semiconductor power data is obtained, and a scattering characteristic data group of the target semiconductor power module is constructed according to the carrier lifetime data and the carrier band structure data.

[0126] The loss analysis module 22 is specifically used for:

[0127] The finite element analysis method is used to perform loss simulation on the material characteristic data to obtain a loss curve reflecting the power loss data of the semiconductor power module at different operating temperatures.

[0128] A mapping relationship table between the total power loss value of the semiconductor power module at each temperature and each material characteristic data is established according to each of the loss curves at different temperatures.

[0129] The method of using the finite element analysis method to perform loss simulation on the material characteristic data to obtain a loss curve reflecting the power loss data of the semiconductor power module at different operating temperatures includes: constructing a finite element basic grid according to the material characteristic data, determining the unit density distribution of the finite element basic grid through hot spot pre-judgment, and obtaining an initial grid of the temperature field.

[0130] Boundary temperature and external load current data are imported into the initial grid of the temperature field, and the temperature distribution value in the initial grid of the temperature field is calculated by a thermal field finite element solver. The temperature gradient value of the hot spot area in the initial grid of the temperature field is obtained through a heat source term boundary function according to the temperature distribution value.

[0131] The temperature stress initial value field in the temperature field initial grid is constructed according to the temperature gradient value, and the temperature stress initial value field is iteratively updated by using a multi-grid iteration method to obtain a stress distribution map.

[0132] The power loss density is calculated according to the failure area data in the stress distribution map and the current density distribution of the stress distribution map.

[0133] The power loss density obtained according to different time sequences is used to obtain the loss curve of the semiconductor power module changing with time at different operating temperatures.

[0134] The method comprises constructing a finite element basic grid according to the material characteristic data, determining the unit density distribution of the finite element basic grid through hot spot pre-judgment, and obtaining the initial grid of the temperature field, including constructing a finite element basic grid according to the material characteristic data, determining the unit density distribution of the finite element basic grid through hot spot pre-judgment, and obtaining the initial grid of the temperature field using a tetrahedron to hexahedron grid converter.

[0135] The optimization evaluation module 23 is specifically used to construct a multi-objective function with the objectives of minimizing the temperature value peak, minimizing the temperature gradient, minimizing the current density peak and minimizing the stress peak of the semiconductor power module at each operating temperature.

[0136] The design parameters of the semiconductor power module are optimized using a particle optimization algorithm according to the multi-objective function.

[0137] The optimization evaluation module 23 is further used to construct a three-dimensional model of a semiconductor power module according to the optimized design parameters and to perform meshing on the three-dimensional model of the semiconductor power module.

[0138] Obtain the temperature field distribution data and stress distribution data in each grid at each temperature.

[0139] According to the temperature field distribution data and the stress distribution data, a multi-physical field coupling simulation analysis method is used to perform loss simulation analysis on the semiconductor power module three-dimensional model, and the anti-loss performance of the semiconductor power module three-dimensional model at different operating temperatures is evaluated based on the simulation results and the mapping relationship table.

[0140] The weight adjustment module 24 is specifically used for:

[0141] The preset standards for the anti-loss performance of the three-dimensional model of the semiconductor power module include that the hot spot temperature is higher than a preset hot spot temperature threshold, the current distribution uniformity is lower than the current distribution uniformity level, or the stress peak is higher than the material strength limit.

[0142] If the anti-loss performance of the three-dimensional model of the semiconductor power module does not meet any of the preset standards, the weight coefficient of each objective in the multi-objective function of the particle swarm optimization algorithm is adjusted;

[0143] The design parameters in the three-dimensional model of the semiconductor power module whose anti-loss performance does not meet any of the preset standards are re-optimized according to the adjusted particle swarm optimization algorithm.

[0144] The semiconductor power module three-dimensional model is reconstructed according to the optimized design parameters and the anti-loss performance of the semiconductor power module three-dimensional model is evaluated until the corresponding anti-loss performance of the semiconductor power module three-dimensional model reaches a preset standard.

[0145] The system also includes a reliability assessment module 26, which is used to design semiconductor power module samples based on each of the semiconductor power module three-dimensional models that meet preset standards, put the semiconductor power module samples into use, and perform reliability assessment on the semiconductor power module three-dimensional models based on operating data of the semiconductor power module samples.

[0146] The technical features and technical effects of the device proposed in the embodiment of the present invention are the same as those of the method proposed in the embodiment of the present invention, and are not described in detail here. Each module in the above device can be implemented in whole or in part by software, hardware, and a combination thereof. Each of the above modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each of the above modules.

[0147] Yet another embodiment of the present invention provides a computer-readable storage medium storing a computer program, wherein when the device where the computer-readable storage medium is located executes the computer program, the semiconductor power module optimization method based on loss calculation as described above is implemented.

[0148] Another embodiment of the present invention provides a computer device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor implements the semiconductor power module optimization method based on loss calculation as described above when executing the computer program.

[0149] Preferably, the computer program can be divided into one or more modules / units (such as computer program 1, computer program 2, ...), and the one or more modules / units are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units can be a series of computer program instruction segments that can complete specific functions, and the instruction segments are used to describe the execution process of the computer program in the computer device.

[0150] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor, or the processor may be any conventional processor. The processor is the control center of the computer device, and various parts of the computer device are connected using various interfaces and lines.

[0151] The memory mainly includes a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function, etc., and the data storage area can store related data, etc. In addition, the memory can be a high-speed random access memory, or a non-volatile memory, such as a plug-in hard disk, a smart memory card (SmartMedia Card, SMC), a secure digital (Secure Digital, SD) card, and a flash card (Flash Card), etc., or the memory can also be other volatile solid-state storage devices.

[0152] It should be noted that the above-mentioned computer device may include, but is not limited to, a processor and a memory. Those skilled in the art will understand that the above-mentioned computer device description is only an example of a computer device and does not constitute a limitation on the computer device. It may include more or less components than shown in the figure, or a combination of certain components, or different components.

[0153] The above-mentioned embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A semiconductor power module optimization method based on loss calculation, characterized in that: include: Measuring intrinsic characteristic data of a target semiconductor power module, and obtaining material characteristic data of the target semiconductor power module according to the intrinsic characteristic data; The finite element analysis method is used to simulate the loss of the material characteristic data, and a mapping relationship table between the total power loss value of the semiconductor power module at each temperature and each of the material characteristic data is established according to the simulation results; specifically, it includes: constructing a finite element basic grid according to the material characteristic data, determining the unit density distribution of the finite element basic grid through hot spot pre-judgment, and obtaining the temperature field initial grid; importing boundary temperature and external load current data into the temperature field initial grid, using a thermal field finite element solver to calculate the temperature distribution value in the temperature field initial grid, and obtaining the temperature field initial grid through a heat source term boundary function according to the temperature distribution value. The temperature gradient value of the hot spot area; construct the temperature stress initial value field in the initial grid of the temperature field according to the temperature gradient value, and iteratively update the temperature stress initial value field using the multi-grid iteration method to obtain a stress distribution map; calculate the power loss density according to the failure area data in the stress distribution map and the current density distribution of the stress distribution map; obtain the loss curve of the semiconductor power module changing with time at different operating temperatures according to the power loss density obtained at different time sequences; establish a mapping relationship table between the total power loss value of the semiconductor power module at each temperature and each material characteristic data according to each loss curve at different temperatures; A particle swarm optimization algorithm is used to optimize the design parameters of the semiconductor power module at each operating temperature, and the optimization process includes: constructing a multi-objective function with the goal of minimizing the temperature numerical peak, minimizing the temperature gradient, minimizing the current density peak and minimizing the stress peak of the semiconductor power module at each operating temperature; optimizing the design parameters of the semiconductor power module using a particle optimization algorithm according to the multi-objective function; constructing a three-dimensional model of the semiconductor power module using the optimized design parameters and evaluating the anti-loss performance of the three-dimensional model of the semiconductor power module, wherein the anti-loss performance evaluation is designed to mesh the three-dimensional model of the semiconductor power module and perform multi-physical field coupling simulation analysis, and evaluating the anti-loss performance of the three-dimensional model of the semiconductor power module at different operating temperatures according to the analysis results and the mapping relationship table; If the anti-loss performance of the three-dimensional model of the semiconductor power module does not meet the preset standard, adjust the target weight of the particle swarm optimization algorithm to re-optimize the design parameters, and return to the step of using the optimized design parameters to construct the three-dimensional model of the semiconductor power module and evaluate the anti-loss performance of the three-dimensional model of the semiconductor power module, until the corresponding anti-loss performance of the three-dimensional model of the semiconductor power module reaches the preset standard; The semiconductor power module is optimized and improved by using the three-dimensional model of the semiconductor power module that meets the preset standard.

2. The semiconductor power module optimization method based on loss calculation according to claim 1, characterized in that: The measuring of the intrinsic characteristic data of the target semiconductor power module and obtaining the material characteristic data of the target semiconductor power module according to the intrinsic characteristic data include at least one of the following: Acquire lattice distortion data and dislocation density distribution map of the target semiconductor power module, and establish a defect feature association data group according to the lattice distortion data and the dislocation density distribution map; obtain interface roughness parameters of the target semiconductor power module according to the defect feature association data group, and obtain interface bonding state data of the target semiconductor power module using a synchrotron radiation photoelectron spectrometer according to the interface roughness parameters; The carrier concentration data and resistivity data of the target semiconductor power module are obtained, and a carrier concentration gradient mapping association table is constructed according to the carrier concentration data and the resistivity data; the carrier lifetime data is calculated according to the carrier concentration gradient mapping association table using the photoconductivity attenuation method, and the carrier band structure data of the target semiconductor power data is obtained, and a scattering characteristic data group of the target semiconductor power module is constructed according to the carrier lifetime data and the carrier band structure data.

3. The semiconductor power module optimization method based on loss calculation according to claim 1, characterized in that: The step of constructing a finite element basic grid according to the material characteristic data, determining the unit density distribution of the finite element basic grid by pre-judgment of hot spots, and obtaining an initial grid of the temperature field includes: A finite element basic grid is constructed according to the material characteristic data, the unit density distribution of the finite element basic grid is determined by hot spot prejudgment, and an initial temperature field grid is obtained by using a tetrahedron to hexahedron grid converter.

4. The semiconductor power module optimization method based on loss calculation according to claim 1, characterized in that: The step of constructing a three-dimensional model of a semiconductor power module using the optimized design parameters and evaluating the anti-loss performance of the three-dimensional model of the semiconductor power module includes: Constructing a three-dimensional model of a semiconductor power module according to the optimized design parameters and meshing the three-dimensional model of the semiconductor power module; Obtain temperature field distribution data and stress distribution data in each grid at each temperature; According to the temperature field distribution data and the stress distribution data, a multi-physical field coupling simulation analysis method is used to perform loss simulation analysis on the semiconductor power module three-dimensional model, and the anti-loss performance of the semiconductor power module three-dimensional model at different operating temperatures is evaluated based on the simulation results and the mapping relationship table.

5. The semiconductor power module optimization method based on loss calculation according to claim 1, characterized in that: The preset standard for the anti-loss performance of the semiconductor power module three-dimensional model includes that the hot spot temperature is higher than a preset hot spot temperature threshold, the current distribution uniformity is lower than the current distribution uniformity level, or the stress peak is higher than the material strength limit; if the anti-loss performance of the semiconductor power module three-dimensional model does not meet the preset standard, the target weight of the particle swarm optimization algorithm is adjusted to re-optimize the design parameters, and the step of returning to construct the semiconductor power module three-dimensional model using the optimized design parameters and evaluating the anti-loss performance of the semiconductor power module three-dimensional model, until the corresponding anti-loss performance of the semiconductor power module three-dimensional model reaches the preset standard, including: If the anti-loss performance of the three-dimensional model of the semiconductor power module does not meet any of the preset standards, the weight coefficient of each objective in the multi-objective function of the particle swarm optimization algorithm is adjusted; The design parameters in the three-dimensional model of the semiconductor power module whose anti-loss performance does not meet any of the preset standards are re-optimized according to the adjusted particle swarm optimization algorithm, and the step of constructing the three-dimensional model of the semiconductor power module using the optimized design parameters and evaluating the anti-loss performance of the three-dimensional model of the semiconductor power module is returned to, until the corresponding anti-loss performance of the three-dimensional model of the semiconductor power module reaches the preset standard.

6. The semiconductor power module optimization method based on loss calculation according to claim 1, characterized in that: The method further comprises: A semiconductor power module sample is designed according to each of the semiconductor power module three-dimensional models that meet preset standards, the semiconductor power module sample is put into use, and the reliability of the semiconductor power module three-dimensional model is evaluated based on the operating data of the semiconductor power module sample.

7. A semiconductor power module optimization system based on loss calculation, characterized in that: include: A data acquisition module, used to measure the intrinsic characteristic data of the target semiconductor power module, and obtain the material characteristic data of the target semiconductor power module according to the intrinsic characteristic data; The loss analysis module is used to perform loss simulation on the material characteristic data by using the finite element analysis method, and establish a mapping relationship table between the total power loss value of the semiconductor power module at each temperature and each material characteristic data according to the simulation results; specifically includes: constructing a finite element basic grid according to the material characteristic data, determining the unit density distribution of the finite element basic grid by hot spot pre-judgment, and obtaining the temperature field initial grid; importing boundary temperature and external load current data into the temperature field initial grid, using a thermal field finite element solver to calculate the temperature distribution value in the temperature field initial grid, and obtaining the temperature according to the temperature distribution value through the heat source term boundary function. The temperature gradient value of the hot spot area in the initial grid of the field; construct the temperature stress initial value field in the initial grid of the temperature field according to the temperature gradient value, and iteratively update the temperature stress initial value field using the multi-grid iteration method to obtain a stress distribution map; calculate the power loss density according to the failure area data in the stress distribution map and the current density distribution of the stress distribution map; obtain the loss curve of the semiconductor power module changing with time at different operating temperatures according to the power loss density obtained at different time sequences; establish a mapping relationship table between the total power loss value of the semiconductor power module at each temperature and each material characteristic data according to each loss curve at different temperatures; An optimization and evaluation module is used to optimize the design parameters of the semiconductor power module at each operating temperature using a particle swarm optimization algorithm, wherein the optimization process includes: constructing a multi-objective function with the goal of minimizing the temperature numerical peak, minimizing the temperature gradient, minimizing the current density peak, and minimizing the stress peak of the semiconductor power module at each operating temperature; optimizing the design parameters of the semiconductor power module using a particle optimization algorithm according to the multi-objective function; constructing a three-dimensional model of the semiconductor power module using the optimized design parameters and evaluating the anti-loss performance of the three-dimensional model of the semiconductor power module, wherein the anti-loss performance evaluation is designed to mesh the three-dimensional model of the semiconductor power module and perform multi-physical field coupling simulation analysis, and evaluating the anti-loss performance of the three-dimensional model of the semiconductor power module at different operating temperatures according to the analysis results and the mapping relationship table; A weight adjustment module, for adjusting the target weight of the particle swarm optimization algorithm to re-optimize the design parameters if the anti-loss performance of the semiconductor power module three-dimensional model does not meet the preset standard, and returning to the step of using the optimized design parameters to construct the semiconductor power module three-dimensional model and evaluating the anti-loss performance of the semiconductor power module three-dimensional model, until the corresponding anti-loss performance of the semiconductor power module three-dimensional model reaches the preset standard; The optimization and improvement module is used to optimize and improve the semiconductor power module based on the three-dimensional model of the semiconductor power module that meets the preset standard.

Images