Deterioration diagnosis method and deterioration diagnosis device

Abstract

This deterioration diagnosis method is a method for diagnosing deterioration of a power module including a semiconductor element and a copper pattern joined to the semiconductor element by a sintered material. The method includes: a first step of acquiring an X-ray CT image of the power module actually used for a predetermined period by using an X-ray CT apparatus; and a second step of diagnosing deterioration of the actually used power module on the basis of the image and an X-ray CT image obtained by photographing an undulation-shaped deformation generated in the copper pattern when a temperature change is repeatedly applied to a power module having the same specification as that of the power module.

Description

Deterioration diagnosis method and deterioration diagnosis device

[0001] This application claims priority to Japanese Patent Application No. 2023-207258, filed on December 7, 2023, the contents of which are incorporated herein by reference.

[0002] Patent Document 1 describes the following power module. That is, the power module described in Patent Document 1 includes a first solder part that joins a first semiconductor chip and a wiring part, and a second solder part that is more susceptible to cracking due to thermal history than the first solder part and joins a second semiconductor chip and a wiring part. With this power module, for example, if a crack occurs in the second solder part, it can be estimated that there is an increased possibility that a crack will occur in the first solder part, and it is said that it is possible to detect signs of a crack occurring in the first solder part.

[0003] Furthermore, Non-Patent Document 1 discloses the following matter related to the present disclosure: That is, Non-Patent Document 1 describes that, as a result of evaluating sintered joints using silver nanoparticles in power modules through a thermal shock cycle test, the copper electrodes joined to the sintered joints were deformed in a large undulating manner, which expanded microvoids in the joints, and in some areas where deformation was large, the microvoids could connect and cause cracks that penetrated between the semiconductor chip and the substrate.

[0004] Japanese Patent Application Laid-Open No. 2022-174923

[0005] Hiroaki Tatsumi, Sho Kumada, Nobuo Yokomura, Yasunari Hino, Yoshihiro Kashiba, "Reliability Evaluation of Sintered Joints Using Silver Nanoparticles in Thermal Shock Cycling Tests," Abstract of the Japan Welding Society National Convention, Japan Welding Society, 2014, pp. 104-105

[0006] In the power module described in Patent Document 1, for example, if the usage environment of the second solder part, such as the ambient temperature, is more severe than the usage environment of the first solder part, even if a crack occurs in the second solder part, it is possible that the possibility of a crack occurring in the first solder part does not necessarily increase, and there is a problem in that it may not be possible to properly detect signs of a crack occurring in the first solder part.

[0007] The present disclosure has been made to solve the above-mentioned problems, and aims to provide a degradation diagnosis method and a degradation diagnosis device that can appropriately detect degradation of a power module.

[0008] In order to solve the above-mentioned problems, the degradation diagnosis method according to the present disclosure is a method for diagnosing degradation of a power module including a semiconductor element and a copper pattern joined to the semiconductor element with a sintered material, and includes a first step of obtaining, using an X-ray CT device, an X-ray CT image of the power module that has actually been used for a predetermined period of time, and a second step of diagnosing degradation of the actually used power module based on the first step and an X-ray CT image capturing wavy deformation that occurs in the copper pattern when a power module having the same specifications as the power module is subjected to repeated temperature changes.

[0009] The degradation diagnosis device according to the present disclosure is a device for diagnosing degradation of a power module including a semiconductor element and a copper pattern joined to the semiconductor element with a sintered material, and includes an acquisition unit that acquires X-ray CT images of the power module that has actually been used for a predetermined period of time, taken using an X-ray CT device, and a diagnosis unit that diagnoses degradation of the power module that has actually been used based on the images and X-ray CT images that capture wavy deformation that occurs in the copper pattern when a power module with the same specifications as the power module is subjected to repeated temperature changes.

[0010] According to the degradation diagnosis method and degradation diagnosis device of the present disclosure, degradation of a power module can be appropriately detected.

[0011] FIG. 1 is a block diagram showing an example configuration of a degradation diagnosis device according to an embodiment of the present disclosure. FIG. 2 is a flowchart showing an example operation of a degradation diagnosis device according to an embodiment of the present disclosure. FIG. 3 is a side view schematically showing a schematic cross-sectional configuration of a power module according to an embodiment of the present disclosure. FIG. 4 is a schematic diagram for explaining a power cycle test according to an embodiment of the present disclosure. FIG. 5 is a schematic diagram for explaining a trained machine learning model according to an embodiment of the present disclosure. FIG. 6 is a perspective view showing an example configuration of a semiconductor element or the like according to an embodiment of the present disclosure. FIG. 7 is a diagram showing an example X-ray CT image according to an embodiment of the present disclosure. FIG. 8 is a diagram showing an example X-ray CT image according to an embodiment of the present disclosure. FIG. 9 is a diagram showing an example X-ray CT image according to an embodiment of the present disclosure. FIG. 10 is a diagram showing an example X-ray CT image according to an embodiment of the present disclosure. FIG. 11 is a diagram showing an example X-ray CT image according to an embodiment of the present disclosure. FIG. 12 is a diagram showing an example X-ray CT image according to an embodiment of the present disclosure. FIG. 13 is a schematic diagram for explaining another trained machine learning model according to an embodiment of the present disclosure. FIG. 14 is a schematic block diagram showing the configuration of a computer according to an embodiment of the present disclosure.

[0012] Degradation diagnosis methods and degradation diagnosis devices according to embodiments of the present disclosure will be described below with reference to FIGS. 1 to 14 . FIG. 1 is a block diagram illustrating an example configuration of a degradation diagnosis device according to an embodiment of the present disclosure. FIG. 2 is a flowchart illustrating an example operation of the degradation diagnosis device according to an embodiment of the present disclosure. FIG. 3 is a side view schematically illustrating a cross-sectional configuration of a power module according to an embodiment of the present disclosure. FIG. 4 is a schematic diagram illustrating a power cycle test according to an embodiment of the present disclosure. FIG. 5 is a schematic diagram illustrating a trained machine learning model according to an embodiment of the present disclosure. FIG. 6 is a perspective view illustrating an example configuration of a semiconductor element or the like according to an embodiment of the present disclosure. FIGS. 7 to 12 are diagrams illustrating example X-ray CT images according to an embodiment of the present disclosure. FIG. 13 is a schematic diagram illustrating another trained machine learning model according to an embodiment of the present disclosure. FIG. 14 is a schematic block diagram illustrating the configuration of a computer according to an embodiment of the present disclosure. Note that the same or corresponding components in each drawing are designated by the same reference numerals, and descriptions thereof will be omitted as appropriate.

[0013] (Configuration of Degradation Diagnosis Device) The degradation diagnosis device 1 according to an embodiment of the present disclosure shown in FIG. 1 can be configured using one or more computers, such as servers or personal computers. Furthermore, some or all of the one or more computers constituting the degradation diagnosis device 1 may be virtually configured on a network. The degradation diagnosis device 1 includes an acquisition unit 11, a diagnosis unit 12, a machine learning unit 13, and a storage unit 14 as a functional configuration configured by a combination of hardware such as a computer constituting the degradation diagnosis device 1, peripheral devices of the computer, and software such as a program executed by the computer. The storage unit 14 also stores a trained machine learning model 15. The machine learning unit 13 and the trained machine learning model 15 may be configured as an integrated unit, for example.

[0014] The degradation diagnosis device 1 shown in Fig. 1 is a device for diagnosing degradation of a power module (also referred to as a power semiconductor module, etc.) 3 actually used in the field. The degradation diagnosis device 1 diagnoses degradation of the power module 3 based on an X-ray CT (Computed Tomography) image 4 of the power module 3 actually used in the field, which is non-destructively captured by an X-ray CT device 2. When capturing the X-ray CT image 4, the power module 3 is temporarily removed from the facility, etc., and then attached to the X-ray CT device 2, for example, during periodic inspection of a mobile object such as a vehicle or equipment, etc., on which the power module 3 is mounted. Furthermore, if no problems are found with regard to degradation, etc., the power module 3 for which the X-ray CT image 4 has been captured is returned to the original facility, etc.

[0015] FIG. 3 schematically illustrates an exemplary configuration of a power module 3 targeted for degradation diagnosis in this embodiment. The power module 3 illustrated in FIG. 3 includes one or more semiconductor elements 31 and one or more copper patterns 33 bonded to each semiconductor element 31 with a sintered material 32. The semiconductor elements 31 are, for example, wide bandgap (WBG) power semiconductor elements made of silicon carbide (SiC) or gallium nitride (GaN). Each semiconductor element 31 constitutes, for example, a transistor such as a metal-oxide semiconductor field-effect transistor (MOSFET), an insulated gate bipolar transistor (IGBT), or a bipolar transistor, or a diode. The sintered material 32 is a bonding material and is, for example, a nanoparticle sintered material made of silver (Ag) or copper (Cu). The copper pattern 33 is, for example, a copper electrode formed on the surface of an insulating substrate 35. The copper patterns 34 and 36 are also, for example, copper electrodes formed on the surface of the insulating substrate 35. The heat dissipation member 37 is inserted between the copper pattern 36 and the base plate 38 and is a member that ensures heat dissipation from the copper pattern 36 to the base plate 38. Terminals 41 are bonded to the copper pattern 33. The copper pattern 34 is connected to the upper surface electrodes of the semiconductor element 31 by wires 40 and is also bonded to terminals 42. A case 39 is fitted onto the base plate 38. The case 39 is filled with a sealing material 43. The power module 3 may also include components not shown, such as a control circuit, a protection circuit, and a temperature sensor.

[0016] 1 acquires X-ray CT images 4 of the power module 3 that has actually been used in the field for a predetermined period of time, which are taken using the X-ray CT device 2, using, for example, a communication line, a removable recording medium, etc. The predetermined period of time is not limited and can be set appropriately depending on, for example, the usage environment of the power module 3.

[0017] The diagnostic unit 12 diagnoses the deterioration of a power module 3 to be diagnosed (hereinafter also referred to as the power module 3) that has actually been used in the field for a predetermined period of time, based on an X-ray CT image 4 of the power module 3 to be diagnosed, and an X-ray CT image capturing undulating deformations in the copper pattern 33 of a power module 3 having the same specifications as the power module 3 when repeated temperature changes are applied. The undulating deformations in the copper pattern 33 are deformations that occur in the copper pattern 33 due to thermal stress caused by repeated temperature changes, and include, for example, wavy deformations, uneven deformations, void (vacancy) expansion, crack propagation, etc. Furthermore, the repeated temperature changes applied to the power module 3 can be applied, for example, by a power cycle test or a temperature cycle test. In this embodiment, as an example, the power cycle test applies repeated temperature changes to the power module 3. The power cycle test is a test in which a constant current is repeatedly applied and interrupted to the power module 3.

[0018] 4 shows an example of the change over time in the junction temperature of the semiconductor element 31 during a power cycle test, and the change over time in the current flowing through the semiconductor element 31. Time Tcycle is the time for one cycle, and during time Tcycle, a constant (almost constant) current Ion is flowed for time Ton, and the flowing current is interrupted for time Toff.

[0019] In this embodiment, the diagnostic unit 12 uses the trained machine learning model 15 to compare the X-ray CT image 4 of the power module 3 with an X-ray CT image (assumed to be an X-ray CT image with waviness) taken of the power module 3 in a state in which degradation has occurred due to a power cycle test, and calculates the degree of degradation based on the degree of agreement between the X-ray CT image 4 of the power module 3 and the X-ray CT image with waviness. In this embodiment, the degree of degradation means the presence or absence of degradation or the degree of degradation.

[0020] As shown in FIG. 5 , the trained machine learning model 15 is a model trained by the machine learning unit 13. When an X-ray CT image 4 of the power module 3 is used as a first image and an X-ray CT image of a power module 3 with the same specifications that has been subjected to repeated temperature changes is used as a second image, the trained machine learning model 15 trains a plurality of second images, each labeled with information indicating the presence or absence of undulating deformation in the copper pattern 33 or the degree of deformation, as training data 5. The trained model inputs the first image and outputs a value indicating the presence or absence or degree of degradation of the power module 3 (a value representing the degree of degradation). If the second image is labeled with either the presence or absence of degradation, the value representing the degree of degradation can be, for example, a value representing the probability (degree of coincidence) that the image is determined to be degraded (matches the image with degradation). Alternatively, if the second image is labeled with a value (e.g., an index) representing the degree of undulating deformation, the value representing the degree of degradation can be, for example, a value representing the degree of deformation.

[0021] The trained machine learning model 15 is a model that uses, for example, a neural network as an element, and the weighting coefficients between neurons in each layer of the neural network are optimized by machine learning so that a desired solution (degree of degradation) is output for a large amount of input data (for example, data representing each pixel value of an image). The trained machine learning model 15 is configured, for example, by a combination of a program that performs calculations from input to output and weighting coefficients (parameters) used for the calculations.

[0022] In the example shown in Figure 5, the training data 5 includes X-ray CT images in which undulation-like deformation is confirmed after a power cycle test of a predetermined number of cycles is performed, including X-ray CT images with waviness 51 labeled with a value indicating the presence or degree of waviness, and X-ray CT images without waviness 52 in which no undulation-like deformation is confirmed and labeled with no waviness (the degree of waviness is zero). Note that the X-ray CT images with waviness 51 and the X-ray CT images without waviness 52 include images with different numbers of cycles in the power cycle test. Furthermore, the labeling of each image can be performed, for example, according to the judgment and instructions of an operator, or automatically or semi-automatically using existing image processing technology.

[0023] FIGS. 7 to 12 show examples of X-ray CT images taken after a power cycle test of a predetermined number of cycles was conducted. The correspondence between the directions in FIGS. 6 and 7 to 12 is indicated by arrows X, Y, and Z. FIG. 7 shows an X-ray CT image 4a taken of region A1 (surface region of copper pattern 33) shown in FIG. 6. No undulating deformation is observed in copper pattern 33 in X-ray CT image 4a. FIG. 8 shows an X-ray CT image 4b ​​taken of region A1 shown in FIG. 6. In X-ray CT image 4b, undulating deformation RA11 is observed in copper pattern 33. FIG. 9 shows an X-ray CT image 4c taken of region A1 shown in FIG. 6. In X-ray CT image 4c, undulating deformation RA12 is observed in copper pattern 33.

[0024] Fig. 10 shows an X-ray CT image 4d of the area A2 shown in Fig. 6. In the X-ray CT image 4d, no undulating deformation is observed in the copper pattern 33. Fig. 11 shows an X-ray CT image 4e of the area A2 shown in Fig. 6. In the X-ray CT image 4e, a undulating deformation RA21 is observed in the copper pattern 33. Fig. 12 shows an X-ray CT image 4f of the area A3 shown in Fig. 6. In the X-ray CT image 4f, a undulating deformation RA31 is observed in the copper pattern 33.

[0025] The diagnostic unit 12 inputs the X-ray CT image 4 of the power module 3 into the trained machine learning model 15, compares the value representing the degree of deterioration output by the trained machine learning model 15 with a predetermined threshold value, and determines that the power module 3 is deteriorated if the value representing the degree of deterioration exceeds the threshold value.

[0026] (Example of Operation of Degradation Diagnosis Device) Fig. 2 shows an example of operation of the degradation diagnosis device 1 shown in Fig. 1. It is assumed that the degradation diagnosis device 1 already stores a trained machine learning model 15 that has been machine-learned by the machine learning unit 13 as described with reference to Fig. 5.

[0027] In the operation example shown in FIG. 2, first, an X-ray CT image 4 of the power module 3 that has actually been used for a predetermined period of time is taken using the X-ray CT device 2, for example, in accordance with the operation of an operator, and the taken X-ray CT image 4 is acquired by the acquisition unit 11 (step S1).

[0028] Next, the diagnostic unit 12 diagnoses the deterioration of the power module 3 that has actually been used based on the acquired X-ray CT image 4 and an X-ray CT image (X-ray CT image with waviness 51) that captures wavy deformation that occurs in the copper pattern 33 when repeated temperature changes are applied by performing a power cycle test on a power module 3 of the same specifications (step S2).

[0029] According to the above operation, the degradation diagnosis device 1 of this embodiment can detect degradation based on wavy deformation that occurs in the copper pattern 33 in the early stage of degradation of the sintered material 32 or as a sign of degradation of the sintered material 32. Therefore, degradation of the power module 3 can be appropriately detected.

[0030] (Another Configuration Example of Trained Machine Learning Model) When a power cycle test was conducted in accordance with the present disclosure, a phenomenon was observed in which the temperature of the semiconductor element 31 increased depending on the number of cycles when the current flowing through the semiconductor element 31 was kept constant. Examples of the temperature increase include a maximum temperature of 225°C immediately after the start of the test, 250°C at 1000 cycles, and 270°C at 2000 cycles when the same current was applied. From this phenomenon, it is inferred that the progression of deterioration of the power module 3 with an increase in the number of cycles affects the temperature increase of the semiconductor element 31. Therefore, with regard to a machine learning model that calculates the degree of deterioration of the power module 3, it is expected that the degree of deterioration, which is the objective variable, can be more appropriately calculated by using the temperature of the semiconductor element 31 in addition to the X-ray CT image as an explanatory variable of the machine learning model.

[0031] 13 shows a model (as trained machine learning model 15a) in which the temperature of the semiconductor element is added to the input of the trained machine learning model 15 described above. The trained machine learning model 15a is a model trained by machine learning by the machine learning unit 13, and when an X-ray CT image 4 of the power module 3 is used as a first image and an X-ray CT image of a power module 3 with the same specifications that has been subjected to repeated temperature changes is used as a second image, the trained machine learning model 15a is trained by machine learning using a plurality of second images (images 51a or 52a) in which information indicating the presence or absence of wavy deformation occurring in the copper pattern 33 or the degree of deformation and information related to the temperature of the semiconductor element 31 obtained during the temperature change are respectively labeled, and the trained model receives as input the first image and information related to the temperature of the semiconductor element obtained during actual use, and outputs a value indicating the presence or absence of degradation or the degree of degradation of the power module that has actually been used.

[0032] The information regarding the temperature of the semiconductor element obtained during actual use may be, for example, a value representing the maximum temperature (or average temperature) of the junction temperature of the semiconductor element 31, calculated from temperature values ​​obtained during a predetermined period immediately prior to diagnosis using a temperature sensor in the power module 3 or a temperature sensor in a unit such as an inverter that includes the power module 3.

[0033] According to this trained machine learning model 15a, deterioration of the power module 3 can be appropriately detected by taking into account information related to the temperature of the semiconductor element 31 in addition to the X-ray CT image 4.

[0034] (Effects) According to the degradation diagnosis method and device of this embodiment, an X-ray CT image 4 of a power module 3 that has actually been used for a predetermined period is obtained using an X-ray CT device 2, and degradation of the power module 3 that has actually been used is diagnosed based on that image and an X-ray CT image 4 that captures wavy deformation that occurs in the copper pattern 33 when a power module 3 with the same specifications as the power module 3 is repeatedly subjected to temperature changes, so that degradation of the power module 3 can be appropriately detected.

[0035] (Other Embodiments) While the embodiments of the present disclosure have been described above in detail with reference to the drawings, the specific configurations are not limited to these embodiments, and design modifications and the like that do not depart from the gist of the present disclosure are also included. For example, the calculation of the degree of degradation (calculation of the degree of match with an image including degradation) may be performed using an image processing technique such as pattern matching to calculate the degree of match with an X-ray CT image with waviness, instead of or in combination with a configuration using a trained machine learning model.

[0036] <Computer Configuration> Fig. 14 is a schematic block diagram showing the configuration of a computer according to at least one embodiment. The computer 90 includes a processor 91, a main memory 92, a storage 93, and an interface 94. The degradation diagnosis device 1 described above is implemented in the computer 90. The operations of the above-described processing units are stored in the storage 93 in the form of a program. The processor 91 reads the program from the storage 93, loads it into the main memory 92, and executes the above-described processing in accordance with the program. The processor 91 also allocates storage areas in the main memory 92 corresponding to the above-described storage units in accordance with the program.

[0037] The program may be for realizing some of the functions to be performed by the computer 90. For example, the program may be combined with other programs already stored in storage or implemented in other devices to perform the functions. In other embodiments, the computer may include a custom LSI (Large Scale Integrated Circuit) such as a PLD (Programmable Logic Device) in addition to or instead of the above configuration. Examples of PLDs include PAL (Programmable Array Logic), GAL (Generic Array Logic), CPLD (Complex Programmable Logic Device), and FPGA (Field Programmable Gate Array). In this case, some or all of the functions realized by the processor may be realized by the integrated circuit.

[0038] Examples of storage 93 include a hard disk drive (HDD), a solid state drive (SSD), a magnetic disk, a magneto-optical disk, a compact disc read-only memory (CD-ROM), a digital versatile disc read-only memory (DVD-ROM), and a semiconductor memory. Storage 93 may be an internal medium directly connected to the bus of computer 90, or an external medium connected to computer 90 via interface 94 or a communication line. Furthermore, if this program is distributed to computer 90 via a communication line, computer 90 that receives the program may load the program into main memory 92 and execute the above-described processing. In at least one embodiment, storage 93 is a non-transitory tangible storage medium.

[0039] <Additional Notes> The degradation diagnosis method and degradation diagnosis device 1 according to the embodiment of the present disclosure can be understood, for example, as follows.

[0040] (1) A degradation diagnosis method according to a first aspect is a method for diagnosing degradation of a power module 3 including a semiconductor element 31 and a copper pattern 33 joined to the semiconductor element 31 with a sintered material 32, and includes a first step (S1) of acquiring an X-ray CT image 4 of the power module 3 that has actually been used for a predetermined period of time using an X-ray CT device 2, and a second step (S2) of diagnosing degradation of the actually used power module 3 based on the acquired image and an X-ray CT image 4 capturing undulating deformation that has occurred in the copper pattern 33 when a power module 3 having the same specifications as the power module 3 has been subjected to repeated temperature changes. According to this aspect and each of the following aspects, degradation can be detected based on undulating deformation that occurs in the copper pattern 33 in the early stages of degradation of the sintered material 32 or as a sign of degradation of the sintered material 32, and therefore degradation of the power module 3 can be appropriately detected.

[0041] (2) A degradation diagnosis method according to a second aspect is the degradation diagnosis method of (1), wherein, when an X-ray CT image of the power module that has actually been used for the predetermined period is used as the first image and an X-ray CT image of the power module with the same specifications that has been repeatedly subjected to the temperature change is used as the second image, in the second step (S2), a plurality of the second images, each labeled with information indicating the presence or absence of wavy deformation occurring in the copper pattern 33 or the degree of the deformation, are machine-learned as training data 5, and degradation of the actually used power module is diagnosed using a trained machine learning model 15 that inputs the first image and outputs a value indicating the presence or absence of degradation of the power module or the degree of degradation. According to this aspect, degradation of the power module 3 can be appropriately detected by appropriately generating the trained machine learning model 15.

[0042] (3) A degradation diagnosis method according to a third aspect is the degradation diagnosis method according to (2), wherein the temperature change is imparted by repeatedly passing and interrupting a constant current through the power module 3 of the same specifications, and the machine learning model (trained machine learning model 15 a) is trained by machine learning using a plurality of second images labeled with information indicating the presence or absence of undulating deformation in the copper pattern 33 or the degree of the deformation and information related to the temperature of the semiconductor element 31 obtained during the temperature change, and the machine learning model receives as input the first image and information related to the temperature of the semiconductor element 31 obtained during actual use, and outputs a value indicating the presence or absence or degree of degradation of the actually used power module 3. According to this aspect, degradation of the power module 3 can be appropriately detected by taking into account the information related to the temperature of the semiconductor element 31 in addition to the X-ray CT image 4.

[0043] According to the degradation diagnosis method and degradation diagnosis device of the present disclosure, degradation of a power module can be appropriately detected.

[0044] 1... Deterioration diagnosis device 11... Acquisition unit 12... Diagnosis unit 13... Machine learning unit 14... Memory unit 15, 15a... Trained machine learning model 3... Power module 31... Semiconductor element 32... Sintered material 33, 34, 36... Copper pattern 35... Insulating substrate 37... Heat dissipation member 38... Base plate 4, 4a to 4f... X-ray CT image 5, 5a... Training data

Claims

1. A method for diagnosing deterioration of a power module comprising a semiconductor element and a copper pattern bonded to the semiconductor element with a sintered material, comprising: a first step of obtaining, by an X-ray CT device, an X-ray CT image of the power module that has actually been used for a predetermined period of time; and a second step of diagnosing deterioration of the power module that has actually been used based on the obtained image and an X-ray CT image of wavy deformation that has occurred in the copper pattern when a power module having the same specifications as the power module is subjected to repeated temperature changes.

2. The degradation diagnosis method according to claim 1, wherein, when an X-ray CT image of a power module that has actually been used for the specified period is used as a first image, and an X-ray CT image of a power module of the same specifications to which the temperature change has been repeatedly applied is used as a second image, in the second step, a plurality of the second images, each labeled with information indicating the presence or absence of wavy deformation in the copper pattern or the degree of the deformation, are machine-learned as training data, and degradation of the actually used power module is diagnosed using a trained machine learning model that inputs the first image and outputs a value indicating the presence or absence of degradation of the power module or the degree of degradation.

3. The degradation diagnosis method according to claim 2, wherein the temperature change is imparted by repeatedly passing and interrupting a constant current through the power module of the same specifications, and the machine learning model is machine-trained using as training data a plurality of the second images in which information indicating the presence or absence of wavy deformation in the copper pattern or the degree of the deformation and information relating to the temperature of the semiconductor element obtained during the temperature change are respectively labeled, the first images and information relating to the temperature of the semiconductor element obtained during actual use are input, and a value indicating the presence or absence of deterioration or the degree of deterioration of the power module actually used is output.

4. A device for diagnosing the deterioration of a power module comprising a semiconductor element and a copper pattern bonded to the semiconductor element with a sintered material, comprising: an acquisition unit for acquiring X-ray CT images of the power module that has been actually used for a predetermined period of time, taken with an X-ray CT device; and a diagnosis unit for diagnosing the deterioration of the power module that has actually been used, based on the images and X-ray CT images of wavy deformation that has occurred in the copper pattern when a power module having the same specifications as the power module is subjected to repeated temperature changes.

Images