A method for predicting the lifetime of metal-oxide-semiconductor field-effect transistors based on multi-scale features

Abstract

The present invention relates to the field of semiconductor technology, and discloses a method for predicting the life of a metal-oxide-semiconductor field-effect transistor based on multi-scale features, comprising: confirming a prediction environment, wherein the prediction environment includes a target field-effect transistor and a life prediction system, wherein the life prediction system includes a data acquisition unit, a data conversion unit, a data analysis unit, and a result feedback unit, using the life prediction system to confirm a life prediction database, parsing prediction parameters based on the target field-effect transistor, wherein the prediction parameters include: a target operating voltage group, a target operating temperature, a target packaging standard, a search temperature mean, and a target junction temperature, using the prediction parameters, searching in the life prediction database to obtain a target predicted life. The main purpose of the present invention is to improve the accuracy of life prediction for metal-oxide-semiconductor field-effect transistors.

Description

Technical Field

[0001] The invention relates to a metal-oxide-semiconductor field-effect transistor life prediction method based on multi-scale features, belonging to the technical field of semiconductors. Background Art

[0002] With the rapid development of electronic devices, the reliability requirements for semiconductor devices are getting higher and higher. As the core component of semiconductor devices, the life prediction of MOSFET is crucial to ensure the stability and reliability of electronic devices. Correspondingly, how to accurately predict the life of MOSFET has become an urgent problem to be solved.

[0003] Currently, a single scale is usually used to predict the lifetime of metal-oxide-semiconductor field-effect transistors.

[0004] Although the above method can realize the life prediction of MOSFET, the multi-scale characteristics of MOSFET are not taken into consideration when making the life prediction, resulting in poor accuracy in the life prediction of MOSFET. Summary of the invention

[0005] The present invention provides a method, device and computer-readable storage medium for predicting the life of a metal oxide semiconductor field effect transistor based on multi-scale features, the main purpose of which is to improve the accuracy of predicting the life of a metal oxide semiconductor field effect transistor.

[0006] To achieve the above object, the present invention provides a method for predicting the lifetime of a metal oxide semiconductor field effect transistor based on multi-scale features, comprising:

[0007] receiving a prediction instruction, and confirming a prediction environment based on the prediction instruction, wherein the prediction environment includes a target field effect transistor and a life prediction system, wherein the life prediction system includes a data acquisition unit, a data conversion unit, a data analysis unit, and a result feedback unit;

[0008] Using the life prediction system to identify a life prediction database;

[0009] Analyzing prediction parameters based on the target field effect transistor, wherein the prediction parameters include: a target operating voltage group, a target operating temperature, a target packaging standard, a search temperature mean, and a target junction temperature, and using the prediction parameters to search in a life prediction database to obtain a target predicted life;

[0010] The target predicted lifetime is sent to the initiator of the prediction instruction by using the result feedback unit, so as to realize the lifetime prediction of the target field effect transistor.

[0011] Optionally, the confirming a life prediction database using the life prediction system includes:

[0012] Confirming receipt of a data acquisition instruction from a data acquisition unit, acquiring a test sample set based on the data acquisition instruction, wherein the test sample set includes a plurality of test samples, and acquiring test sample parameters based on the test sample set, wherein the test sample parameters include a reference packaging standard and a reference size structure;

[0013] According to the reference packaging standard and the reference size structure, the test samples are respectively aggregated to obtain one or more initial test sample sets;

[0014] The following operations are performed on each of the one or more initial test sample sets:

[0015] Extracting multiple grouped test sample sets from the initial test sample set using a pre-constructed grouping threshold, wherein the number of test samples in the grouped test sample sets is the same as the grouping threshold;

[0016] The life prediction database is acquired based on the multiple grouped test sample sets.

[0017] Optionally, acquiring the life prediction database based on the multiple grouped test sample sets includes:

[0018] Obtain one or more operating voltage ranges and operating temperature ranges of the metal oxide semiconductor field effect transistor, extract a plurality of first operating temperatures from the operating temperature range using a preset first temperature division value, and respectively extract one or more first operating voltage sets from the one or more operating voltage ranges, wherein the first operating voltage set corresponds to the operating voltage range one by one, and the first operating voltage set includes a plurality of first operating voltages;

[0019] In a combined form, one or more first operating voltage sets are used to obtain multiple operating voltage groups, and the operating voltage groups include one or more first operating voltages, and the number corresponding to the one or more first operating voltages is the same as the number corresponding to the one or more first operating voltage sets;

[0020] Associating, in a combined form, an operating voltage group in a plurality of operating voltage groups, a first operating temperature in a plurality of first operating temperatures, and a grouped test sample set in a plurality of grouped test sample sets to obtain a plurality of initial test groups, wherein the initial test group includes a working voltage group, a first operating temperature, and a test sample set;

[0021] For each of the multiple initial test groups, perform the following operations:

[0022] Identify the working mode of the initial test group to obtain an identified working mode, wherein the identified working mode is a forward working mode, a reverse working mode or a body diode mode;

[0023] Using the identification working mode, performing a marking operation on the initial test group to obtain a matching test group, and summarizing the matching test groups to obtain multiple matching test groups;

[0024] A life prediction database is acquired based on the plurality of matching test groups.

[0025] Optionally, acquiring a life prediction database based on the multiple matching test groups includes:

[0026] Sequentially extract matching test groups from the multiple matching test groups, and perform the following operations on the extracted matching test groups:

[0027] The following operations are performed on each test sample in the test sample set corresponding to the matching test group:

[0028] Setting the temperature of the environment in which the test sample is located to match the first operating temperature corresponding to the test group to match the operating voltage group corresponding to the test group, starting the test sample, and obtaining an on-resistance timing sequence based on the started test sample, wherein the on-resistance timing sequence includes multiple on-resistances;

[0029] Summarizing the on-resistance timing to obtain an on-resistance timing set, and using the on-resistance timing set to obtain an on-resistance group timing, wherein the on-resistance group timing includes multiple on-resistance timing groups, and the on-resistance timing group includes multiple on-resistances, and the number corresponding to the multiple on-resistances in the on-resistance timing group is the grouping threshold;

[0030] Extracting a first on-resistance timing group from the on-resistance group timing to obtain an analysis resistance timing group, and obtaining an on-resistance variance using a plurality of on-resistances corresponding to the analysis resistance timing group, wherein the on-resistance variance is a variance of the plurality of on-resistances;

[0031] comparing the on-resistance variance with a preset on-resistance variance threshold;

[0032] If the on-resistance variance is greater than or equal to the on-resistance variance threshold, obtaining an updated test group using the operating voltage group and the first operating temperature corresponding to the matching test group, and using the updated test group as the matching test group, returning to the step of performing the following operation on each test sample in the test sample set corresponding to the matching test group;

[0033] If the on-resistance variance is less than the on-resistance variance threshold, the on-resistance mean time sequence is obtained by using the on-resistance group time sequence, wherein the on-resistance mean time sequence includes multiple on-resistance mean values, the on-resistance mean values ​​correspond to the on-resistance time sequence groups one by one, and the on-resistance mean value is the mean value of the multiple on-resistances corresponding to the on-resistance time sequence groups;

[0034] A life prediction database is obtained based on the on-resistance mean time series.

[0035] Optionally, acquiring a life prediction database based on the on-resistance mean time series includes:

[0036] The first on-resistance average values ​​are sequentially extracted in the on-resistance average value time sequence, and the following operations are performed on the extracted first on-resistance average values:

[0037] Based on the extracted first on-resistance average value, a second on-resistance average value is confirmed in the on-resistance average value time sequence, wherein the second on-resistance average value is adjacent to the first on-resistance average value, and the second on-resistance average value lags behind the first on-resistance average value in the on-resistance average value time sequence;

[0038] The on-resistance change value is obtained by using the first on-resistance average value and the second on-resistance average value, and the sorting time is confirmed by using the on-resistance change value, wherein the calculation formula of the sorting time is as follows:

[0039] ;

[0040] in, Indicates the sorting time, Indicates taking the minimum value, represents the time corresponding to the first on-resistance average value, represents the time corresponding to the second on-resistance average value;

[0041] Using the sorting time to identify the on-resistance change value to obtain a sorting resistance change value, summarizing the sorting resistance change values ​​to obtain a sorting resistance change value set, and sorting the sorting resistance change values ​​in the sorting resistance change value set in the order of sorting time from earliest to later to obtain a sorting resistance change sequence;

[0042] Obtaining a first predicted lifespan using the ranked resistance change sequence, a pre-constructed variational mode decomposition algorithm, and a pre-constructed neural network model;

[0043] A second test life is obtained according to the test sample, and the second test life is summarized to obtain a second test life set. A second test life mean is obtained based on the second test life set, wherein the second test life mean is the mean of multiple second test lives in the second test life set. The life deviation ratio is calculated using the first predicted life and the second test life mean, and the calculation formula is as follows:

[0044] ;

[0045] in, represents the life deviation ratio, represents the first predicted life span, represents the second test life mean;

[0046] comparing the life deviation ratio with a preset life deviation threshold;

[0047] If the life deviation ratio is greater than the life deviation threshold, returning to the step of obtaining an updated test group by using the working voltage group and the first working temperature corresponding to the matching test group;

[0048] If the life deviation ratio is less than or equal to the life deviation threshold, the reference life is calculated based on the first predicted life and the average of the second tested life, and the calculation formula is as follows:

[0049] ;

[0050] in, represents the reference life, All are preset coefficients;

[0051] A life prediction database is acquired based on the reference life.

[0052] Optionally, acquiring a life prediction database based on the reference life includes:

[0053] The monitoring temperature time sequence is obtained by using the test sample after startup and the pre-built temperature sensor, and the time when the test sample is started is used as the starting point to record the time in real time to obtain the real-time time, wherein the monitoring temperature time sequence includes multiple monitoring temperature groups, wherein the monitoring temperature group is composed of monitoring time and monitoring temperature, and based on the pre-built sliding window, the analysis temperature time sequence is sequentially extracted from the monitoring temperature time sequence, wherein the step size of the sliding window is preset, the analysis temperature time sequence includes multiple monitoring temperature groups, and the extracted analysis temperature time sequence is subjected to the following operations:

[0054] Acquire an analysis temperature variance based on the analysis temperature time series, wherein the analysis temperature variance is a variance of a plurality of monitored temperatures in the analysis temperature time series, and compare the analysis temperature variance with a preset temperature variance threshold;

[0055] If the analysis temperature variance is greater than or equal to the temperature variance threshold, then the analysis temperature mean is obtained based on the analysis temperature time series, wherein the analysis temperature mean is the mean of multiple monitoring temperatures in the analysis temperature time series, and the following operations are performed for each monitoring temperature in the analysis temperature time series:

[0056] Obtaining an analysis temperature difference value by using the monitoring temperature and the analysis temperature mean, wherein the analysis temperature difference value is the absolute difference between the monitoring temperature and the analysis temperature mean, summarizing the analysis temperature difference values ​​to obtain an analysis temperature difference value set, and sorting the analysis temperature difference values ​​in the analysis temperature difference value set in descending order to obtain an analysis temperature difference value sequence;

[0057] Extract the first analysis temperature difference value from the analysis temperature difference sequence to obtain a rejection temperature difference value, obtain the position sequence of the rejection temperature difference value in the analysis temperature sequence, obtain a rejection position sequence, and if the rejection position sequence is 1, return to the step of sequentially extracting the analysis temperature sequence in the monitoring temperature sequence;

[0058] If the elimination bit sequence is not 1, the monitoring temperature corresponding to the elimination temperature difference is eliminated in the analysis temperature sequence to obtain an updated temperature sequence, and the updated temperature sequence is used as the analysis temperature sequence. Return to the step of obtaining the analysis temperature variance based on the analysis temperature sequence, and record the number of monitoring temperatures corresponding to the elimination temperature difference eliminated in the analysis temperature sequence to obtain the elimination number.

[0059] When the number of rejections is greater than or equal to the preset rejection threshold or the analysis temperature variance is greater than or equal to the temperature variance threshold, returning to the step of sequentially extracting the analysis temperature time series in the monitoring temperature time series until the analysis temperature variance is less than the temperature variance threshold, and obtaining the target temperature time series;

[0060] A life prediction database is obtained based on the target temperature time series, the reference life and the real time.

[0061] Optionally, acquiring a life prediction database based on the target temperature time series, reference life and real time includes:

[0062] Confirming a start time in the monitoring temperature sequence, wherein the start time is the monitoring time corresponding to the first monitoring temperature group in the monitoring temperature sequence, and confirming an end time using the target temperature sequence, wherein the end time is the monitoring time corresponding to the last monitoring temperature group in the target temperature sequence;

[0063] Obtaining a target temperature average value by using the target temperature time series, wherein the target temperature average value is an average value of multiple monitoring temperatures corresponding to the target temperature time series;

[0064] Determine a detection time based on the start time and the end time, wherein the detection time is an absolute difference between the start time and the end time, and when the real time reaches the detection time, obtain a gate threshold voltage using a test sample after startup;

[0065] Acquiring an operating junction temperature based on the gate threshold voltage;

[0066] A life prediction database is obtained using the operating junction temperature, target temperature average and reference life.

[0067] Optionally, the acquiring a life prediction database by using the operating junction temperature, the target temperature mean and the reference life includes:

[0068] Correlating the operating junction temperature, the reference packaging standard, the target temperature average, the first operating temperature, the operating voltage group, the identified operating mode and the reference life to obtain associated data;

[0069] Confirm receipt of the data conversion instruction from the data conversion unit, obtain a standard data template based on the data conversion instruction, import the associated data into the standard data template, and obtain detection data, wherein the detection data is as follows:

[0070] ;

[0071] in, represents the detection data, Indicates the reference packaging standard, represents the identification working mode, and The value of is 1, 2 or 3. When it is 1, it indicates the forward working mode. When it is 2, it indicates the reverse working mode. When it is 3, it indicates body diode mode. represents the first operating temperature, Represents the operating junction temperature, represents the target temperature mean, represents the reference life, represents the operating voltage group, and ,in, Respectively represent the first working voltage, the second working voltage and the A first operating voltage, Indicates that there are a total of a first operating voltage;

[0072] The detection data are respectively aggregated using the working modes corresponding to the detection data to obtain three detection data sets;

[0073] The three detection data sets are aggregated to obtain a life prediction database.

[0074] Optionally, using the prediction parameters to search in a life prediction database to obtain a target predicted lifespan includes:

[0075] Identify the working mode corresponding to the target working voltage group in the prediction parameters to obtain the target working mode;

[0076] Using the target packaging standard and the target working mode in the prediction parameters, a search life data set is identified in the life prediction database, wherein the working mode corresponding to the search life data set is the target working mode, and the reference packaging standard corresponding to the search life data set is the target packaging standard, and the search life data set includes a plurality of search life data;

[0077] The following operations are performed for each of the multiple retrieved life span data:

[0078] Using the retrieved life data, reference retrieval data is identified in the detection data corresponding to the retrieved life data, wherein the reference retrieval data is as follows:

[0079] ;

[0080] in, Indicates reference retrieval data;

[0081] Acquire target retrieval data using prediction parameters, aggregate the reference retrieval data and the target retrieval data, and obtain a life prediction data set;

[0082] confirming receipt of a data analysis instruction from a data analysis unit, acquiring a k-means clustering algorithm based on the data analysis instruction, clustering the life prediction data set into a plurality of life retrieval data sets using the k-means clustering algorithm, and confirming a target life retrieval data set from the plurality of life retrieval data sets, wherein the target life retrieval data set is a life retrieval data set including the target retrieval data;

[0083] A target predicted lifespan is calculated based on the target lifespan retrieval data set.

[0084] Optionally, the calculating the target predicted lifespan based on the target lifespan retrieval data set comprises:

[0085] The reference weight value is calculated using the target retrieval data in the target life retrieval data set and each of the multiple reference retrieval data corresponding to the target life retrieval data set. The calculation formula is as follows:

[0086] ;

[0087] in, represents the reference weight value, are preset coefficients, and , They respectively represent the target operating temperature, target junction temperature and average retrieval temperature corresponding to the target retrieval data. They respectively represent the first operating temperature, operating junction temperature and target temperature average corresponding to the reference retrieval data, Respectively represent the target operating voltage group The target operating voltage and the a first operating voltage;

[0088] The reference weight values ​​are summarized to obtain a reference weight value set, and the target predicted life is calculated based on the reference weight value set. The calculation formula is as follows:

[0089] ;

[0090] in, represents the target predicted lifespan, Indicates that the reference weight value set has A reference weight value, Respectively represent the reference weight value set reference weight values ​​and the reference weight value set A reference weight value, Indicates the reference weight value set The reference life span corresponding to a reference weight value.

[0091] In order to solve the above problem, the present invention further provides an electronic device, the electronic device comprising:

[0092] at least one processor; and,

[0093] a memory communicatively connected to the at least one processor; wherein,

[0094] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to implement the above-mentioned metal oxide semi-conductor field effect transistor life prediction method based on multi-scale features.

[0095] In order to solve the above problems, the present invention also provides a computer-readable storage medium, in which at least one instruction is stored. The at least one instruction is executed by a processor in an electronic device to implement the above-mentioned metal-oxide-semiconductor field-effect transistor life prediction method based on multi-scale features.

[0096] Compared with the problems described in the background technology, the present invention first uses the life prediction system to confirm the life prediction database, and considers the reference packaging standard, first operating temperature, target temperature average, operating voltage group and operating junction temperature of the metal oxide semiconductor field effect transistor when confirming the life prediction database. It can be seen that the embodiment of the present invention takes into account multiple characteristics of the metal oxide semiconductor field effect transistor when constructing the life prediction database, thereby laying the foundation for accurately obtaining the target predicted life, and when obtaining the relevant parameters of the test sample, wherein the relevant parameters include: target temperature average, operating junction temperature and reference life, different screening requirements are set respectively to improve the accuracy of obtaining the relevant parameters, thereby improving the accuracy of obtaining the target predicted life. The present invention parses prediction parameters based on the target field effect transistor, wherein the prediction parameters include: target operating voltage group, target operating temperature, target packaging standard, retrieval temperature mean and target junction temperature, and uses the prediction parameters to search in the life prediction database to obtain the target predicted life. It can be seen that the present invention first uses the target operating voltage group to identify the target operating mode, and then combines the target packaging standard and the target operating mode to realize the retrieval of the life prediction database to obtain the detection life database set, and then combines the target operating voltage group, target operating temperature, retrieval temperature mean and target junction temperature of the target field effect transistor according to the k-means clustering algorithm to retrieve the target life retrieval data set in the detection life database set, and calculates different reference weight values ​​according to different reference detection data in the target life retrieval data set, and then, sets the proportion of the reference life corresponding to the reference weight value through different reference weight values, thereby improving the accuracy of the obtained target predicted life. Therefore, the metal oxide semi-conductor field effect transistor life prediction method, device, electronic device and computer-readable storage medium based on multi-scale features proposed by the present invention mainly aim to improve the accuracy of life prediction for metal oxide semi-conductor field effect transistors. BRIEF DESCRIPTION OF THE DRAWINGS

[0097] Figure 1 A schematic diagram of a process flow of a method for predicting the lifetime of a metal oxide semi-conductor field effect transistor based on multi-scale features provided by an embodiment of the present invention;

[0098] Figure 2 A schematic diagram of the structure of an electronic device for implementing the method for predicting the lifetime of a metal oxide semiconductor field effect transistor based on multi-scale features provided in one embodiment of the present invention.

[0099] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings in conjunction with the embodiments. DETAILED DESCRIPTION

[0100] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0101] The embodiment of the present application provides a method for predicting the life of a metal oxide semiconductor field effect transistor based on multi-scale features. The execution subject of the method for predicting the life of a metal oxide semiconductor field effect transistor based on multi-scale features includes but is not limited to at least one of the electronic devices such as a server and a terminal that can be configured to execute the method provided in the embodiment of the present application. In other words, the method for predicting the life of a metal oxide semiconductor field effect transistor based on multi-scale features can be executed by software or hardware installed in a terminal device or a server device. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc.

[0102] Example 1: Reference Figure 1 FIG. 1 is a flow chart of a method for predicting the lifetime of a metal oxide semiconductor field effect transistor based on multi-scale features according to an embodiment of the present invention. In this embodiment, the method for predicting the lifetime of a metal oxide semiconductor field effect transistor based on multi-scale features includes:

[0103] S1. Receive a prediction instruction, and confirm a prediction environment based on the prediction instruction, wherein the prediction environment includes a target field effect transistor and a life prediction system, wherein the life prediction system includes a data acquisition unit, a data conversion unit, a data analysis unit, and a result feedback unit.

[0104] It should be explained that the prediction instruction is an instruction issued by a person who wants to realize the life prediction of the metal oxide semiconductor field effect transistor. The prediction environment refers to the necessary environment that can realize the life prediction of the metal oxide semiconductor field effect transistor. The target field effect transistor refers to the metal oxide semiconductor field effect transistor for which the life prediction is required. Generally speaking, metal oxide semiconductor field effect transistor and field effect transistor both refer to metal-oxide semiconductor field effect transistor, and the field effect transistor includes a gate, a source, a drain, a substrate, an oxide insulating layer and a semiconductor channel. The life prediction system refers to an app or a small program used to realize the life prediction of the metal oxide semiconductor field effect transistor, wherein the life prediction system includes: a data acquisition unit, a data conversion unit, a data analysis unit and a result feedback unit.

[0105] For example, Xiao Zhang, as a designer of MOSFET, issues the prediction instruction and confirms the prediction environment in order to optimize the MOSFET. When Xiao Zhang tests the relevant parameters of the MOSFET and inputs the relevant parameters into the life prediction system, the life prediction system can predict the life of the MOSFET.

[0106] S2. Using the life prediction system to identify a life prediction database.

[0107] Further, the using the life prediction system to identify the life prediction database includes:

[0108] Confirming receipt of a data acquisition instruction from a data acquisition unit, acquiring a test sample set based on the data acquisition instruction, wherein the test sample set includes a plurality of test samples, and acquiring test sample parameters based on the test sample set, wherein the test sample parameters include a reference packaging standard and a reference size structure;

[0109] According to the reference packaging standard and the reference size structure, the test samples are respectively aggregated to obtain one or more initial test sample sets;

[0110] The following operations are performed on each of the one or more initial test sample sets:

[0111] Extracting multiple grouped test sample sets from the initial test sample set using a pre-constructed grouping threshold, wherein the number of test samples in the grouped test sample sets is the same as the grouping threshold;

[0112] The life prediction database is acquired based on the multiple grouped test sample sets.

[0113] It should be explained that the test sample refers to a MOSFET with known MOSFET-related parameters. The reference packaging standard refers to the standard for packaging MOSFETs. For example: TO packaging, SOT packaging, etc. The reference dimension structure refers to the structure and dimension parameters designed when designing a MOSFET, wherein the reference dimension structure includes but is not limited to oxide layer thickness, channel length, channel width, threshold voltage, gate voltage, source voltage, etc. Ideally, MOSFETs with the same reference packaging standard and the same reference dimension structure are the same.

[0114] Furthermore, the acquiring the life prediction database based on the multiple grouped test sample sets includes:

[0115] Obtain one or more operating voltage ranges and operating temperature ranges of the metal oxide semiconductor field effect transistor, extract a plurality of first operating temperatures from the operating temperature range using a preset first temperature division value, and respectively extract one or more first operating voltage sets from the one or more operating voltage ranges, wherein the first operating voltage set corresponds to the operating voltage range one by one, and the first operating voltage set includes a plurality of first operating voltages;

[0116] In a combined form, one or more first operating voltage sets are used to obtain multiple operating voltage groups, and the operating voltage groups include one or more first operating voltages, and the number corresponding to the one or more first operating voltages is the same as the number corresponding to the one or more first operating voltage sets;

[0117] Associating, in a combined form, an operating voltage group in a plurality of operating voltage groups, a first operating temperature in a plurality of first operating temperatures, and a grouped test sample set in a plurality of grouped test sample sets to obtain a plurality of initial test groups, wherein the initial test group includes a working voltage group, a first operating temperature, and a test sample set;

[0118] For each of the multiple initial test groups, perform the following operations:

[0119] Identify the working mode of the initial test group to obtain an identified working mode, wherein the identified working mode is a forward working mode, a reverse working mode or a body diode mode;

[0120] Using the identification working mode, performing a marking operation on the initial test group to obtain a matching test group, and summarizing the matching test groups to obtain multiple matching test groups;

[0121] A life prediction database is acquired based on the plurality of matching test groups.

[0122] It is understandable that the operating voltage range refers to the voltage range in which the MOS field effect transistor can work normally, and usually, there is more than one operating voltage range for the MOS field effect transistor, and the operating voltage range can be set in combination with specific problems. For example: select the MOS field effect transistor of model XYZ123, the operating voltage range corresponding to its threshold voltage is 1V to 2V, the operating voltage range corresponding to the gate-source voltage is 0V to 10V, the operating voltage range corresponding to the drain-source voltage is 0V to 30V, etc. The operating temperature range refers to the temperature range in which the MOS field effect transistor can work normally, for example: the operating temperature range corresponding to the industrial-grade MOS field effect transistor is: minus 40 degrees Celsius to 85 degrees Celsius.

[0123] Further, the first temperature division value refers to a value that can extract multiple temperature values ​​in the operating temperature range. For example, if the operating temperature range is 0 degrees Celsius to 70 degrees Celsius, and the preset first temperature division value is 10, then the first temperature division value can be used to extract 8 first operating temperatures in the operating temperature range, wherein the 8 first operating temperatures are respectively 0 degrees Celsius, 10 degrees Celsius, 20 degrees Celsius, 30 degrees Celsius, 40 degrees Celsius, 50 degrees Celsius, 60 degrees Celsius and 70 degrees Celsius. The method for extracting the first operating voltage set in the operating voltage range is the same as the method for extracting the first operating temperature in the operating temperature range, and will not be repeated here. The forward working mode means that the drain of the half field effect transistor is forward biased relative to the source, that is, the drain voltage is higher than the source voltage, the reverse working mode means that the drain is reverse biased relative to the source, that is, the source voltage is higher than the drain voltage, and the body diode mode means that the parasitic body diode inside the half field effect transistor is in a conducting state, and the parasitic body diode refers to a PN junction formed between the source and the substrate.

[0124] Exemplarily, the MOSFET has three operating voltage ranges, and three first operating voltages can be extracted from each of the three operating voltage ranges, and the three first operating voltages corresponding to each operating voltage range constitute a first operating voltage set. In combination, 27 operating voltage groups can be obtained using the three first operating voltage sets, and each of the 27 operating voltage groups includes three first operating voltages.

[0125] It is understandable that the method for obtaining the initial test group is the same as the method for obtaining the working voltage group, and can achieve the same effect, which will not be repeated here. For example: there are 4 initial test groups in total, among which the identification working mode corresponding to the first initial test group is: forward working mode, then the first initial test group is identified as: forward working mode-initial test group using the forward working mode, the identification working mode corresponding to the second initial test group is: reverse working mode, and the second initial test group is identified as: reverse working mode-initial test group using the reverse working mode, and the method for marking the third initial test group and the fourth initial test group is the same as the method for marking the first initial test group and the second initial test group, and can achieve the same effect, which will not be repeated here.

[0126] It should be understood that different operating modes will affect the life of the MOSFET. Therefore, predicting the life of the MOSFET according to different operating modes can improve the accuracy of the life prediction of the target MOSFET and reduce the energy consumption required for predicting the life of the target MOSFET. Please refer to the subsequent embodiments for specific steps.

[0127] It should be explained that the step of acquiring a life prediction database based on the multiple matching test groups includes:

[0128] Sequentially extract matching test groups from the multiple matching test groups, and perform the following operations on the extracted matching test groups:

[0129] The following operations are performed on each test sample in the test sample set corresponding to the matching test group:

[0130] Setting the temperature of the environment in which the test sample is located to match the first operating temperature corresponding to the test group to match the operating voltage group corresponding to the test group, starting the test sample, and obtaining an on-resistance timing sequence based on the started test sample, wherein the on-resistance timing sequence includes multiple on-resistances;

[0131] Summarizing the on-resistance timing to obtain an on-resistance timing set, and using the on-resistance timing set to obtain an on-resistance group timing, wherein the on-resistance group timing includes multiple on-resistance timing groups, and the on-resistance timing group includes multiple on-resistances, and the number corresponding to the multiple on-resistances in the on-resistance timing group is the grouping threshold;

[0132] Extracting a first on-resistance timing group from the on-resistance group timing to obtain an analysis resistance timing group, and obtaining an on-resistance variance using a plurality of on-resistances corresponding to the analysis resistance timing group, wherein the on-resistance variance is a variance of the plurality of on-resistances;

[0133] comparing the on-resistance variance with a preset on-resistance variance threshold;

[0134] If the on-resistance variance is greater than or equal to the on-resistance variance threshold, obtaining an updated test group using the operating voltage group and the first operating temperature corresponding to the matching test group, and using the updated test group as the matching test group, returning to the step of performing the following operation on each test sample in the test sample set corresponding to the matching test group;

[0135] If the on-resistance variance is less than the on-resistance variance threshold, the on-resistance mean time sequence is obtained by using the on-resistance group time sequence, wherein the on-resistance mean time sequence includes multiple on-resistance mean values, the on-resistance mean values ​​correspond to the on-resistance time sequence groups one by one, and the on-resistance mean value is the mean value of the multiple on-resistances corresponding to the on-resistance time sequence groups;

[0136] A life prediction database is obtained based on the on-resistance mean time series.

[0137] It should be understood that the on-resistance refers to the resistance of the test sample when the test sample is turned on under the working voltage. Optionally, the on-resistance is obtained by measuring the current value of the test sample in each component, dividing the first working voltage in the working voltage group by the current value corresponding to the first working voltage, and summing them. Optionally, the on-resistance timing sequence is obtained by: using a preset detection frequency, monitoring the on-resistance of the test sample after startup and the time corresponding to the on-resistance, and sorting the on-resistance in order from the first to the last time corresponding to the on-resistance, so as to obtain the on-resistance timing sequence.

[0138] It should be explained that due to the existence of errors in process, materials, etc., the test samples in the matching test group may be inconsistent, which may in turn cause the constructed life prediction database to be inaccurate. Therefore, before constructing the life prediction database, it is necessary to combine the specific parameters of each test sample to improve the accuracy of constructing the life prediction database.

[0139] Furthermore, the on-resistances of the same bit sequence in different on-resistance timing sequences are summarized to obtain the on-resistance timing sequence group. For example, if the matching test group includes 3 test samples, 3 on-resistance timing sequences can be obtained using the 3 test samples, wherein the first on-resistance timing sequence is: 3, 3.5, 4, the second on-resistance timing sequence is: 3.1, 3.4, 3.9, and the third on-resistance timing sequence is: 2.9, 3.3, 3.7, then the bit sequences corresponding to 3, 3.1 and 2.9 are all 1, and 3, 3.1 and 2.9 constitute the first on-resistance timing sequence group in the on-resistance group timing sequence, wherein the on-resistance group timing sequence is composed of: (3, 3.1, 2.9), (3.4, 3.5, 3.3) and (4, 3.9, 3.7).

[0140] It should be explained that the first on-resistance timing group in the on-resistance group timing can characterize the basic situation of the test samples in the matching test group. When the on-resistance variance is less than the on-resistance variance threshold, it indicates that the test samples in the matching test group are relatively consistent. The operating voltage group and the first operating temperature corresponding to the update test group and the matching test group are consistent, and the acquisition method of the update test group and the matching test group is the same, which will not be repeated here.

[0141] Furthermore, the acquiring of the life prediction database based on the on-resistance mean time series includes:

[0142] The first on-resistance average values ​​are sequentially extracted in the on-resistance average value time sequence, and the following operations are performed on the extracted first on-resistance average values:

[0143] Based on the extracted first on-resistance average value, a second on-resistance average value is confirmed in the on-resistance average value time sequence, wherein the second on-resistance average value is adjacent to the first on-resistance average value, and the second on-resistance average value lags behind the first on-resistance average value in the on-resistance average value time sequence;

[0144] The on-resistance change value is obtained by using the first on-resistance average value and the second on-resistance average value, and the sorting time is confirmed by using the on-resistance change value, wherein the calculation formula of the sorting time is as follows:

[0145] ;

[0146] in, Indicates the sorting time, Indicates taking the minimum value, represents the time corresponding to the first on-resistance average value, represents the time corresponding to the second on-resistance average value;

[0147] Using the sorting time to identify the on-resistance change value to obtain a sorting resistance change value, summarizing the sorting resistance change values ​​to obtain a sorting resistance change value set, and sorting the sorting resistance change values ​​in the sorting resistance change value set in the order of sorting time from earliest to later to obtain a sorting resistance change sequence;

[0148] Obtaining a first predicted lifespan using the ranked resistance change sequence, a pre-constructed variational mode decomposition algorithm, and a pre-constructed neural network model;

[0149] A second test life is obtained according to the test sample, and the second test life is summarized to obtain a second test life set. A second test life mean is obtained based on the second test life set, wherein the second test life mean is the mean of multiple second test lives in the second test life set. The life deviation ratio is calculated using the first predicted life and the second test life mean, and the calculation formula is as follows:

[0150] ;

[0151] in, represents the life deviation ratio, represents the first predicted life span, represents the second test life mean;

[0152] comparing the life deviation ratio with a preset life deviation threshold;

[0153] If the life deviation ratio is greater than the life deviation threshold, returning to the step of obtaining an updated test group by using the working voltage group and the first working temperature corresponding to the matching test group;

[0154] If the life deviation ratio is less than or equal to the life deviation threshold, the reference life is calculated based on the first predicted life and the average of the second tested life, and the calculation formula is as follows:

[0155] ;

[0156] in, represents the reference life, All are preset coefficients;

[0157] A life prediction database is acquired based on the reference life.

[0158] It should be explained that the second on-resistance mean value is subtracted from the first on-resistance to obtain the on-resistance change value. The method of using the sorting time to mark the on-resistance change value is the same as the method of using the identification working mode to perform the marking operation on the initial test group, and can achieve the same effect, which will not be repeated here.

[0159] Exemplarily, the on-resistance average time sequence is: 1, 2, 4, 7 and 11. The first on-resistance average extracted in the on-resistance average time sequence is 1, and the second on-resistance average confirmed by the first on-resistance average is 2. Then the on-resistance change value obtained by 1 and 2 is 1. When the extracted first on-resistance average is 2, the second on-resistance average confirmed by 2 is 4, and the on-resistance change value calculated by 2 and 4 is 2. And so on. It can be seen from the on-resistance average time sequence that the time corresponding to the on-resistance change value of 1 is before the time corresponding to the on-resistance change value of 2. Therefore, the obtained sorted resistance change sequence is: 1, 2, 3 and 4.

[0160] It should be explained that the use of the sorted resistance change sequence, the pre-constructed variational modal decomposition algorithm and the pre-constructed neural network model to obtain the first predicted life refers to: using a preset sliding window, sequentially extracting the change value sequence in the sorted resistance change sequence, using the variational modal decomposition algorithm and the change value sequence to obtain the modal classification and residual components, inputting the modal classification and residual classification into the neural network model to obtain the predicted modal components, superimposing the predicted modal components, and obtaining the first predicted life. The technology of obtaining the first predicted life using the sorted resistance change sequence, the variational modal decomposition algorithm and the neural network model is a prior art and will not be repeated here. Optionally, the NARX neural network model is used as the neural network model.

[0161] It should be noted that the first predicted life obtained is used to characterize the predicted life under the working conditions corresponding to the matching test group, wherein the parameters used to describe the working conditions are the first operating temperature and operating voltage group corresponding to the matching test group. The second test life refers to the life of the test sample actually tested under the working conditions. Optionally, the second test life is obtained by setting a power cycle experiment. When the life deviation ratio is greater than the life deviation threshold, it indicates that an accidental situation may have occurred in the test sample, for example: the bond is disconnected due to an accidental situation. Therefore, under this working condition, there is a large error in the life prediction of the matching test group.

[0162] It should be explained that the obtaining of the life prediction database based on the reference life includes:

[0163] The monitoring temperature time sequence is obtained by using the test sample after startup and the pre-built temperature sensor, and the time when the test sample is started is used as the starting point to record the time in real time to obtain the real-time time, wherein the monitoring temperature time sequence includes multiple monitoring temperature groups, wherein the monitoring temperature group is composed of monitoring time and monitoring temperature, and based on the pre-built sliding window, the analysis temperature time sequence is sequentially extracted from the monitoring temperature time sequence, wherein the step size of the sliding window is preset, the analysis temperature time sequence includes multiple monitoring temperature groups, and the extracted analysis temperature time sequence is subjected to the following operations:

[0164] Acquire an analysis temperature variance based on the analysis temperature time series, wherein the analysis temperature variance is a variance of a plurality of monitored temperatures in the analysis temperature time series, and compare the analysis temperature variance with a preset temperature variance threshold;

[0165] If the analysis temperature variance is greater than or equal to the temperature variance threshold, then the analysis temperature mean is obtained based on the analysis temperature time series, wherein the analysis temperature mean is the mean of multiple monitoring temperatures in the analysis temperature time series, and the following operations are performed for each monitoring temperature in the analysis temperature time series:

[0166] Obtaining an analysis temperature difference value by using the monitoring temperature and the analysis temperature mean, wherein the analysis temperature difference value is the absolute difference between the monitoring temperature and the analysis temperature mean, summarizing the analysis temperature difference values ​​to obtain an analysis temperature difference value set, and sorting the analysis temperature difference values ​​in the analysis temperature difference value set in descending order to obtain an analysis temperature difference value sequence;

[0167] Extract the first analysis temperature difference value from the analysis temperature difference sequence to obtain a rejection temperature difference value, obtain the position sequence of the rejection temperature difference value in the analysis temperature sequence, obtain a rejection position sequence, and if the rejection position sequence is 1, return to the step of sequentially extracting the analysis temperature sequence in the monitoring temperature sequence;

[0168] If the elimination bit sequence is not 1, the monitoring temperature corresponding to the elimination temperature difference is eliminated in the analysis temperature sequence to obtain an updated temperature sequence, and the updated temperature sequence is used as the analysis temperature sequence. Return to the step of obtaining the analysis temperature variance based on the analysis temperature sequence, and record the number of monitoring temperatures corresponding to the elimination temperature difference eliminated in the analysis temperature sequence to obtain the elimination number.

[0169] When the number of rejections is greater than or equal to the preset rejection threshold or the analysis temperature variance is greater than or equal to the temperature variance threshold, returning to the step of sequentially extracting the analysis temperature time series in the monitoring temperature time series until the analysis temperature variance is less than the temperature variance threshold, and obtaining the target temperature time series;

[0170] A life prediction database is obtained based on the target temperature time series, the reference life and the real time.

[0171] It should be understood that the temperature sensor refers to a sensor for detecting temperature. Optionally, an infrared temperature sensor is used as the temperature sensor, and the monitored temperature is the temperature of the chip surface where the metal oxide semiconductor field effect transistor is located monitored by the temperature sensor. The monitoring temperature sequence is a sequence obtained by sorting the monitoring temperatures corresponding to the monitoring temperature group in order from the earliest to the latest according to the monitoring time. Optionally, a fixed sliding window with a preset window size is used as the sliding window, and the analysis temperature sequence refers to the sequence selected in the monitoring temperature sequence using the sliding window.

[0172] Further, when the analysis temperature variance is greater than or equal to the temperature variance threshold, it indicates that the state of heat exchange between the MOSFET and the environment is not yet stable or the measured monitoring temperature has a large noise. For example: the analysis temperature sequence is: 4, 3.7 and 3.4, and the monitoring temperature corresponding to the eliminated temperature difference is 4, then the bit sequence of the eliminated temperature difference in the analysis temperature sequence is 1, which indicates that the state of heat exchange between the MOSFET and the environment is not yet stable, so a new analysis temperature sequence needs to be extracted. When the elimination bit sequence is not 1, there are the following two situations, one is that the measured monitoring temperature has a large noise, and the other is that the temperature of the MOSFET when it is heat exchanged with the environment has a large change. The elimination number is greater than or equal to the preset elimination threshold or the analysis temperature variance is greater than or equal to the temperature variance threshold, indicating that the reason why the MOSFET is unstable during operation is that the state of heat exchange between the MOSFET and the environment is not yet stable, so a relatively stable heat exchange state can be confirmed by obtaining a new analysis temperature sequence.

[0173] It should be explained that the acquisition of the life prediction database based on the target temperature time series, reference life and real time includes:

[0174] Confirming a start time in the monitoring temperature sequence, wherein the start time is the monitoring time corresponding to the first monitoring temperature group in the monitoring temperature sequence, and confirming an end time using the target temperature sequence, wherein the end time is the monitoring time corresponding to the last monitoring temperature group in the target temperature sequence;

[0175] Obtaining a target temperature average value by using the target temperature time series, wherein the target temperature average value is an average value of multiple monitoring temperatures corresponding to the target temperature time series;

[0176] Determine a detection time based on the start time and the end time, wherein the detection time is an absolute difference between the start time and the end time, and when the real time reaches the detection time, obtain a gate threshold voltage using a test sample after startup;

[0177] Acquiring an operating junction temperature based on the gate threshold voltage;

[0178] A life prediction database is obtained using the operating junction temperature, target temperature average and reference life.

[0179] It should be understood that the target temperature average represents the temperature at which the MOSFET operates under stable heat exchange conditions. This temperature is used to represent the reference packaging standard, the first operating temperature, etc., and can represent the heat dissipation characteristics of the MOSFET. The operating junction temperature is used to represent the temperature of the MOSFET under design, material, and operating conditions. Optionally, the gate threshold voltage is obtained by a static transfer characteristic curve method.

[0180] Furthermore, the method of obtaining a life prediction database by using the operating junction temperature, the target temperature mean and the reference life includes:

[0181] Correlating the operating junction temperature, the reference packaging standard, the target temperature average, the first operating temperature, the operating voltage group, the identified operating mode and the reference life to obtain associated data;

[0182] Confirm receipt of the data conversion instruction from the data conversion unit, obtain a standard data template based on the data conversion instruction, import the associated data into the standard data template, and obtain detection data, wherein the detection data is as follows:

[0183] ;

[0184] in, represents the detection data, Indicates the reference packaging standard, represents the identification working mode, and The value of is 1, 2 or 3. When it is 1, it indicates the forward working mode. When it is 2, it indicates the reverse working mode. When it is 3, it indicates body diode mode. represents the first operating temperature, Represents the operating junction temperature, represents the target temperature mean, represents the reference life, represents the operating voltage group, and ,in, Respectively represent the first working voltage, the second working voltage and the A first operating voltage, Indicates that there are a total of a first operating voltage;

[0185] The detection data are respectively aggregated using the working modes corresponding to the detection data to obtain three detection data sets;

[0186] The three detection data sets are aggregated to obtain a life prediction database.

[0187] It should be explained that the detection data are summarized respectively using the working modes corresponding to the detection data, which means: the detection data are distinguished according to the working modes corresponding to the detection data. For example: there are 5 detection data, among which the working modes corresponding to the first detection data and the second detection data are both forward working modes, the working modes corresponding to the third detection data and the fourth detection data are both reverse working modes, and the working mode corresponding to the fifth detection data is the body diode mode, then the detection data are summarized respectively using the working modes corresponding to the detection data to obtain three detection data sets, among which the first detection data set is composed of the first detection data and the second detection data set, the second detection data set is composed of the third detection data and the fourth detection data set, and the third detection data set is composed of the fifth detection data. The life prediction database is composed of three detection data sets.

[0188] Further, the use of the prediction parameters to search in a life prediction database to obtain a target predicted life includes:

[0189] Identify the working mode corresponding to the target working voltage group in the prediction parameters to obtain the target working mode;

[0190] Using the target packaging standard and the target working mode in the prediction parameters, a search life data set is identified in the life prediction database, wherein the working mode corresponding to the search life data set is the target working mode, and the reference packaging standard corresponding to the search life data set is the target packaging standard, and the search life data set includes a plurality of search life data;

[0191] The following operations are performed for each of the multiple retrieved life span data:

[0192] Using the retrieved life data, reference retrieval data is identified in the detection data corresponding to the retrieved life data, wherein the reference retrieval data is as follows:

[0193] ;

[0194] in, Indicates reference retrieval data;

[0195] Acquire target retrieval data using prediction parameters, aggregate the reference retrieval data and the target retrieval data, and obtain a life prediction data set;

[0196] confirming receipt of a data analysis instruction from a data analysis unit, acquiring a k-means clustering algorithm based on the data analysis instruction, clustering the life prediction data set into a plurality of life retrieval data sets using the k-means clustering algorithm, and confirming a target life retrieval data set from the plurality of life retrieval data sets, wherein the target life retrieval data set is a life retrieval data set including the target retrieval data;

[0197] A target predicted lifespan is calculated based on the target lifespan retrieval data set.

[0198] It should be understood that the method for obtaining the target working mode is the same as the method for obtaining the reference working mode, and can achieve the same effect. The target packaging standard is the same as the definition of the reference packaging standard, and can achieve the same effect, which will not be repeated here. Generally, the data in the life prediction data set can be clustered into multiple clusters using the k-means clustering algorithm. Here, the data contained in each of the multiple clusters constitutes the life retrieval data set, and the data includes: reference retrieval data and target retrieval data. Therefore, the target life retrieval data set can be identified in the clustered multiple life detection data sets through the target detection data.

[0199] S3. Analyze prediction parameters based on the target field effect transistor, wherein the prediction parameters include: a target operating voltage group, a target operating temperature, a target packaging standard, a search temperature mean, and a target junction temperature. Use the prediction parameters to search in a life prediction database to obtain a target predicted life.

[0200] It should be explained that the target operating voltage group has the same definition as the operating voltage group and can achieve the same effect. The target operating temperature has the same definition as the first operating temperature. Optionally, multiple temperature values ​​are obtained by repeatedly detecting the temperature in the environment where the metal oxide semi-field effect transistor is to work, and the average of the multiple temperature values ​​is taken as the target operating temperature. The target temperature junction has the same definition as the operating temperature junction and can achieve the same effect. The method for obtaining the retrieval temperature average is the same as the method for obtaining the target temperature average, and can achieve the same effect.

[0201] Further, the calculating the target predicted lifespan based on the target lifespan retrieval data set includes:

[0202] The reference weight value is calculated using the target retrieval data in the target life retrieval data set and each of the multiple reference retrieval data corresponding to the target life retrieval data set. The calculation formula is as follows:

[0203] ;

[0204] in, represents the reference weight value, are preset coefficients, and , They respectively represent the target operating temperature, target junction temperature and average retrieval temperature corresponding to the target retrieval data. They respectively represent the first operating temperature, operating junction temperature and target temperature average corresponding to the reference retrieval data, Respectively represent the target operating voltage group The target operating voltage and the a first operating voltage;

[0205] The reference weight values ​​are summarized to obtain a reference weight value set, and the target predicted life is calculated based on the reference weight value set. The calculation formula is as follows:

[0206] ;

[0207] in, represents the target predicted lifespan, Indicates that the reference weight value set has A reference weight value, Respectively represent the reference weight value set reference weight values ​​and the reference weight value set A reference weight value, Indicates the reference weight value set The reference life span corresponding to a reference weight value.

[0208] It is understandable that the coefficient can be set by the degree of influence. Optionally, the coefficient is obtained by an orthogonal hierarchical method. When the reference weight value is smaller, it indicates that the target retrieval data is closer to the reference retrieval data. Therefore, when calculating the target predicted life, the smaller the reference weight value is set, the larger the weight value of the reference life corresponding to the reference weight value is, thereby improving the accuracy of calculating the target predicted life.

[0209] S4. Using a result feedback unit, the target predicted lifetime is sent to an initiator of a prediction instruction to achieve lifetime prediction of a target field effect transistor.

[0210] Exemplarily, Xiao Wang measures the designed metal oxide semiconductor field effect transistor according to the parameters required by the life prediction system, obtains the prediction parameters for realizing life prediction, inputs the prediction parameters into the life prediction system, and the life prediction system sends the target predicted life of the metal oxide semiconductor field effect transistor to Xiao Wang, thereby realizing the life prediction of the target field effect transistor.

[0211] Compared with the problems described in the background technology, the present invention first uses the life prediction system to confirm the life prediction database, and considers the reference packaging standard, first operating temperature, target temperature average, operating voltage group and operating junction temperature of the metal oxide semiconductor field effect transistor when confirming the life prediction database. It can be seen that the embodiment of the present invention takes into account multiple characteristics of the metal oxide semiconductor field effect transistor when constructing the life prediction database, thereby laying the foundation for accurately obtaining the target predicted life, and when obtaining the relevant parameters of the test sample, wherein the relevant parameters include: target temperature average, operating junction temperature and reference life, different screening requirements are set respectively to improve the accuracy of obtaining the relevant parameters, thereby improving the accuracy of obtaining the target predicted life. The present invention parses prediction parameters based on the target field effect transistor, wherein the prediction parameters include: target operating voltage group, target operating temperature, target packaging standard, retrieval temperature mean and target junction temperature, and uses the prediction parameters to search in the life prediction database to obtain the target predicted life. It can be seen that the present invention first uses the target operating voltage group to identify the target operating mode, and then combines the target packaging standard and the target operating mode to realize the retrieval of the life prediction database to obtain the detection life database set, and then combines the target operating voltage group, target operating temperature, retrieval temperature mean and target junction temperature of the target field effect transistor according to the k-means clustering algorithm to retrieve the target life retrieval data set in the detection life database set, and calculates different reference weight values ​​according to different reference detection data in the target life retrieval data set, and then, sets the proportion of the reference life corresponding to the reference weight value through different reference weight values, thereby improving the accuracy of the obtained target predicted life. Therefore, the metal oxide semi-conductor field effect transistor life prediction method, device, electronic device and computer-readable storage medium based on multi-scale features proposed by the present invention mainly aim to improve the accuracy of life prediction for metal oxide semi-conductor field effect transistors.

[0212] Example 2: Figure 2 , is a schematic diagram of the structure of an electronic device for implementing a method for predicting the lifetime of a metal oxide semi-conductor field effect transistor based on multi-scale features provided by an embodiment of the present invention.

[0213] The electronic device 1 may include a processor 10, a memory 11, a bus 12 and a communication interface 13, and may also include a computer program stored in the memory 11 and executable on the processor 10, such as a metal oxide semiconductor field effect transistor lifetime prediction program based on multi-scale features.

[0214] Among them, the memory 11 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, mobile hard disk, multimedia card, card-type memory (for example: SD or DX memory, etc.), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 11 can be an internal storage unit of the electronic device 1, such as a mobile hard disk of the electronic device 1. In other embodiments, the memory 11 can also be an external storage device of the electronic device 1, such as a plug-in mobile hard disk, a smart memory card (SmartMediaCard, SMC), a secure digital (SecureDigital, SD) card, a flash card (FlashCard), etc. equipped on the electronic device 1. Further, the memory 11 can also include both an internal storage unit of the electronic device 1 and an external storage device. The memory 11 can not only be used to store application software and various types of data installed in the electronic device 1, such as the code of the metal oxide semiconductor field effect transistor life prediction program based on multi-scale features, but also can be used to temporarily store data that has been output or is to be output.

[0215] In some embodiments, the processor 10 may be composed of an integrated circuit, for example, a single packaged integrated circuit, or a plurality of packaged integrated circuits with the same or different functions, including one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and combinations of various control chips. The processor 10 is the control core (Control Unit) of the electronic device, and uses various interfaces and lines to connect various components of the entire electronic device, and executes or executes programs or modules stored in the memory 11 (such as a metal oxide semiconductor field effect transistor life prediction program based on multi-scale features, etc.), and calls data stored in the memory 11 to execute various functions of the electronic device 1 and process data.

[0216] The bus may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. The bus is configured to realize connection and communication between the memory 11 and at least one processor 10, etc.

[0217] Figure 2 Only an electronic device with components is shown, and those skilled in the art will understand that Figure 2The structure shown does not constitute a limitation on the electronic device 1, and may include fewer or more components than shown in the figure, or combine certain components, or arrange the components differently.

[0218] For example, although not shown, the electronic device 1 may also include a power source (such as a battery) for supplying power to various components. Preferably, the power source may be logically connected to the at least one processor 10 through a power management device, so that the power management device can realize functions such as charging management, discharging management, and power consumption management. The power source may also include one or more DC or AC power sources, recharging devices, power failure detection circuits, power converters or inverters, power status indicators, and other arbitrary components. The electronic device 1 may also include a variety of sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be repeated here.

[0219] Furthermore, the electronic device 1 may also include a network interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a WI-FI interface, a Bluetooth interface, etc.), which is generally used to establish a communication connection between the electronic device 1 and other electronic devices.

[0220] Optionally, the electronic device 1 may further include a user interface, which may be a display (Display), an input unit (such as a keyboard (Keyboard)), and optionally, the user interface may also be a standard wired interface or a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, and an OLED (Organic Light-Emitting Diode) touch device, etc. The display may also be appropriately referred to as a display screen or a display unit, which is used to display information processed in the electronic device 1 and to display a visual user interface.

[0221] It should be understood that the embodiment is for illustration only and the scope of the patent application is not limited to this structure.

[0222] The MOSFET lifetime prediction program based on multi-scale features stored in the memory 11 in the electronic device 1 is a combination of multiple instructions. When running in the processor 10, the following can be achieved:

[0223] receiving a prediction instruction, and confirming a prediction environment based on the prediction instruction, wherein the prediction environment includes a target field effect transistor and a life prediction system, wherein the life prediction system includes a data acquisition unit, a data conversion unit, a data analysis unit, and a result feedback unit;

[0224] Using the life prediction system to identify a life prediction database;

[0225] Analyzing prediction parameters based on the target field effect transistor, wherein the prediction parameters include: a target operating voltage group, a target operating temperature, a target packaging standard, a search temperature mean, and a target junction temperature, and using the prediction parameters to search in a life prediction database to obtain a target predicted life;

[0226] The target predicted lifetime is sent to the initiator of the prediction instruction by using the result feedback unit, so as to realize the lifetime prediction of the target field effect transistor.

[0227] Specifically, the specific implementation method of the processor 10 for the above instructions can refer to Figure 1 to Figure 2 The description of the relevant steps in the corresponding embodiments will not be repeated here.

[0228] Furthermore, if the module / unit integrated in the electronic device 1 is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. The computer-readable storage medium can be volatile or non-volatile. For example, the computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a mobile hard disk, a magnetic disk, an optical disk, a computer memory, and a read-only memory (ROM).

[0229] The present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor of an electronic device, the computer program can implement:

[0230] receiving a prediction instruction, and confirming a prediction environment based on the prediction instruction, wherein the prediction environment includes a target field effect transistor and a life prediction system, wherein the life prediction system includes a data acquisition unit, a data conversion unit, a data analysis unit, and a result feedback unit;

[0231] Using the life prediction system to identify a life prediction database;

[0232] Analyzing prediction parameters based on the target field effect transistor, wherein the prediction parameters include: a target operating voltage group, a target operating temperature, a target packaging standard, a search temperature mean, and a target junction temperature, and using the prediction parameters to search in a life prediction database to obtain a target predicted life;

[0233] The target predicted lifetime is sent to the initiator of the prediction instruction by using the result feedback unit, so as to realize the lifetime prediction of the target field effect transistor.

[0234] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0235] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0236] It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0237] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for predicting the lifetime of a metal-oxide-semiconductor field-effect transistor based on multi-scale features, characterized in that: The method comprises: receiving a prediction instruction, and confirming a prediction environment based on the prediction instruction, wherein the prediction environment includes a target field effect transistor and a life prediction system, wherein the life prediction system includes a data acquisition unit, a data conversion unit, a data analysis unit, and a result feedback unit; Using the life prediction system to identify a life prediction database; Wherein, the method of using the life prediction system to identify the life prediction database includes: Confirming receipt of a data acquisition instruction from a data acquisition unit, acquiring a test sample set based on the data acquisition instruction, wherein the test sample set includes a plurality of test samples, and acquiring test sample parameters based on the test sample set, wherein the test sample parameters include a reference packaging standard and a reference size structure; According to the reference packaging standard and the reference size structure, the test samples are respectively aggregated to obtain one or more initial test sample sets; The following operations are performed on each of the one or more initial test sample sets: Extracting multiple grouped test sample sets from the initial test sample set using a pre-constructed grouping threshold, wherein the number of test samples in the grouped test sample sets is the same as the grouping threshold; Acquire the life prediction database based on the multiple grouped test sample sets; The acquiring the life prediction database based on the multiple grouped test sample sets includes: Obtain one or more operating voltage ranges and operating temperature ranges of the metal oxide semiconductor field effect transistor, extract a plurality of first operating temperatures from the operating temperature range using a preset first temperature division value, and respectively extract one or more first operating voltage sets from the one or more operating voltage ranges, wherein the first operating voltage set corresponds to the operating voltage range one by one, and the first operating voltage set includes a plurality of first operating voltages; In a combined form, one or more first operating voltage sets are used to obtain multiple operating voltage groups, and the operating voltage groups include one or more first operating voltages, and the number corresponding to the one or more first operating voltages is the same as the number corresponding to the one or more first operating voltage sets; Associating, in a combined form, an operating voltage group in a plurality of operating voltage groups, a first operating temperature in a plurality of first operating temperatures, and a grouped test sample set in a plurality of grouped test sample sets to obtain a plurality of initial test groups, wherein the initial test group includes a working voltage group, a first operating temperature, and a test sample set; For each of the multiple initial test groups, perform the following operations: Identify the working mode of the initial test group to obtain an identified working mode, wherein the identified working mode is a forward working mode, a reverse working mode or a body diode mode; Using the identification working mode, performing a marking operation on the initial test group to obtain a matching test group, and summarizing the matching test groups to obtain multiple matching test groups; acquiring a life prediction database based on the plurality of matching test groups; The acquiring of a life prediction database based on the multiple matching test groups comprises: Sequentially extract matching test groups from the multiple matching test groups, and perform the following operations on the extracted matching test groups: The following operations are performed on each test sample in the test sample set corresponding to the matching test group: Setting the temperature of the environment in which the test sample is located to match the first operating temperature corresponding to the test group to match the operating voltage group corresponding to the test group, starting the test sample, and obtaining an on-resistance timing sequence based on the started test sample, wherein the on-resistance timing sequence includes multiple on-resistances; Summarizing the on-resistance timing to obtain an on-resistance timing set, and using the on-resistance timing set to obtain an on-resistance group timing, wherein the on-resistance group timing includes multiple on-resistance timing groups, and the on-resistance timing group includes multiple on-resistances, and the number corresponding to the multiple on-resistances in the on-resistance timing group is the grouping threshold; Extracting a first on-resistance timing group from the on-resistance group timing to obtain an analysis resistance timing group, and obtaining an on-resistance variance using a plurality of on-resistances corresponding to the analysis resistance timing group, wherein the on-resistance variance is a variance of the plurality of on-resistances; comparing the on-resistance variance with a preset on-resistance variance threshold; If the on-resistance variance is greater than or equal to the on-resistance variance threshold, obtaining an updated test group using the operating voltage group and the first operating temperature corresponding to the matching test group, and using the updated test group as the matching test group, returning to the step of performing the following operation on each test sample in the test sample set corresponding to the matching test group; If the on-resistance variance is less than the on-resistance variance threshold, the on-resistance mean time sequence is obtained by using the on-resistance group time sequence, wherein the on-resistance mean time sequence includes multiple on-resistance mean values, the on-resistance mean values ​​correspond to the on-resistance time sequence groups one by one, and the on-resistance mean value is the mean value of the multiple on-resistances corresponding to the on-resistance time sequence groups; Acquire a life prediction database based on the on-resistance mean time series; The method of acquiring a life prediction database based on the on-resistance mean time series includes: The first on-resistance average values ​​are sequentially extracted in the on-resistance average value time sequence, and the following operations are performed on the extracted first on-resistance average values: Based on the extracted first on-resistance average value, a second on-resistance average value is confirmed in the on-resistance average value time sequence, wherein the second on-resistance average value is adjacent to the first on-resistance average value, and the second on-resistance average value lags behind the first on-resistance average value in the on-resistance average value time sequence; The on-resistance change value is obtained by using the first on-resistance average value and the second on-resistance average value, and the sorting time is confirmed by using the on-resistance change value, wherein the calculation formula of the sorting time is as follows: ； in, Indicates the sorting time, Indicates taking the minimum value, represents the time corresponding to the first on-resistance average value, represents the time corresponding to the second on-resistance average value; Using the sorting time to identify the on-resistance change value to obtain a sorting resistance change value, summarizing the sorting resistance change values ​​to obtain a sorting resistance change value set, and sorting the sorting resistance change values ​​in the sorting resistance change value set in the order of sorting time from earliest to later to obtain a sorting resistance change sequence; Obtaining a first predicted life using the ranked resistance change sequence, a pre-constructed variational mode decomposition algorithm, and a pre-constructed neural network model; A second test life is obtained according to the test sample, and the second test life is summarized to obtain a second test life set. A second test life mean is obtained based on the second test life set, wherein the second test life mean is the mean of multiple second test lives in the second test life set. The life deviation ratio is calculated using the first predicted life and the second test life mean, and the calculation formula is as follows: ； in, represents the life deviation ratio, represents the first predicted life span, represents the second test life mean; comparing the life deviation ratio with a preset life deviation threshold; If the life deviation ratio is greater than the life deviation threshold, returning to the step of obtaining an updated test group by using the working voltage group and the first working temperature corresponding to the matching test group; If the life deviation ratio is less than or equal to the life deviation threshold, the reference life is calculated based on the first predicted life and the average of the second tested life, and the calculation formula is as follows: ； in, represents the reference life, All are preset coefficients; Acquire a life prediction database based on the reference life; The acquiring of a life prediction database based on the reference life includes: The monitoring temperature time sequence is obtained by using the test sample after startup and the pre-built temperature sensor, and the time when the test sample is started is used as the starting point to record the time in real time to obtain the real-time time, wherein the monitoring temperature time sequence includes multiple monitoring temperature groups, wherein the monitoring temperature group is composed of monitoring time and monitoring temperature, and based on the pre-built sliding window, the analysis temperature time sequence is sequentially extracted from the monitoring temperature time sequence, wherein the step size of the sliding window is preset, the analysis temperature time sequence includes multiple monitoring temperature groups, and the extracted analysis temperature time sequence is subjected to the following operations: Acquire an analysis temperature variance based on the analysis temperature time series, wherein the analysis temperature variance is a variance of a plurality of monitored temperatures in the analysis temperature time series, and compare the analysis temperature variance with a preset temperature variance threshold; If the analysis temperature variance is greater than or equal to the temperature variance threshold, then the analysis temperature mean is obtained based on the analysis temperature time series, wherein the analysis temperature mean is the mean of multiple monitoring temperatures in the analysis temperature time series, and the following operations are performed for each monitoring temperature in the analysis temperature time series: Obtaining an analysis temperature difference value by using the monitoring temperature and the analysis temperature mean, wherein the analysis temperature difference value is the absolute difference between the monitoring temperature and the analysis temperature mean, summarizing the analysis temperature difference values ​​to obtain an analysis temperature difference value set, and sorting the analysis temperature difference values ​​in the analysis temperature difference value set in descending order to obtain an analysis temperature difference value sequence; Extract the first analysis temperature difference value from the analysis temperature difference sequence to obtain a rejection temperature difference value, obtain the position sequence of the rejection temperature difference value in the analysis temperature sequence, obtain a rejection position sequence, and if the rejection position sequence is 1, return to the step of sequentially extracting the analysis temperature sequence in the monitoring temperature sequence; If the elimination bit sequence is not 1, the monitoring temperature corresponding to the elimination temperature difference is eliminated in the analysis temperature sequence to obtain an updated temperature sequence, and the updated temperature sequence is used as the analysis temperature sequence. Return to the step of obtaining the analysis temperature variance based on the analysis temperature sequence, and record the number of monitoring temperatures corresponding to the elimination temperature difference eliminated in the analysis temperature sequence to obtain the elimination number. When the number of rejections is greater than or equal to the preset rejection threshold or the analysis temperature variance is greater than or equal to the temperature variance threshold, returning to the step of sequentially extracting the analysis temperature time series in the monitoring temperature time series until the analysis temperature variance is less than the temperature variance threshold, and obtaining the target temperature time series; Acquire a life prediction database based on the target temperature time series, reference life and real time; The acquiring of the life prediction database based on the target temperature time series, the reference life and the real-time time comprises: Confirming a start time in the monitoring temperature sequence, wherein the start time is the monitoring time corresponding to the first monitoring temperature group in the monitoring temperature sequence, and confirming an end time using the target temperature sequence, wherein the end time is the monitoring time corresponding to the last monitoring temperature group in the target temperature sequence; Obtaining a target temperature average value by using the target temperature time series, wherein the target temperature average value is an average value of multiple monitoring temperatures corresponding to the target temperature time series; Determine a detection time based on the start time and the end time, wherein the detection time is an absolute difference between the start time and the end time, and when the real time reaches the detection time, obtain a gate threshold voltage using a test sample after startup; Acquiring an operating junction temperature based on the gate threshold voltage; Acquire a life prediction database using the operating junction temperature, target temperature average and reference life; The method of obtaining a life prediction database by using the working junction temperature, the target temperature mean and the reference life includes: Correlating the operating junction temperature, the reference packaging standard, the target temperature average, the first operating temperature, the operating voltage group, the identified operating mode and the reference life to obtain associated data; Confirm receipt of the data conversion instruction from the data conversion unit, obtain a standard data template based on the data conversion instruction, import the associated data into the standard data template, and obtain detection data, wherein the detection data is as follows: ； in, represents the detection data, Indicates the reference packaging standard, represents the identification working mode, and The value of is 1, 2 or 3. When it is 1, it indicates the forward working mode. When it is 2, it indicates the reverse working mode. When it is 3, it indicates body diode mode. represents the first operating temperature, Indicates the operating junction temperature, represents the target temperature mean, represents the reference life, represents the operating voltage group, and ,in, Respectively represent the first working voltage, the second working voltage and the A first operating voltage, Indicates that there are a total of a first operating voltage; The detection data are respectively aggregated using the working modes corresponding to the detection data to obtain three detection data sets; Summarizing the three detection data sets to obtain a life prediction database; Analyzing prediction parameters based on the target field effect transistor, wherein the prediction parameters include: a target operating voltage group, a target operating temperature, a target packaging standard, a search temperature mean, and a target junction temperature, and using the prediction parameters to search in a life prediction database to obtain a target predicted life; The target predicted lifetime is sent to the initiator of the prediction instruction by using the result feedback unit, so as to realize the lifetime prediction of the target field effect transistor.

2. The method for predicting the lifetime of a metal oxide semiconductor field effect transistor based on multi-scale features according to claim 1, characterized in that: The method of using the prediction parameters to search in a life prediction database to obtain a target predicted lifespan includes: Identify the working mode corresponding to the target working voltage group in the prediction parameters to obtain the target working mode; Using the target packaging standard and the target working mode in the prediction parameters, a search life data set is identified in the life prediction database, wherein the working mode corresponding to the search life data set is the target working mode, and the reference packaging standard corresponding to the search life data set is the target packaging standard, and the search life data set includes a plurality of search life data; The following operations are performed for each of the multiple retrieved life span data: Using the retrieved life data, reference retrieval data is identified in the detection data corresponding to the retrieved life data, wherein the reference retrieval data is as follows: ； in, Indicates reference retrieval data; Acquire target retrieval data using prediction parameters, aggregate the reference retrieval data and the target retrieval data, and obtain a life prediction data set; confirming receipt of a data analysis instruction from a data analysis unit, acquiring a k-means clustering algorithm based on the data analysis instruction, clustering the life prediction data set into a plurality of life retrieval data sets using the k-means clustering algorithm, and confirming a target life retrieval data set from the plurality of life retrieval data sets, wherein the target life retrieval data set is a life retrieval data set including the target retrieval data; A target predicted lifespan is calculated based on the target lifespan retrieval data set.

3. The method for predicting the lifetime of a metal oxide semiconductor field effect transistor based on multi-scale features according to claim 2, characterized in that: The calculating the target predicted lifespan based on the target lifespan retrieval data set comprises: The reference weight value is calculated using the target retrieval data in the target life retrieval data set and each of the multiple reference retrieval data corresponding to the target life retrieval data set. The calculation formula is as follows: ； in, represents the reference weight value, are preset coefficients, and , They respectively represent the target operating temperature, target junction temperature and average retrieval temperature corresponding to the target retrieval data. They respectively represent the first operating temperature, operating junction temperature and target temperature average corresponding to the reference retrieval data, Respectively represent the target operating voltage group The target operating voltage and the a first operating voltage; The reference weight values ​​are summarized to obtain a reference weight value set, and the target predicted life is calculated based on the reference weight value set. The calculation formula is as follows: ； in, represents the target predicted lifespan, Indicates that the reference weight value set has A reference weight value, Respectively represent the reference weight value set reference weight values ​​and the reference weight value set A reference weight value, Indicates the reference weight value set The reference life span corresponding to a reference weight value.

Images