Failure sample data generation method, device and equipment and readable storage medium

Abstract

The invention discloses a failure sample data generation method and device, equipment and a readable storage medium. The method comprises the steps that basic sample data are obtained based on a solving result of a second-order partial differential equation, the basic sample data comprise a plurality of first data pairs, and each first data pair comprises a set of input data of the second-order partial differential equation and a corresponding numerical solution of the input data; each group of input data is a physical quantity influencing the remaining available life of the optoelectronic device, the numerical solution corresponding to each group of input data is the theoretical remaining available life of the optoelectronic device, and the second-order partial differential equation is constructed based on a failure mechanism of the optoelectronic device in a physical field; and obtaining invalid sample data according to the basic sample data. Through the method and the device, the data quality of the generated failure sample data is improved.

Description

Technical Field

[0001] This application relates to the field of optical network operation and maintenance technology, specifically to a method, apparatus, device, and computer-readable storage medium for generating failure sample data. Background Technology

[0002] High-quality failure sample data of optoelectronic devices is a fundamental element for training large-scale failure analysis models. Existing technologies mainly use electrical simulation, finite element modeling, and Goumer iteration to create failure sample data, but the data quality of failure sample data obtained by existing technologies is low. Summary of the Invention

[0003] This application provides a method, apparatus, device, and computer-readable storage medium for generating failure sample data, which can solve the technical problem of low data quality of failure sample data obtained in the prior art.

[0004] In a first aspect, embodiments of this application provide a method for generating failure sample data, the method comprising: Based on the solution results of the second-order partial differential equation, basic sample data is obtained. The basic sample data includes multiple sets of first data pairs. Each set of first data pairs includes a set of input data of the second-order partial differential equation and its corresponding numerical solution. Each set of input data is a physical quantity that affects the remaining usable lifetime of the optoelectronic device. The numerical solution corresponding to each set of input data is the theoretical remaining usable lifetime of the optoelectronic device. The second-order partial differential equation is constructed based on the failure mechanism of the optoelectronic device in the physical field. Failure sample data is obtained based on the basic sample data.

[0005] In conjunction with the first aspect, in one implementation, obtaining failure sample data based on the basic sample data includes: Acquire actual sample data and detect whether the difference between the actual sample data and the basic sample data meets the correction conditions. The actual sample data includes multiple sets of second data pairs. Each set of second data pairs includes a physical quantity that affects the remaining usable lifetime of the optoelectronic device and its corresponding actual remaining usable lifetime of the optoelectronic device. If the correction conditions are met, the basic sample data is corrected to obtain the corrected sample data. Failed sample data is obtained based on the revised sample data.

[0006] In conjunction with the first aspect, in one implementation, the step of acquiring actual sample data and detecting whether the difference between the actual sample data and the basic sample data satisfies the correction condition includes: Obtain actual sample data, and for each second data pair in the actual sample data, determine the first data pair corresponding to each second data pair from the basic sample data. Each second data pair and its corresponding first data pair contain the same physical quantity that affects the remaining usable lifetime of the optoelectronic device. Calculate the difference between the actual remaining usable lifetime of the optoelectronic devices included in each second data pair and the theoretical remaining usable lifetime of the optoelectronic devices included in the corresponding first data pair. If the number of differences in the remaining available lifetime of the target is greater than the preset number, then the difference between the actual sample data and the basic sample data is determined to meet the correction condition, wherein the difference in the remaining available lifetime of the target is greater than or equal to the preset difference.

[0007] In conjunction with the first aspect, in one implementation, the basic sample data is corrected to obtain corrected sample data, including: The correction value is obtained by combining the differences in the remaining available lifetime of all targets; The theoretical remaining available lifetime of the optoelectronic devices included in the target first data corresponding to the difference in the remaining available lifetime of each target in the basic sample data is increased by the correction value.

[0008] In conjunction with the first aspect, in one implementation, obtaining the failed sample data based on the revised sample data includes: Based on the difference between the material properties of the optoelectronic device to be analyzed in the failure detection scenario and the material properties of the optoelectronic device, the revised sample data is adjusted to obtain failure sample data.

[0009] In conjunction with the first aspect, in one implementation, based on the characteristic difference between the characteristic value of the theoretical remaining usable lifetime of the optoelectronic device under the same physical field characteristic index and the characteristic value of the theoretical remaining usable lifetime of the optoelectronic device, the difference between the material properties of the optoelectronic device targeted by the failure detection scenario and the material properties of the optoelectronic device is characterized. The revised sample data is adjusted based on the difference between the material properties of the optoelectronic device targeted by the failure detection scenario and the material properties of the optoelectronic device, resulting in failure sample data including: Based on the characteristic differences and the operation and maintenance statistics of the optical communication network, the mean and variance parameters of the probability distribution are calculated. Using the mean parameter and variance parameter as the mathematical expectation and variance of the normal distribution, the adjustment amount of the theoretical remaining usable lifetime of each optoelectronic device included in the revised sample data is obtained; The theoretical remaining usable lifetime of each optoelectronic device included in the revised sample data is increased by a corresponding adjustment amount to obtain the failure sample data.

[0010] In conjunction with the first aspect, in one implementation, obtaining failure sample data based on the basic sample data includes: Based on the difference between the material properties of the optoelectronic device to be analyzed in the failure detection scenario and the material properties of the optoelectronic device, the basic sample data is adjusted to obtain failure sample data.

[0011] Secondly, embodiments of this application provide a failure sample data generation apparatus, the failure sample data generation apparatus comprising: The first construction module is used to obtain basic sample data based on the solution results of the second-order partial differential equation. The basic sample data includes multiple sets of first data pairs. Each set of first data pairs includes a set of input data of the second-order partial differential equation and its corresponding numerical solution. Each set of input data is a physical quantity that affects the remaining usable lifetime of the optoelectronic device. The numerical solution corresponding to each set of input data is the theoretical remaining usable lifetime of the optoelectronic device. The second-order partial differential equation is constructed based on the failure mechanism of the optoelectronic device in the physical field. The second construction module is used to obtain failed sample data based on the basic sample data.

[0012] Thirdly, embodiments of this application provide a failure sample data generation device, which includes a processor, a memory, and a failure sample data generation program stored in the memory and executable by the processor. When the failure sample data generation program is executed by the processor, it implements the steps of the failure sample data generation method as described in the first aspect.

[0013] Fourthly, embodiments of this application provide a computer-readable storage medium storing a failure sample data generation program, wherein when the failure sample data generation program is executed by a processor, it implements the steps of the failure sample data generation method as described in the first aspect.

[0014] The beneficial effects of the technical solutions provided in this application include: In this embodiment, basic sample data is obtained based on the solution results of the second-order partial differential equation. The basic sample data includes multiple sets of first data pairs. Each set of first data pairs includes a set of input data for the second-order partial differential equation and its corresponding numerical solution. Each set of input data represents a physical quantity affecting the remaining usable lifetime of the optoelectronic device, and the numerical solution corresponding to each set of input data represents the theoretical remaining usable lifetime of the optoelectronic device. The second-order partial differential equation is constructed based on the failure mechanism of the optoelectronic device in a physical field. Failure sample data is then obtained based on the basic sample data. Through this embodiment, the second-order partial differential equation is constructed based on the failure mechanism of the optoelectronic device in a physical field, and the basic sample data is obtained by combining the solution results. Failure sample data is then obtained based on the basic sample data, thus improving the data quality of the generated failure sample data. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating an embodiment of the method for generating failed sample data according to this application; Figure 2 A schematic diagram of a numerical solution method for the partial differential equations of failure in optoelectronic devices; Figure 3 A schematic diagram of the architecture of the system for generating failure sample data; Figure 4 This is a schematic diagram of the functional modules of an embodiment of the failure sample data generation device of this application; Figure 5 This is a schematic diagram of the hardware structure of the failure sample data generation device involved in the embodiments of this application. Detailed Implementation

[0016] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present application.

[0017] Optoelectronic devices, which convert optical signals into electrical signals, are crucial components of optical communication systems. During operation, various external factors, such as metal electromigration, ohmic contact degradation, hot carrier effects, chip breakage, and external lead corrosion, can cause failures, leading to parameter drift, functional degradation, and even device malfunction. Failure analysis of optoelectronic devices using data mining and large-scale artificial intelligence models to identify failure risks, determine the degree of failure, and proactively eliminate potential failure hazards to prevent major problems such as optical communication network outages caused by optoelectronic device failures has become an important research direction for intelligent network operation and maintenance. However, due to the lack of high-quality optoelectronic device failure sample data, there are currently no mature failure analysis systems or related application cases. Generating failure samples using data processing techniques remains a key research area in the industry.

[0018] In related technologies, electrical simulation methods construct geometric models and discretize them to obtain parameters of conductive paths and current-carrying components, establish stress-strain relationships and failure models, and generate data related to current, magnetic fields, and heat dissipation. This method can achieve coupled analysis of electromagnetic and temperature fields, but it cannot achieve incremental correction, resulting in relatively large sample errors. Finite element modeling and simulation methods establish failure problem regions, use variational methods to generate stable solutions, simulate electromigration voids, and predict migration failure locations. This method can analyze failure causes and establish failure models, but the analysis time is long, and it does not address the overfitting problem that may arise from failure sample data in artificial intelligence models. The Gummel iterative analysis method constructs a nonlinear equation system for the failure mechanism of optoelectronic devices, formally describing the relationships between various physical quantities, and approximates the solution of the equation system through repeated iterations. This method obtains functional expressions through parameter substitution of independent variables and iteratively solves until the termination condition is met, generating failure simulation data for devices. However, in the process of generating simulation data for certain types of optoelectronic devices, problems such as non-convergence of the iterative algorithm and computational complexity exist.

[0019] High-quality failure sample data for optoelectronic devices is a key element supporting the training and fine-tuning of large-scale failure analysis models and the intelligent operation and maintenance of optical communication networks. Traditional techniques suffer from problems such as large errors, long processing times, and non-convergence. They have not constructed efficient multi-physics coupled failure mechanism models, nor have they proposed optimization methods to improve the accuracy of sample data. Currently, there is an urgent need to propose new technologies and methods to generate high-quality failure sample data.

[0020] This application proposes a method for generating failure sample data of optoelectronic devices, including three specific implementation methods: basic sample data generation, data increment correction, and sample feature adjustment. In using these three methods, the basic sample data generation method must first be called to construct a multiphysics failure mechanism model and perform numerical solutions to obtain basic sample data. Then, either the data increment correction method or the sample feature adjustment method is called to improve the quality of the basic sample data. The data increment correction method and the sample feature adjustment method can be called in any order; they can be called individually or in combination. Wherein: The basic sample data is generated as follows: This embodiment proposes a method for generating basic sample data based on multi-physics failure mechanism modeling. It deeply analyzes the failure mechanisms of optoelectronic devices in physical fields such as electric, temperature, and force fields, establishes second-order partial differential equations to describe the failure process of optoelectronic devices, obtains boundary values ​​through detection instruments, establishes a system of linear equations using the difference substitution method, obtains the numerical solution of the second-order partial differential equations, uses the numerical solution as output data, and uses the values ​​of the independent variables of each physical field as input data. The output and input data are used as the basic sample data for optoelectronic device failure. This invention uses symbols... This represents the dependent variable, describing the remaining usable time of the optoelectronic device, in weeks (7 days). It is represented by the symbol... , , ..., The independent variables of each physical field are represented, such as bias current, temperature, and impact force, with units of milliamperes, degrees Celsius, and newtons. This embodiment analyzes in depth the physical processes of semiconductor materials and optical chips in optoelectronic devices under multiple physical fields, such as metal electromigration, ohmic contact degradation, secondary breakdown, and chip breakage. A second-order partial differential equation describing the failure mechanism of optoelectronic devices is proposed, as shown in formula (1):

[0021] In the above formula (1), This represents the weight of the second-order partial derivative term. The weight of the first-order partial derivative term is indicated by the equation (1). It should be noted that the forms of the second-order partial differential equations for optoelectronic devices made of different semiconductor materials are all as shown in equation (1), but... as well as The value varies depending on the semiconductor material. It is a polynomial representing the calculation function of the remaining available time after the combined effect of various external operation and maintenance factors during the operation of optoelectronic devices; in formula (1), the two terms on the left side of the equation describe the intrinsic mechanism of multiphysics failure, and the right side is the external factor, which is considered as an external source; the independent variable of formula (1) The value of corresponds to the input data of the basic sample data of the optoelectronic device, and the dependent variable The value corresponds to the output data. Obtaining the analytical solution to the above equation is very difficult. In this embodiment, a numerical solution method is used to divide the high-dimensional space enclosed by the independent variables into discrete cells. Values ​​of the boundary cells are collected using a detection instrument, and the difference equation (1) is converted into a system of linear equations. The dependent variable at each cell node is then iteratively calculated. The numerical value. Specifically: This embodiment uses formula (2) to achieve the difference substitution of the second-order differential:

[0022] This embodiment uses formula (3) to achieve the difference substitution of the first-order differential:

[0023] In formulas (2) and (3) above, the symbols , , ..., Represents n independent variables in a multiphysics field. Indicates the first The nth independent variable. The nth... After dividing the independent variables into equal parts, the length of each part is indicated by the symbol. In other words, in a multiphysics context, the first... Each independent variable along its own dimension The dependent variables corresponding to the next, itself, and the previous are respectively , , .

[0024] The numerical values ​​at the boundary of the second-order partial differential equation of the multiphysics field are collected using a detection instrument. The values ​​are then substituted into the above formulas (2) and (3) and rearranged to obtain the linear equation system. Symbols in the system of equations Indicates the dependent variable that needs to be solved numerically, symbol This indicates the result obtained after performing linear operations using formulas (2) and (3) and the numerical values ​​at the boundary. Solving the above system of linear equations yields the numerical solution of the dependent variable of the second-order partial differential equation.

[0025] The data incremental correction method is as follows: This embodiment proposes an incremental data correction method that uses state data reported during the actual operation of optoelectronic devices to correct the basic sample data obtained by numerically solving second-order partial differential equations. Based on the failure mechanism of optoelectronic devices, second-order partial differential equations are constructed to generate basic sample data, which theoretically describes the quantitative relationships between independent and dependent variables of the optoelectronic device in various physical fields. The optoelectronic device is packaged and integrated into communication equipment, deployed in a data center, powered on, and activated for service, entering the operational state. During operation, due to various external factors such as electrostatic discharge, data center temperature, and collisions, the quantitative relationships between dependent and independent variables may change in certain regions of the high-dimensional space formed by the independent variables of multiple physical fields. This embodiment proposes an incremental correction method that uses the state data reported by the optoelectronic device to calculate the deviation between the theoretical and actual values ​​of the dependent variable, adding the deviation to the dependent variable in the positive direction along the high-dimensional space, thus achieving incremental correction of the basic sample data. This embodiment utilizes symbolic... This represents the vector of deviations between the theoretical and actual values ​​of the dependent variable, with the first deviation being... , No. The deviation is The deviation vector and calculation method are shown in formula (4).

[0026]

[0027] In formula (4), the deviation is The serial number in With the right In These can be different values, because the serial number... This corresponds to the ordinal number in the one-dimensional deviation vector, and the ordinal number... This corresponds to the sequence number in the higher-dimensional space. In formula (4), This represents the actual value of the dependent variable. This represents the theoretical value of the dependent variable.

[0028] In the actual operation of optoelectronic devices, errors in numerical accuracy can occur due to the acquisition and reporting of status data. Furthermore, incremental corrections for even small deviations would be a significant waste of computational resources. Therefore, this embodiment sets a threshold. Calculate within a set period The deviation between the theoretical and actual values ​​is used to obtain the deviation vector. If the number of elements in the vector that exceed the threshold is greater than the set value. If so, then an incremental correction of the base sample size is performed. When performing the incremental correction task, the bias vector is first calculated. The average value is then added to the current theoretical value along the positive direction in the high-dimensional space formed by the multi-physics independent variables, as shown in formula (5).

[0029]

[0030] In formula (5), the symbol The average value of the deviation vector is represented by the symbol. The corresponding current deviation vector in high-dimensional space The index of the forward grid point with the largest index element. For example, in electric field, temperature field, and force field, bias current is selected as the first dimension, temperature as the second dimension, collision force as the third dimension, and the remaining available time of optoelectronic device as the dependent variable. For example, the high-dimensional space elements corresponding to the current actual value of the dependent variable are... , , Then, in the process of incremental correction of sample data using formula (5), the values ​​of elements with indices greater than 4 in the first dimension, greater than 5 in the second dimension, and greater than 5 in the third dimension are added to the average value of the deviation vector. .

[0031] The sample feature adjustment method is as follows: This embodiment proposes a sample feature adjustment method. It uses failure sample data as the source training set and adjusts the dependent variable of the sample data using the statistical features of the target training set. This addresses the problem of artificial intelligence models fitting well on the source training set but poorly on the target training set. For example, lasers can use gallium arsenide (GaAs) or indium phosphide (IPT) as semiconductor materials. Suppose that for a GaAs-based laser, failure sample data is generated using a basic sample data generation method or a basic sample data generation method combined with data increment correction, and then used to train a large-scale artificial intelligence model. If this model is directly applied to an IPT-based laser for failure analysis, the error may be significant. This is because GaAs and IPT have different decomposition temperatures, melting points, and thermal conductivity. Under the same temperature field, their failure responses, such as evaporation and thermal damage, differ. Therefore, it is necessary to adjust the source training set using some statistical features of the target training set. This embodiment utilizes symbolic... This represents the difference in statistical features between the source training set and the target training set. The statistical feature calculation function is: In this embodiment, the statistical feature calculation function refers to the expected value, variance, and other functions of the training set. According to the previous description, the formula for calculating the statistical feature difference between the source training set and the target training set is shown in (6).

[0032]

[0033] In formula (6), the function Calculate the statistical characteristics of the target training set, function Calculate the statistical features of the source training set, and subtract the two calculation results to obtain the statistical feature difference. The feature adjustment vector is calculated using this difference, and then added to the source training set to obtain the adjusted failure sample data. The calculation formula is shown in (7).

[0034]

[0035] In formula (7), the symbol This represents the feature adjustment vector, which is processed by a function. This indicates that the adjustment vector is obtained. For example, the adjustment vector is calculated by the expected value of the dependent variables of the target training set and the source training set. The right side of formula (7) represents the first step of adjusting the feature vector. Each element value and the dependent variable element in the source training set The features of the element are adjusted by adding them together. In actual operation, the sample feature adjustment method proposed in this embodiment selects features such as expectation and variance based on the failure sample data of optoelectronic devices, and chooses an appropriate number of features as the number of elements in the feature adjustment vector based on the computing power provided by the current operating environment, thus calculating the feature adjustment vector. Feature adjustment is performed on the training set of failure sample sources for optoelectronic devices.

[0036] This embodiment generates optoelectronic device failure sample data based on the basic sample data generation method, data increment correction method, and sample feature adjustment method described above. Overall, compared with the prior art, the technical solution conceived in this embodiment has the following beneficial effects: This embodiment proposes a basic data generation method, which uses a multi-physics coupling modeling method to generate failure sample data, solving the problem of large errors in single-physics failure samples. Furthermore, it uses a numerical solution method to calculate the solution set of the second-order partial differential equation of the failure process of optoelectronic devices, thus solving the problem of difficulty in obtaining analytical solutions.

[0037] This embodiment proposes a data incremental correction method, which uses the status data reported during the operation of optoelectronic devices to incrementally correct the basic data of failure samples. By collecting data from the current network operation and optimizing the mode of failure samples in a timely manner, the quality of failure sample data of optoelectronic devices can be improved.

[0038] This embodiment proposes a sample feature adjustment method, which uses the difference in statistical features between the target training set and the source training set to adjust the dependent variable of the sample basic data. This solves the problem that the failure data of optoelectronic devices fits well on one type of sample training set but poorly on another type of training set.

[0039] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0040] In a first aspect, embodiments of this application provide a method for generating failure sample data.

[0041] In one embodiment, reference is made to Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the method for generating failed sample data according to this application. Figure 1 As shown, the methods for generating failure sample data include: Step S10: Based on the solution results of the second-order partial differential equation, basic sample data is obtained. The basic sample data includes multiple sets of first data pairs. Each set of first data pairs includes a set of input data of the second-order partial differential equation and its corresponding numerical solution. Each set of input data is a physical quantity that affects the remaining usable lifetime of the optoelectronic device. The numerical solution corresponding to each set of input data is the theoretical remaining usable lifetime of the optoelectronic device. The second-order partial differential equation is constructed based on the failure mechanism of the optoelectronic device in the physical field. In this embodiment, physical fields with three dimensions—electric field, temperature field, and force field—are used as examples for illustration. Generating basic sample data includes the following steps: S11. Establish a three-dimensional physical field space model of the electric field, temperature field, and force field of optoelectronic device failure, and determine the value range of the three physical fields. A multiphysics coupling failure model was constructed by selecting three representative physical fields: electric field, temperature field, and force field. Taking a laser as an example, the characteristic index of the electric field is the bias current, the characteristic index of the temperature field is the temperature reported by the laser during operation, and the characteristic index of the force field is the impact force experienced by the laser during the insertion and removal of the optical module. The units of measurement for bias current, temperature, and impact force are milliamperes, degrees Celsius, and newtons, respectively. Based on the relevant maintenance data of the coherent optical module laser on the line side of the optical transmission equipment, the value ranges of the above three characteristic indices are determined as shown in Table 1.

[0042] Table 1

[0043] S12, establish the second-order partial differential equation of coupling failure of optoelectronic devices in three physical field environments: electric field, temperature field and force field, perform network partitioning on the high-dimensional space model of the three-dimensional physical field, and create a three-dimensional discrete mesh model. Using bias current, laser temperature, and impact force on the laser as independent variables. , , Based on the remaining usable time of the laser As the dependent variable, a second-order partial differential equation for the coupling failure of the laser in the three-dimensional physical fields of electric field, temperature field and force field is established, as shown in Equation (8).

[0044]

[0045] It should be noted that Equation (1) is a general formula applicable to multiphysics failure coupling modeling of all optoelectronic devices. Equation (8) above is a special case of Equation (1), used to illustrate the quantitative relationship between the remaining usable time of the laser in the three physical fields of electric field, temperature field, and force field and the bias current, laser temperature, and collision force experienced by the laser. In practical applications, based on the failure mechanism of each optoelectronic device, a second-order partial differential equation applicable to each device can be derived from Equation (1).

[0046] Reference Figure 2 , Figure 2 This diagram illustrates a numerical solution method for the partial differential equations governing the failure of optoelectronic devices. (Example:) Figure 2 As shown, in this embodiment, the value ranges of the three independent variables—bias current, laser temperature, and impact force on the laser—are divided into three equal parts. The bias current ranges from 20 mA to 350 mA, and after division, the discrete grid increment is 110 mA. The laser temperature ranges from 35°C to 65°C, and after division, the discrete grid increment is 10°C. The impact force on the laser ranges from 100 N to 1600 N, and after division, the discrete grid increment is 500 N. , , This represents the step size of three dimensions: bias current, laser temperature, and impact force experienced by the laser. Based on the above description, the equation can be obtained. , , .

[0047] S13, calculate the boundary values ​​of the three-dimensional discrete mesh model, replace the differential terms in the second-order partial differential equations with finite differences, and then rearrange to obtain a system of linear equations; The parameters of the boundary points of a three-dimensional discrete mesh model are collected by detection instruments, and the values ​​of the dependent variables of the second-order partial differential equation at the boundary points are obtained through processing and calculation. Currently, the technology and methods for detecting the operating status of optoelectronic devices in electric, temperature, and force fields using relevant instruments are relatively mature. For example, a high-temperature aging chamber can be used to detect the changes of optoelectronic devices in the temperature field through accelerated aging and other means, obtain the minimum and maximum values ​​of the high-temperature aging chamber during the test, and calculate the boundary point values. Figure 2 This is an example of boundary points in a 3D discrete mesh model, containing a total of eight boundary points: 362, 213, 120, 38, 126, 278, 510, and 630. Figure 2These eight points correspond to the black dots around the rectangle in the lower right corner, while the four circles in the middle of the matrix represent the dependent variable of the remaining usable time of the optoelectronic device that needs to be solved, i.e.: , , , Using formulas (2) and (3), the second and first differentials are replaced by differences, and the system of linear equations is obtained after simplification, as shown in formula (9).

[0048]

[0049] S14. Solve the linear equation system to obtain the numerical solution of the second-order partial differential equation. Use the values ​​of each dimension of the three-dimensional discrete grid model as input data and the numerical solution as output data to obtain the basic sample data. Solving formula (9) yields Using the values ​​of each dimension of the three-dimensional discrete mesh model as input data and the numerical solution as output data, the basic sample data is obtained as shown in (10).

[0050] In formula (10), the symbol This matrix represents the basic data for optoelectronic device failure samples. The right side of the equals sign is a 4x10 matrix. The first row corresponds to the bias current of the 3D discrete mesh model, the second row to the laser temperature, the third row to the impact force experienced by the laser, and the fourth row to the remaining usable time of the optoelectronic device. This matrix represents the basic sample data for the 3D discrete mesh model when the laser experiences an impact force of 100 Newtons. Similarly, using the four steps described above, data can be obtained when the laser experiences impact forces of 600 Newtons, 1100 Newtons, and 1600 Newtons. This constitutes the basic sample data for optoelectronic device failure due to the coupling of physical fields in the electric, temperature, and force fields.

[0051] Step S20: Obtain failed sample data based on the basic sample data.

[0052] In this embodiment, as shown above, after obtaining the basic sample data, the basic sample data can be further processed using the data increment correction method and the sample feature adjustment method to obtain the final failed sample data.

[0053] In this embodiment, basic sample data is obtained based on the solution results of the second-order partial differential equation. The basic sample data includes multiple sets of first data pairs. Each set of first data pairs includes a set of input data for the second-order partial differential equation and its corresponding numerical solution. Each set of input data represents a physical quantity affecting the remaining usable lifetime of the optoelectronic device, and the numerical solution corresponding to each set of input data represents the theoretical remaining usable lifetime of the optoelectronic device. The second-order partial differential equation is constructed based on the failure mechanism of the optoelectronic device in a physical field. Failure sample data is then obtained based on the basic sample data. Through this embodiment, the second-order partial differential equation is constructed based on the failure mechanism of the optoelectronic device in a physical field, and the basic sample data is obtained by combining the solution results. Failure sample data is then obtained based on the basic sample data, thus improving the data quality of the generated failure sample data.

[0054] Further, in one embodiment, step S20 includes: Acquire actual sample data and check whether the difference between the actual sample data and the basic sample data meets the correction conditions. The actual sample data includes multiple sets of second data pairs. Each set of second data pairs includes a physical quantity that affects the remaining usable lifetime of the optoelectronic device and the corresponding actual remaining usable lifetime of the optoelectronic device. If the correction conditions are met, the basic sample data is corrected to obtain corrected sample data. Failure sample data is obtained based on the corrected sample data.

[0055] In this embodiment, actual measurement data of the optoelectronic device is collected. This data includes two categories: one is the status data generated during the operation of the optoelectronic device, which is reported by the device to the network management system via a single disk; the second is the data collected by the test personnel using instruments. For example, in one scenario, one day constitutes a collection cycle, with data collected every 288 minutes, and five measurement data points collected per cycle. The actual sample data constructed based on the actual measurement data is as follows:

[0056] in, Each column represents a second set of data pairs. The first three rows of each column represent the physical quantities affecting the remaining usable lifetime of the optoelectronic device, and the fourth row represents the actual remaining usable lifetime of the corresponding optoelectronic device. For example, the first column represents bias current, laser temperature, and laser impact force values ​​of 310 mA, 51 degrees Celsius, and 128 N, respectively, corresponding to an actual remaining usable lifetime of 276 weeks for the optoelectronic device.

[0057] Search for related data in the basic sample data. The corresponding data pairs are identified, and the difference between them is determined. This difference is a set of real numbers, each representing the difference between the actual test data of the remaining usable time of the optoelectronic device and the theoretical data generated using second-order partial differential equations.

[0058] Further, in one embodiment, the step of acquiring actual sample data and detecting whether the difference between the actual sample data and the basic sample data meets the correction condition includes: Obtain actual sample data, and for each second data pair in the actual sample data, determine the first data pair corresponding to each second data pair from the basic sample data. Each second data pair and its corresponding first data pair contain the same physical quantity that affects the remaining usable lifetime of the optoelectronic device. Calculate the difference between the actual remaining usable lifetime of the optoelectronic devices included in each second data pair and the theoretical remaining usable lifetime of the optoelectronic devices included in the corresponding first data pair. If the number of differences in the remaining available lifetime of the target is greater than the preset number, then the difference between the actual sample data and the basic sample data is determined to meet the correction condition, wherein the difference in the remaining available lifetime of the target is greater than or equal to the preset difference.

[0059] In this embodiment, it is assumed that The combined representation of the first data pair corresponding to the second data pair in each group is as follows:

[0060] The difference between the actual remaining usable lifetime of the optoelectronic devices included in each set of second data pairs and the theoretical remaining usable lifetime of the optoelectronic devices included in the corresponding set of first data pairs is: .

[0061] If the preset difference value is set to 3 and the preset number is 3 according to the application scenario, then the number of target remaining usable lifetime differences greater than or equal to 3 is 4, which is greater than the preset number. Therefore, it is determined that the difference between the actual sample data and the basic sample data meets the correction condition.

[0062] Furthermore, in one embodiment, the basic sample data is corrected to obtain corrected sample data, including: A correction value is obtained by combining the differences in remaining available lifetime of all targets; the theoretical remaining available lifetime of the optoelectronic devices included in the target first data corresponding to the difference in remaining available lifetime of each target in the basic sample data is increased by the correction value.

[0063] In this embodiment, in conjunction with the description of the above embodiments, the remaining usable lifetime difference is... Among them, the difference in target remaining usable lifespan that is greater than the preset difference is... The average value is calculated as a correction value of 4. Then, the theoretical remaining usable lifetime of the optoelectronic devices included in the target first data pair is increased by the correction value of 4. The target first data pair corresponding to each difference in the target remaining usable lifetime in the base sample data is... In columns 2 through 5 of the original sample data, the other first data pairs remain unchanged, thus obtaining the revised sample data. Among them, for... The modified example for columns 2 through 5 is as follows:

[0064] Further, in one embodiment, step S20 includes: Based on the difference between the material properties of the optoelectronic device to be analyzed in the failure detection scenario and the material properties of the optoelectronic device, the revised sample data is adjusted to obtain failure sample data.

[0065] In this embodiment, the failure sample data is used for model training, so that the trained failure detection model can be applied to failure detection scenarios to predict the remaining usable lifetime of the optoelectronic device under analysis. Assuming that the semiconductor material of the optoelectronic device under analysis in the failure detection scenario is different from the semiconductor material of the optoelectronic device on which the revised sample data is based, directly using the revised sample data for model training would result in insufficient prediction accuracy of the trained failure detection model. Therefore, it is necessary to adjust the revised sample data based on the differences in material properties to obtain the final failure sample data.

[0066] Further, in one embodiment, based on the characteristic difference between the characteristic value of the theoretical remaining usable lifetime of the optoelectronic device under the same physical field characteristic index and the characteristic value of the theoretical remaining usable lifetime of the optoelectronic device, which characterizes the difference between the material properties of the optoelectronic device under the failure detection scenario and the material properties of the optoelectronic device, step S20 includes: Based on the characteristic difference and the operation and maintenance statistics of the optical communication network, the mean and variance parameters of the probability distribution are calculated; using the mean and variance parameters as the expected value and variance of the normal distribution, the adjustment amount of the theoretical remaining usable lifetime of each optoelectronic device included in the revised sample data is obtained; the theoretical remaining usable lifetime of each optoelectronic device included in the revised sample data is increased by the corresponding adjustment amount to obtain the failure sample data.

[0067] In this embodiment, for ease of explanation, it is assumed that the revised sample data is obtained based on a laser made of gallium arsenide semiconductor material, as shown in the following example:

[0068] The first row of formula (12) represents the bias current, the second row represents the temperature of the laser, the third row represents the impact force on the laser, and the fourth row represents the theoretical remaining usable lifetime of the laser. The average value of the fourth row is used as the characteristic value, that is, the characteristic value of the theoretical remaining usable lifetime of the optoelectronic device is 271.

[0069] Assuming the optoelectronic device to be analyzed is a laser based on indium phosphide semiconductor material, according to the operation and maintenance statistics of optical communication networks and the physicochemical properties of the two semiconductor materials, the decomposition temperature and thermal conductivity of indium phosphide are higher than those of gallium arsenide. Under the same physical field characteristic indicators, the characteristic value of the theoretical remaining usable lifetime of the laser based on indium phosphide semiconductor material is 279.

[0070] Based on the description above, the feature difference is... =279 271 = 8.

[0071] Based on feature difference ( =8), referring to the operation and maintenance statistics of optical communication network, the mean parameter of the probability distribution is calculated to be 8 and the variance parameter is 1; taking these two parameters as the mathematical expectation and variance of the normal distribution, the adjustment amounts of the five sample points in formula (12) are calculated as (7, 8, 9, 9, 7).

[0072] The final failure sample data obtained is as follows:

[0073] In summary, we can obtain failure sample data that is suitable for failure detection scenarios.

[0074] Further, in one embodiment, step S20 includes: Based on the difference between the material properties of the optoelectronic device to be analyzed in the failure detection scenario and the material properties of the optoelectronic device, the basic sample data is adjusted to obtain failure sample data.

[0075] In this embodiment, as described above, the data incremental correction method and the sample feature adjustment method are not called in any particular order; they can be called individually or in combination. That is, the sample feature adjustment method can be directly applied to the basic sample data to obtain the invalid sample data. Specific embodiments can be found in the above description of using the sample feature adjustment method on the corrected sample data, and will not be repeated here.

[0076] Furthermore, in one embodiment, reference is made to Figure 3 , Figure 3 A schematic diagram of the system architecture for generating failure sample data. (Example) Figure 3 As shown, the system comprises three modules: a basic sample data generation module (executes the basic sample data generation method), a data increment correction module (executes the data increment correction method), and a sample feature adjustment module (executes the sample feature adjustment method). These three modules are located in... Figure 3 Within the dashed rectangle in the middle, the task of generating failure sample data is executed.

[0077] Figure 3The top left corner shows the physical field failure mechanism acquisition process of the failure sample data generation system. The failure mechanisms of electric field, temperature field, and force field are first described in the form of high-order partial differential equations, and then uniformly converted into the multi-physics field second-order partial differential equations proposed in this patent. Through algorithm design and software coding process, it finally becomes the basic data generation module. This module performs numerical solution calculation tasks and uses boundary values ​​and differences to obtain the basic data of optoelectronic device failure samples.

[0078] exist Figure 3 In the middle left section, the failure sample data generation system collects engineering operation data of optoelectronic devices or uses instruments to test the devices to obtain actual measurement data. This data is then input into the incremental correction module to calculate the difference between the basic data and the actual measurement data, and to perform incremental correction on the dependent variable of the failure sample.

[0079] Figure 3 The bottom left corner shows the sample feature adjustment process of the failure sample data generation system. The adjustment vector is calculated using the difference between the sample features of the source training set and the target training set, and then added to the dependent variable of the dataset to obtain the adjusted failure sample data.

[0080] exist Figure 3 In the upper right corner, the failure sample data generation system uses second-order partial differential equations of multiphysics to create basic data and saves it to the database. The samples generated by this process are the most basic data. In practical engineering applications, the data increment correction module or sample feature adjustment module can be flexibly called to improve the quality of the failure sample data. The corrected or adjusted data obtained using these two modules is then updated to the dependent variables corresponding to the multiphysics characteristic indicators in the failure sample database, such as... Figure 3 The two processes are shown in the middle and bottom right corners.

[0081] The failure sample data generation system provided in this embodiment can operate as a standalone system or as a subsystem of the optical communication network management system. During operation, the failure sample data generation system collects optoelectronic device engineering operation data through an external interface, imports instrument detection data and calculated target training set feature data, and utilizes this data... Figure 3 The three functional modules shown in the dashed rectangle in the middle generate high-quality failure sample data of optoelectronic devices. This data is called by an external intelligent operation and maintenance system to realize the training or fine-tuning of the large failure model of optoelectronic devices.

[0082] Secondly, embodiments of this application also provide a device for generating failure sample data.

[0083] In one embodiment, reference is made to Figure 4 , Figure 4This is a schematic diagram of the functional modules of an embodiment of the failed sample data generation device of this application. Figure 4 As shown, the failure sample data generation device includes: The first construction module 10 is used to obtain basic sample data based on the solution results of the second-order partial differential equation. The basic sample data includes multiple sets of first data pairs. Each set of first data pairs includes a set of input data of the second-order partial differential equation and its corresponding numerical solution. Each set of input data is a physical quantity that affects the remaining usable lifetime of the optoelectronic device. The numerical solution corresponding to each set of input data is the theoretical remaining usable lifetime of the optoelectronic device. The second-order partial differential equation is constructed based on the failure mechanism of the optoelectronic device in the physical field. The second construction module 20 is used to obtain failed sample data based on the basic sample data.

[0084] Furthermore, in one embodiment, the second building module 20 is used for Acquire actual sample data and detect whether the difference between the actual sample data and the basic sample data meets the correction conditions. The actual sample data includes multiple sets of second data pairs. Each set of second data pairs includes a physical quantity that affects the remaining usable lifetime of the optoelectronic device and its corresponding actual remaining usable lifetime of the optoelectronic device. If the correction conditions are met, the basic sample data is corrected to obtain the corrected sample data. Failed sample data is obtained based on the revised sample data.

[0085] Furthermore, in one embodiment, the second building module 20 is used for: Obtain actual sample data, and for each second data pair in the actual sample data, determine the first data pair corresponding to each second data pair from the basic sample data. Each second data pair and its corresponding first data pair contain the same physical quantity that affects the remaining usable lifetime of the optoelectronic device. Calculate the difference between the actual remaining usable lifetime of the optoelectronic devices included in each second data pair and the theoretical remaining usable lifetime of the optoelectronic devices included in the corresponding first data pair. If the number of differences in the remaining available lifetime of the target is greater than the preset number, then the difference between the actual sample data and the basic sample data is determined to meet the correction condition, wherein the difference in the remaining available lifetime of the target is greater than or equal to the preset difference.

[0086] Furthermore, in one embodiment, the second building module 20 is used for: The correction value is obtained by combining the differences in the remaining available lifetime of all targets; The theoretical remaining available lifetime of the optoelectronic devices included in the target first data corresponding to the difference in the remaining available lifetime of each target in the basic sample data is increased by the correction value.

[0087] Furthermore, in one embodiment, the second building module 20 is used for: Based on the difference between the material properties of the optoelectronic device to be analyzed in the failure detection scenario and the material properties of the optoelectronic device, the revised sample data is adjusted to obtain failure sample data.

[0088] Further, in one embodiment, based on the characteristic difference between the characteristic value of the theoretical remaining usable lifetime of the optoelectronic device under the same physical field characteristic index and the characteristic value of the theoretical remaining usable lifetime of the optoelectronic device, which characterizes the difference between the material properties of the optoelectronic device targeted by the failure detection scenario and the material properties of the optoelectronic device, the second construction module 20 is used to: Based on the characteristic differences and the operation and maintenance statistics of the optical communication network, the mean and variance parameters of the probability distribution are calculated. Using the mean parameter and variance parameter as the mathematical expectation and variance of the normal distribution, the adjustment amount of the theoretical remaining usable lifetime of each optoelectronic device included in the revised sample data is obtained; The theoretical remaining usable lifetime of each optoelectronic device included in the revised sample data is increased by a corresponding adjustment amount to obtain the failure sample data.

[0089] Furthermore, in one embodiment, the second building module 20 is used for Based on the difference between the material properties of the optoelectronic device to be analyzed in the failure detection scenario and the material properties of the optoelectronic device, the basic sample data is adjusted to obtain failure sample data.

[0090] The functions of each module in the above-mentioned failure sample data generation device correspond to the steps in the above-mentioned failure sample data generation method embodiment, and their functions and implementation processes will not be described in detail here.

[0091] Thirdly, embodiments of this application provide a failure sample data generation device, which can be a personal computer (PC), laptop computer, server, or other device with data processing capabilities.

[0092] Reference Figure 5 , Figure 5 This is a schematic diagram of the hardware structure of the failure sample data generation device involved in the embodiments of this application. In the embodiments of this application, the failure sample data generation device may include a processor, a memory, a communication interface, and a communication bus.

[0093] The communication bus can be of any type and is used to interconnect the processor, memory, and communication interface.

[0094] The communication interface includes input / output (I / O) interfaces, physical interfaces, and logical interfaces used for interconnecting devices within the failure sample data generation device, as well as interfaces used for interconnecting the failure sample data generation device with other devices (such as other computing devices or user equipment). Physical interfaces can be Ethernet interfaces, fiber optic interfaces, ATM interfaces, etc.; user equipment can be displays, keyboards, etc.

[0095] Memory can be various types of storage media, such as random access memory (RAM), read-only memory (ROM), non-volatile RAM (NVRAM), flash memory, optical storage, hard disk, programmable ROM (PROM), erasable PROM (EPROM), electrically erasable PROM (EEPROM), etc.

[0096] The processor can be a general-purpose processor, which can call the failure sample data generation program stored in memory and execute the failure sample data generation method provided in the embodiments of this application. For example, the general-purpose processor can be a central processing unit (CPU). The method executed when the failure sample data generation program is called can be referred to in various embodiments of the failure sample data generation method of this application, and will not be repeated here.

[0097] Those skilled in the art will understand that Figure 5 The hardware structure shown does not constitute a limitation of this application and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0098] Fourthly, embodiments of this application also provide a computer-readable storage medium.

[0099] The present application has a computer-readable storage medium storing a failure sample data generation program, wherein when the failure sample data generation program is executed by a processor, it implements the steps of the failure sample data generation method described above.

[0100] The method implemented when the failure sample data generation program is executed can be referred to in various embodiments of the failure sample data generation method of this application, and will not be repeated here.

[0101] It should be noted that the sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0102] The terms "comprising" and "having," and any variations thereof, in the specification, claims, and accompanying drawings of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to such process, method, product, or apparatus. The terms "first," "second," and "third," etc., are used to distinguish different objects, etc., and do not indicate a sequence, nor do they limit "first," "second," and "third" to different types.

[0103] In the description of the embodiments of this application, terms such as "exemplary," "for example," or "for instance" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as "exemplary," "for example," or "for instance" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of terms such as "exemplary," "for example," or "for instance" is intended to present the relevant concepts in a concrete manner.

[0104] In the description of the embodiments of this application, unless otherwise stated, " / " means "or". For example, A / B can mean A or B. The "and / or" in the text is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist simultaneously, and B exists alone. In addition, in the description of the embodiments of this application, "multiple" means two or more.

[0105] In some processes described in the embodiments of this application, multiple operations or steps are included in a specific order. However, it should be understood that these operations or steps may not be executed in the order they appear in the embodiments of this application, or they may be executed in parallel. The sequence number of the operation is only used to distinguish different operations, and the sequence number itself does not represent any execution order. In addition, these processes may include more or fewer operations, and these operations or steps may be executed sequentially or in parallel, and these operations or steps may be combined.

[0106] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device to execute the methods described in the various embodiments of this application.

[0107] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A method for generating failure sample data, characterized in that, The method for generating failure sample data includes: Based on the solution results of the second-order partial differential equation, basic sample data is obtained. The basic sample data includes multiple sets of first data pairs. Each set of first data pairs includes a set of input data of the second-order partial differential equation and its corresponding numerical solution. Each set of input data is a physical quantity that affects the remaining usable lifetime of the optoelectronic device. The numerical solution corresponding to each set of input data is the theoretical remaining usable lifetime of the optoelectronic device. The second-order partial differential equation is constructed based on the failure mechanism of the optoelectronic device in the physical field. Failure sample data is obtained based on the basic sample data.

2. The method for generating failure sample data as described in claim 1, characterized in that, The failure sample data obtained based on the aforementioned basic sample data includes: Acquire actual sample data and detect whether the difference between the actual sample data and the basic sample data meets the correction conditions. The actual sample data includes multiple sets of second data pairs. Each set of second data pairs includes a physical quantity that affects the remaining usable lifetime of the optoelectronic device and its corresponding actual remaining usable lifetime of the optoelectronic device. If the correction conditions are met, the basic sample data is corrected to obtain the corrected sample data. Failed sample data is obtained based on the revised sample data.

3. The method for generating failure sample data as described in claim 2, characterized in that, The step of obtaining actual sample data and detecting whether the difference between the actual sample data and the basic sample data meets the correction conditions includes: Obtain actual sample data, and for each second data pair in the actual sample data, determine the first data pair corresponding to each second data pair from the basic sample data. Each second data pair and its corresponding first data pair contain the same physical quantity that affects the remaining usable lifetime of the optoelectronic device. Calculate the difference between the actual remaining usable lifetime of the optoelectronic devices included in each second data pair and the theoretical remaining usable lifetime of the optoelectronic devices included in the corresponding first data pair. If the number of differences in the remaining available lifetime of the target is greater than the preset number, then the difference between the actual sample data and the basic sample data is determined to meet the correction condition, wherein the difference in the remaining available lifetime of the target is greater than or equal to the preset difference.

4. The method for generating failure sample data as described in claim 3, characterized in that, The basic sample data is corrected to obtain the corrected sample data, which includes: The correction value is obtained by combining the differences in the remaining available lifetime of all targets; The theoretical remaining available lifetime of the optoelectronic devices included in the target first data corresponding to the difference in the remaining available lifetime of each target in the basic sample data is increased by the correction value.

5. The method for generating failure sample data as described in claim 2, characterized in that, The process of obtaining the failed sample data based on the revised sample data includes: Based on the difference between the material properties of the optoelectronic device to be analyzed in the failure detection scenario and the material properties of the optoelectronic device, the revised sample data is adjusted to obtain failure sample data.

6. The method for generating failure sample data as described in claim 5, characterized in that, Based on the characteristic difference between the characteristic value of the theoretical remaining usable lifetime of the optoelectronic device under the same physical field characteristic index and the characteristic value of the theoretical remaining usable lifetime of the optoelectronic device, the difference between the material properties of the optoelectronic device under the failure detection scenario and the material properties of the optoelectronic device is characterized. The revised sample data is adjusted based on the difference between the material properties of the optoelectronic device under the failure detection scenario and the material properties of the optoelectronic device, resulting in failure sample data including: Based on the characteristic differences and the operation and maintenance statistics of the optical communication network, the mean and variance parameters of the probability distribution are calculated. Using the mean parameter and variance parameter as the mathematical expectation and variance of the normal distribution, the adjustment amount of the theoretical remaining usable lifetime of each optoelectronic device included in the revised sample data is obtained; The theoretical remaining usable lifetime of each optoelectronic device included in the revised sample data is increased by a corresponding adjustment amount to obtain the failure sample data.

7. The method for generating failure sample data as described in claim 1, characterized in that, The failure sample data obtained based on the aforementioned basic sample data includes: Based on the difference between the material properties of the optoelectronic device to be analyzed in the failure detection scenario and the material properties of the optoelectronic device, the basic sample data is adjusted to obtain failure sample data.

8. A device for generating failure sample data, characterized in that, The failure sample data generation device includes: The first construction module is used to obtain basic sample data based on the solution results of the second-order partial differential equation. The basic sample data includes multiple sets of first data pairs. Each set of first data pairs includes a set of input data of the second-order partial differential equation and its corresponding numerical solution. Each set of input data is a physical quantity that affects the remaining usable lifetime of the optoelectronic device. The numerical solution corresponding to each set of input data is the theoretical remaining usable lifetime of the optoelectronic device. The second-order partial differential equation is constructed based on the failure mechanism of the optoelectronic device in the physical field. The second construction module is used to obtain failed sample data based on the basic sample data.

9. A device for generating failure sample data, characterized in that, The failure sample data generation device includes a processor, a memory, and a failure sample data generation program stored in the memory and executable by the processor, wherein when the failure sample data generation program is executed by the processor, it implements the steps of the failure sample data generation method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a failure sample data generation program, wherein when the failure sample data generation program is executed by a processor, it implements the steps of the failure sample data generation method as described in any one of claims 1 to 7.

Images