Avionics product reliability test work performance test case automatic generation method

Abstract

The invention relates to the field of system reliability, and discloses an avionics product reliability test work performance test case automatic generation method comprising the following steps: S1, determining product work performance test items; s2, determining importance of product work performance test items; s3, generating a product work performance test path; s4, generating product work performance test excitation; s5, generating a product work performance test criterion; s6, determining a product work performance test opportunity; and S7, automatically generating a success performance test case. According to the invention, different coverage requirements are given according to the importance of the work performance test items to generate test paths and different value requirements to generate test excitation, so that the automatically generated work performance test case can fully test sensitive key faults in a reliability test.

Description

Technical Field

[0001] This invention relates to the field of electronic product reliability testing, and specifically to a method for automatically generating performance test cases for reliability testing of avionics products. Background Technology

[0002] In reliability testing of airborne electronic products, multiple functional performance tests need to be conducted before, during, and after the test to determine whether the product has malfunctioned and to support the conclusion of the reliability test. The functional performance test requirements specify the methods and approaches for detecting whether the product functions normally. Each test requirement includes elements such as test stimulus, test path, test criteria, and test timing. Currently, in the actual process of reliability testing, the determination of test requirements mainly faces the following problems: First, the testing requirements in product reliability testing are basically the same as those in performance testing. They do not take into account the probability of different faults occurring under reliability testing conditions or the impact of different faults on the execution of tasks in the upper-level system. Consequently, different testing strategies are designed for test items of different associated faults, which may lead to the risk that some key and important faults may not be identified.

[0003] Second, the number of performance test paths is too small. Usually, only a single path under typical and important working conditions is selected, and some unconventional and multi-working-condition paths are not fully covered, resulting in insufficient test coverage.

[0004] Third, the performance test stimuli are usually selected based on experience, and boundary values ​​and outliers are not fully considered, resulting in insufficient test sufficiency. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides an automatic generation method for performance test cases in reliability testing of avionics products. It establishes a technical process for determining test requirements such as test paths and test stimuli based on importance analysis. The method proposed in this invention fully considers the characteristics of reliability testing conditions and fault identification requirements for airborne electronic products. It analyzes the importance of different test items from dimensions such as the probability of fault occurrence and severity level, and designs different test paths and stimulus generation strategies based on importance, achieving a more comprehensive and thorough identification of product faults under reliability testing conditions.

[0006] This application discloses a method for automatically generating test cases for reliability testing of avionics products, including the following steps: S1, Determine the product performance test items; Identify or import product performance test items. Each product performance test item has a corresponding state diagram or activity diagram, and each state diagram or activity diagram provides a set of main paths. S2, Determine the importance of product performance test items; The overall failure probability of the device is: ; in, The overall failure probability of the device. This represents the probability of temperature cycling fatigue failure. This represents the probability of failure due to high-temperature aging. This represents the probability of vibration fatigue failure. This represents the probability of electromigration failure. The importance of the performance test items is as follows: ; in, The importance of performance test items, In order to be with the first The overall device failure probability of the i-th device related to each function. For the first The dependence of each function on the i-th device. For the first The severity level of functional failure for each function. For the test item and the first Functional relevance; Indicates the total number of functions related to the test item; I indicates the number of functions related to the first test item. The total number of devices related to each function; The importance level of each performance test item is determined based on its importance. S3, Product performance test path generation; Based on the importance level of the power performance test items, different path coverage type requirements are adopted, and each power performance test path is generated based on the activity diagram or state diagram associated with the power performance test items in S1. S4, Product performance test stimulus generation; Based on the different importance levels of the test items obtained in S2, different value requirements are given for single variable values, multivariate combination coverage, and multivariate combination values ​​for the test stimuli. S5, Product performance test criteria generation; S6, the timing for product performance testing is determined; S7 automatically generates successful performance test cases; For each performance test item, test cases are generated based on the test path, stimulus, criteria, and test timing.

[0007] Preferably, the specific method for obtaining the temperature cycle fatigue failure probability in S2 is as follows: Based on the number of temperature cycles and the frequency of temperature cycles in the test profile, and the range of component junction temperature changes obtained from thermal simulation, the probability of temperature cycle fatigue failure is calculated using the following formula: ; In the formula, N is the number of temperature cycles, and f is the frequency of temperature cycles. This refers to the range of junction temperature variation of the device. , , These are three temperature cycling fatigue parameters, obtained by fitting the device's fatigue life experiment.

[0008] Preferably, the specific method for obtaining the high-temperature aging failure probability in S2 is as follows: Based on the duration of high temperatures in the experimental profile and the device junction temperature obtained from thermal simulation, the probability of high-temperature aging failure is calculated using the following formula: ; In the formula, The duration of the high temperature is given by t, where t is the junction temperature of the device. To activate energy, Boltzmann's constant, These are aging parameters, obtained by fitting the data based on high-temperature aging experiments of the device.

[0009] Preferably, the specific method for obtaining the vibration fatigue failure probability in S2 is as follows: Based on the number of cycles at different vibration stress levels in the test profile, and the local stress of the device obtained from stress simulation, the probability of vibration fatigue failure is calculated using the following formula: ; In the formula, The number of vibration stress cycles, For local stress in the device, , These are two vibration fatigue parameters, obtained by fitting based on the vibration fatigue experiments of the device.

[0010] Preferably, the specific method for obtaining the electromigration failure probability in S2 is as follows: Based on the duration of electrical stress in the experimental profile, the device current density obtained from current simulation, and the device junction temperature obtained from thermal simulation, the probability of electromigration failure is calculated using the following formula: ; In the formula, Let j be the duration of electrical stress, j be the device current density, and t be the device junction temperature. To activate energy, Boltzmann's constant, , These are the electromigration parameters, obtained by fitting based on the electromigration failure experiments of the device.

[0011] Preferably, in step S3, different path coverage type requirements are adopted according to the importance level of the performance test items, specifically as follows: S31, Test item coverage type requirements must be determined; State graph test path coverage types include transition coverage, state coverage, and main path coverage; activity graph test path coverage types include edge coverage, node coverage, and main path coverage. When the importance level of the performance test item is high, the state graph test path coverage requirement is transition coverage, and the activity graph test path coverage requirement is edge coverage. When the importance level of the performance test item is medium, the state graph test path coverage requirement is transition coverage or state coverage, and the activity graph test path coverage requirement is edge coverage or node coverage. When the importance level of the performance test item is low, the state graph test path coverage requirement is state coverage or main path coverage, and the activity graph test path coverage requirement is node coverage or main path coverage.

[0012] Preferably, the generation of each power performance test path in S3 based on the activity graph or state graph associated with the power performance test item in S1 specifically involves: S32, Test path generation based on state diagram; First, a complete state transition model is constructed based on all state nodes and transition edges in the state graph. Second, according to different test coverage requirements, corresponding graph traversal strategies are used to generate an initial test path set, including: for state coverage, a breadth-first search algorithm is used to generate the shortest path set to access all states; or for transition coverage, a depth-first search algorithm is used to ensure that each state transition is executed at least once. Finally, path merging and Eulerian path search optimization algorithms are used to process the initial test path set, eliminating redundant paths while ensuring coverage, thus forming the final test path set. The test path set for main path coverage is the main path set of the state graph given in S1.

[0013] S33, Test path generation based on activity graph; First, a complete control flow model is constructed based on all activity nodes, decision nodes, control edges, and concurrent structures in the activity graph. Second, according to different test coverage requirements, a graph traversal strategy is used to generate an initial test path set, including: for node coverage, a breadth-first search algorithm is used to generate the minimum path set that traverses all activity nodes; for edge coverage, a depth-first search algorithm is used to ensure that each control flow edge is executed at least once, and for concurrent structures, only each concurrent branch is guaranteed to be executed at least once. Finally, path merging and sequence optimization algorithms are used to process the initial test path set, eliminating redundant paths while ensuring coverage, forming a test path set; the test path set with main path coverage is the main path set of the activity graph given in S1.

[0014] Preferably, in step S4, based on the importance level of the same test item, different value requirements are given for single-variable values, multi-variable combination coverage, and multi-variable combination values ​​for the test stimulus, specifically as follows: The requirements for single-variable values ​​include three types: typical values, boundary values, and outliers; the requirements for multivariate combination coverage include two types: modified conditional coverage and decision coverage; the requirements for multivariate combination values ​​include three types: all typical values, all boundary values, and typical values ​​plus any outlier. When the importance level of the performance test item is high, the single variable value requirement is three types of value selection: typical value, boundary value, and outlier value. The multivariate combination coverage requirement is modified condition judgment coverage or decision coverage. The multivariate combination value requirement is three types of value selection: all typical values, all boundary values, and typical value and any outlier value. When the importance level of the performance test item is medium, the single variable value is required to be either typical value or boundary value. The multivariate combination coverage requirement is either decision coverage or no requirement. The multivariate combination value is required to be either full typical value or full boundary value. When the importance level of the performance test item is low, the single variable value is required to be a typical value, the multivariate combination coverage requirement is no requirement, and the multivariate combination value is required to be a full typical value.

[0015] Preferably, the generation of the S5 product performance test criteria specifically involves: By leveraging a large language model combined with prompt word engineering, performance test criteria are automatically extracted.

[0016] Preferably, the timing of the S6 product performance test is determined as follows: The testing methods are divided into continuous testing and periodic testing. Based on the criteria in S5, it is determined whether the corresponding testing conditions are met to achieve continuous testing. If they are met, continuous testing is used to cover the entire test process for the functional performance test items. If they are not met, periodic testing is used for the functional performance test items.

[0017] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) Analyze the importance of different test items from the dimensions of fault provocation probability and severity level, support the formulation of test strategies for test items of different importance, and realize the full testing of sensitive key faults.

[0018] (2) Generate test paths based on different coverage requirements to solve the problem of missing test paths under unconventional and multi-condition conditions and improve test coverage.

[0019] (3) Generate test stimuli based on different value requirements to solve the problem of missing test stimuli under boundary value and outlier conditions and improve test sufficiency. Attached Figure Description

[0020] Figure 1 A schematic diagram of a method for automatically generating performance test cases for reliability testing of avionics products; Figure 2 This is a functional activity diagram of the controller current loop for a certain actuator controller; Figure 3 This is a functional activity diagram of the secondary power supply of an actuator controller; Figure 4 This is a state transition diagram for the working state of a certain actuator controller. Figure 5 This is an activity diagram of the self-test function of the communication interface of an actuator controller; Figure 6 Flowchart of a method for automatically generating performance test cases for reliability testing of avionics products. Detailed Implementation

[0021] The embodiments of the present invention will be described below with reference to the accompanying drawings. The present invention proposes an automatic generation method for functional performance test cases in reliability testing of avionics products. It fully considers the characteristics of reliability testing conditions and fault identification requirements for airborne electronic products, analyzes the importance of different test items from dimensions such as the probability of fault occurrence and severity level, and designs different test paths and stimulus generation strategies based on importance. This achieves a more comprehensive and thorough identification of product faults under reliability testing conditions. Test requirements were generated using an aircraft actuator controller as an example, initially demonstrating the feasibility and effectiveness of generating functional performance test requirements for reliability testing of airborne electronic products based on this technical method.

[0022] To provide a more detailed explanation of the automatic generation method for reliability testing performance test cases of avionics products according to the present invention, such as... Figure 1As shown, this process includes clearly defining the product's functional performance test items, assigning the importance of each test item, generating test paths and test stimuli based on the importance, and determining test criteria and timing to generate complete product reliability test functional performance test cases. The following example uses an actuator controller. Figure 6 As shown, the specific process is as follows: S1, Determine the product performance test items; Product performance refers to the specific, measurable, and verifiable standards and requirements for the functions and performance of a product.

[0023] All functional performance requirements of a product can be extracted from documents such as product development requirements, technical agreements, and design specifications. Taking into account constraints such as testing scenarios, testing equipment, and testing duration, the product's functional performance test items are determined with the goal of covering all functional performance requirements to the greatest extent possible. Alternatively, existing product functional performance test items can be directly imported.

[0024] Each product performance test item has a corresponding state diagram or activity diagram, and each state diagram or activity diagram provides a set of main paths. Test items related to the transition of working states of airborne electronic products are associated with state diagrams, and test items related to the execution of working activities of airborne electronic products are associated with activity diagrams.

[0025] S2, Determine the importance of product performance test items; To balance test path coverage, test stimulus sufficiency, and test cost, effectively identify faults triggered in reliability testing, avoid missed detections, and improve testing efficiency, it is necessary to determine the importance of each test item.

[0026] S21, Obtain the overall failure probability of the device; Considering the potential temperature cycling fatigue failure, high-temperature aging failure, vibration fatigue failure, and electromigration failure of airborne electronic products under temperature, vibration, and electrical stress conditions in reliability test profiles, the overall failure probability of the device is obtained based on the load values ​​and durations of the test profiles, combined with simulation methods.

[0027] Devices are the basic units that realize a product. Multiple devices are combined to achieve a function. Therefore, when determining the importance of each test item, it is necessary to consider the devices related to it.

[0028] S211, Calculation of probability of failure due to temperature cycling fatigue; Based on the number of temperature cycles and the frequency of temperature cycles in the test profile, and the range of component junction temperature changes obtained from thermal simulation, the probability of temperature cycle fatigue failure is calculated using the following formula: ; In the formula, N is the number of temperature cycles, and f is the frequency of temperature cycles. This refers to the range of junction temperature variations of the device. , , These are three temperature cycling fatigue parameters, obtained by fitting the device's fatigue life experiment.

[0029] S212, Calculation of high-temperature aging failure probability; Based on the duration of high temperatures in the experimental profile and the device junction temperature obtained from thermal simulation, the probability of high-temperature aging failure is calculated using the following formula: ; In the formula, The duration of the high temperature is given by t, where t is the junction temperature of the device. To activate energy, Boltzmann's constant, These are aging parameters, obtained by fitting the data based on high-temperature aging experiments of the device.

[0030] S213, Calculation of vibration fatigue failure probability; Based on the number of cycles at different vibration stress levels in the test profile, and the local stress of the device obtained from stress simulation, the probability of vibration fatigue failure is calculated using the following formula: ; In the formula, The number of vibration stress cycles, For local stress in the device, , These are vibration fatigue parameters, obtained by fitting based on vibration fatigue experiments of the device.

[0031] S214, Calculation of electromigration failure probability; Based on the duration of electrical stress in the experimental profile, the device current density obtained from current simulation, and the device junction temperature obtained from thermal simulation, the probability of electromigration failure is calculated using the following formula: ; In the formula, Let j be the duration of electrical stress, j be the device current density, and t be the device junction temperature. To activate energy, Boltzmann's constant, , These are electromigration parameters, obtained by fitting based on electromigration failure experiments of the device.

[0032] S215, Calculation of overall device failure probability; Calculate the overall failure probability of the device using the following formula: .

[0033] S22, Obtain function-device dependency; The dependence of product functionality on components is obtained through the failure mode effect probability in FMECA (Failure Mode, Effects, and Criticality Analysis). The value range is 0-1.

[0034] S23, obtain the severity level of the functional failure; Obtain the severity level of product functional failures by analyzing the severity level of failure mode impact in FMECA. The severity level is set at 1.0 for Class I, 0.7 for Class II, 0.4 for Class III, and 0.1 for Class IV.

[0035] S24, Obtain the functional relevance of the test item; Obtain the test item-functionality correlation based on the input functional requirements-test item traceability matrix. The functional requirements-test item traceability matrix contains the correlation between each function and test item. When a test item fully validates the function... The value is 1.0, and it is used when the test item partially verifies this function. The value is 0.5, when the test item does not verify this function. The value is 0.0. A test item is usually associated with one function, but sometimes it is associated with multiple functions. When a test item is associated with only one function... The value must be 1.0.

[0036] S25, Calculation of the importance of performance test items; Calculate the importance of performance test items using the following formula: ; in, The importance of performance test items, In order to be with the first The overall device failure probability of the i-th device related to each function. For the first The dependence of each function on the i-th device. For the first The severity level of functional failure for each function. For the test item and the first Functional relevance; Indicates the total number of functions related to the test item; I indicates the number of functions related to the first test item. The total number of devices related to each function.

[0037] S26, Determination of the importance level of performance test items; like If the importance level of the test item is high, then the importance level of the test item is high; if If the importance level of the test item is medium; If the score is low, then the importance level of the test item is low.

[0038] S3, Product performance test path generation; Based on the importance level of the power performance test items, different path coverage type requirements are adopted, and each power performance test path is generated based on the activity graph or state graph associated with the power performance test items in S1.

[0039] S31, Test item coverage type requirements must be determined; Based on the different importance levels of the test items obtained in S2, the test path coverage type is determined for the corresponding state diagram or activity diagram of different test items.

[0040] State diagrams are primarily used to generate test paths for test items related to the transition of operating states in airborne electronic products. The test path coverage types include transition coverage, state coverage, and main path coverage. Transition coverage means that all state transitions are executed at least once; state coverage means that all states are executed at least once; and main path coverage means that the main state transition path is executed at least once.

[0041] Activity graphs are primarily used to generate test paths for test items related to the execution of operational activities in airborne electronic products. Test path coverage types include edge coverage, node coverage, and main path coverage. Edge coverage means that all migration edges are executed at least once; node coverage means that all nodes are executed at least once; and main path coverage means that the main functional activity path is executed at least once.

[0042] The correspondence between the test path coverage types and the importance levels of the test items is shown in Table 1.

[0043] Table 1. Correspondence between test items and test path coverage types

[0044] S32, Test path generation based on state diagram; First, all state nodes and transition edges in the state graph are systematically extracted and analyzed to construct a complete state transition model. Second, based on different test coverage requirements, corresponding graph traversal strategies are used to generate an initial test path set. This includes: for state coverage, algorithms such as Breadth-First Search (BFS) are used to efficiently generate the shortest path set to all states; or for transition coverage, algorithms such as Depth-First Search (DFS) are used to ensure that each state transition is executed at least once. Finally, optimization algorithms such as path merging and Eulerian path search are comprehensively applied to simplify and integrate the initial test path set, eliminating redundant paths while maintaining coverage, thus forming the final test path set. The test path set for main path coverage is the main path set of the state graph given in S1, which does not require algorithmic generation.

[0045] S33, Test path generation based on activity graph; First, a systematic approach is used to extract all activity nodes, decision nodes, control edges, and concurrent structures from the activity graph, constructing a complete control flow model. Concurrent structures typically include branching and merging. Second, based on different test coverage requirements, targeted graph traversal strategies are employed to generate an initial test path set. This includes: for node coverage, algorithms such as Breadth-First Search (BFS) are used to generate the minimum set of paths traversing all activity nodes; for edge coverage, algorithms such as Depth-First Search (DFS) are used to ensure that every control flow edge, especially all output branches of decision nodes, is executed at least once. For concurrent structures, a strategy that ensures each concurrent branch is executed at least once is used to avoid path combinatorial explosion. Finally, path merging and sequence optimization algorithms are comprehensively applied to simplify and integrate the initial test path set, eliminating redundant paths while maintaining coverage, thus forming the final test path set. The test path set for main path coverage is the main path set of the activity graph given in S1, which does not require algorithmic generation.

[0046] S4, Product performance test stimulus generation; Based on the different importance levels of the test items obtained in S2, different value requirements are given for single variable values, multivariate combination coverage, and multivariate combination values ​​for the test stimuli.

[0047] The single-variable value requirements mainly target variables such as voltage, current, and bus data, and include three value types: typical values, boundary values, and outliers. Typical values ​​refer to default values, initial values, or high-frequency values ​​in actual product use; boundary values ​​are the maximum or minimum values ​​that meet the criteria and range; and outliers are values ​​slightly greater than the maximum or slightly less than the minimum. (Single-variable value requirements.)

[0048] Multivariate combination coverage refers to the requirement that when decision conditions in an activity diagram or state transition conditions in a state diagram are composed of multiple variable condition decisions through logical operators, each variable condition decision must be true or false. This includes two types of coverage: Modified Decision Coverage (MC / DC) and Decision Coverage. MC / DC requires that all branches of the decision or state transition result be covered, that each variable decision (true or false) constituting the decision or state transition condition be covered, and that each variable decision independently changes the decision or state transition result. Decision Coverage only requires that all branches of the decision or state transition result be covered. For example, T=(A&&B) || C is the decision or state transition condition. Only four combinations can satisfy MC / DC: (A=true, B=true, C=false, T=true), (A=false, B=true, C=false, T=false), (A=true, B=false, C=false, T=false), and (A=true, B=false, C=true, T=true). Only two combinations can satisfy decision coverage: (A=true, B=true, C=false, T=true) and (A=false, B=true, C=false, T=false).

[0049] Multivariate combination requirements refer to the combination of typical values, boundary values, and outliers for multiple variables, including three methods: all typical values, all boundary values, and typical values ​​plus any outlier. All typical values ​​mean that all variables in the decision or state transition conditions take typical values; all boundary values ​​mean that all variables in the decision or state transition conditions take boundary values; typical values ​​plus any outlier mean that any variable in the decision or state transition conditions takes an outlier value, while the remaining variables take typical values.

[0050] The requirements for generating test stimuli for test items of different importance are shown in Table 2.

[0051] Table 2 Test Item Test Incentive Generation Requirements

[0052] S5, Product performance test criteria generation; By utilizing a large language model and combining prompt word engineering, the functional performance requirements in product development requirements, technical agreements, design specifications and other documents are automatically extracted and converted into test pass criteria for each test path in each functional performance test item.

[0053] S6, the timing for product performance testing is determined; The testing methods are divided into continuous testing and periodic testing. Based on the criteria in S5, it is determined whether the corresponding testing conditions are met to achieve continuous testing. If so, continuous testing is used to cover the entire test process for the performance test items. If not, periodic testing is used for the performance test items. The timing of the test should be selected at the most severe point of each stress in the test profile, such as the high temperature extreme point, the low temperature extreme point, the vibration maximum point, etc. The testing frequency is determined according to the importance of the performance test items.

[0054] S7 automatically generates successful performance test cases; Each test case includes a test path, stimulus, criterion, and test timing. Because the importance of each performance test item is different, there will be different numbers of paths and stimulus requirements. The paths and stimuli are combined and arranged, and each combination constitutes a test case, thus generating a different number of test cases.

[0055] To ensure the reliability of test cases, the automatically generated test cases can be manually verified, which will greatly reduce the complexity of manually writing test cases and improve the efficiency of test case generation.

[0056] The following section uses a certain actuator controller as an example to illustrate in detail the method for automatically generating reliability test requirements for airborne electronic products established by this invention: S1, Determine the product performance test items; The test items for a certain actuator controller include controller current loop test, working state transition test, secondary power supply test, and communication interface self-test test.

[0057] S2, Determine the importance of product performance test items; Using the reliability test profile as input, the comprehensive failure probability of each device is analyzed through simulation calculation. The device-function dependency and functional failure severity level are obtained through FMECA analysis. The test item-function correlation degree in the functional requirement-test item traceability matrix is ​​extracted. According to the formula, the importance of the controller current loop test item is 0.73 (high importance), the importance of the working state transition test item is 0.5 (medium importance), the importance of the secondary power supply test item is 0.449 (medium importance), and the importance of the communication interface self-test item is 0.033 (low importance).

[0058] Tables 3-6 list the parameter values ​​involved in obtaining the importance of the controller current loop test item, the working state transition test item, the secondary power supply test item, and the communication interface self-test item based on S2.

[0059] Table 3 Importance Analysis of Controller Current Loop Test Items

[0060] Table 4 Importance Analysis of Work State Transition Test Items

[0061] Table 5 Importance Analysis of Secondary Power Supply Test Items

[0062] Table 6 Importance Analysis of Communication Interface Self-Test Items

[0063] According to S26, if If the importance level of the test item is high, then the importance level of the test item is high; if If the importance level of the test item is medium; If the test result is negative, then the importance level of that test item is low. Therefore, the controller current loop test has a high importance level, the secondary power supply test has a medium importance level, the operating state transition test has a medium importance level, and the communication interface self-test test has a low importance level.

[0064] S3, Product performance test path generation; Based on importance, the test path generation requirements for each test item of a certain actuator controller are shown in Table 3.

[0065] The controller current loop test correlation activity diagram and the controller current loop functional activity diagram are attached. Figure 2 Based on the importance level, the coverage type is determined to be edge coverage, and two test paths are generated according to the edge coverage requirements: L1: A1 → A2 → D1 → A3 → A4; L2: A1 → A2 → D1 → A5.

[0066] The secondary power supply test association activity diagram and the secondary power supply function activity diagram are attached. Figure 3 Based on the importance level, the coverage type is determined to be node coverage; two test paths are generated according to the node coverage requirements: L3: A1 → D1 → A2 → A3 → D2 → A4 → A3 → D2; L4: A1 → D1 → A5.

[0067] The working state transition test associated state diagram is attached. Figure 4 Based on the importance level, the coverage type is determined to be transfer coverage, and two test paths are generated according to the transfer coverage requirements: L5: B1 → B2 → B3 → B2; L6: B1 → B2 → B4 → B2.

[0068] The activity diagram for the communication interface self-test is attached. Figure 5 Based on the importance level, the coverage type is determined to be main path coverage, and a test path is directly obtained according to the main path coverage requirements: L7: A1 → A2 → A3 → D1 → A4 → A5.

[0069] S4, Product performance test stimulus generation; Based on importance, the test excitation generation requirements for each test item of a certain actuator controller are shown in Table 3.

[0070] The controller current loop test, based on a high importance level, specifies single-variable value requirements as typical, boundary, and outlier values. Multi-variable combination coverage requires modified conditional decision coverage (MC / DC), with value requirements including all typical values, all boundary values, and a typical value plus any outlier. The controller current loop function generates two test paths, L1 and L2. According to the MC / DC requirements, four multi-variable combination methods are generated for decision point D1: for test path L1, (current value within range = true, current change rate normal = true, valid sampling = true); for test path L2, (current value within range = false, current change rate normal = true, valid sampling = true), (current value within range = true, current change rate normal = false, valid sampling = true), (current value within range = true, current change rate normal = true, valid sampling = false). Then, according to the requirements for multivariate combination values ​​and single variable values, specific test stimulus values ​​are generated. The normal range for current value is [-2, 2], and the normal range for current change rate is [-0.5, 0.5]. Valid sampling is defined as (true, false). Taking (current value within range = true, current change rate normal = false, valid sampling = true) as an example, the full typical value test stimulus is (current value = 1, current change rate = 3, valid sampling = true), the full boundary value test stimulus is (current value = -2, current change rate = 0.6, valid sampling = true), and the typical value and any outlier test stimulus is (current value = -2.1, current change rate = 3, valid sampling = true). Considering four multivariate combination methods, this test item has a total of 12 test stimuli, covering two test paths.

[0071] The secondary power supply test, based on a medium importance level, determines the single-variable value requirements as typical and boundary values, and the multi-variable combination coverage requirement as decision coverage, with the multi-variable combination value requirements being full typical and full boundary values. The secondary power supply function generates two test paths, L3 and L4. Since decision conditions D1 and D2 each involve only one variable, the specific values ​​of the test stimuli are directly generated according to the multi-variable combination and single-variable value requirements. The normal range requirement for input voltage is [24, 32], and the normal range requirement for output voltage is [14, 16]. Taking path L3 as an example, the full typical value test stimuli are (input voltage = 28, output voltage (first time) = 18, output voltage (second time) = 15), and the full boundary value test stimuli are (input voltage = 32, output voltage (first time) = 16.1, output voltage (second time) = 16). This test item has a total of 4 test stimuli, covering 2 test paths.

[0072] The job state transition test, based on a medium importance level, specifies that the single-variable value requirements are typical and boundary values, and the multivariate combination coverage requirement is decision coverage, with all typical and boundary values ​​required. Two test paths, L5 and L6, are generated for the job state transition. However, since the variables in the state transition conditions are all Boolean values, there are no boundary values; therefore, only multivariate combination decision coverage needs to be satisfied. In state B2, the test stimuli are set to cover the two paths respectively: (self-test command = true, communication command = false) and (self-test command = false, communication command = true). This test item has a total of 2 test stimuli, covering 2 test paths.

[0073] The communication interface self-test, based on its low importance level, determines the single-variable value requirement as typical, the multi-variable combination coverage requirement as none, and the multi-variable combination value requirement as full typical. A test path L7 is generated for the communication interface self-test. Since there is no multi-variable combination coverage requirement, the specific values ​​of the test stimuli are generated directly according to the multi-variable combination value requirement and the single-variable value requirement. The normal range for response time is less than 100, and a correct response content is a Boolean value. The full typical value test stimulus is (response time = 50, correct response content = true). This test item has a total of 1 test stimulus, covering 1 test path.

[0074] Table 7 shows the specific requirements for the test paths and excitation generation requirements of each test item of a certain actuator controller.

[0075] Table 7 Test Paths and Excitation Generation Requirements for Each Test Item of an Actuator Controller

[0076] S5, Product test criterion generation; Using a large language model and combining prompt word engineering, the system automatically extracts the power performance requirements from documents such as product development requirements, technical agreements, and design specifications, and converts them into test pass criteria for each test path in each power performance test item. The pass criteria for the controller current loop test are: current control accuracy is ±1%; the pass criteria for the working state transition test are: normal state transition; the pass criteria for the secondary power supply test are: output voltage greater than or equal to 14V and less than or equal to 16V; and the pass criteria for the communication interface self-test test are: report self-test success.

[0077] S6, Product testing timing determined; Based on the stress sensitivity and severity of the product test items, in the reliability test profile, the controller current loop test should be carried out periodically at the highest and lowest temperatures of the temperature cycling stress, and when the operating state is under control / management. The operating state transition test should be carried out periodically under different operating states. The secondary power supply test should be carried out periodically before and after the operating state changes or when voltage fluctuations occur. The communication interface self-test should be carried out periodically when the operating state is under communication.

[0078] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made by those skilled in the art to the technical solutions of the present invention without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A method for automatically generating performance test cases for reliability testing of avionics products, characterized in that, It includes the following steps: S1, Determine the product performance test items; Identify or import product performance test items. Each product performance test item has a corresponding state diagram or activity diagram, and each state diagram or activity diagram provides a set of main paths. S2, Determine the importance of product performance test items; The overall failure probability of the device is: ； in, The overall failure probability of the device. This represents the probability of temperature cycling fatigue failure. This represents the probability of failure due to high-temperature aging. This represents the probability of vibration fatigue failure. This represents the probability of electromigration failure. The importance of the performance test items is as follows: ； in, The importance of performance test items, In order to be with the first The overall device failure probability of the i-th device related to each function. For the first The dependence of each function on the i-th device. For the first The severity level of functional failure for each function. For the test item and the first Functional relevance; Indicates the total number of functions related to the test item; I indicates the number of functions related to the first test item. The total number of devices related to each function; The importance level of each performance test item is determined based on its importance. S3, Product performance test path generation; Based on the importance level of the power performance test items, different path coverage type requirements are adopted, and each power performance test path is generated based on the activity diagram or state diagram associated with the power performance test items in S1. S4, Product performance test stimulus generation; Based on the different importance levels of the test items obtained in S2, different value requirements are given for single variable values, multivariate combination coverage, and multivariate combination values ​​for the test stimuli. S5, Product performance test criteria generation; S6, the timing for product performance testing is determined; S7 automatically generates successful performance test cases; For each performance test item, test cases are generated based on the test path, stimulus, criteria, and test timing.

2. The method for automatically generating test cases for reliability testing of avionics products according to claim 1, characterized in that: The specific method for obtaining the probability of temperature cycle fatigue failure in S2 is as follows: Based on the number of temperature cycles and the frequency of temperature cycles in the test profile, and the range of component junction temperature changes obtained from thermal simulation, the probability of temperature cycle fatigue failure is calculated using the following formula: ； In the formula, N is the number of temperature cycles, and f is the frequency of temperature cycles. This refers to the range of junction temperature variation of the device. , , These are three temperature cycling fatigue parameters, obtained by fitting the device's fatigue life data based on experiments.

3. The method for automatically generating test cases for reliability testing of avionics products according to claim 1, characterized in that: The specific method for obtaining the high-temperature aging failure probability in S2 is as follows: Based on the duration of high temperatures in the experimental profile and the device junction temperature obtained from thermal simulation, the probability of high-temperature aging failure is calculated using the following formula: ； In the formula, The duration of the high temperature is given by t, where t is the junction temperature of the device. To activate energy, Boltzmann's constant, These are aging parameters, obtained by fitting the data based on high-temperature aging experiments of the device.

4. The method for automatically generating test cases for reliability testing of avionics products according to claim 1, characterized in that: The specific method for obtaining the vibration fatigue failure probability in S2 is as follows: Based on the number of cycles at different vibration stress levels in the test profile, and the local stress of the device obtained from stress simulation, the probability of vibration fatigue failure is calculated using the following formula: ； In the formula, The number of vibration stress cycles, For local stress in the device, , These are vibration fatigue parameters, obtained by fitting based on vibration fatigue experiments of the device.

5. The method for automatically generating test cases for reliability testing of avionics products according to claim 1, characterized in that: The specific method for obtaining the electromigration failure probability in S2 is as follows: Based on the duration of electrical stress in the experimental profile, the device current density obtained from current simulation, and the device junction temperature obtained from thermal simulation, the probability of electromigration failure is calculated using the following formula: ； In the formula, Let j be the duration of electrical stress, j be the device current density, and t be the device junction temperature. To activate energy, Boltzmann's constant, , These are electromigration parameters, obtained by fitting based on electromigration failure experiments of the device.

6. The method for automatically generating test cases for reliability testing of avionics products according to claim 1, characterized in that: In S3, different path coverage type requirements are adopted according to the importance level of the performance test items, specifically: S31, Test item coverage type requirements must be determined; State graph test path coverage types include transition coverage, state coverage, and main path coverage; activity graph test path coverage types include edge coverage, node coverage, and main path coverage. When the importance level of the performance test item is high, the state graph test path coverage requirement is transition coverage, and the activity graph test path coverage requirement is edge coverage. When the importance level of the performance test item is medium, the state graph test path coverage requirement is transition coverage or state coverage, and the activity graph test path coverage requirement is edge coverage or node coverage. When the importance level of the performance test item is low, the state graph test path coverage requirement is state coverage or main path coverage, and the activity graph test path coverage requirement is node coverage or main path coverage.

7. The method for automatically generating performance test cases for reliability testing of avionics products according to claim 6, characterized in that: The specific steps for generating each power performance test path in S3 based on the activity graph or state graph associated with the power performance test item in S1 are as follows: S32, Test path generation based on state diagram; First, a complete state transition model is constructed based on all state nodes and transition edges in the state graph. Second, according to different test coverage requirements, corresponding graph traversal strategies are used to generate an initial test path set, including: for state coverage, a breadth-first search algorithm is used to generate the shortest path set to access all states; or for transition coverage, a depth-first search algorithm is used to ensure that each state transition is executed at least once. Finally, the initial test path set is processed by combining path merging and Euler path search optimization algorithms to eliminate redundant paths while ensuring coverage, forming a test path set; the test path set with main path coverage is the main path set of the state graph given in S1. S33, Test path generation based on activity graph; First, a complete control flow model is constructed based on all activity nodes, decision nodes, control edges, and concurrent structures in the activity graph. Second, according to different test coverage requirements, a graph traversal strategy is used to generate an initial test path set, including: for node coverage, a breadth-first search algorithm is used to generate the minimum path set that traverses all activity nodes; for edge coverage, a depth-first search algorithm is used to ensure that each control flow edge is executed at least once, and for concurrent structures, only each concurrent branch is guaranteed to be executed at least once. Finally, path merging and sequence optimization algorithms are used to process the initial test path set, eliminating redundant paths while ensuring coverage, forming a test path set; the test path set with main path coverage is the main path set of the activity graph given in S1.

8. The method for automatically generating test cases for reliability testing of avionics products according to claim 1, characterized in that: In S4, based on the importance level of the same test item, different value requirements are given for single-variable values, multi-variable combination coverage, and multi-variable combination values ​​for the test stimulus, specifically: The requirements for single-variable values ​​include three types: typical values, boundary values, and outliers; the requirements for multivariate combination coverage include two types: modified conditional coverage and decision coverage; the requirements for multivariate combination values ​​include three types: all typical values, all boundary values, and typical values ​​plus any outlier. When the importance level of the performance test item is high, the single variable value requirement is three types of value selection: typical value, boundary value, and outlier value. The multivariate combination coverage requirement is modified condition judgment coverage or decision coverage. The multivariate combination value requirement is three types of value selection: all typical values, all boundary values, and typical value and any outlier value. When the importance level of the performance test item is medium, the single variable value is required to be either typical value or boundary value. The multivariate combination coverage requirement is either decision coverage or no requirement. The multivariate combination value is required to be either full typical value or full boundary value. When the importance level of the performance test item is low, the single variable value is required to be a typical value, the multivariate combination coverage requirement is no requirement, and the multivariate combination value is required to be a full typical value.

9. The method for automatically generating test cases for reliability testing of avionics products according to claim 1, characterized in that: The specific method for generating the S5 product performance test criteria is as follows: By leveraging a large language model combined with prompt word engineering, performance test criteria are automatically extracted.

10. The method for automatically generating test cases for reliability testing of avionics products according to claim 1, characterized in that: The specific timing for the performance testing of the S6 product is as follows: The testing methods are divided into continuous testing and periodic testing. Based on the criteria in S5, it is determined whether the corresponding testing conditions are met to achieve continuous testing. If they are met, continuous testing is used to cover the entire test process for the functional performance test items. If they are not met, periodic testing is used for the functional performance test items.

Images