A method and system for dynamic optimization of a batch production satellite test matrix based on successive pass thresholds

Abstract

The application discloses a kind of based on continuous passing threshold batch satellite test matrix dynamic optimization method and system, it is related to spacecraft general assembly integrated test and environmental test technical field, including the following steps: constructing the test and test project set covering the whole process of spacecraft general assembly integrated test, test and test project set are divided into general assembly must measure class project, ESS screening class project and design verification class project according to function.The test system covering AIT whole process is constructed, the key test of general assembly is included into unified tailoring framework, and the test frequency dynamic adjustment and defect backout are realized by continuous passing threshold.Combined with design change trigger and sampling bottom mechanism, the quality bottom line in tailoring process is guaranteed.Risk constraint model is introduced to control the risk of missed detection, and the results are output in R / ER / ---matrix form, which improves the efficiency and executability of batch test.

Description

Technical Field

[0001] This invention relates to the field of spacecraft assembly, integration testing and environmental testing technology, specifically to a dynamic optimization method and system for batch satellite test matrix based on continuous pass threshold. Background Technology

[0002] With the rapid development of large-scale low-Earth orbit satellite constellation systems, spacecraft development is shifting from traditional single-satellite customized production to high-speed mass production. Against this backdrop, spacecraft assembly, integration, and testing (AIT) and environmental testing, as crucial links in ensuring the overall functionality and on-orbit reliability of the satellite, still employ the traditional testing system of graded qualification and component-by-component acceptance. This means that every satellite undergoes comprehensive testing and environmental trials, including electrical performance testing, thermal vacuum testing, vibration testing, and electromagnetic compatibility testing. While this model effectively ensures product quality in small-batch development, it faces significant challenges in mass production scenarios involving hundreds or even thousands of satellites. The component-by-component, full-coverage testing approach puts considerable pressure on testing costs, production cycles, and production line resource usage, becoming a major bottleneck restricting the improvement of mass production efficiency. Therefore, how to rationally optimize the execution strategy of testing projects and achieve efficient allocation of testing resources while ensuring satellite product quality and reliability has become a key technical problem that needs to be solved.

[0003] Existing techniques for tailoring spacecraft tests still have many shortcomings: First, most methods only optimize certain environmental test items (such as thermal vacuum or vibration tests), lacking a unified test item management system covering the entire AIT process (including assembly testing and environmental testing); second, existing methods mostly adopt static sampling strategies with fixed rules (such as full inspection every N satellites or batch sampling), lacking a dynamic iterative tailoring mechanism based on single-satellite test results, making it difficult to adjust the test frequency in real time according to the product quality status; third, when design changes occur or new failure modes are exposed in orbit, existing methods lack targeted test recovery triggering mechanisms, and also lack effective periodic fallback inspection methods after a large number of test reductions, posing potential quality risks; in addition, existing tailoring decision results are mostly presented in parameter or text form, lacking an intuitive and standardized test matrix output format, making it difficult to directly guide AIT field execution.

[0004] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for dynamic optimization of batch satellite test matrix based on continuous pass threshold, so as to solve the problems in the background art mentioned above.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a dynamic optimization method for a batch satellite test matrix based on a continuous pass threshold, comprising the following steps: A set of test and experiment items covering the entire process of spacecraft assembly, integration and testing is constructed. The set of test and experiment items is divided into mandatory assembly test items, ESS screening items and design verification items according to function. Among them, the mandatory assembly test items are always executed, the ESS screening items are as customizable items, and the design verification items are not executed by default in the mass production stage. Based on the set of test and experimental items, establish a set of correlation relationships between test items and stress types and detectable fault types, and limit the statistical range of effectiveness of each test item to its corresponding detectable fault type; Based on the set of relationships, a comprehensive technology maturity index is obtained, and combined with the cumulative number of executions and the cumulative number of defects detected for each ESS screening project, the corresponding test effectiveness index is calculated to characterize the defect detection capability of the corresponding test project. For each ESS screening item, an interval level and consecutive pass count are introduced. The interval level and consecutive pass count are updated based on the test effectiveness indicators and test execution results. The interval level determines whether the current satellite should execute the corresponding test item. If the test is passed, the interval level is increased. If the test is failed, the interval level is reverted to the initial state. Based on the interval level, test execution decisions are generated for each satellite. The initial satellite executes all tests and experimental items. Subsequent satellites continue to execute mandatory assembly test items and determine whether to execute or skip ESS screening items according to the interval level. When a design change occurs, the execution of the corresponding test items is restored according to the change type and the interval level is simultaneously rolled back. For each satellite, the risk of missed detection is calculated based on the test effectiveness index, the comprehensive technology maturity index, and the skipped ESS screening items. When the risk of missed detection exceeds the preset threshold, the execution of the corresponding test items is restored in order of risk contribution. Combined with the periodic sampling rules, a test matrix is ​​generated with the results of execution, evaluation execution, or non-execution as identifiers.

[0007] Preferably, the pilot projects are organized in a tiered manner based on differences in execution attributes during the mass production process, in order to achieve stable execution of key projects and flexible adjustment of tailorable projects. The steps are as follows: All test and experiment items are divided into mandatory assembly test items, ESS screening items, and design verification items according to their functional attributes, and corresponding classification labels and management parameters are established for each type of item. Fixed execution rules are configured for mandatory test items in the final assembly, and the execution status is directly marked during the generation of each satellite test plan; Set a consecutive pass threshold for ESS filter items and maximum interval level And establish an execution interval mapping relationship corresponding to the interval level; For design verification projects, set the initial execution scope and default state for the mass production stage, and adjust the execution flag according to the design state during the test matrix generation process.

[0008] Preferably, constraint rules are established around the correspondence between test items and defect exposure capabilities to improve the accuracy and independence of test effectiveness statistics. The steps are as follows: For each test item, a set of corresponding stress types is established, and each stress type is classified and labeled. Determine the corresponding set of detectable fault types based on the stress type, and establish a mapping relationship between test items and fault types; During the defect recording process, each defect is categorized and identified, and it is determined whether it falls within the detectable fault range of the corresponding test item. In the validity statistics, only defect data that meets the mapping relationship is included in the statistics, and defect data that does not meet the conditions are removed.

[0009] Preferably, the test frequency is adaptively adjusted by combining the test execution results with historical status information to achieve a balance between dynamic trimming and quality control. The steps are as follows: Initialize interval levels for each ESS screening test item and number of consecutive passes And establish a corresponding status record table; Determine the test execution interval based on the current interval level. And calculate the execution decision of the corresponding test item according to the satellite serial number. ; After the test is completed, the consecutive pass count is updated based on the test results. When the test result is a pass, [the following is applied]. Accumulate the values ​​and compare them with the consecutive thresholds. When the number of consecutive passes reaches a threshold, the interval level is increased; when the test result is a failure, the interval level is reset and the number of consecutive passes is cleared to zero.

[0010] Preferably, a preset mapping relationship is established between the interval level and the execution interval, and periodic determination is performed based on the execution interval corresponding to the interval level during the execution decision generation process. At the same time, the interval level is triggered to increase when the number of consecutive passes meets the consecutive pass threshold condition, and the interval level is rolled back and the number of consecutive passes is cleared simultaneously when the test result is a failure, thus forming a stable state update rule.

[0011] Preferably, response control is implemented to address the impact of design state changes on the test execution strategy, ensuring the integrity of verification under changed conditions. The steps are as follows: Receive design change information and classify and label the changes as thermal control, structural or electrical changes; Based on the pre-defined correspondence between change types and test items, determine the set of test items that will be affected; Adjust the execution decision status of the affected test items and mark the corresponding test items as execution status; The interval levels of the above test items were rolled back, and their consecutive pass records were updated simultaneously.

[0012] Preferably, a periodic full inspection strategy is introduced to compensate for the long-term cutting process, in order to prevent the accumulation of potential quality risks. The steps are as follows: Set sampling interval parameters based on satellite production serial numbers and establish corresponding serial number judgment rules; For each satellite, determine whether its sequence number meets the sampling interval conditions. If the conditions are met, mark all test items as being in execution status. Set sampling cycle parameters based on the time dimension and record production time information; The current production time is assessed, and all test items are executed when the preset time cycle conditions are met, and the sampling inspection records are updated.

[0013] Preferably, the sampling interval parameter and the time period parameter correspond to independent triggering conditions. When any triggering condition is met, all test items are uniformly marked as execution status, and the execution records and status identifiers of the corresponding test items are updated synchronously to maintain the consistency between the sampling triggering process and the test execution status.

[0014] Preferably, quantitative control is implemented around potential risks during the experimental tailoring process to ensure the safety boundary of the overall experimental strategy. The steps are as follows: Calculate the risk of missed detection based on the current satellite test execution status and relevant parameters. And record the calculation results; Compare the risk of missed detection with the preset risk threshold to determine whether the current test plan meets the constraints. When the risk exceeds the threshold, the corresponding risk contribution value is calculated for the unexecuted test items. And sort them according to their contribution value; The execution status of each test item is adjusted according to the ranking results, and the risk is recalculated after each adjustment until the constraints are met.

[0015] A dynamic optimization system for batch satellite test matrix based on continuous pass thresholds includes a project management module, a status tracking module, an interval upgrade engine, a change triggering module, a risk constraint module, and a matrix output module. The project management module is used to manage parameter information for all star-level tests and experiments. Parameter information includes the project number. Project name, priority sorting, function positioning, consecutive pass threshold Maximum Interval Level Experimental costs Test cycle and consequences weight ; The status tracking module is used to maintain the interval level for ESS screening test items. and number of consecutive passes And record the execution status and test results of the corresponding test items for each satellite; An interval upgrade engine is used to update test results based on the state tracking module and the corresponding consecutive pass thresholds. For interval levels Perform upgrade or rollback calculations and output the execution decision for the next satellite's corresponding test project. ; The change triggering module is used to receive design change information, identify the affected test items according to the change type, adjust the execution decision of the corresponding test item to the execution state, and revert the corresponding interval level to the initial state. The risk constraint module is used to calculate the current satellite's risk of missed detections. And when the risk of missed detection exceeds the preset threshold At that time, according to the risk contribution value of each test item Based on the sorting results, the execution status of each test item is restored until the risk constraints are met; The matrix output module is used to generate an experimental matrix based on the execution decisions of each experimental item and output it in a preset identifier format to support experimental execution.

[0016] The technical effects and advantages provided by the present invention in the above technical solution are as follows: This invention constructs a test project system covering the entire process of spacecraft assembly, integration, and testing. It integrates key assembly-stage projects such as electrical performance testing, quality characteristic testing, and accuracy characteristic testing into a unified tailored management framework, achieving a shift from traditional single-environment test optimization to full-scale, star-level process optimization. Based on this, it introduces an interval upgrade and rollback mechanism based on a continuous pass threshold. This allows the test execution strategy to be dynamically adjusted star-by-star according to changes in product quality status. Under continuous pass conditions, the execution frequency is automatically reduced, and when defects are detected, it quickly rolls back to a full inspection state, thus balancing test efficiency and quality safety, demonstrating excellent adaptability and engineering applicability.

[0017] This invention further achieves refined control of the test recovery strategy through a design change triggering mechanism. It automatically identifies affected test items based on different types of design changes and recovers the relevant tests accordingly, avoiding the problem of indiscriminate recovery of all test items in traditional methods. Simultaneously, by introducing a sampling inspection fallback mechanism, it constrains test execution from two dimensions: production quantity and time cycle. Even with significant reductions in test volume, it can still periodically perform full-item inspections, thereby establishing a stable quality assurance baseline and ensuring that the testing system does not accumulate risks during long-term operation.

[0018] This invention also establishes a risk constraint mechanism based on a probabilistic model, quantitatively assessing the risk of missed detections for each satellite and ensuring that the risk remains within a controllable range through threshold control. When the risk exceeds a set threshold, the execution of high-risk test items is automatically resumed, forming a closed-loop safety control system. Simultaneously, all decision results are output in a standardized R / ER / - test matrix format, transforming complex test optimization strategies into intuitive and executable operational instructions that can be directly applied to the production site, significantly reducing execution complexity and improving consistency and efficiency in mass production processes. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0020] Figure 1 This is a flowchart of a method for dynamic optimization of a batch satellite test matrix based on a continuous pass threshold, according to the present invention.

[0021] Figure 2 This is a schematic diagram of a module of a batch satellite test matrix dynamic optimization system based on a continuous pass threshold according to the present invention. Detailed Implementation

[0022] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that the description of this disclosure will be more complete and fully convey the concept of the exemplary embodiments to those skilled in the art.

[0023] This invention provides, for example Figure 1 The method for dynamic optimization of a batch satellite test matrix based on a continuous pass threshold, as shown, includes the following steps: For the entire process of spacecraft assembly, integration, and testing (AIT), a systematic review and structured modeling of all satellite-wide testing and experimental items were conducted to establish a testing and experimental item system covering all key quality characteristics and environmental adaptability. The system includes 11 representative testing and experimental items, prioritized according to their impact on overall satellite performance, reliability, and on-orbit safety, thus forming a unified standardized test set. This provides a basic data framework and constraints for subsequent dynamic tailoring and optimization decisions.

[0024] Specifically, the testing and experimentation system is organized according to numbering. Identify and maintain the consistency and uniqueness of the number throughout the entire process, including: Electrical performance testing is used to verify the electrical functional integrity and interface compatibility of the entire satellite under different operating modes, including subsystem joint testing and overall satellite comprehensive testing. This test is directly related to whether the overall satellite function meets the design requirements, and therefore serves as the most basic verification method. The ambient temperature cycling test has a temperature range of -20℃ to +60℃ and a cycle count of no less than 8 times. Through repeated temperature stress loading, it is used to expose welding defects and early failure problems of components. For thermal vacuum testing, the vacuum level shall not exceed [a certain value]. The temperature range is -25℃ to +65℃, and the number of cycles is no less than 4. This test is used to simulate the on-orbit space environment and to comprehensively verify the thermal design and sealing performance of the entire satellite. For leak detection testing, helium mass spectrometry was used, and the leak rate was controlled within a certain range. Within this range, measures are taken to ensure that the structural sealing meets the requirements of space applications; For mass characteristic testing, including the measurement of the overall satellite mass, center of mass position, and moment of inertia parameters, to ensure that the overall satellite dynamic characteristics meet the design specifications; For accuracy characteristic testing, including antenna pointing accuracy and payload installation accuracy measurement, to ensure the pointing and imaging accuracy of the entire satellite when performing its mission; It is a random vibration test with a frequency range of 20-2000Hz and a total root mean square acceleration of not less than 6.0 grms, used to simulate the random vibration environment during the launch process; Electromagnetic compatibility (EMC) testing includes radiated emission (RE), radiated susceptibility (RS), and conducted emission (CE) tests, used to verify the stability of the entire satellite in an electromagnetic environment. For noise testing, the total sound pressure level is not less than 140dB, and the frequency range is 31.5-10000Hz, used to simulate the impact of the emission acoustic environment on the structure; For the thermal equilibrium test, used to verify the accuracy of the thermal analysis model, the deviation between the model-calculated temperature and the actual experimental temperature should not exceed ±3℃; The test is a sinusoidal vibration test with a frequency range of 5-100Hz. The acceleration amplitude is determined according to the specific mechanical conditions and is used to verify the response characteristics of the structure under low-frequency excitation.

[0025] After the above 11 test and experimental projects were established, in order to achieve differentiated management and dynamic tailoring control, based on the functional role and risk contribution of each experimental project in the whole satellite verification process, they were divided into three different levels, and different execution strategies and tailoring rules were set for each level.

[0026] First, for the mandatory test items in final assembly, including Electrical performance testing Leak detection test Quality characteristic testing and Accuracy characteristic tests, totaling four items. These items are primarily used to ensure the basic functional integrity, safety, and accuracy of key physical parameters of the entire satellite, and are fundamental test components that cannot be omitted under any circumstances. Therefore, these items are executed 100% at any stage of mass production, without any form of trimming or optimization. In the test matrix, these items are always marked as R, indicating that they must be executed, and their execution status is not affected by subsequent algorithms or state variables.

[0027] Secondly, for ESS filters, including Atmospheric pressure temperature cycling test Thermal vacuum test Random vibration test Electromagnetic compatibility testing and Noise testing, totaling 5 items. The main purpose of this type of test is to expose manufacturing defects, process fluctuations, and potential failure modes by applying environmental stress, which is a typical environmental stress screening method. In the early stages of product development, this type of test needs to be performed frequently to establish a quality baseline. However, as the stability of the production process improves and the technology matures, its frequency can be gradually reduced. Therefore, as a primary target for dynamic tailoring, the execution of this type of test will be jointly determined by the successive pass threshold mechanism and risk constraint mechanism in subsequent steps. In the test matrix, this type of test is marked according to the decision results. (Execute) or - (Skip), and mark it as ER when it is in the strategy transition phase to indicate that the project is in the evaluation decision state.

[0028] Finally, for design / model validation items, including Thermal equilibrium test and Two sinusoidal vibration tests are conducted. These tests primarily verify the accuracy of the design model and the rationality of the structural analysis model, with the verification objective focused on the design phase rather than the mass production phase. Therefore, these tests are only performed in the initial launch phase, i.e. and Executed on the first satellite. This type of test is executed in the first two stars. Once mass production begins, this type of test is generally discontinued to reduce unnecessary resource consumption. In the test matrix, this type of item is marked as "-" by default, indicating it is not executed, and is only restored to the correct mark when a design change occurs. .

[0029] The three-tiered classification structure described above enables differentiated management of different test items while ensuring the overall satellite safety and functional integrity. The mandatory assembly test items constitute the quality baseline, the ESS (Effective Service Detection) screening items handle defect screening and dynamic optimization, and the design / model verification items support design correctness verification. This hierarchical system provides clear controllable objects and decision boundaries for the subsequent interval upgrade mechanism based on continuous pass thresholds, thus ensuring good controllability, interpretability, and engineering feasibility of the entire test matrix optimization process.

[0030] After completing the construction of the testing and experimental project system, in order to reasonably evaluate the effectiveness of various experimental projects and dynamic tailoring strategies, it is necessary to further establish the mapping relationship between experimental projects and their corresponding stress types and detectable fault types, thereby forming a structured stress-fault correlation matrix. This correlation matrix is ​​used to connect experimental behavior with defect exposure capability, and plays a role in providing basic data support and logical constraints throughout the entire dynamic optimization process. Its construction quality directly affects the accuracy and reliability of subsequent experimental effectiveness evaluation and risk control results.

[0031] In the specific construction process, 11 pre-defined test and experimental items were used as row dimensions, and stress type and detectable fault type were used as column dimensions for organization. Among them, stress type is used to describe the environmental excitation conditions applied to the spacecraft structure, electronic system and various functional units during the test. The stress categories are divided into six categories: temperature stress, vacuum stress, vibration stress, electromagnetic stress, acoustic stress and electrical stress. These six types of stress cover the main physical load forms that the spacecraft may experience during the ground test phase and in the on-orbit operation environment, and have strong engineering representativeness and completeness.

[0032] In terms of failure type, the focus is on typical defect types that can be directly triggered or significantly amplified by the aforementioned stresses, including incomplete weld joints, sealing failures, structural resonance, electromagnetic interference, material fatigue cracks, thermal design defects, and electrical short circuits or open circuits. These failure types are derived from long-term engineering practice experience, failure mode and effects analysis (FMEA), and statistical results of historical test data, and can comprehensively reflect the main risk points that may exist in the design, manufacturing, and assembly of the entire satellite.

[0033] In establishing the correlation, it is necessary to clarify the type of stress applied to each test item j, and further determine the types of faults that can be directly triggered or detected under this stress. This process is based on FMEA analysis, systematically identifying potential failure paths under different stress conditions, and statistically summarizing the relationship between stress and faults by combining the distribution of detected defects in historical test data. Simultaneously, it is verified with engineering experience to ensure that the mapping relationship has both theoretical basis and conforms to actual engineering applications.

[0034] For example, thermal vacuum test ( Simultaneously applying temperature stress and vacuum stress, under the combined effects of temperature cycling and a low-pressure environment, can induce connection failure at the weld joints due to thermal expansion and contraction. It can also expose leakage problems in the sealing structure under vacuum conditions, as well as abnormal temperature distribution caused by unreasonable thermal design. Therefore, the detectable fault types corresponding to this test include poor weld joints, sealing failures, and thermal design defects. Another example is the random vibration test (…). It simulates the mechanical environment during launch by using broadband vibration excitation, which can effectively induce structural resonance. At the same time, it accelerates the material fatigue process under repetitive vibration loads. Therefore, the detectable fault types include structural resonance and material fatigue cracks.

[0035] In this correlation matrix, there is a many-to-many relationship between test items and stress types, meaning that a single test item can apply multiple stress types, and the same stress type may be provided by multiple test items. Similarly, there is also a many-to-many relationship between test items and fault types, meaning that a single test can correspond to multiple detectable fault types, and the same fault may be detected by multiple test items. This multidimensional mapping relationship can comprehensively characterize the test system's coverage of different fault types.

[0036] In the subsequent evaluation of test effectiveness, this correlation matrix plays a crucial role in screening and attribution constraints. Specifically, a defect is only included in the effectiveness statistics of a test project if it is exposed during the execution of the corresponding test project and the defect type falls within the detectable fault range corresponding to that test project. If a defect is discovered during a test but its nature does not belong to the fault type that can be directly induced by the stress applied in the test, then the defect is not included in the effectiveness calculation of the test project.

[0037] This constraint mechanism effectively avoids statistical interference between different tests, ensuring that the effectiveness evaluation of each test reflects only its true defect detection capability, without bias caused by indirect findings from other tests. This is significant for subsequent adjustments to test execution frequency based on effectiveness indicators. When a test item fails to detect defects within its detectable fault range for an extended period, it indicates that the test's marginal detection capability at the current product quality level is low, serving as an important basis for reducing its execution frequency. Conversely, when a test consistently detects related types of defects, it indicates that the test remains highly necessary at the current stage and should maintain a high execution frequency.

[0038] Furthermore, this correlation matrix provides fundamental support for subsequent risk assessment. In calculating the risk of missed detections, the contribution of each test item to the overall risk is directly related to the range of failures it can cover, and this range is defined by the correlation matrix. Therefore, the correlation matrix not only determines the data boundaries of the validity statistics but also directly affects the physical meaning of each parameter in the risk model and the rationality of the calculation results.

[0039] By constructing a complete stress-fault correlation matrix, a systematic mapping between test items and defect types is achieved, providing a unified data foundation for subsequent test effectiveness modeling, execution interval adjustment, and risk constraint control. This enables the entire dynamic optimization process to have a clear logical structure, reliable data support, and good engineering applicability.

[0040] After completing the construction of the test project system and establishing the stress-fault correlation matrix, it is necessary to obtain a comprehensive technology maturity index. This index is derived from the technology maturity assessment results of the previous stage and is used to quantify the overall technical status of mass-produced satellites. Technology Maturity Index This indicator reflects the maturity of the overall product design, manufacturing, and processes. A higher value indicates a more stable overall quality and a lower probability of defects, while a lower value indicates that the product is still in an unstable stage and faces higher risks. This indicator serves as a benchmark for the overall quality level and provides crucial input for subsequent calculations of test effectiveness.

[0041] Based on this, the effectiveness index of the trial is calculated for each ESS screening trial item. This is used to characterize the defect detection capability of the test in the current batch production stage. The test effectiveness index is calculated using the Bayesian estimation method, and its formula is as follows: The meanings of each parameter are as follows: As of the When the satellite is launched, the test project The cumulative number of defects detected that fall within the detectable range of the test; Experimental Project The cumulative number of executions; and Prior parameters are used to incorporate historical data or empirical knowledge.

[0042] In the prior parameter settings, for experimental projects with abundant historical data, take... 、 The corresponding prior validity is approximately For projects lacking historical data, take the no-information prior. 、 This is to ensure that the initial estimate is not biased towards any conclusion.

[0043] The physical meaning of the test effectiveness index is: in the execution of a single test item Under the given conditions, if a defect exists in the product that can be detected by this test, this is an estimated probability that the defect will be detected by the test. (Index value) The larger the value, the more defects the test can detect, indicating that it still has high value under the current product quality condition; the smaller the value, the fewer defects are detected under the current product quality level, and the lower the risk of the test.

[0044] It is important to emphasize that in the calculation of test validity, only defects falling within the detectable failure range of the test are included in the validity statistics. These ranges are provided by the stress-failure correlation matrix established in the previous stage. This ensures that the validity index reflects the true value of the test itself, without being influenced by defects indirectly discovered by other tests.

[0045] Integrated Technology Readiness Index As a global input, and with the experimental validity indicators These factors work together to influence subsequent dynamic pruning decisions. higher and When the level is low, the frequency of this test can be safely reduced; when lower or When the level is high, it is necessary to maintain or increase the frequency of test execution to ensure that critical defects can be exposed in a timely manner.

[0046] This modeling method exhibits good dynamic adaptability: as the number of batch-produced satellites increases, the cumulative number of executions also increases. With growth and increasingly rich statistical data, the effectiveness indicators of the tests gradually converge to the actual detection capability. At the same time, the technology maturity index, as the global quality input obtained in the previous stage, ensures that the tailoring decision for each satellite can take into account the overall product status, and achieve coordination between the local test value and the global quality level.

[0047] By introducing technology maturity indicators With test effectiveness indicators The dual modeling not only quantifies the defect detection capability of individual tests, but also enables dynamic adjustment of the overall technical status of batch-produced satellites, providing complete data support and decision-making basis for subsequent continuous threshold pruning, interval upgrades, and risk constraints.

[0048] During the dynamic optimization of the batch satellite test matrix, for each ESS screening test item (including the number) This introduces a bivariate management system using interval level and consecutive pass count to dynamically adjust the test execution frequency and pruning strategy. Each test item maintains two key state variables: interval level. Number of consecutive passes This is used to record the historical execution status and consecutive successes of experiments within the current production sequence. Interval levels. The initial value is set to 0, and the range of values ​​is [value range missing]. This represents the interval level of the test execution. Number of consecutive passes The initial value is 0, used to count the number of consecutive defects not found at the current interval level. These two state variables work together to support the interval escalation decision logic based on the consecutive pass threshold, enabling dynamic adjustment of the test execution frequency.

[0049] The mapping relationship between interval levels and execution intervals is defined as follows: (This is performed on every satellite) (Administer 1 out of every 2) (Administer 1 out of every 5) (Administer 1 out of every 10) (Administer 1 dose for every 25 doses) (Administer 1 dose for every 50 doses) (Administer 1 dose for every 100 doses) Under this mapping relationship, the higher the interval level, the longer the test execution interval, thus achieving a smooth transition from low-risk, high-frequency tests to low-frequency, economical tests. Each test item also has a continuously pass threshold set. and the maximum allowed number of intervals These two parameters control the upgrade conditions and the maximum trimming range. The specific parameter settings are as follows: (Maximum spacing 50) (Maximum spacing 25) (Maximum spacing 50) (Maximum spacing 100) (Maximum spacing 100) Decision-making for the execution of tests on each satellite The interval level and the number of consecutive passes are jointly determined. The specific rules are as follows: 1. Execute judgment: in This is the execution interval corresponding to the current interval level.

[0050] 2. Continuously pass the update and upgrade conditions when Furthermore, when the test passes (no defects found), the consecutive pass count is updated: If the upgrade conditions are met Then the level will be upgraded at intervals: Simultaneously reset the consecutive pass count: When the test fails the treatment And if the test fails (a defect is found), it should be rolled back immediately: Revert to the state of full inspection for each chip, record defects in the stress-failure correlation matrix, and update the test validity. .

[0051] Skip processing when At that time, the state variables remain unchanged: The core idea of ​​this mechanism is to utilize the continuous passing threshold. The speed of upgrading the interval level can be controlled. A larger threshold requires more consecutive passes to upgrade the interval level, resulting in a more conservative strategy and higher safety margin, but a slower load-shear speed. Conversely, a smaller threshold allows for faster upgrades and cost savings, but reduces the safety margin. In engineering practice, high-consequence projects (such as thermal vacuum tests) are given larger thresholds to ensure sufficient safety redundancy, while low-consequence projects (such as noise tests) can be given smaller thresholds to improve efficiency and save costs.

[0052] This mechanism allows the execution frequency of each ESS screening test project to be adaptively adjusted based on historical consecutive pass rates, achieving closed-loop control of dynamic pruning while balancing safety and resource optimization. In mass production, this mechanism can significantly reduce unnecessary repeated tests while ensuring that high-risk projects are continuously and adequately validated, enabling refined management and automated optimization of test execution.

[0053] After completing the division of the test project system, establishing the stress-fault correlation, and determining the test effectiveness and interval escalation mechanism, it is necessary to further apply the above model to the actual mass production process, making decisions on the test execution plan for each satellite on a satellite-by-satellite basis, thereby forming a dynamic tailoring rule that can be implemented. This rule takes satellite-by-satellite decision-making as its core, and by combining test category attributes, interval level status, and execution judgment results, it realizes the dynamic generation and real-time updating of the test matrix.

[0054] In the initial production phase, both the first and second satellites underwent all 11 tests and experiments without any modifications. The primary purpose of this phase was to establish an initial baseline for test effectiveness, providing raw data support for subsequent statistical analysis. By fully executing all tests, the defect detection rates for various tests in the initial batch could be obtained, thus providing data for subsequent test effectiveness indicators. The calculations provide reliable input while ensuring that the interval upgrade mechanism has a sufficient data foundation in its initial state. During this phase, all experimental items are marked as being in an execution state, used to form a complete experimental matrix baseline.

[0055] Starting with the third satellite, the satellite-by-satellite pruning rule officially took effect. At this point, the test execution plan for each satellite was no longer fixed, but dynamically determined based on the test project category and the current status of the interval upgrade mechanism. The processing logic for different categories of test projects differed significantly.

[0056] For mandatory assembly tests, including electrical performance testing, leak detection testing, quality characteristic testing, and accuracy characteristic testing, these tests bear the fundamental responsibility for verifying the overall satellite's functional integrity and safety, and are irreplaceable key test contents. Therefore, throughout the entire mass production process, these tests are always executed 100%, unaffected by any tailoring strategies or state variables. During the satellite-by-satellite decision-making process, these tests are directly marked as executed, and their existence constitutes the quality baseline of the testing system, ensuring that the basic functional verification capability will not be weakened by tailoring under any circumstances.

[0057] For ESS screening items, including ambient temperature cycling, thermal vacuum, random vibration, electromagnetic compatibility, and noise testing, their execution is dynamically determined by the interval escalation mechanism. On each satellite, for each test item j, based on the current interval level... Calculate the corresponding execution interval The execution status is determined based on the following decision relationship: when When, the corresponding test item is executed; when When the interval changes, the corresponding test item is skipped. This rule allows the test execution frequency to be automatically adjusted according to the interval level, thereby achieving a gradual transition from high-frequency screening to low-frequency verification.

[0058] In actual implementation, a special state exists where the experimental project has not yet accumulated enough consecutive pass records after completing the interval upgrade; at this point, the experiment is in a strategy transition phase. During this phase, although skipping may be allowed according to the interval rules, sufficient statistical stability has not yet been achieved, and judgment still needs to be made based on engineering experience. Therefore, an evaluated execution state is introduced to identify this type of experiment as being in an intermediate state where it can be executed or skipped. The existence of this state allows the model to maintain a balance between rigorous algorithmic decision-making and practical engineering needs, improving the overall strategy's flexibility and controllability.

[0059] Design verification projects, including thermal equilibrium tests and sinusoidal vibration tests, are primarily used to verify the correctness of the design model and are typically no longer performed during mass production to reduce unnecessary resource consumption. During the star-by-star tailoring process, these projects are in a default "not executed" state, meaning they are not included in the regular testing process. However, these tests are not permanently cancelled but are conditionally reinstated. When design changes occur during subsequent production, the relevant test projects will be reinstated to ensure that design changes do not introduce new risks.

[0060] By employing the aforementioned star-by-star tailoring rules, differentiated control can be achieved for different categories of test items: mandatory assembly test items ensure basic quality baselines; ESS screening items achieve cost optimization and risk balance through a dynamic interval mechanism; and design verification items can be resumed when necessary to ensure design consistency. These three types of items operate collaboratively within a unified framework, giving the test system both rigid constraints and dynamic adjustment capabilities.

[0061] The core advantage of this satellite-by-satellite decision-making mechanism lies in transforming static experimental plans into dynamic evolutionary processes. The experimental configuration for each satellite is decided independently based on historical execution results, current state variables, and experimental effectiveness indicators, thus avoiding the inefficiencies or uncontrolled risks associated with a one-size-fits-all, fixed sampling strategy. Simultaneously, this mechanism can automatically adjust the experimental intensity as mass production progresses, maintaining high-intensity verification in the initial stages and gradually reducing the experimental frequency in later stages, maximizing resource utilization efficiency.

[0062] Furthermore, by introducing three status indicators—execution, skipping, and evaluated execution—into the experiment matrix, the final output results possess excellent visualization and operability. On the production floor, corresponding experiments can be executed directly based on the matrix indicators without the need for additional analysis of complex parameters, thereby significantly reducing execution complexity and improving overall production efficiency.

[0063] In summary, by constructing a star-by-star iterative pruning rule, an effective connection between model decision-making and engineering execution is achieved, enabling the experimental optimization strategy to operate stably in a mass production environment and achieving continuous optimization of experimental resources while ensuring quality and safety.

[0064] During the dynamic optimization of the mass production satellite test matrix, the test execution strategy mainly relies on historical data and statistical patterns for adjustment. However, when design-level changes occur, the original tailoring strategy based on historical stability will lose its effectiveness. Therefore, it is necessary to introduce a design change triggering mechanism to forcibly intervene in the test execution status to ensure that quality risks under the new design status can be fully identified and verified.

[0065] When satellite design or configuration changes, the system automatically identifies affected test items based on pre-established association rules between change types and test items. These test items are then restored from their current pruned state to an operational state, and relevant state variables are reset, allowing the test to re-enter a high-intensity verification phase. This mechanism essentially provides interruption and restart control for the interval upgrade strategy, enabling the test strategy to quickly revert to a conservative mode when design changes occur, thereby avoiding the risk of missed detections due to invalid historical data.

[0066] Different types of design changes are handled using a categorized triggering approach. For thermal control-related changes, including adjustments to thermal control coatings, changes in heat sink area, optimization of heat pipe paths, and adjustments to heater power, these changes directly affect the overall thermal environment distribution and thermal equilibrium state of the satellite. Therefore, it is necessary to resume key thermal environment-related tests, including thermal equilibrium tests, thermal vacuum tests, and atmospheric pressure temperature cycling tests. Simultaneously with the resumption, the interval levels of the corresponding test items are reset to their initial states. This operation allows the relevant experiments to be performed on every satellite, thereby quickly establishing the experimental data foundation under the new design conditions.

[0067] For structural changes, including modifications to the main load-bearing structure, adjustments to the solar array mechanism, optimization of the deployment and locking mechanism, and redesign of the mounting bracket, these changes significantly affect the overall satellite's dynamic characteristics and structural response. Therefore, it is necessary to reinstate test items related to structural reliability, including sinusoidal vibration tests, random vibration tests, and noise tests. Simultaneously, the interval levels for random vibration and noise tests will be reverted, i.e.: This allows it to re-enter a high-frequency execution state, in order to fully verify the impact of the structural changes.

[0068] For electrical changes, including replacement of individual equipment, cable network adjustments, changes to power supply and distribution schemes, and modifications to radio frequency links, these changes directly impact electromagnetic compatibility (EMC) and electrical performance. Therefore, it is necessary to reinstate EMC testing and electrical performance testing. While electrical performance testing is a standard procedure, additional retesting is required in this case to strengthen the verification of the changes' impact. Simultaneously, the interval level for EMC testing should be rolled back. This ensures that electromagnetic-related risks can be fully exposed within a short period of time.

[0069] In practical engineering, design changes often involve multiple categories simultaneously, such as structural adjustments and electrical modifications. In such cases, it is necessary to merge the test items corresponding to each type of change, that is, to take the union of all triggered test items and uniformly resume their execution. This approach can avoid overlooking any potential risk points and ensure the completeness of test coverage.

[0070] The execution strategy following the change is not to maintain a high-intensity state permanently, but rather to be phased. Specifically, the resumed test projects are only enforced on the satellite in which the change occurred and its first subsequent satellite, in order to quickly obtain initial verification data under the new design state. Starting with the second subsequent satellite, the interval upgrade process based on continuous pass thresholds is reinstated, i.e., the normal dynamic pruning mechanism is restored. Through this strategy of short-term reinforcement plus long-term recovery, safety is ensured while avoiding the waste of long-term test resources.

[0071] By introducing a design change triggering mechanism, the experimental strategy achieves rapid response to design changes, enabling the dynamic tailoring system to adapt not only to changes in quality status but also to changes in design status, thus forming a complete closed-loop control system. This mechanism, in conjunction with the interval upgrade strategy, experimental effectiveness evaluation, and risk constraint mechanism, ensures that the entire experimental matrix optimization process maintains good stability and reliability even in complex engineering environments.

[0072] During the dynamic optimization of the mass production satellite test matrix, the execution frequency of ESS screening tests will gradually decrease as the interval upgrade mechanism continues to operate. Some tests may be skipped multiple times over a long period or within a large production volume. Although this process is reasonable based on historical data and continuous pass status, without additional constraints, it may lead to a lack of comprehensive verification for a long period, thus creating potential systemic risks. Therefore, it is necessary to introduce a spot check fallback mechanism to periodically enforce all tests and experimental items, ensuring that the necessary full-coverage verification capability is maintained even under long-term reduction conditions.

[0073] The mechanism employs a dual constraint strategy, imposing constraints on the comprehensive testing from both the production quantity and time dimensions to ensure the stability and safety of the testing system under different production rhythms.

[0074] In terms of production quantity, a fixed-interval sampling rule is introduced. In a continuously produced satellite sequence, at each interval... Of the 11 satellites, at least one must be selected to perform all 11 tests and experiments. Parameters The value range is set between 20 and 100, with a recommended value of [value missing]. Under this rule, a full inspection is performed on the corresponding satellite when the production serial number satisfies the following relationship: This means that for every 50 satellites produced, the 50th, 100th, 150th, and so on satellites undergo a complete test. This rule ensures that as production scales up, the system is consistently validated at a fixed frequency, allowing for the timely detection of potential defects that might not have been exposed during the tailoring process.

[0075] In terms of time, a periodic sampling rule is introduced. Regardless of how the current test interval level changes, every interval... Each month, at least one satellite will be selected to perform all tests and experiments. Parameters The value range is set between 3 and 12 months, with a recommended value of [value missing]. Months. This rule can cover situations where production rhythm is unstable or production is interrupted, and can ensure regular and comprehensive verification over time, even when satellite output is low or production cycle fluctuates greatly.

[0076] The two rules mentioned above are combined using an OR logic, meaning that when either rule is satisfied, the corresponding satellite performs all tests and experiments. This combination approach simultaneously addresses constraints from both production scale and timeframe, making the sampling strategy more robust. When production is fast-paced, the quantity rule plays a dominant role; when production is slow-paced, the time rule provides supplementary assurance, thus preventing situations where comprehensive verification is neglected for extended periods.

[0077] During the spot check as a fallback, all tests and trials resume execution, essentially temporarily reverting the test matrix to full inspection mode. This full inspection is not isolated but continues to participate in the interval upgrade mechanism's state update process. When the spot check result is a pass, the consecutive pass count for the corresponding test item continues to accumulate, i.e.: If a defect is found during the sampling inspection, a rollback mechanism is triggered, and the interval level of the corresponding test item is reset to the initial state, that is: Simultaneously, the number of consecutive passes is reset to zero, allowing the experiment to re-enter the high-frequency execution phase. In this way, the sampling results not only serve a verification purpose but also directly influence subsequent experiment execution strategies, achieving a closed-loop linkage with the interval upgrade mechanism.

[0078] The introduction of a spot-check fallback mechanism ensures a stable safety boundary for the dynamic trimming system during long-term operation. Even with a large number of trials being trimmed, potential problems can still be identified promptly through periodic full inspections, preventing risk accumulation. Furthermore, this mechanism complements the continuous threshold mechanism: the former reduces costs and improves efficiency, while the latter controls long-term risks. Together, they achieve a balance between optimizing experimental resources and ensuring quality and safety.

[0079] By constructing a sampling strategy based on both quantity and time constraints, the blind spots that may be caused by a single pruning mechanism can be effectively avoided, ensuring that the test matrix maintains the necessary verification strength during dynamic changes, thereby guaranteeing the reliability and consistency of mass-produced satellites throughout their entire life cycle.

[0080] With the test intervals gradually increasing and some ESS screening tests being skipped, it is necessary to uniformly constrain and control the overall missed detection risk of each satellite under the current test configuration to prevent key defects from going unidentified due to excessive pruning. A risk assessment model based on probability combination is introduced to quantitatively verify the current test execution plan. The missed detection risk is calculated using the following expression: This expression combines the risk contributions of multiple test items in a product form, reflecting the overall probability of missed detections under different test combinations. When some tests are skipped, the risk contribution of the corresponding item is activated and accumulated through a product, thus reflecting the cumulative effect when multiple tests are simultaneously missing. To ensure system security, a uniform upper limit is set for the risk of missed detections: This means that the risk of missed detections for a single satellite under the current experimental execution plan must not exceed 2%. This threshold, as a hard constraint, must be verified after each experimental tailoring decision. When the calculation results meet... At that time, the current experimental plan was considered to have controllable risks and could be implemented directly; when If this occurs, it indicates that the current degree of experiment trimming is too high, and the experiment execution status needs to be backtracked and corrected.

[0081] The risk correction process is based on the ranking of individual trial risk contributions. First, the independent risk contribution value is calculated for all skipped trials: Independent Risk Contribution Value This is used to measure the impact of a single test item on the overall risk of missed detections in the current state. Then, according to... Sort them from largest to smallest to identify the pilot projects with the greatest risk impact.

[0082] Based on the sorting, starting with the project with the largest contribution value, the execution status of the corresponding experimental projects is restored one by one, that is... Adjust to execution status. Recalculate the overall risk after each trial is resumed. And continuously determine whether the constraints are satisfied: When the risk decreases to within a certain threshold, the recovery operation is stopped, ensuring that only necessary test items are recovered. This avoids resource waste caused by excessive rollback. Recovered test items are marked as "execution status" in the test matrix and distinguished by additional identifiers to indicate that the execution behavior stems from risk constraint triggering, rather than being a natural result of the interval escalation mechanism itself. This identifier method facilitates subsequent data analysis and strategy optimization.

[0083] Through the above mechanism, real-time risk closed-loop control can be introduced into the dynamic pruning process, enabling the experimental execution strategy to not only rely on historical statistical results but also to be adjusted in real time according to the current risk level. This method avoids the over-pruning problem that may result from simply relying on interval upgrades, while not disrupting the continuity of the overall optimization strategy, thus maintaining a stable balance between efficiency and safety in experimental execution.

[0084] From an overall operational perspective, the risk constraint mechanism is equivalent to adding a safety verification layer outside the dynamic tailoring system. All experimental tailoring results must be verified by this layer before they can take effect. This design ensures that regardless of changes in the frequency of experiment execution, the final output experiment matrix always meets the predetermined risk control requirements, thereby improving the reliability and controllability of the batch production process.

[0085] After completing various experimental execution decisions, interval upgrade status updates, and risk constraint verifications, all decision results need to be uniformly summarized and output in the form of a standardized experimental matrix. This matrix serves as the final expression of the dynamic pruning strategy, transforming complex decision logic into directly executable operational instructions, thereby achieving effective connection between algorithmic decision-making and engineering implementation.

[0086] The matrix structure is organized in a two-dimensional table format. Rows are arranged according to test and experiment items, containing 11 items, sorted by priority from highest to lowest, making key test items more visible within the matrix. Columns are expanded according to satellite production serial numbers, such as #1, #2, #3, etc., with each column corresponding to a specific satellite. This structure allows for a clear visual display of the execution status of different satellites on different test items, enabling unified management across batches and projects.

[0087] Each cell in the matrix represents the execution decision of a satellite in a certain test project. Three standardized identifiers are used to express this information, thereby achieving simplification and consistency.

[0088] The first type of identifier is R (Required), indicating that it must be executed. This identifier indicates that the test item is a mandatory item on the current satellite and must be executed regardless of changes in other state variables or pruning strategies. This type of identifier typically corresponds to critical tests such as mandatory assembly test items, design change triggered items, and risk constraint recovery items.

[0089] The second type of label is ER (Evaluated Required), indicating that the project has been evaluated and implemented. This label describes pilot projects in the transition phase, where the project meets the conditions for skipping certain steps but has not yet fully met the stable cutoff criteria. Therefore, judgment needs to be made based on engineering experience. This type of label reflects the coordination between model decisions and human experience, and can be flexibly adjusted according to the actual situation during implementation.

[0090] The third type of identifier is — (NotRequired), indicating that execution is not required. This identifier is used to indicate that the experimental project has been cut from the current satellite and can be skipped. This status usually originates from the interval upgrade mechanism or the default cut-off status of design verification projects during the mass production phase.

[0091] By combining the above three types of identifiers, the execution status of all test items can be fully expressed in a single matrix, while maintaining conciseness and easy identification, which is conducive to rapid understanding and execution on site.

[0092] Along with the matrix output, key status information should be appended below the matrix to aid in interpreting the current decision results. This information includes the current interval level status and consecutive pass count for each ESS screening test item, reflecting the dynamic process of test tailoring. Additionally, the current satellite's missed detection risk assessment result should be indicated, along with whether special status information such as design change recovery mechanisms, spot check fallback mechanisms, or risk constraint mechanisms have been triggered. These additional explanations allow for the correlation between the execution results in the matrix and their causes, enhancing the traceability and interpretability of the decision.

[0093] This matrix has excellent engineering applicability and can be directly output to the AIT site as a production execution document. Operators do not need to understand complex algorithm models; they only need to perform the corresponding tests and experiments based on the corresponding identifiers in the matrix to complete the entire satellite testing process. This approach significantly lowers the execution threshold, avoids operational errors caused by human misunderstanding, and improves the consistency and efficiency of test execution.

[0094] Furthermore, this matrix format supports historical data recording and retrospective analysis. By comparing matrices from different batches, the evolution of experimental pruning strategies can be analyzed, and changes in quality status at different stages can be assessed, providing data support for subsequent strategy optimization.

[0095] By outputting the dynamic trimming results in a standardized matrix form, an effective transformation from a multi-dimensional decision-making model to an engineering execution interface is achieved, enabling complex experimental optimization strategies to have a clear, intuitive, and operable representation, thereby ensuring the feasibility and stability of the entire batch production testing system in practical applications.

[0096] To more clearly illustrate the implementation process of this method, a specific application scenario will be used as an example. The following explanation uses the mass production of low-Earth orbit communication constellations as an example.

[0097] The constellation plan calls for the mass production of 1,000 satellites, at a production rate of 10 per month. (Based on thermal vacuum testing...) This example illustrates the interval upgrade mechanism and overall process.

[0098] I. Test Item System and Parameter Settings A complete experimental project system was constructed, comprising 11 tests and experiments, with specific parameters as follows: Table 1: Test Item Parameters; II. Stress-Fault Correlation The stress-fault mapping relationship is established for ESS screening projects as follows: Table 2: Stress-fault correlation (partial); III. Example of Interval Upgrade Mechanism Operation (Thermal Vacuum Test) Setting parameters: , , ; Initial state: 、 .

[0099] The execution process of the first 20 satellites (key milestones); #1–#5: Executed consecutively and passed; →The 5th one is satisfied ,upgrade; → , ; #6: Execute (6mod2=0), via → ; #7: Skip (7mod2≠0) #8: Execute → #10: Execution → #12: Execute → #14: Execute → Meet the upgrade requirements → , ; #15: Execute (15mod5=0) → #16–#19: Skip #20: Execution → Defect Found (Seal Leakage) Rollback: 、 Full inspection will resume starting with #21.

[0100] Table 3: Complete process of thermal vacuum interval upgrade; Rollback: 、 Full inspection will resume starting with #21.

[0101] IV. Example of Risk Constraint Check Taking satellite #16 as an example: Given: Weight: Execution status: thermal vacuum skip, random vibration skip, noise skip.

[0102] Risk contribution calculation: Total Risk: Exceeded the threshold of 0.02; Correction strategy: Restore ; The constraints are satisfied.

[0103] V. Example of Design Change Triggering Assuming a change occurs in the thermal control design of satellite #35: Heat dissipation area: ; Triggering result: Thermal balance restoration is performed; , 、 Go back to: ; And restart the interval upgrade.

[0104] VI. Example of Dynamic Clipping Matrix Output Table 4: (#1–#10); VII. Cost Savings: Original Full Inspection Cost: After optimization: ESS cost: 152 million → 15 million; Total cost: Approximately 29 million; save: ; at the same time: .

[0105] This invention constructs a test project system covering the entire process of spacecraft assembly, integration, and testing. It integrates key assembly-stage projects such as electrical performance testing, quality characteristic testing, and accuracy characteristic testing into a unified tailored management framework, achieving a shift from traditional single-environment test optimization to full-scale, star-level process optimization. Based on this, it introduces an interval upgrade and rollback mechanism based on a continuous pass threshold. This allows the test execution strategy to be dynamically adjusted star-by-star according to changes in product quality status. Under continuous pass conditions, the execution frequency is automatically reduced, and when defects are detected, it quickly rolls back to a full inspection state, thus balancing test efficiency and quality safety, demonstrating excellent adaptability and engineering applicability.

[0106] This invention further achieves refined control of the test recovery strategy through a design change triggering mechanism. It automatically identifies affected test items based on different types of design changes and recovers the relevant tests accordingly, avoiding the problem of indiscriminate recovery of all test items in traditional methods. Simultaneously, by introducing a sampling inspection fallback mechanism, it constrains test execution from two dimensions: production quantity and time cycle. Even with significant reductions in test volume, it can still periodically perform full-item inspections, thereby establishing a stable quality assurance baseline and ensuring that the testing system does not accumulate risks during long-term operation.

[0107] This invention also establishes a risk constraint mechanism based on a probabilistic model, quantitatively assessing the risk of missed detections for each satellite and ensuring that the risk remains within a controllable range through threshold control. When the risk exceeds a set threshold, the execution of high-risk test items is automatically resumed, forming a closed-loop safety control system. Simultaneously, all decision results are output in a standardized R / ER / - test matrix format, transforming complex test optimization strategies into intuitive and executable operational instructions that can be directly applied to the production site, significantly reducing execution complexity and improving consistency and efficiency in mass production processes.

[0108] This invention provides, for example Figure 2 The system described is a dynamic optimization system for a batch satellite test matrix based on a continuous pass threshold. It is characterized by including a project management module, a status tracking module, an interval upgrade engine, a change triggering module, a risk constraint module, and a matrix output module. The project management module is used to manage parameter information for all star-level tests and experiments. Parameter information includes the project number. Project name, priority sorting, function positioning, consecutive pass threshold Maximum Interval Level Experimental costs Test cycle and consequences weight ; The status tracking module is used to maintain the interval level L_j(n) and the number of consecutive passes for ESS screening test items. And record the execution status and test results of the corresponding test items for each satellite; An interval upgrade engine is used to update test results based on the state tracking module and the corresponding consecutive pass thresholds. For interval levels Perform upgrade or rollback calculations and output the execution decision for the next satellite's corresponding test project. ; The change triggering module is used to receive design change information, identify the affected test items according to the change type, adjust the execution decision of the corresponding test item to the execution state, and revert the corresponding interval level to the initial state. The risk constraint module is used to calculate the current satellite's risk of missed detections. And when the risk of missed detection exceeds the preset threshold At that time, according to the risk contribution value of each test item Based on the sorting results, the execution status of each test item is restored until the risk constraints are met; The matrix output module is used to generate an experimental matrix based on the execution decisions of each experimental item and output it in a preset identifier format to support experimental execution.

[0109] The present invention provides a method for dynamic optimization of a batch satellite test matrix based on a continuous pass threshold, which is implemented by the aforementioned dynamic optimization system for a batch satellite test matrix based on a continuous pass threshold. For details of the specific method and process of the dynamic optimization system for a batch satellite test matrix based on a continuous pass threshold, please refer to the aforementioned embodiment of the method for dynamic optimization of a batch satellite test matrix based on a continuous pass threshold, which will not be repeated here.

[0110] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A dynamic optimization method for a batch satellite test matrix based on a continuous pass threshold, characterized in that, Includes the following steps: Construct a set of test and experiment items covering the entire process of spacecraft assembly, integration and testing, and divide the set of test and experiment items into mandatory assembly test items, ESS screening items and design verification items according to their functions; Establish a set of associations between test items and stress types and detectable fault types based on the set of test and experimental items; Based on the set of relationships, a comprehensive technology maturity index is obtained, and the corresponding test effectiveness index is calculated by combining the cumulative number of executions and the cumulative number of defects detected for each ESS screening project. For each ESS screening item, an interval level and consecutive pass count are introduced. The interval level and consecutive pass count are updated based on the test effectiveness indicators and test execution results. The interval level determines whether the current satellite should execute the corresponding test item. If the test is passed, the interval level is increased. If the test is failed, the interval level is reverted to the initial state. Based on the interval level, test execution decisions are generated for each satellite. The initial satellite executes all tests and experimental items. Subsequent satellites continue to execute mandatory assembly test items and determine whether to execute or skip ESS screening items according to the interval level. When a design change occurs, the execution of the corresponding test items is restored according to the change type and the interval level is simultaneously rolled back. For each satellite, the risk of missed detection is calculated based on the test effectiveness index, the comprehensive technology maturity index, and the skipped ESS screening items. When the risk of missed detection exceeds the preset threshold, the execution of the corresponding test items is restored in order of risk contribution. Combined with the periodic sampling rules, a test matrix is ​​generated with the results of execution, evaluation execution, or non-execution as identifiers. The experimental frequency is adaptively adjusted by combining the experimental results with historical status information. The steps are as follows: Initialize the interval level and consecutive pass count for each ESS screening test item, and establish a corresponding status record table; The test execution interval is determined based on the current interval level, and the execution decision for the corresponding test item is calculated according to the satellite serial number; After the test is completed, the consecutive pass count is updated according to the test results. When the test result is a pass, the consecutive pass count is accumulated and compared with the consecutive pass threshold. When the number of consecutive passes reaches a threshold, the interval level is increased; when the test result is a failure, the interval level is reset and the number of consecutive passes is cleared to zero.

2. The method for dynamic optimization of a batch satellite test matrix based on a continuous pass threshold as described in claim 1, characterized in that, To address the differences in execution attributes of pilot projects during the batch production process, a stratified organization was implemented, following these steps: All test and experiment items are divided into mandatory assembly test items, ESS screening items, and design verification items according to their functional attributes, and corresponding classification labels and management parameters are established for each type of item. Fixed execution rules are configured for mandatory test items in the final assembly, and the execution status is directly marked during the generation of each satellite test plan; Set consecutive pass thresholds and maximum interval levels for ESS filtering items, and establish a mapping relationship between the execution interval and the interval level; For design verification projects, set the initial execution scope and default state for the mass production stage, and adjust the execution flag according to the design state during the test matrix generation process.

3. The method for dynamic optimization of a batch satellite test matrix based on a continuous pass threshold as described in claim 2, characterized in that, Establish constraint rules based on the correspondence between test items and defect exposure capabilities, following these steps: For each test item, a set of corresponding stress types is established, and each stress type is classified and labeled. Determine the corresponding set of detectable fault types based on the stress type, and establish a mapping relationship between test items and fault types; During the defect recording process, each defect is categorized and identified, and it is determined whether it falls within the detectable fault range of the corresponding test item. In the validity statistics, only defect data that meets the mapping relationship is included in the statistics, and defect data that does not meet the conditions are removed.

4. The method for dynamic optimization of a batch satellite test matrix based on a continuous pass threshold as described in claim 1, characterized in that, A preset mapping relationship is established between the interval level and the execution interval. During the execution decision generation process, the execution interval corresponding to the interval level is periodically determined. At the same time, when the number of consecutive passes meets the consecutive pass threshold condition, the interval level is incremented. When the test result is a failure, the interval level is rolled back and the number of consecutive passes is cleared simultaneously, forming a stable state update rule.

5. The method for dynamic optimization of a batch satellite test matrix based on a continuous pass threshold as described in claim 1, characterized in that, The steps for responding to the impact of design state changes on the test execution strategy are as follows: Receive design change information and classify and label the changes as thermal control, structural or electrical changes; Based on the pre-defined correspondence between change types and test items, determine the set of test items that will be affected; Adjust the execution decision status of the affected test items and mark the corresponding test items as execution status; The interval levels of the above test items were rolled back, and their consecutive pass records were updated simultaneously.

6. The method for dynamic optimization of a batch satellite test matrix based on a continuous pass threshold as described in claim 5, characterized in that, The long-term cutting process is compensated by introducing a periodic full inspection strategy, and the steps are as follows: Set sampling interval parameters based on satellite production serial numbers and establish corresponding serial number judgment rules; For each satellite, determine whether its sequence number meets the sampling interval conditions. If the conditions are met, mark all test items as being in execution status. Set sampling cycle parameters based on the time dimension and record production time information; The current production time is assessed, and all test items are executed when the preset time cycle conditions are met, and the sampling inspection records are updated.

7. The method for dynamic optimization of a batch satellite test matrix based on a continuous pass threshold as described in claim 6, characterized in that, The sampling interval parameter and the time period parameter each correspond to independent triggering conditions. When any triggering condition is met, all test items will be marked as executed, and the execution records and status indicators of the corresponding test items will be updated synchronously.

8. The method for dynamic optimization of a batch satellite test matrix based on a continuous pass threshold as described in claim 6, characterized in that, The following steps are taken to quantitatively control potential risks during the experimental cutting process: Calculate the risk of missed detection based on the current satellite test execution status and relevant parameters, and record the calculation results; Compare the risk of missed detection with the preset risk threshold to determine whether the current test plan meets the constraints. When the risk exceeds the threshold, the corresponding risk contribution value is calculated for the unexecuted test items, and they are sorted according to the contribution value; The execution status of each test item is adjusted according to the ranking results, and the risk is recalculated after each adjustment until the constraints are met.

9. A dynamic optimization system for a batch satellite test matrix based on a continuous pass threshold, used to implement the dynamic optimization method for a batch satellite test matrix based on a continuous pass threshold as described in any one of claims 1-8, characterized in that, It includes a project management module, a status tracking module, an interval upgrade engine, a change triggering module, a risk constraint module, and a matrix output module: The project management module is used to manage parameter information for full-star testing and experimental projects; The status tracking module is used to maintain the interval level and consecutive pass count for ESS screening test items, and to record the execution status and test results of the corresponding test items for each satellite; The interval upgrade engine is used to perform upgrade or rollback calculations on the interval level based on the test results recorded by the state tracking module and the corresponding continuous pass threshold, and output the execution decision for the test item corresponding to the next satellite. The change triggering module is used to receive design change information, identify the affected test items according to the change type, adjust the execution decision of the corresponding test item to the execution state, and revert the corresponding interval level to the initial state. The risk constraint module is used to calculate the current satellite's missed detection risk, and when the missed detection risk exceeds the preset threshold, it restores the execution status of each test item one by one according to the ranking of the risk contribution values ​​of each test item until the risk constraint conditions are met. The matrix output module is used to generate an experimental matrix based on the execution decisions of each experimental item and output it in a preset identifier format to support experimental execution.

Images