Aircraft equipment calendar life evaluation method based on multi-data source fusion

Abstract

The invention provides an aircraft equipment calendar life evaluation method based on multi-data source fusion, and relates to the technical field of aircraft equipment life evaluation, and the method comprises the following steps: S1, data sources are classified and collected; s2, constructing key degradation parameters; s3, establishing a multi-source fusion degradation model; s4, establishing a life arrival judgment basis of the aircraft equipment; s5, predicting the remaining life of the aircraft equipment; according to the method, the limitation of a single data source is avoided by fusing multi-dimensional information such as external field use data of aircraft equipment, detection data of a disassembly part during overhaul, test data obtained by a simulation test, design parameter data and the like; therefore, the state change and performance degradation condition of the aircraft in the actual use process can be known more comprehensively and accurately, a more comprehensive and accurate aircraft use environment and degradation model is constructed, and the accuracy and reliability of aircraft life evaluation are improved. Therefore, the problem of excessive maintenance or insufficient maintenance caused by inaccurate evaluation of the aircraft equipment is avoided.

Description

Technical Field

[0001] This invention relates to the field of aircraft equipment life assessment technology, and specifically to an aircraft equipment calendar life assessment method based on multi-data source fusion. Background Technology

[0002] As aircraft service life increases, the aging problem becomes increasingly prominent. Conducting aircraft life extension work and fully exploring the remaining lifespan value has significant strategic and economic value. Aircraft lifespan typically includes two key indicators: operational lifespan and calendar lifespan. Life extension projects require a comprehensive evaluation of these two lifespan indicators. Calendar lifespan is calculated based on service years, such as a 10-year first overhaul period and a total lifespan of 30 years, characterizing the aircraft's degradation characteristics under environmental factors.

[0003] Calendar lifespans can extend to decades, and traditional assessment methods primarily rely on accelerated life testing in laboratories. Current mainstream accelerated life testing methods accelerate the aging process of materials by applying high-stress conditions such as extreme temperatures, humidity, and vibrations in the laboratory. However, this method has several significant limitations: first, the simplification of the environmental stress spectrum leads to significant extrapolation errors; second, the acceleration factor is difficult to calibrate and has limited applicability; third, it cannot reflect individual usage differences; and fourth, the testing cycle is long. These problems severely affect the confidence level of lifespan assessment results and limit its application in practical engineering.

[0004] It is worth noting that aircraft nearing or reaching the end of their service life have accumulated a wealth of real-world usage data, including field usage data recorded by flight recorders, periodic overhaul inspection data, and structural health monitoring data. This data accurately reflects the comprehensive environmental stresses experienced by the aircraft throughout its entire service life. Effectively integrating this multi-source data, supplemented by necessary historical laboratory data and additional experimental data, to construct a calendar life assessment method based on real-world service environments will significantly overcome the limitations of traditional accelerated life testing. Summary of the Invention

[0005] To address the shortcomings of the existing technologies, this invention integrates multi-dimensional information such as field usage data of aircraft equipment, inspection data of disassembled parts during overhauls, test data obtained from simulation tests, and design parameter data. By establishing a multi-source fusion degradation model, it achieves accurate assessment of the calendar life of aircraft equipment, significantly reducing the reliance on laboratory accelerated aging tests for calendar life assessment of aircraft equipment.

[0006] Specifically, the present invention provides a method for assessing the calendar life of aircraft equipment based on the fusion of multiple data sources, which includes the following steps:

[0007] S1: Data source classification and collection: Based on the characteristics of multi-source data of aircraft equipment, the multi-source data of aircraft equipment to be evaluated is classified. The classification includes field usage data of aircraft equipment, inspection data of disassembled parts during major overhaul of aircraft equipment, test data of aircraft equipment or its components obtained in simulation tests, and design parameter data of aircraft equipment.

[0008] S2: Construct key degradation parameters: Statistically analyze historical experimental results, select parameters whose weights reflect the decline in aircraft equipment performance exceeding a set threshold as key degradation parameters, and extract key degradation parameters from the detection data of disassembled parts obtained during the overhaul of aircraft equipment in step S1.

[0009] S3: Establish a multi-source fusion degradation model: Based on the classification results of multi-source data in step S1, and combined with the key degradation parameters extracted in step S2, a multi-source fusion degradation model is established by weighting and fusing the physical model and the data model according to data credibility through multi-source fusion. This includes the following sub-steps:

[0010] S31: Establish a physical degradation model: Establish a physical degradation model for the key degradation parameters of the model that can be built;

[0011] S32: Establish a data-driven degradation model: For the remaining key degradation parameters in step S31, use disassembly sample data to train and fit the degradation curve to establish a data-driven degradation model.

[0012] S33: Establish a multi-source fusion degradation model: Based on the physical degradation model established in step S31 and the data-driven degradation model established in step S32, the physical degradation model and the data-driven degradation model are fused through multiple fusion methods to obtain a multi-source fusion degradation model;

[0013] S4: Establish criteria for judging the end of service of aircraft equipment: Determine the criteria for judging the end of service of aircraft equipment based on the absolute degradation threshold, the relative performance degradation ratio and / or the performance limits defined in the specifications.

[0014] S5: Predict the remaining lifespan of the aircraft equipment: Based on the multi-source fusion degradation model established in step S3 and the lifespan assessment criteria for the aircraft equipment determined in step S4, combined with the current usage data of the aircraft equipment, the remaining lifespan of the aircraft equipment is obtained:

[0015] ;

[0016] in, Indicates the remaining service life of the aircraft's equipment. Indicates the calendar life of aircraft equipment. This indicates the current service life of the aircraft equipment.

[0017] Furthermore, the specific data sources for the aircraft equipment include: material properties, structural design, operating conditions, flight time, environmental history, and experience curves related to life assessment.

[0018] Furthermore: the key feature parameters and corresponding aircraft equipment in step S2 include:

[0019] Rubber seals: compression set;

[0020] Valves: opening pressure drift, increased leakage flow, spring deformation;

[0021] Electrical components: decreased insulation resistance, conduction delay, loss factor;

[0022] Structural components: wear amount, crack length, corrosion area.

[0023] Furthermore, the various fusion methods in step S33 include: a weighted average-based fusion method, a Bayesian update-based fusion method, a state space-based fusion method, and a multi-layer LSTM-based fusion method.

[0024] Furthermore, the weighted average-based fusion method involves integrating the model results generated from different types of data using a weighted approach.

[0025] ;

[0026] in, The weights reflect the credibility or applicability of each source data. This refers to the design parameters of the aircraft's equipment. This refers to test data obtained from simulation tests of aircraft equipment or its components. This refers to the inspection data of disassembled parts during major overhauls of aircraft equipment.

[0027] Furthermore, the fusion method based on Bayesian updates involves using design data or experimental models as prior distributions, and disassembly-based measured data for Bayesian posterior updates to optimize parameter estimation.

[0028] ;

[0029] in: Prior knowledge based on design / experimentation experience; Let be the likelihood function of the measured data; This represents the posterior distribution after combining the data.

[0030] Furthermore, the state-update-based fusion method involves modeling the degradation of aircraft equipment as a state transition system, fusing multi-source data as observations at different times, and using filters to estimate health status and remaining lifetime.

[0031] ;

[0032] ;

[0033] Where A is the state transition matrix. The current health status of the system; These are observations from disassembly or testing; H is the observation matrix. For process noise, To observe noise.

[0034] Furthermore, the fusion method based on multi-layer LSTM captures the degradation features of time series and is used to model the temporal dependencies in long sequences. It includes single-input LSTM model, multi-source LSTM fully connected fusion model and multi-task LSTM model.

[0035] Furthermore, in step S5, the visualization methods include the intersection of the degradation curve and the criterion line, and the lifetime distribution map and uncertainty band map of the multi-machine sample.

[0036] Furthermore, in step S1, the testing methods for obtaining the testing data of disassembled parts during aircraft equipment overhaul include: non-destructive testing, performance testing, and microstructure analysis.

[0037] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0038] 1. This invention integrates multi-dimensional information such as field usage data of aircraft equipment, inspection data of disassembled parts during overhaul, test data obtained from simulation tests, and design parameter data, avoiding the limitations of a single data source. This allows for a more comprehensive and accurate understanding of the aircraft's state changes and performance degradation during actual use, constructing a more comprehensive and accurate aircraft operating environment and degradation model. This improves the accuracy and reliability of aircraft life assessment, thereby avoiding the problems of over-maintenance or under-maintenance caused by inaccurate assessment of aircraft equipment.

[0039] 2. This invention utilizes field data accumulated during actual use of the aircraft to truly reflect the changes in the aircraft's state under actual environmental and operational conditions. This data is naturally accumulated during normal use and maintenance of the aircraft, without the need for additional testing costs. At the same time, it combines historical laboratory data and supplementary experimental data to conduct a more comprehensive analysis and evaluation of the actual situation, thereby more comprehensively and accurately reflecting the true lifespan of the aircraft during actual use, making the evaluation results closer to reality.

[0040] 3. When constructing the degradation model of aircraft equipment, this invention adopts a multi-source data fusion approach to effectively integrate the physical degradation model and the data-driven degradation model, and can dynamically adjust the evaluation parameters and model structure. This dynamic evaluation mechanism can reflect the actual damage and remaining life of aircraft components in a timely manner. Attached Figure Description

[0041] Figure 1 This is a graph showing the relationship between the O-ring degradation model and calendar life in Embodiment 1 of the present invention. Detailed Implementation

[0042] The multi-source data involved in this invention is a rich and comprehensive collection, covering information from multiple key stages and dimensions throughout the product's entire lifecycle, such as... Figure 1 As shown, the details are as follows:

[0043] Field usage data of aircraft equipment includes key indicators such as flight hours, mission completion rate, failure rate, and maintenance cycle. These data directly affect combat effectiveness and logistical support capabilities.

[0044] Inspection data from disassembled parts during aircraft overhauls: This data is obtained through meticulous inspection of replaced components during regular aircraft overhauls. These inspections include non-destructive testing (such as ultrasonic testing, radiographic testing, and corrosion morphology inspection), performance testing (such as compression set, elasticity, leakage current, power consumption, strength testing, and fatigue testing), and microstructure analysis. This data directly reflects the damage and performance changes of components during actual service, providing valuable practical cases for life assessment.

[0045] Performance degradation model over time: A model established based on prior experience, historical test results and theoretical analysis to describe the changes in product performance over time. This model can predict the performance status of a product at different points in time and is one of the core tools for life assessment.

[0046] Laboratory test data: Data obtained from various simulated tests conducted on products or their components in a laboratory environment. These tests include accelerated life tests, environmental simulation tests, etc., which aim to quickly obtain information on the performance degradation of products under specific conditions and provide data support for the establishment and verification of degradation models.

[0047] Product design parameter data: Information such as various parameters, specifications and performance requirements determined during the product design phase. This information includes the product's material properties, structural design, and operating conditions, and is the basis for determining product failure criteria and conducting life assessments.

[0048] The following are the specific steps of the aircraft equipment calendar life assessment method based on multi-data source fusion:

[0049] S1: Data Source Classification and Collection: Based on the characteristics of multi-source data of aircraft equipment, the multi-source data of aircraft equipment to be evaluated is classified. The classification includes field usage data of aircraft equipment, inspection data of disassembled parts during major overhaul of aircraft equipment, test data of aircraft equipment or its components obtained from simulation tests, and design parameter data of aircraft equipment.

[0050] The details are shown in the table below:

[0051]

[0052] Table 1

[0053] S2: Construct key degradation parameters: Statistically analyze historical experimental results, select parameters whose weights reflect the decline in aircraft equipment performance exceeding a set threshold as key degradation parameters, and extract key degradation parameters from the inspection data of disassembled parts obtained during the overhaul of aircraft equipment in step S1.

[0054] Extract key degradation indicators reflecting performance decline from disassembly and testing data, such as:

[0055] Rubber seals: compression set.

[0056] Valves: opening pressure drift, leakage flow increase, spring deformation.

[0057] Electrical components: decreased insulation resistance, conduction delay, loss factor.

[0058] Structural components: wear amount, crack length, corrosion area.

[0059] Establish the following structural formula:

[0060] ;

[0061] Among them: Among them, This represents the current degradation value of the key degradation parameter. These are the initial values ​​for the key degradation parameters. This represents the time-related degradation.

[0062] S3: Establish a multi-source fusion degradation model: Based on the classification results of multi-source data in step S1, and combined with the key degradation parameters extracted in step S2, a multi-source fusion degradation model is established by weighting and fusing the physical model and the data model according to data credibility through a multi-source fusion method.

[0063] First, focusing on the product to be evaluated, determine its performance degradation model over time based on prior experience or historical test results. Prior experience comes from long-term use and research of similar products, while historical test results are obtained by recording and analyzing the product's performance changes under different conditions through scientific testing methods.

[0064] For a critical component of an aircraft, referencing the performance changes of the component in previous aircraft of the same model under the same operating environment, and combining the results of accelerated life tests conducted in the laboratory on similar materials and structures, a mathematical model is constructed that can accurately describe the performance degradation of the product over time, taking into account various factors such as environmental factors like temperature, humidity, and vibration, as well as operating condition factors like workload and frequency of use. This model can be a linear model, a nonlinear model, or a model based on physical mechanisms.

[0065] For example, assuming the evaluation object is an engine blade, based on historical data, its performance degradation model can be initially set as follows:

[0066] ;

[0067] in, This indicates the performance at time t. Indicates initial performance. This indicates the degradation rate.

[0068] Then, based on the service data regression degradation model, during aircraft overhauls, the disassembled parts are fully utilized for performance testing, and the test results are recorded in detail. These disassembled parts are real samples of the product during actual service, carrying information on various stresses and damage experienced by the product in actual use. By collecting disassembled parts from different service years, comprehensive performance testing is conducted on them to obtain values ​​for their key performance indicators. Based on this data, mathematical methods such as regression analysis are used to revise and improve the previously determined degradation model. If the performance degradation rate of some disassembled parts is found to be inconsistent with the initial model prediction, the parameters in the model can be dynamically adjusted to make the model more consistent with the actual situation. This allows the degradation model to more accurately reflect the performance change patterns of the product during actual service.

[0069] For example: Suppose we collected engine blade data from 10 different service years, and through regression analysis we obtained:

[0070] ;

[0071] Among them, degradation rate The initial value was adjusted to 0.05.

[0072] Establishing a multi-source fusion degradation model specifically includes the following sub-steps:

[0073] S31: Establish a physical degradation model: Establish a physical degradation model for the key degradation parameters of the model that can be built; applicable to parts of the model that can be built, such as material aging, wear, and stress corrosion.

[0074] For example: Modeling engine blades:

[0075] ;

[0076] Where P(t) represents the performance at time t, Let λ represent the initial performance and λ represent the degradation rate.

[0077] For example, the degradation model of the compression set rate of an O-ring is:

[0078] ;

[0079] Where ε is the compressive permanent deformation rate, t is the calendar year, and a and b are model parameters.

[0080] Friction component wear:

[0081] .

[0082] S32: Establish a data-driven degradation model: For the remaining key degradation parameters in step S31, use disassembly sample data to train and fit degradation curves, such as exponential, logarithmic, and polynomial curves, to establish a data-driven degradation model.

[0083] Regression / time series models can also be built, such as GPR, LSTM / GRU, RNN + Attention, etc.

[0084] For example, the modeling form is represented as:

[0085] ;

[0086] in, These are design parameters. It is the degenerate function obtained through training.

[0087] S33: Establish a multi-source fusion degradation model: Based on the physical degradation model established in step S31 and the data-driven degradation model established in step S32, the physical degradation model and the data-driven degradation model are fused through multiple fusion methods to obtain a multi-source fusion degradation model.

[0088] The methods for multi-source melting are shown in the table below:

[0089]

[0090] Table 2

[0091] The fusion method based on weighted average is as follows:

[0092] The model results from data from different sources are integrated using a weighted average method:

[0093] ;

[0094] in, The weights reflect the credibility or applicability of each source data.

[0095] The Bayesian fusion-based method is as follows:

[0096] Using design data or experimental models as the prior distribution, and disassembly-based measured data for Bayesian posterior updates, parameter estimation is progressively optimized:

[0097] ;

[0098] in:

[0099] As a priori based on design / experiment experience, Let be the likelihood function of the measured data. This represents the posterior distribution after combining the data.

[0100] The state-space fusion-based method is as follows:

[0101] Equipment degradation is modeled as a state transition system, multi-source data is fused as observations at different times, and filters are used to estimate health status and remaining lifetime.

[0102] ;

[0103] ;

[0104] Where A is the state transition matrix. Based on the current system health status, H represents the observations from disassembly or testing, and H is the observation matrix. For process noise, To observe noise.

[0105] The fusion method based on multi-layer LSTM is as follows:

[0106] Capture time series degradation features to model time dependencies in long series.

[0107] S4: Establish criteria for determining the end of service life of aircraft equipment: Determine the criteria for determining the end of service life of aircraft equipment based on the absolute degradation threshold, the relative performance degradation rate, and / or the performance limits defined in the specifications.

[0108] Commonly used methods include: based on absolute degradation thresholds, based on the relative percentage of performance degradation, and based on performance limits defined by specifications.

[0109] Product design information is crucial for determining product failure or end-of-life criteria. During the product design phase, engineers determine the product's performance indicators and permissible ranges under normal operating conditions based on usage requirements, performance targets, and safety standards. For example, for aircraft engines, the design information specifies permissible ranges for performance indicators such as maximum thrust, fuel consumption rate, and vibration level. When these indicators exceed these ranges, it means the product may have failed or reached the end of its service life. By carefully studying product design documents and technical specifications, key parameters and criteria related to product failure or end-of-life can be extracted. These criteria can be single performance indicator thresholds or comprehensive evaluation standards for multiple performance indicators. Once the end-of-life criteria are clearly defined, it is possible to accurately determine whether the product has reached failure or the end of its service life in subsequent life assessments.

[0110] For example, according to design specifications, an engine blade is considered to have reached the end of its service life when its performance drops to 70% of its initial performance. Therefore, the failure / end-of-service criterion is:

[0111] .

[0112] After determining the product's degradation model and failure / life-end criteria, the product's failure / life-end criteria are substituted into the degradation model for calculation. By solving the model equations, the time required for the product to reach failure or life-end state is obtained, which is the product's life assessment result.

[0113] For example: criteria for determining failure / end of service. Substitute into the degradation model :

[0114] ;

[0115] Solving for: Year.

[0116] S5: Predict the remaining lifespan of the aircraft equipment: Based on the multi-source fusion degradation model established in step S3 and the lifespan assessment criteria for the aircraft equipment determined in step S4, combined with the current usage data of the aircraft equipment, the remaining lifespan of the aircraft equipment is obtained:

[0117] ;

[0118] in Indicates the remaining service life of the aircraft's equipment. Indicates the calendar life of aircraft equipment. This indicates the current service life of the aircraft equipment.

[0119] Visualization methods include the intersection of degradation curves and criterion lines, multi-machine sample lifetime distribution maps, and uncertainty band maps.

[0120] The lifespan assessment results are compared and analyzed with the pre-set lifespan extension expectations. Lifespan extension expectations are typically determined based on a combination of factors, including product usage requirements, economic costs, and technical feasibility. For example, if a product's design lifespan is 10 years, but for economic and usage needs, the desired lifespan is extended to 15 years, then 15 years is the lifespan extension expectation.

[0121] By comparing the lifespan assessment results with the expected lifespan extension, we can conclude that: if the assessment results indicate that the product's remaining lifespan meets or exceeds the expected lifespan extension, then lifespan extension measures can be considered, and a corresponding lifespan extension plan should be developed; if the assessment results show that the product's remaining lifespan cannot meet the expected lifespan extension, then the product's usage strategy needs to be reassessed, such as early replacement or technical improvements. When making a lifespan extension conclusion, it is also necessary to comprehensively consider the impact of various factors, such as product maintenance costs and safety, to ensure the rationality and feasibility of the conclusion.

[0122] For example, assuming an expected lifespan extension of 12 years, the estimated lifespan is 10.24 years. Since the estimated lifespan is less than the expected lifespan extension, direct lifespan extension is not possible. Further analysis of the causes of performance degradation is needed, and consideration should be given to whether lifespan can be extended through improved design or maintenance strategies.

[0123] The following examples illustrate the implementation methods and steps of the present invention in detail:

[0124] Example 1: O-ring calendar life assessment:

[0125] Determine the product's degradation model over time:

[0126] As the calendar years increase, the compression set of O-rings increases. When the compression set exceeds a certain threshold, leakage occurs. Based on prior experience, the degradation model of the compression set of O-rings is as follows:

[0127] ;

[0128] in, The compression permanent deformation rate is given by t, where t is the calendar year, and a and b are model parameters.

[0129] Based on service data regression degradation model:

[0130] Obtain disassembled O-rings with service lives of 4.2, 6.5, and 8.8 years, and measure their compression set. The following table was obtained:

[0131]

[0132] Table 3

[0133] The degradation model is as follows:

[0134] ;

[0135] Clearly define the criteria for determining the product's lifespan:

[0136] Based on historical test data, A leak occurred when the concentration was 65%, therefore it was determined that... The product reaches the end of its life when 65% of the product is used.

[0137] Life assessment:

[0138] ;

[0139] t = 20.60 years;

[0140] The conclusion regarding life extension is as follows:

[0141] The life extension target is 15 years. The life assessment result is 20.6 years, which meets the life extension requirement.

[0142] 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 assessing the calendar life of aircraft equipment based on multi-data source fusion, characterized in that, It includes the following steps: S1: Data source classification and collection: Based on the characteristics of multi-source data of aircraft equipment, the multi-source data of aircraft equipment to be evaluated is classified. The classification includes field usage data of aircraft equipment, inspection data of disassembled parts during major overhaul of aircraft equipment, test data of aircraft equipment or its components obtained in simulation tests, and design parameter data of aircraft equipment. S2: Construct key degradation parameters: Statistically analyze historical experimental results, select parameters whose weights reflect the decline in aircraft equipment performance exceeding a set threshold as key degradation parameters, and extract key degradation parameters from the detection data of disassembled parts obtained during the overhaul of aircraft equipment in step S1. S3: Establish a multi-source fusion degradation model: Based on the classification results of multi-source data in step S1, and combined with the key degradation parameters extracted in step S2, a multi-source fusion degradation model is established by weighting and fusing the physical model and the data model according to data credibility through multi-source fusion. This includes the following sub-steps: S31: Establish a physical degradation model: Establish a physical degradation model for the key degradation parameters of the model that can be built; S32: Establish a data-driven degradation model: For the remaining key degradation parameters in step S31, use disassembly sample data to train and fit the degradation curve to establish a data-driven degradation model. S33: Establish a multi-source fusion degradation model: Based on the physical degradation model established in step S31 and the data-driven degradation model established in step S32, the physical degradation model and the data-driven degradation model are fused through multiple fusion methods to obtain a multi-source fusion degradation model; S4: Establish criteria for judging the end of service of aircraft equipment: Determine the criteria for judging the end of service of aircraft equipment based on the absolute degradation threshold, the relative performance degradation ratio and / or the performance limits defined in the specifications. S5: Predict the remaining lifespan of the aircraft equipment: Based on the multi-source fusion degradation model established in step S3 and the lifespan assessment criteria for the aircraft equipment determined in step S4, combined with the current usage data of the aircraft equipment, the remaining lifespan of the aircraft equipment is obtained: ； in, Indicates the remaining service life of the aircraft's equipment. Indicates the calendar life of aircraft equipment. This indicates the current service life of the aircraft equipment.

2. The aircraft equipment calendar life assessment method based on multi-data source fusion as described in claim 1, characterized in that: The specific data sources for aircraft equipment include: material properties, structural design, operating conditions, flight time, environmental history, and experience curves related to life assessment.

3. The aircraft equipment calendar life assessment method based on multi-data source fusion as described in claim 1, characterized in that: The key feature parameters and corresponding aircraft equipment in step S2 include: Rubber seals: compression set; Valves: opening pressure drift, increased leakage flow, spring deformation; Electrical components: decreased insulation resistance, conduction delay, loss factor; Structural components: wear amount, crack length, corrosion area.

4. The aircraft equipment calendar life assessment method based on multi-data source fusion as described in claim 1, characterized in that: The various fusion methods in step S33 include: weighted average-based fusion method, Bayesian update-based fusion method, state space-based fusion method, and multi-layer LSTM-based fusion method.

5. The aircraft equipment calendar life assessment method based on multi-data source fusion as described in claim 4, characterized in that: The weighted average-based fusion method integrates the model results generated from different types of data using a weighted approach. ； in, The weights reflect the credibility or applicability of each source data. This refers to the design parameters of the aircraft's equipment. This refers to test data obtained from simulation tests of aircraft equipment or its components. This refers to the inspection data of disassembled parts during major overhauls of aircraft equipment.

6. The aircraft equipment calendar life assessment method based on multi-data source fusion as described in claim 4, characterized in that: The fusion method based on Bayesian updates is as follows: Design data or experimental models are used as prior distributions, and disassembly-based measured data are used for Bayesian posterior updates to optimize parameter estimation. ； in: Prior knowledge based on design / experimentation experience; Let be the likelihood function of the measured data; This represents the posterior distribution after combining the data.

7. The aircraft equipment calendar life assessment method based on multi-data source fusion as described in claim 4, characterized in that: The state-update-based fusion method involves modeling the degradation of aircraft equipment as a state transition system, fusing multi-source data as observations at different times, and using filters to estimate health status and remaining lifespan. ； ； Where A is the state transition matrix. The current health status of the system; These are observations from disassembly or testing; H is the observation matrix. For process noise, To observe noise.

8. The aircraft equipment calendar life assessment method based on multi-data source fusion as described in claim 4, characterized in that: The fusion method based on multi-layer LSTM is to capture the degradation features of time series and to model the temporal dependencies in long series. It includes single-input LSTM model, multi-source LSTM fully connected fusion model and multi-task LSTM model.

9. The aircraft equipment calendar life assessment method based on multi-data source fusion as described in claim 1, characterized in that: In step S5, the visualization methods include the intersection of the degradation curve and the criterion line, and the lifetime distribution map and uncertainty band map of the multi-machine sample.

10. The method for assessing the calendar life of aircraft equipment based on multi-data source fusion as described in any one of claims 1-9, characterized in that: In step S1, the testing methods for obtaining the testing data of disassembled parts during aircraft equipment overhaul include: non-destructive testing, performance testing, and microstructure analysis.

Images