A UAV propeller reliability modeling method, system, medium and program product

Abstract

The present invention relates to the field of drone design technology, and in particular to a method, system, medium and program product for modeling the reliability of drone propellers. The method first constructs a requirement model, system context, use case and activity diagram of drone propellers based on SysML, and defines the mathematical function relationship of core performance parameters such as thrust, torque and speed; then statistical fitting is performed in combination with experimental data to construct a joint performance margin equation, and the reliability theory is introduced to calculate the reliability value of the propeller through the standard normal cumulative distribution function; finally, the above constraint equation is constrained and bound to the system structure model to realize dynamic response and model linkage under parameter changes. The present invention realizes the integrated integration of reliability analysis in the MBSE model, significantly improving the automation level, evaluation accuracy and engineering application efficiency of propeller system reliability modeling.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle (UAV) design, and in particular to a UAV propeller reliability modeling method, system, medium and program product. Background Art

[0002] With the widespread deployment of multi-rotor drones in scenarios such as urban delivery, inspection and monitoring, and emergency response, their structural complexity and uncertain operating environments are increasingly demanding the reliability of core components. Under the high-intensity, multi-environment, and multi-mission operating conditions, the reliability of the power system has become a bottleneck limiting the industry's large-scale application and expansion into high-end applications. The power system typically consists of propellers (rotors), brushless motors, electronic speed controllers (ESCs), batteries, and control algorithms. The rotors, in particular, are both the direct component for generating lift and the point of interaction for aerodynamic loads, structural vibration, and external disturbances. Numerous accident statistics show that rotor structural failure, dynamic imbalance, aeroelastic coupling, and the resulting insufficient thrust or excessive airframe vibration are among the primary causes of drone loss of control and crashes.

[0003] Traditionally, blade geometry (diameter, pitch, aspect ratio, etc.) has been used as independent variables, with the Betz criterion, blade element momentum theory (BEMT), or CFD iterative calculations used to derive thrust-torque curves. Safety margins are then verified during the prototype phase through static and fatigue testing. This process has a long design cycle and high iteration costs. More importantly, it mitigates system-level interactions with the propeller (such as current limiting with electronically controlled current controllers, flight control rate adjustment, and airframe attitude disturbances). This approach only provides static strength or lifespan estimates for isolated components, failing to reflect real-time reliability changes across the entire aircraft under multiple operating conditions.

[0004] Furthermore, engineering practice often uses statistical manuals such as FMEA / FMECA, Fault Tree Analysis (FTA), or MIL-HDBK-217 to derive propeller failure rates; or employs empirical safety factors of 1.5–2.0 to account for load uncertainty. However, these methods fail to fully quantify the coupling effects between thrust, torque, and speed, and lack a comprehensive assessment of model errors and environmental disturbances (temperature, air pressure, and air shear). The results are often overly conservative or underestimate risks, making them difficult to sustain the rapid pace of drone product updates.

[0005] Although there have been studies in recent years that combine finite element analysis, CFD, hardware-in-the-loop (HIL) testing of propellers with digital twins to perform multi-physics field simulation of material fatigue and aerodynamic loads, most platforms focus on post-processing visualization or online fault diagnosis, and lack a unified modeling language and automatic parameter constraint mechanism for the early stages of design; within the system engineering dimension, the data interfaces between different simulation tools often rely on manual script maintenance, making it difficult to achieve end-to-end model consistency.

[0006] Since its inception, Model-Based Systems Engineering (MBSE) has been vigorously promoted by organizations such as NASA, INCOSE, and Dassault. Its core concept is to use digital models throughout the entire system lifecycle, unifying requirements, structure, behavior, parameters, constraints, and verification activities, while ensuring information consistency within a single digital thread (Single Source of Truth). MBSE provides a new paradigm for addressing the challenges of UAV power system reliability assessment, such as complex coupling, diverse mechanisms, dispersed data, and fragmented models. However, the MBSE framework alone is insufficient to directly output reliability metrics; further integration of uncertainty modeling and confidence into the model is necessary. MBSE is gaining increasing attention within the domestic academic community, with numerous universities and research institutes developing single-machine MBSE frameworks and conducting pilot projects in areas such as avionics, engine control, and overall UAV systems.

[0007] A Chinese invention patent application (publication number: CN114781183 A, publication date: July 22, 2022) proposes a MBSE-based approach to modeling perception systems from logical models to physical simulation models, and constructs system reliability metrics using state transition and parameter models. However, this approach primarily focuses on intelligent perception systems and does not yet address the modeling requirements of power actuators with well-defined mechanical output characteristics (such as thrust, torque, and speed), such as drone propellers. Furthermore, a Chinese invention patent application (publication number: CN116776656 A, publication date: September 19, 2023) points out that traditional approaches often focus on a single component or function, making it difficult to systematically assess the reliability transfer and interactive failure paths between different functions in unmanned systems. Furthermore, they fail to accurately characterize the impact of each module on reliability within the "perception-decision-execution" chain. In addition, by constructing a multi-layer hypernetwork and Petri net, the interaction relationship between the perception, collaboration and driving layers is formally modeled, and reliability measurement indicators of the functional dimension are defined. However, this method focuses more on the task completion capability evaluation of complex unmanned systems, does not introduce a physical parameter margin model based on statistical distribution characteristics, and does not consider the quantitative impact of model errors on reliability results. Summary of the Invention

[0008] To address the aforementioned technical issues, the present invention aims to provide a method for modeling assured reliability of UAV propellers based on MBSE. This method embeds assured reliability theory into the system structure model as constraints, proposing a unified modeling framework applicable to a wide range of physical systems. This method not only reflects the dynamic coupling between performance parameters and system structure but also enables the visualization of system reliability characteristics under multiple operating conditions. Through the synergistic effect of model hierarchy, parameter dependency, and constraint propagation mechanisms, the applicability and engineering feasibility of assured reliability theory in MBSE scenarios are further verified, demonstrating the scientific, universal, and rigorous nature of this methodology.

[0009] In order to achieve the above-mentioned purpose, the present invention adopts the following technical solutions:

[0010] A reliable reliability modeling method for UAV propellers based on MBSE is implemented as follows:

[0011] 1) Demand Analysis Model Generation: The propeller in the UAV power system is identified as the target research unit; propeller parameters are input into the demand analysis module to generate the demand analysis model;

[0012] 2) System context and use case modeling:

[0013] Draw the drone system context model at the system level to define the interface relationships between the propeller and the flight control module, electronic speed controller, motor, and battery components. Draw the context model and flight use case diagram at the power subsystem level to indicate the closed-loop execution logic for the propeller to achieve attitude adjustment by adjusting thrust and angular velocity.

[0014] 3) Performance function and statistical distribution construction:

[0015] The experimental data of propeller thrust F and torque τ at different speeds were collected. After removing outliers using BOX-COX transformation, the thrust and torque functions were established.

[0016] 4) Fault threshold setting: Set the thrust threshold F according to the flight mission requirements. th , torque threshold τ th and speed threshold ω th ;

[0017] 5) Joint margin calculation: Calculate the dimensionless performance margin m of thrust, torque and speed through Monte Carlo simulation F 、m τ 、m ω ; The weight m is obtained based on minimum-maximum normalization and variance analysis F 、m τ 、w ω ; The joint margin M is obtained by calculation:

[0018] ;

[0019] 6) Calculation of confidence reliability: The performance fluctuation variance σ m ² and model error variance σ e ²Substitution

[0020] ;

[0021] in Φ N (.) is the standard normal cumulative distribution function, M th is the joint margin threshold, and the propeller reliability R is obtained MB .

[0022] Preferably, the propeller parameters in step 1) include propeller diameter, pitch, center hole diameter, material, weight, self-locking mechanism, blade cross-sectional thickness, factory balance and parameters of the adaptation platform.

[0023] As a preference, in step 3), the square of the speed ω 2 Normal distribution fitting is performed to obtain the mean μ and standard deviation σ. The normal distribution fitting adopts maximum likelihood estimation and the normality is verified by Shapiro-Wilk test. The thrust function is PF(ω 2 )=CFω 2 , the torque function is Pτ(ω 2 )=Cτω 2 , the coefficients CF and Cτ are obtained by least squares linear fitting, and the goodness of fit R² is not less than 0.98.

[0024] Preferably, the threshold values Fth, τth, and ωth in step 4) are jointly determined by the hovering thrust requirement, the minimum torque starting value, and the thrust attenuation inflection point; and / or, the number of Monte Carlo simulations in step 5) is not less than 1000, and the simulation results use a 95% confidence interval to filter out abnormal samples; the weight calculation further includes performing a principal component analysis on the normalized features, and when the contribution rate of the first principal component is greater than 85%, the weights are corrected using its load coefficient.

[0025] As a preference, step 6) calculates the statistical fluctuation variance σ of the performance parameter m 2 The formula is as follows:

[0026] ;

[0027] in:

[0028] σ m ² is the variance of the integrated or joint margin;

[0029] w i is the weight of the i-th performance indicator, reflecting the impact of this indicator on the overall margin;

[0030] σ m,i ² is the variance of the i-th performance indicator itself, that is, the variance of its performance margin;

[0031] To sum n indicators;

[0032] And / or, the model error variance σ e ²Based on expert experience and a comprehensive assessment of historical model-measured deviations, the value range is 0.01-0.10.

[0033] Preferably, the method further includes MBSE constraint binding integration: the mathematical equations of steps 3)-6) are inherited and bound to the propeller module in the internal structure diagram through SysML constraint blocks, so as to realize real-time automatic refresh of reliability when parameters are updated; the constraint binding relationship supports automatic synchronous update in Magic Draw, Cameo Systems Modeler or MBSE tools.

[0034] Furthermore, the present invention also provides a model-based UAV propeller reliability modeling system, which implements the method described above, including:

[0035] A. Requirement analysis module, used to receive propeller geometry, material and assembly parameters and generate a requirement analysis model;

[0036] B. System context modeling module, used to build SysML context models of device hierarchical relationships and interface interactions;

[0037] C. Use case and activity modeling module, used to define propeller attitude adjustment use cases and draw corresponding activity diagrams;

[0038] D. Performance modeling module, including:

[0039] D1. Data acquisition unit, used to obtain propeller thrust and torque experimental data at different speeds;

[0040] D2. Parameter fitting unit, used to fit the normal distribution of the square of the speed ω² and obtain the coefficient C F 、C τ ;

[0041] E. Threshold setting unit, used to store and call thrust, torque and speed fault thresholds;

[0042] F. Joint margin calculation unit, used to calculate the joint margin M based on dimensionless transformation, weight distribution and Monte Carlo simulation;

[0043] G. Reliability evaluation unit, used to execute the confidence reliability formula and output RMB;

[0044] H.MBSE constraint binding unit, used to establish inheritance and constraint binding between performance equations and propeller structure models in the MBSE tool environment, enabling real-time updates of parameter changes;

[0045] I. Processor, used to run instructions of the above modules / units.

[0046] Preferably, the performance modeling module further includes an outlier processing subunit for performing BOX-COX transformation and eliminating data points that deviate from ±3σ; and / or, the joint margin calculation unit is configured to call three algorithms of minimum-maximum normalization, variance analysis and principal component analysis when calculating the weights and automatically select the weight set with the smallest error; and / or, the reliability assessment unit is configured to send a maintenance warning signal to the flight control system when RMB is lower than a set threshold.

[0047] Furthermore, the present invention also provides a computer-readable storage medium having a computer program or instruction stored thereon, which implements the method when the computer program or instruction is executed by a processor.

[0048] Furthermore, the present invention also provides a computer program product, comprising a computer program or instructions, which implement the method when executed by a processor.

[0049] By employing the aforementioned technical solution, this invention embeds a mathematical model of assured reliability within the MBSE framework, enabling a full-process model-driven approach to the reliability of drone propellers, from requirements to reliability. This demonstrates the precise modeling and expression of assured reliability theory within a systems engineering framework for research objects like drone propellers, which possess quantifiable physical performance parameters (such as thrust, torque, and speed). Specifically, the following significant technical benefits are achieved:

[0050] 1. Model consistency and data integration: The three core performance equations of thrust, torque, and speed are bound to the propeller structural model using SysML constraint blocks. Any change in input parameters (such as propeller diameter, pitch, and material density) automatically updates the reliability results in less than 0.5 seconds, eliminating the data fragmentation and manual synchronization errors caused by the traditional "CAD-simulation-spreadsheet" separation. The requirements, function, structure, and parameter layers are unified from a single source, significantly improving model version traceability and configuration accuracy.

[0051] 2. Improved reliability quantification accuracy: Establishing the confidence reliability R by combining the margin M with the standard normal cumulative distribution function Φ MB , for the first time considering the statistical fluctuation variance σ m ² and model error variance σ e Compared to methods using only empirical safety factors, the mean square error of reliability assessment is reduced by approximately 28%. The coupled setting of equipment-level failure thresholds and mission-level thrust requirements can precisely locate failure risk areas and avoid redundant weight caused by overly conservative designs.

[0052] 3. Rapid simulation and decision support for multiple working conditions: Built-in Monte Carlo simulation and weight adaptive algorithm can output margin distribution curve within 3 seconds under 1000 random working condition simulations, providing real-time and callable reliability boundary for the closed-loop design of flight control algorithm. MB When the temperature falls below the set threshold, a maintenance warning instruction can be immediately sent to the upper flight control system, providing a quantitative basis for preventive maintenance decisions.

[0053] In summary, the present invention establishes a new paradigm for propeller reliability design at the system engineering level, achieving a technological breakthrough in high-precision quantification, rapid iteration, and real-time monitoring, significantly improving the safety, economy, and engineering development efficiency of UAV power systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 This is a structural diagram of the self-locking spiral wing of the rotary wing UAV system of the present invention.

[0055] Figure 2 Modeling diagram for the UAV system-level system context.

[0056] Figure 3 Modeling diagram for the subsystem-level context of the UAV power system.

[0057] Figure 4 This is a use case diagram for drone propeller flight.

[0058] Figure 5 This is the system-level activity logic diagram of the drone.

[0059] Figure 6 This is the activity logic diagram of the power module subsystem level.

[0060] Figure 7 This is a graph showing the fitted values of the rotational speed when it is fitted to a normal distribution using the Box-Cox method.

[0061] Figure 8 C F Plot of fitted values.

[0062] Figure 9 Cτ Plot of fitted values. DETAILED DESCRIPTION

[0063] The following is a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts are within the scope of protection of the present invention.

[0064] Selected research units

[0065] A rotary-wing UAV system typically consists of a power module, networking and communication module, flight control module, core computing module, and perception module. The power system, consisting of batteries, electronic speed controllers (ESCs), motors, and propellers, receives and executes control commands from the flight control module to achieve flight maneuvers.

[0066] like Figure 1 As shown, the research object of the present invention is the DJI-1045 self-locking propeller, which is an optimized design of a multi-rotor platform with a specification of 10 × 4.5 inches (1045). It adopts a spline-type self-locking fastening structure, which can achieve tool-free quick assembly and disassembly.

[0067] 2. UAV system modeling

[0068] 2.1 Demand Analysis Modeling

[0069] As shown in Table 1, a demand analysis model is constructed based on the DJI-1045 UAV's rotor design parameters, material requirements, rotation attitude control, structural characteristics and other requirements.

[0070] Table 1 Propeller demand analysis table

[0071]

[0072] 2.2 System Context

[0073] In order to reasonably construct the propeller reliability model, such as Figure 2 、 Figure 3 As shown, it is necessary to reasonably set the logical interactions between each system level and the interaction between the propeller and each system level and other components in the power system to draw a reasonable system context.

[0074] 2.3 Use Cases

[0075] This paper focuses on a key use case in drone propulsion systems: dynamic adjustment of flight attitude by adjusting propeller thrust and angular velocity. This use case is the core execution link in the flight control closed loop, directly related to flight stability and mission responsiveness.

[0076] 2.4 Activity Diagram

[0077] like Figure 5 、 Figure 6 As shown, the present invention constructs an activity logic diagram of the propeller attitude change in the power module at the system level and an activity logic diagram of the propeller execution attitude change at the power module subsystem level.

[0078] 2.5 System Port Definition

[0079] The definition of system ports is the key mechanism to achieve dynamic coupling between UAV systems and is of great significance for modeling the correlation between system elements. Figure 7 As shown, this invention uses SysML's Block Definition Diagram (BDD) and Internal Block Diagram (IBD) to achieve structured integration and relationship modeling of various drone system modules. The system components are no longer isolated functional units, but instead form a highly coupled, collaborative, integrated system.

[0080] During the modeling process, changes in system- and subsystem-level input variables are no longer a single module's response. Instead, through the establishment of ports and connections, all system modules achieve real-time linkage and coordinated responses. This design approach embodies the "integration-centric" philosophy of Model-Driven Systems Engineering (MBSE) and effectively supports the simulation and verification of the dynamic behavior of complex systems.

[0081] 3. Margin equation construction

[0082] 3.1 Performance parameter selection

[0083] Based on the physical experimental data of DJI-1045 self-locking propeller, the present invention establishes a dynamic fitting model for thrust and torque response. As the independent variable, the changing law of thrust and torque is fitted.

[0084] Propeller aerodynamic theory shows that when the inlet velocity and aerodynamic load are constant, the thrust F, torque τ and the square of the speed are is directly proportional, that is Torque Also approximately follows The law of.

[0085] Therefore, the present invention uses thrust (F), torque (τ) and speed square ( ) as the core modeling parameters, the system characterizes the aerodynamic performance and power response law of this type of propeller under typical working conditions, and establishes a basic mathematical model suitable for aerodynamic modeling and simulation optimization of rotorcraft, providing theoretical support for related flight control algorithms and system design.

[0086] 3.2 Performance Function Construction

[0087] P n It means that the propeller is rotated at different speeds The key performance indicators below are mapped to functions of mathematical expressions. Corresponding performance parameters.

[0088] For the DJI-1045 model, we constructed the following based on aerodynamic theory and experimental data:

[0089] (1) Thrust function:

[0090] ，

[0091] (2) Torque function:

[0092] ，

[0093] (3) Speed function:

[0094] 。

[0095] Based on the physical experimental data, the independent variable speed is first Perform a normal distribution fit and determine its mean μ = 794.448 and standard deviation σ = 339.283. These distribution parameters will serve as the statistical limits for the normal distribution input of related variables in subsequent parameter diagrams. Figure 7 The detailed calculation process is shown in Figure 2. The experimental speed data was fitted to a normal distribution using the Box-Cox method, eliminating outliers. The conversion coefficient was 0.341, resulting in a mean of μ = 794.448 and a standard deviation of σ = 339.283.

[0096] Furthermore, by performing linear fitting on the experimental data, the coefficient C that conforms to the formula expression is calculated. F =1.99, C τ =3.5E-5. Figure 8 、 Figure 9 As shown in the figure, the detailed calculation process is as follows: by sorting out the square data of speed and thrust and torque data, and using the linear fitting method, we can get two constant coefficients C F =1.99 and C τ =3.5E-5.

[0097] The experimental data are shown in Table 2.

[0098]

[0099] 3.3 Fault Threshold Setting

[0100] As shown in Table 3, a corresponding fault threshold is set for each performance parameter. The threshold is based on mission requirements or design specifications and is determined by the minimum safe lift, minimum torque starting value, and minimum energy efficiency, respectively.

[0101]

[0102] Table 3 Corresponding fault thresholds for parameter settings

[0103]

[0104] Calculation yields:

[0105] 3.4 UAV propeller joint margin construction

[0106] ,

[0107] in,

[0108] ,

[0109] ,

[0110] .

[0111] m n Indicates the performance margin result of the selected nth performance parameter.

[0112] The definition equation in this step eliminates the influence of the unit of the parameter itself and realizes the dimensionless unified measurement between parameters of different dimensions. F =0.3109>0,m τ =0.0282>0,m ω =0.6824>0.

[0113] Detailed calculation process: By defining the constraint formula, inputting the value of the square of the speed, and using Monte Carlo simulation to calculate 1000 times to obtain the best value.

[0114] Weight Reflects the importance of different performance parameters to the overall performance margin of the system, and the weights meet the normalization conditions:

[0115] ，

[0116] Calculated: w F =0.3637,w τ =0.3589,w ω =0.2774.

[0117] Detailed calculation process:

[0118] 1. Use Min-Max Scaling to normalize each column feature (thrust, torque, RPM²) to [0,1]:

[0119] ;

[0120] 2. Calculate the variance of each column

[0121] ;

[0122] 3. Calculate weights

[0123] ;

[0124] Then get the specific value.

[0125] 3.5 Model-based UAV propeller reliability assessment

[0126] The formula is expressed as;

[0127] ;

[0128] in Φ N (.) is the standard normal cumulative distribution function, M th is the joint margin threshold.

[0129] The result is R MB =0.9838.

[0130] Detailed calculation process:

[0131] 1) Calculate the statistical fluctuation variance of performance parameters :

[0132] ;

[0133] is the variance of the integrated or joint margin (i.e., the overall variance after weighted synthesis);

[0134] is the weight of the i-th performance indicator, reflecting the impact of this indicator on the overall margin;

[0135] is the variance of the i-th performance indicator itself, that is, the variance of its performance margin;

[0136] To sum n indicators;

[0137] Arrange and input all the calculated data to get the result.

[0138] 2) Model error variance This value reflects the structural simplification, parameter uncertainty, and modeling error of the MBSE model. Expert experience is used to assess the reliability of the model to avoid overoptimism, with a value of 0.04 being used here because the drone propeller is a simple physical entity.

[0139] The final calculation result is: m =0.3247,σ e ≈0.04.

[0140] 4. Parametric Graphs for Simultaneous Equations

[0141] In the third step of modeling, all defined performance equations are assigned a generalization and binding relationship with the propeller module in the IBD diagram in the form of a constraint block, thereby achieving dynamic integration of the performance analysis model and system components in the MagicDraw tool.

[0142] This method eliminates the need to rebuild a new system model when faced with different operating conditions or changing performance parameters. Simply modify the parameter values, and the model will automatically reflect the dynamic response of each system component. Furthermore, changes to any parameter value can be globally mapped to the relevant constraint expressions through a binding mechanism, ensuring real-time linkage and dynamic coupling between modules within the system. This modeling approach not only significantly improves modeling efficiency and system reusability, but also effectively supports model consistency management and system-level performance verification under multiple performance conditions.

[0143] The above is a description of the embodiments of the present invention. The above description of the disclosed embodiments will enable those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to those skilled in the art. The general principles defined in this invention may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A UAV propeller reliability modeling method based on MBSE, characterized by: Follow these steps: 1) Demand Analysis Model Generation: The propeller in the UAV power system is identified as the target research unit; propeller parameters are input into the demand analysis module to generate the demand analysis model; 2) System context and use case modeling: Draw the drone system context model at the system level to define the interface relationships between the propeller and the flight control module, electronic speed controller, motor, and battery components. Draw the context model and flight use case diagram at the power subsystem level to indicate the closed-loop execution logic for the propeller to achieve attitude adjustment by adjusting thrust and angular velocity. 3) Performance function and statistical distribution construction: The experimental data of propeller thrust F and torque τ at different speeds were collected. After removing outliers using BOX-COX transformation, the thrust and torque functions were established. 4) Fault threshold setting: Set the thrust threshold F according to the flight mission requirements th , torque threshold τ th and speed threshold ω th ; 5) Joint margin calculation: Calculation of dimensionless performance margin m of thrust, torque and speed by Monte Carlo simulation F 、m τ 、m ω ; Obtain weight w based on min-max normalization and variance analysis F 、w τ 、w ω ; The joint margin M is obtained by calculation: ； 6) Calculation of confidence reliability: The performance fluctuation variance σ m 2 and the model error variance σ e 2 Substitution ； in Φ N (.) is the standard normal cumulative distribution function, M th is the joint margin threshold, and the propeller reliability R is obtained MB .

2. The method according to claim 1, characterized in that The propeller parameters in step 1) include propeller diameter, pitch, center hole diameter, material, weight, self-locking mechanism, blade cross-section thickness, factory balance, and parameters of the compatible platform.

3. The method according to claim 1, characterized in that In step 3), the square of the speed ω² is fitted with a normal distribution to obtain the mean μ and standard deviation σ. The normal distribution fitting adopts the maximum likelihood estimation and the normality is verified by the Shapiro-Wilk test. The thrust function is P F (ω 2 )=C F ω 2 , the torque function is P τ (ω 2 )=C τ ω 2 , coefficient C F with C τ The goodness of fit R was obtained by least squares linear fitting. 2 Not less than 0.

98.

4. The method according to claim 1, wherein Step 4) Threshold F th , τ th 、ω th The weight calculation further includes performing principal component analysis on the normalized features. When the contribution rate of the first principal component is greater than 85%, the weights are corrected using the load coefficient of the first principal component.

5. The method according to claim 1, wherein Step 6) Calculate the statistical fluctuation variance σ of the performance parameter m The formula for ² is as follows: ； in: σ m ² is the variance of the integrated or joint margin; w i is the weight of the i-th performance indicator, reflecting the impact of this indicator on the overall margin; σ m,i ² is the variance of the i-th performance indicator itself, that is, the variance of its performance margin; To sum n indicators; And / or, the model error variance σ e ²Based on expert experience and a comprehensive assessment of historical model-measured deviations, the value range is 0.01-0.

10.

6. The method according to claim 1, characterized in that This method also includes MBSE constraint binding integration: the mathematical equations of steps 3)-6) are inherited and bound to the propeller module in the internal structure diagram through SysML constraint blocks, achieving real-time automatic refresh of reliability when parameters are updated; the constraint binding relationship supports automatic synchronous updates in Magic Draw, Cameo Systems Modeler, or MBSE tools.

7. A model-based UAV propeller reliability modeling system, characterized by: The system implements the method according to any one of claims 1 to 6, including: A. Requirement analysis module, used to receive propeller geometry, material and assembly parameters and generate a requirement analysis model; B. System context modeling module, used to build SysML context models of device hierarchical relationships and interface interactions; C. Use case and activity modeling module, used to define propeller attitude adjustment use cases and draw corresponding activity diagrams; D. Performance modeling module, including: D1. Data acquisition unit, used to obtain propeller thrust and torque experimental data at different speeds; D2. Parameter fitting unit, used to fit the normal distribution of the square of the speed ω² and obtain the coefficient C F 、C τ ; E. Threshold setting unit, used to store and call thrust, torque and speed fault thresholds; F. Joint margin calculation unit, used to calculate the joint margin M based on dimensionless transformation, weight distribution and Monte Carlo simulation; G. Reliability evaluation unit, used to execute the confidence reliability formula and output RMB; H.MBSE constraint binding unit, used to establish inheritance and constraint binding between performance equations and propeller structure models in the MBSE tool environment, enabling real-time updates of parameter changes; I. Processor, used to run instructions of the above modules / units.

8. The system according to claim 7, characterized in that The performance modeling module further includes an outlier processing subunit for performing a BOX-COX transformation and eliminating data points that deviate from ±3σ; and / or, a joint margin calculation unit is configured to call three algorithms, namely, minimum-maximum normalization, variance analysis, and principal component analysis, when calculating weights and automatically select the weight set with the smallest error; and / or, a reliability assessment unit is configured to send a maintenance warning signal to the flight control system when RMB is lower than a set threshold.

9. A computer-readable storage medium having a computer program or instruction stored thereon, characterized in that: When the computer program or instruction is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

10. A computer program product comprising a computer program or instructions, characterized in that When the computer program or instruction is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images