Unmanned aerial vehicle structure fatigue prediction method and system

Abstract

The present application relates to the technical field of unmanned aerial vehicle evaluation, in particular to an unmanned aerial vehicle structure fatigue prediction method and system, the method comprising: acquiring a sensor data set of a key structure part of an unmanned aerial vehicle; dynamically evaluating the data quality of the sensor data set, and dynamically adjusting a data cleaning preprocessing strategy according to the evaluation result; based on the preprocessed sensor data set, combining real-time flight environment parameters, and constructing a stress field model reflecting the global stress distribution of the structure; and calculating the fatigue damage prediction value of each structure part according to the stress field model and the material fatigue characteristics. The purpose is to improve the prediction accuracy of the fatigue damage of the unmanned aerial vehicle structure.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of unmanned aerial vehicle evaluation, in particular to an unmanned aerial vehicle structure fatigue prediction method and system. BACKGROUND

[0002] With the development of low-altitude economy, low-altitude security capability has become a core element of cultivating low-altitude new productivity. For example, in the urban low-altitude environment with high-density buildings, as a key carrier of future urban air transportation and logistics distribution, the structural health status of electric vertical take-off and landing aircraft (eVTOL) is directly related to public safety and urban economic stability. Under this background, the large-scale application of unmanned aerial vehicles in urban low-altitude environment puts forward higher real-time, accuracy and intelligent requirements for structural health monitoring technology.

[0003] At present, the mainstream unmanned aerial vehicle structure fatigue evaluation method in the industry mainly includes periodic visual detection, strain gauge local monitoring and finite element simulation analysis. Although some advanced systems integrate flight data recorders and single strain sensors to realize real-time acquisition of static stress and fatigue life estimation of key parts, there are still limitations. On the one hand, the threshold rules are mostly based on static design parameters, which cannot dynamically adapt to the coupling effect of complex weather conditions and multi-source loads, thereby affecting the structural fatigue prediction accuracy. On the other hand, the existing stress field modeling methods mostly rely on local measurement data of single sensors, which lack effective characterization of the global stress distribution of the structure, resulting in significant error accumulation effect in fatigue damage prediction based on local stress data. It is difficult to meet the high-precision structural fatigue prediction requirements of unmanned aerial vehicles in urban low-altitude environment. SUMMARY

[0004] In order to improve the prediction accuracy of unmanned aerial vehicle structure fatigue damage, the present application provides an unmanned aerial vehicle structure fatigue prediction method and system, and the technical solutions adopted are as follows:

[0005] The technical solution of the first aspect of the present application provides an unmanned aerial vehicle structure fatigue prediction method, which comprises:

[0006] Obtaining a sensor data set of a key structural part of an unmanned aerial vehicle;

[0007] Dynamically evaluating the data quality of the sensor data set, and dynamically adjusting the data cleaning preprocessing strategy according to the evaluation result;

[0008] Based on the preprocessed sensor data set, combining real-time flight environment parameters, a stress field model reflecting the global stress distribution of the structure is constructed;

[0009] According to the stress field model and the material fatigue characteristics, the fatigue damage prediction value of each structural part is calculated.

[0010] Further, dynamically evaluate the data quality of the sensor data set, including:

[0011] Extract the signal-to-noise ratio and battery supply voltage of the sensor data;

[0012] Through a sliding time window, analyze the continuity of the time series of the sensor data, and calculate the data mutation amplitude of adjacent sampling points;

[0013] According to the signal-to-noise ratio, battery supply voltage and data mutation amplitude, calculate the sensor data health index.

[0014] Further, dynamically adjust the data cleaning preprocessing strategy according to the evaluation result, including:

[0015] Compare the data health index at the current time with the preset data quality threshold to calculate the real-time difference;

[0016] Based on the real-time difference, historical difference accumulation and difference change rate, generate a dynamic cleaning decision index;

[0017] According to the dynamic cleaning decision index, adjust the data cleaning frequency in real time.

[0018] Further, based on the preprocessed sensor data set, combine the real-time flight environment parameters to construct a stress field model reflecting the global stress distribution of the structure, including:

[0019] Deploy a sensor group at the target part of the unmanned aerial vehicle to collect static stress and dynamic vibration data;

[0020] Combine the wind speed and temperature parameters in the real-time flight environment to correct the static stress by dynamic vibration amplification, aerodynamic load correction and thermal stress correction;

[0021] Fuse the corrected multi-source stress data to construct a stress field model reflecting the global stress distribution of the structure.

[0022] Further, combine the wind speed and temperature parameters in the real-time flight environment to correct the static stress by dynamic vibration amplification, aerodynamic load correction and thermal stress correction, including:

[0023] Measure the vibration amplitude by an accelerometer and compensate the static stress by a dynamic amplification coefficient based on the material stiffness parameter;

[0024] Based on the real-time wind speed and aerodynamic parameters, calculate the quadratic growth effect of aerodynamic pressure on the structure stress;

[0025] According to the temperature change and the material thermal expansion coefficient, calculate the superposition effect of thermal stress on the structure stress.

[0026] Further, according to the stress field model and material fatigue characteristics, a fatigue damage prediction value of each structure part is calculated, including:

[0027] According to the corresponding relationship between the stress amplitude of the material and the fatigue life, a curve model reflecting the stress amplitude and the life characteristics is established;

[0028] Based on the stress amplitude output by the stress field model for each flight, combined with the cycle number of the historical flight task, the cumulative fatigue damage value of each structure part is calculated.

[0029] Further, based on the stress amplitude output by the stress field model, combined with the cycle number of the historical flight, the cumulative fatigue damage value of each structure part is calculated, including:

[0030] According to the stress amplitude in each flight process, the curve model reflecting the stress amplitude and the life characteristics is matched to obtain the corresponding fatigue life;

[0031] The ratio of the cycle number of each flight to the corresponding fatigue life is accumulated to obtain the cumulative fatigue damage value.

[0032] Further, the method further comprises:

[0033] Based on the fatigue damage prediction value, combined with the flight history data, the environmental parameter record and the structure damage repair record, a multi-dimensional damage evaluation system is constructed;

[0034] According to the real-time monitored flight environment dynamic parameter and task load intensity, a dynamic correction factor is generated to adjust the fatigue damage prediction value;

[0035] According to the adjusted fatigue damage prediction value, a graded maintenance suggestion is generated.

[0036] Further, the method further comprises:

[0037] When the fatigue damage prediction value exceeds a preset safety threshold, an airworthiness deviation warning is triggered;

[0038] The unmanned aerial vehicle of the airworthiness deviation warning is listed in an abnormal aerial vehicle management list, and the unmanned aerial vehicle listed in the management list is implemented with a flight suspension instruction until the management is released after the maintenance verification is completed.

[0039] The technical scheme of the second aspect of the application provides an unmanned aerial vehicle structure fatigue prediction system, which adopts the unmanned aerial vehicle structure fatigue prediction method of the first aspect of the application, and the system comprises:

[0040] A data acquisition module is configured to obtain a sensor data set of a structure part of an unmanned aerial vehicle;

[0041] a data preprocessing module configured to dynamically evaluate the data quality of the sensor data set and dynamically adjust the data cleaning preprocessing strategy according to the evaluation results;

[0042] The stress field distribution module is configured to construct a stress field model reflecting the global stress distribution of the structure based on the preprocessed sensor data set and combined with real-time flight environment parameters;

[0043] The fatigue damage prediction module is configured to calculate the fatigue damage prediction value of each structural part according to the stress field model and the fatigue characteristics of the material.

[0044] The present invention has the following beneficial effects:

[0045] The UAV structural fatigue prediction method provided by the present invention dynamically optimizes the sensor data preprocessing strategy, combines real-time flight environment parameters to construct a global stress field model, and integrates material fatigue characteristics analysis, with the aim of improving the accuracy and reliability of structural fatigue damage prediction. First, through the deployment of multi-source sensors in key structural parts and dynamic data quality assessment, high-precision modeling of the stress field under complex working conditions is achieved. The integration of static stress, dynamic vibration stress, aerodynamic load, and thermal stress coupled corrections breaks through the limitations of traditional single sensor local monitoring and comprehensively perceives the global stress distribution of the structure. Combining material fatigue life characteristics with historical load cycle data, the cumulative damage value is dynamically calculated, and the remaining life of key parts can be accurately predicted. This method can identify potential fatigue risks in real time and support proactive maintenance decisions, thereby effectively extending the service life of UAVs, ensuring the safety of low-altitude urban operations, and providing key technical support for the large-scale application of new low-altitude vehicles such as eVTOL. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0047] Figure 1 A flow chart of a method for predicting structural fatigue of an unmanned aerial vehicle provided by one embodiment of the present invention;

[0048] Figure 2 A schematic diagram of the structure of an unmanned aerial vehicle structural fatigue prediction system provided by one embodiment of the present invention;

[0049] Figure 3 A schematic diagram of a flow chart for handling airworthiness deviations provided in one embodiment of the present invention. DETAILED DESCRIPTION

[0050] To further clarify the technical means and effects taken by the present application to achieve the intended purpose, the following describes in detail the specific implementation, structure, features and effects of a UAV structure fatigue prediction method and system according to the present application, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0051] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0052] The specific scheme of the UAV structure fatigue prediction method and system provided by the present application is described in detail below in combination with the accompanying drawings.

[0053] Please refer to Figure 1 which shows the method flowchart of the UAV structure fatigue prediction method provided by one embodiment of the present application, which comprises:

[0054] Step S100: Obtain a sensor data set of a key structure part of a UAV; Specifically, the sensors include but are not limited to strain gauges, accelerometers, UAV data recording systems, GPS / Beidou receivers, barometers, and laser radars; For strain gauges, according to the stress characteristics of the measurement part, they can be installed at the key structure parts such as the center of gravity, the engine support, the main wing beam, the half-span, the vertical tail root, etc., for measuring the strain of each part and obtaining static stress data; Accelerometers collect dynamic vibration data; UAV data recording systems can be installed in the equipment cabin for collecting, storing and preliminarily processing various sensor data; GPS / Beidou receivers can be installed on the top of the machine body to ensure accurate acquisition of the position, speed and attitude information of the UAV; Barometers are used to measure atmospheric pressure for calculating flight height and other data; Laser radars adjust the installation angle reasonably according to the monitoring requirements and detect the information of obstacles in the flight environment.

[0055] Step S200: Dynamically evaluate the data quality of the sensor data set and dynamically adjust the data cleaning preprocessing strategy according to the evaluation results; Since the traditional data cleaning method uses fixed rules or simple threshold values, it cannot cope with the data changes of the UAV under different flight environments and different health conditions of the sensors, and the adaptive ability is insufficient, therefore, the present embodiment needs to evaluate the sensor data quality before data cleaning, and then dynamically select the cleaning strategy according to the data quality;

[0056] Step S200 specifically includes:

[0057] Step S210: Extract the signal-to-noise ratio of the sensor data and the battery supply voltage; specifically, for the data of each sensor, calculate its signal-to-noise ratio by using a suitable algorithm; taking a strain gauge as an example, compare the average power of the signal with the average power of the noise in a stable measurement period, separate the signal and noise components by filtering the signal, then calculate the power ratio, and take the logarithm to convert it into the signal-to-noise ratio; the battery supply voltage can be directly read from the data recording system of the unmanned aerial vehicle, and the voltage data will be recorded in the data recording system in real time, reflecting the power supply state of the sensor at different times.

[0058] The time series of sensor data is analyzed continuously by a sliding time window, and the data mutation amplitude of adjacent sampling points is calculated; specifically, according to the characteristics and variation frequency of the sensor data, a suitable sliding time window size can be selected, and in each sliding time window, the data mutation amplitude can be represented as:

[0059]

[0060] In the formula, represents the data mutation amplitude, the larger the value, the more unsmooth the data changes, and the worse the continuity; represents the sampling point sequence number in the time series; as the time window slides, the value of each window is calculated in turn.

[0061] According to the signal-to-noise ratio, the battery supply voltage and the data mutation amplitude, the sensor data health index is calculated, and the sensor data health index can be represented as:

[0062]

[0063] In the formula, represents the data health index of the sensor; represents the signal-to-noise ratio; represents the data mutation amplitude; represents the battery supply voltage; represents the time; represents the first weight coefficient, which reflects the influence degree of the signal-to-noise ratio on the data quality; represents the second weight coefficient, which reflects the influence degree of the data mutation amplitude on the data quality; represents the third weight coefficient, which reflects the influence degree of the battery supply voltage on the data quality;

[0064] ​The embodiment can comprehensively and dynamically evaluate the quality of sensor data by extracting the signal-to-noise ratio of sensor data, the battery supply voltage, continuously analyzing the time sequence of sensor data, calculating the data mutation amplitude, and comprehensively calculating the sensor data health index. The data change of the unmanned aerial vehicle in different flight environments and the sensor in different health conditions is fully considered, and the problem of poor self-adaptability of the traditional data cleaning method is overcome. Through obtaining the accurate sensor data health index, a basis is provided for dynamically selecting a suitable data cleaning preprocessing strategy according to the data quality, which helps to improve the pertinence and effectiveness of data cleaning, thereby improving the accuracy and reliability of the unmanned aerial vehicle structure fatigue prediction.

[0065] Step S220: comparing the data health index at the current time with the preset data quality threshold value to calculate the real-time difference; specifically, after obtaining the sensor data health index at each time according to step S210, the sensor data health index at the current time is compared with the preset data quality threshold value , which can be preset according to factors such as unmanned aerial vehicle flight task requirements, sensor performance, and data processing experience, and is used to measure whether the sensor data meets the quality standard for subsequent analysis; the real-time difference is calculated by comparing the sensor data health index at the current time with the preset data quality threshold value ; the difference reflects the gap between the quality of the sensor data at the current time and the expected quality standard;

[0066] Based on the real-time difference, the historical difference accumulation, and the difference change rate, a dynamic cleaning decision index is generated; according to the dynamic cleaning decision index, the data cleaning frequency is adjusted in real time; wherein the dynamic cleaning decision index can be expressed as:

[0067]

[0068] In the formula, , wherein the sensor dynamic cleaning decision index represents that the larger the value is, the more serious the data quality problem is, and a more frequent cleaning strategy needs to be adopted; represents the preset data quality threshold value; represents the step size; represents the proportional coefficient; represents the integral coefficient; represents the differential coefficient; in order to calculate the historical difference accumulation, the real-time difference at each time since the start of monitoring needs to be recorded, and the historical difference accumulation is obtained through integral operation ; represents the difference change rate; the values of the proportional coefficient, the integral coefficient, and the differential coefficient are usually determined according to the characteristics of the unmanned aerial vehicle system, data processing experience, and experimental optimization, and are used to adjust the weights of the real-time difference, the historical difference accumulation, and the difference change rate in the decision index;

[0069] Further, the data cleaning frequency is adjusted in real time according to the calculated dynamic cleaning decision index; specifically, different threshold ranges correspond to different data cleaning frequencies, and the cleaning frequency is dynamically adjusted according to the actual situation of data quality. This dynamic adjustment strategy overcomes the limitations of traditional fixed cleaning rules, reduces unnecessary cleaning operations when the data quality is good, and saves computing resources; when the data quality is problematic or has a deterioration trend, the cleaning frequency is increased in time to ensure the reliability of the data quality entering the subsequent analysis link.

[0070] Step S300: Based on the pre-processed sensor data set, combine the real-time flight environment parameters to construct a stress field model reflecting the global stress distribution of the structure; specifically, the unmanned aerial vehicle will experience various stress cycles in the flight process, such as take-off, landing, and maneuvering actions, which will cause material fatigue; the fatigue damage periodic detection method cannot adapt to the demand for more intelligent and adaptive fatigue prediction methods due to the complex and variable working environment of unmanned aerial vehicles, and the traditional single sensor installed on the body can only measure local stress, which cannot be globally modeled, resulting in misjudgment of the structure failure risk, and the model accuracy is low without considering the coupling effect of flight environment on structural stress.

[0071] Step S300 specifically includes:

[0072] Step S310: Deploy a sensor group at the target part of the unmanned aerial vehicle to collect static stress and dynamic vibration data; specifically, in the sensor array configuration, strain gauges and accelerometers are installed at key positions of the body, and real-time wind speed and temperature during flight are obtained in combination with the body's own sensors; the installation positions of strain gauges and accelerometers include but are not limited to:

[0073] Center of gravity: The center of gravity of the unmanned aerial vehicle is the center point of the mass distribution of the aircraft, and the balance and stability of the unmanned aerial vehicle are closely related to the position of the center of gravity; Engine support: The engine support is a structural part that supports the engine of the unmanned aerial vehicle, and measuring the strain at this point can evaluate the impact of the engine on the structure of the unmanned aerial vehicle; Main wing beam: The main wing beam is the main load-bearing element in the structure of the unmanned aerial vehicle, located in the central wing box, and measuring the strain at this point can help understand the structural state of the wing box during flight; Half wing span: The half wing span refers to the middle position of the wing from the wing root to the wing tip of the unmanned aerial vehicle, and the strain at this point can be used to evaluate the bending and twisting of the wing during flight; Vertical tail root: The vertical tail is the vertical stabilizer of the tail of the unmanned aerial vehicle, and measuring the strain at this point helps to understand the stress condition of the tail structure during flight;

[0074] ​During the flight of the unmanned aerial vehicle, the strain gauge collects the static stress data of each target part in real time, and the accelerometer collects the dynamic vibration data; at the same time, the wind speed sensor and the temperature sensor of the machine body respectively acquire the real-time wind speed and temperature; these data are recorded at a certain sampling frequency and stored in the data acquisition system. In combination with the evaluation and cleaning pretreatment of the sensor data quality in step S200, it is ensured that the collected data are of high quality and reliable, and can truly reflect the actual state of the unmanned aerial vehicle.

[0075] Step S320: In combination with the wind speed and temperature parameters in the real-time flight environment, the static stress is corrected by dynamic vibration amplification, aerodynamic load and thermal stress; specifically, it includes:

[0076] The vibration amplitude is measured by the accelerometer , which reflects the amplification of high-frequency vibration or instantaneous impact on the static stress; in combination with the material stiffness parameter, the dynamic amplification coefficient of the static stress is compensated, which can be expressed as , , which represents the dynamic sensitivity coefficient, which can be determined according to the fatigue characteristics of the material and the structural stiffness; this step considers the stress increase of the aircraft caused by vibration and other factors during flight;

[0077] Based on the real-time wind speed and aerodynamic parameters, the quadratic growth effect of aerodynamic pressure on the structural stress is calculated; specifically, the aerodynamic parameters include air density , drag coefficient and windward area A, and the relevant coefficient of the aerodynamic coupling system is calculated: , which reflects the effect of the square growth of the aerodynamic pressure caused by the increase of the wind speed. The quadratic growth effect of the aerodynamic pressure on the structural stress is calculated, i.e. , which reflects the influence of the wind speed on the structural stress of the aircraft;

[0078] According to the temperature change and the thermal expansion coefficient of the material, the superposition effect of the thermal stress on the structural stress is calculated; specifically, the real-time temperature T is recorded, and the reference temperature is set, and the temperature change is calculated; according to the elasticity and thermal expansion characteristics of the material, the thermal stress coefficient is determined, and the superposition effect of the thermal stress on the structural stress is calculated, i.e. , which considers the thermal stress caused by the expansion or contraction of the material due to temperature change;

[0079] Step S330: Fusion of the corrected multi-source stress data, construction of the stress field model reflecting the global stress distribution of the structure; which can be expressed as:

[0080]

[0081] In the formula, represents the global stress value; Indicates that the static stress is measured by the strain gauge; Represents the dynamic sensitivity coefficient, which is determined by the material fatigue characteristics and structural stiffness; It is measured by accelerometer and reflects the amplification of static stress caused by high-frequency vibration or instantaneous impact; It represents the aerodynamic coupling system and air density , drag coefficient , the correlation coefficient of the windward area A, that is , which is used to reflect that the increase in wind speed causes the aerodynamic pressure to increase in a square manner; Indicates real-time wind speed; represents the thermal stress coefficient, which is determined by the elasticity and thermal expansion of the material; Indicates the temperature change; It is used to correct nonlinear coupling effects not considered in the model and is dynamically adjusted through machine learning or experimental data;

[0082] Finally, based on the calculated total stress at each target location, combined with the structural characteristics and geometric information of the UAV, a stress field model reflecting the global stress distribution of the structure is constructed using methods such as finite element analysis. This model can intuitively display the stress distribution of various parts of the UAV during flight, providing a foundation for subsequent fatigue prediction and structural assessment. This embodiment deploys a sensor array at key locations on the UAV to collect static stress and dynamic vibration data. This data is then combined with wind speed and temperature parameters in the real-time flight environment to perform multi-faceted corrections to the static stresses. Ultimately, this multi-source correction is integrated to construct a stress field model reflecting the global stress distribution of the structure. This method overcomes the limitations of traditional single sensors that can only measure local stress and fail to consider the effects of flight environment coupling, enabling accurate calculation of the stress distribution of key parts of the airframe during flight. By comprehensively considering multiple factors such as dynamic vibration, aerodynamic loads, and thermal stress, the model's accuracy and reliability are improved, providing a more comprehensive and accurate basis for structural fatigue prediction of UAVs.

[0083] Step S400: Calculating fatigue damage prediction values ​​of various structural parts according to the stress field model and material fatigue characteristics;

[0084] Step S400 specifically includes:

[0085] Step S410: According to the corresponding relationship between the stress amplitude of the material and the fatigue life, a curve model reflecting the stress amplitude and the life characteristics is established; specifically, for the key materials used by the unmanned aerial vehicle, the fatigue test can be carried out, the material samples are tested under different stress amplitude levels, and the cycle number of each sample until fatigue failure is recorded. These data reflect the fatigue life of the material under different stress amplitude. After collecting a large amount of fatigue test data, a mathematical fitting method such as least square method is used to establish a curve model of the corresponding relationship between the stress amplitude and the fatigue life. For example, the S-N curve (stress-cycle number curve) can intuitively show the fatigue life variation law of the material under the action of different stress amplitudes. In practical application, according to the characteristics of the material and the test data, a suitable function form can be selected to fit the curve, such as power function, exponential function, etc. By continuously adjusting the function parameters, the model can best fit the test data, so as to accurately reflect the stress amplitude and life characteristics of the material.

[0086] Step S420: Based on the stress amplitude of each flight output by the stress field model, the cycle number of the historical flight task is combined to calculate the cumulative fatigue damage value of each structure part; specifically, the stress amplitude of each structure part in each flight process is extracted from the stress field model constructed in step S300. At the same time, through the data recording system of the unmanned aerial vehicle, the cycle number of each flight is obtained, which can be defined according to the characteristics of the flight task, for example, a complete take-off, flight and landing process can be regarded as a cycle, or the cycle can be defined according to a specific flight action;

[0087] For the stress amplitude of each flight, the corresponding fatigue life is obtained by looking up and matching in the curve model established in step S410 Since the curve model reflects the fatigue life characteristics of the material under different stress amplitudes, the corresponding fatigue life prediction value can be obtained from the model according to the stress amplitude, and the final cumulative fatigue damage value can be expressed as:

[0088]

[0089] In the formula, represents the cumulative fatigue damage; represents the total number of flights; represents the cycle number of the i th flight; represents the cycle number of the i th flight; represents the cycle number of the i th flight; The fatigue life of the material under the secondary flight stress amplitude; in the calculation process, from the first flight, the ratio of the flight cycle number to the corresponding fatigue life is calculated in turn, and the ratios are accumulated. With the increase of flight times, the cumulative fatigue damage value is constantly updated to reflect the fatigue damage degree of each structure part of the unmanned aerial vehicle in real time; if the cumulative damage is close to 1, it means that the structure of a part will soon fatigue fracture. If the cumulative damage is 0.3 after 100 flights according to the calculation, it can be predicted that the remaining life is about 70% of the flight cycle;

[0090] The high-quality sensor data obtained in the embodiment, the stress field model constructed, and the fatigue characteristics of the material can accurately evaluate the fatigue damage state of each structure part of the unmanned aerial vehicle at different flight stages by establishing a curve model of stress amplitude and life characteristics, and combining the stress amplitude and cycle number of each flight to calculate the cumulative fatigue damage value. This enables the operator to master the health status of the key structure of the unmanned aerial vehicle in real time, and predict the structure fatigue fracture risk in advance. In this way, when the cumulative fatigue damage is close to 1, the part that will soon fatigue fracture can be repaired or replaced in time, effectively avoiding flight accidents caused by structure fatigue failure, greatly improving the flight safety and reliability of the unmanned aerial vehicle, ensuring the smooth progress of the flight mission, and also helping to reasonably arrange the maintenance plan and reduce the operating cost.

[0091] Preferably, the method further comprises:

[0092] Step S500: multi-dimensional preventive maintenance evaluation and decision; specifically, the occurrence of fatigue damage is a process of comprehensive action of multiple factors, not only the influence of stress amplitude, but also flight history of the aircraft, flight environment, aging of the material itself, possible collision damage and other factors; the embodiment establishes a multi-dimensional damage evaluation system, comprehensively considers the cumulative fatigue damage caused by stress amplitude, flight environment factors, local damage history of the structure, and considers the accelerated influence of high-load tasks on fatigue through loading analysis of different tasks.

[0093] Step S500 specifically includes:

[0094] Step S510: Based on the fatigue damage prediction value, combined with flight history data, environmental parameter records and structure damage repair records, a multi-dimensional damage assessment system is constructed; Specifically, the fatigue damage prediction value of each structure part is obtained from step S400, and flight history data is collected, which comes from the data recording system of the unmanned aerial vehicle, covering flight time, flight mileage, flight mode (such as the time proportion of different stages such as take-off, cruising, landing, etc.) and other information. Obtain environmental parameter records, including temperature, humidity, wind speed, etc., which are collected and stored by various environmental sensors on the machine. In addition, the structure damage repair record is sorted out, including the location of each damage, damage type, repair time and repair method, etc. Detailed content; Then, through the establishment of a comprehensive evaluation model, the fatigue damage prediction value, flight history data, environmental parameter records and structure damage repair records are used as the input parameters of the model. For example, use the neural network model in the machine learning algorithm to fuse these multi-source data; For example, the flight time and mileage in the flight history data can reflect the use intensity of the aircraft, which is a characteristic dimension of the model; The temperature and humidity in the environmental parameters will affect the fatigue performance of the material, which is another characteristic dimension; The structure damage repair record is converted into a feature recognizable by the model through coding and other methods; Through training this model, it can learn the complex relationship between these factors and fatigue damage, so as to construct a multi-dimensional damage assessment system.

[0095] Step S520: According to the real-time monitoring of flight environment dynamic parameters and task load intensity, a dynamic correction factor is generated to adjust the fatigue damage prediction value; Specifically, various sensors on the unmanned aerial vehicle, such as meteorological sensors, accelerometers, stress sensors, etc., can be used to monitor flight environment dynamic parameters and task load intensity in real time. For example, through the meteorological sensor, it is detected whether it encounters strong turbulence, and if strong turbulence is detected, the real-time stress response of the aircraft is obtained by using the stress sensor; By analyzing these data, the influence degree of the current flight condition on fatigue damage is judged. For unpredictable operation limit conditions, according to the characteristics of extreme flight tasks, such as high overload maneuvering flight, the additional fatigue damage caused by the task to the structure is analyzed, and the corresponding correction factor is calculated, so that the fatigue damage prediction value is more in line with the actual situation; Then, according to different flight conditions, the corresponding dynamic correction factor calculation rules are formulated; Finally, the calculated dynamic correction factor is applied to the fatigue damage prediction value to adjust it.

[0096] Step S530: generating a graded maintenance suggestion according to the adjusted fatigue damage prediction value; specifically, the unmanned aerial vehicle structure is accurately fatigue damage evaluated through the multi-dimensional damage evaluation system, and the maintenance decision suggestion is immediately obtained, and the maintenance suggestion is notified to the unmanned aerial vehicle operator and the city low-altitude supervision department. The maintenance suggestion generation can be made according to the adjusted fatigue damage prediction value, and the graded maintenance standard is formulated. For example, according to the value range of the fatigue damage prediction value, it is determined that the damage is slight and regular inspection and maintenance are suggested; it is determined that the damage is moderate and local structure inspection and repair are suggested; it is determined that the damage is severe and it is suggested to stop flying immediately for overall maintenance or replacement of related parts. Further, through wireless communication technology such as 5G network, the maintenance suggestion is sent to the management platform of the unmanned aerial vehicle operator and the monitoring system of the city low-altitude supervision department. The unmanned aerial vehicle operator can arrange the maintenance work of the unmanned aerial vehicle in time according to the maintenance suggestion, and ensure the safe operation of the aerial vehicle; the city low-altitude supervision department can supervise the maintenance of the unmanned aerial vehicle, and ensure the safety of the city low-altitude flight.

[0097] Please refer to Figure 3 The method further comprises:

[0098] Step S600: airworthiness deviation disposal; specifically, when the fatigue damage prediction value exceeds the preset safety threshold, an airworthiness deviation early warning is triggered; the unmanned aerial vehicle of the airworthiness deviation early warning is listed in the abnormal aerial vehicle management list, and the unmanned aerial vehicle listed in the management list is implemented stop flying instruction until the management control is released after the maintenance verification is completed. Please refer to Figure 3 Before the unmanned aerial vehicle is put into use, the preset safety threshold of the fatigue damage prediction value can be determined according to the design standard of the aerial vehicle, the material characteristics, the flight task requirement and the related airworthiness regulations, combined with the calculation method of the fatigue damage prediction value in step S400, which is a key index for ensuring the safe flight of the unmanned aerial vehicle, and represents the maximum fatigue damage degree that each structure part of the aerial vehicle can withstand under normal flight conditions.

[0099] During the flight of the UAV, the fatigue damage prediction calculation of step S400 is continuously performed to obtain real-time fatigue damage prediction values of each structural part. The urban low-altitude integrated information platform receives these data in real time and compares them with the preset safety threshold. Once the fatigue damage prediction value exceeds the preset safety threshold, the platform immediately triggers the airworthiness deviation warning. After triggering the warning, the urban low-altitude integrated information platform automatically lists the UAV with airworthiness deviation in the abnormal UAV control list (Capability Deviation UAV List, CDUL). The platform records relevant information of the UAV, such as UAV number, model, operator, current location, specific part and degree of fatigue damage warning, etc., to ensure accurate control of the UAV. Then, the information platform uses wireless communication technology to send information to the UAV operator and the urban low-altitude supervision department at the same time. The information sent to the UAV operator explicitly includes a grounding instruction, which requires it to immediately stop the flight mission of the UAV; after receiving the grounding instruction, the UAV operator arranges maintenance personnel to repair the UAV. During the repair process, the maintenance personnel can repair the fatigue damage part or replace the damaged parts according to the repair suggestions generated by the multi-dimensional preventive maintenance evaluation and decision-making in step S500. After the repair is completed, the UAV needs to be verified for repair, and the verification methods can include ground testing, simulated flight testing, etc., to ensure that the structural strength and performance of the UAV are restored to the safety standard. After verification, the operator submits a control removal application to the urban low-altitude integrated information platform, and the platform removes the UAV from the abnormal UAV control list after verifying that the repair verification data is correct, and restores its normal flight qualification. If the urban low-altitude integrated information platform issues a notice, and does not receive a repair response from the UAV operator within the specified time, the urban low-altitude supervision department will continue to implement measures to stop the UAV in the CDUL from operating in the urban low-altitude, and only after receiving the repair response and completing the repair verification, the operation restriction on the UAV will be lifted.

[0100] The embodiment forms a complete safety guarantee closed loop, effectively avoids the UAV with serious fatigue damage continuing to fly, reduces the risk of flight accidents, ensures the safety order of urban low-altitude flight, and protects the personal safety of residents and urban infrastructure. Through the clear control process and repair verification requirements, the overall safety and reliability of the UAV are improved.

[0101] Please refer to Figure 2 , which shows a structural schematic diagram of a UAV structure fatigue prediction system provided by an embodiment of the application, the system comprising:

[0102] a data acquisition module configured to obtain a sensor data set of a structural part of a UAV;

[0103] a data preprocessing module configured to dynamically evaluate data quality of the sensor data set and dynamically adjust a data cleaning preprocessing strategy according to an evaluation result;

[0104] a stress field distribution module configured to construct a stress field model reflecting global stress distribution of the structure based on the preprocessed sensor data set and in combination with real-time flight environment parameters;

[0105] a fatigue damage prediction module configured to calculate fatigue damage prediction values of each structure part according to the stress field model and material fatigue characteristics.

[0106] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description, and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or can be advantageous.

[0107] Each of the embodiments in the specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other. Each embodiment focuses on the difference from other embodiments.

Claims

1. A method for predicting structural fatigue of an unmanned aerial vehicle, characterized in that: The method comprises: Obtain sensor data sets of key structural parts of UAVs; Dynamically evaluate the data quality of the sensor data set and dynamically adjust the data cleaning preprocessing strategy based on the evaluation results; Based on the preprocessed sensor data set and combined with real-time flight environment parameters, a stress field model reflecting the global stress distribution of the structure is constructed, including: Deploy a sensor group at the target location of the UAV to collect static stress and dynamic vibration data; Combined with wind speed and temperature parameters in the real-time flight environment, dynamic vibration amplification correction, aerodynamic load correction and thermal stress correction are performed on static stress, including: The vibration amplitude is measured by an accelerometer, and the static stress is compensated by the dynamic amplification factor in combination with the material stiffness parameters; Calculate the quadratic growth effect of aerodynamic pressure on structural stress based on real-time wind speed and aerodynamic parameters; Calculate the superposition effect of thermal stress on structural stress based on temperature change and material thermal expansion coefficient; The corrected multi-source stress data are integrated to construct a stress field model reflecting the global stress distribution of the structure; The fatigue damage prediction value of each structural part is calculated based on the stress field model and the fatigue characteristics of the material.

2. The method for predicting structural fatigue of an unmanned aerial vehicle according to claim 1, wherein: Dynamically evaluating the data quality of the sensor data set includes: Extract the signal-to-noise ratio and battery supply voltage of sensor data; The continuity analysis of the time series of sensor data is performed through a sliding time window to calculate the data mutation amplitude of adjacent sampling points; A sensor data health index is calculated according to the signal-to-noise ratio, the battery supply voltage, and the data mutation amplitude.

3. The method for predicting structural fatigue of an unmanned aerial vehicle according to claim 2, wherein: Dynamically adjust data cleaning and preprocessing strategies based on evaluation results, including: Compare the current data health index with the preset data quality threshold and calculate the real-time difference; Generate a dynamic cleaning decision index based on the real-time difference, the historical difference accumulation and the difference change rate; The data cleaning frequency is adjusted in real time according to the dynamic cleaning decision index.

4. The method for predicting structural fatigue of an unmanned aerial vehicle according to claim 1, wherein: According to the stress field model and material fatigue properties, the fatigue damage prediction value of each structural part is calculated, including: According to the corresponding relationship between the stress amplitude and fatigue life of the material, a curve model reflecting the stress amplitude and life characteristics is established; Based on the stress amplitude of each flight output by the stress field model and the number of cycles of historical flight missions, the cumulative fatigue damage value of each structural part is calculated.

5. The method for predicting structural fatigue of an unmanned aerial vehicle according to claim 4, wherein: Based on the stress amplitude output by the stress field model and the number of historical flight cycles, the cumulative fatigue damage value of each structural part is calculated, including: According to the stress amplitude during each flight, the curve model reflecting the stress amplitude and life characteristics is matched to obtain the corresponding fatigue life; The cumulative fatigue damage value is obtained by adding up the ratio of the number of flight cycles to the corresponding fatigue life.

6. The method for predicting structural fatigue of an unmanned aerial vehicle according to any one of claims 1 to 5, wherein: The method further comprises: Based on the fatigue damage prediction value, combined with flight history data, environmental parameter records and structural damage repair records, a multi-dimensional damage assessment system is constructed; generating a dynamic correction factor to adjust the fatigue damage prediction value based on real-time monitored flight environment dynamic parameters and mission load intensity; Generate graded maintenance recommendations based on adjusted fatigue damage predictions.

7. The method for predicting structural fatigue of an unmanned aerial vehicle according to claim 6, wherein: The method further comprises: When the fatigue damage prediction value exceeds a preset safety threshold, an airworthiness deviation warning is triggered; The unmanned aerial vehicles with the airworthiness deviation warning will be included in the abnormal aircraft control list, and grounding orders will be issued to the unmanned aerial vehicles included in the control list until the control is lifted after the maintenance verification is completed.

8. Unmanned aerial vehicle structural fatigue prediction system, characterized by: The method for predicting structural fatigue of an unmanned aerial vehicle according to any one of claims 1 to 7 is adopted, wherein the system comprises: a data acquisition module configured to acquire a sensor data set of a structural part of the UAV; a data preprocessing module configured to dynamically evaluate the data quality of the sensor data set and dynamically adjust the data cleaning preprocessing strategy according to the evaluation results; The stress field distribution module is configured to construct a stress field model reflecting the global stress distribution of the structure based on the preprocessed sensor data set and combined with real-time flight environment parameters; The fatigue damage prediction module is configured to calculate the fatigue damage prediction value of each structural part according to the stress field model and the fatigue characteristics of the material.

Images