Fault detection method and system for rotor head in dynamic work

Abstract

The invention discloses a fault detection method and system for a rotor head under dynamic work, and relates to the technical field of fault detection methods for rotor heads, and the method comprises the steps: determining a plurality of dynamic abnormal regions according to the recognition of a multi-dimensional state diagram of the rotor head; and predicting a plurality of fault features according to the region position of each dynamic abnormal region and the corresponding region form, and determining a plurality of key fault features of the rotor head based on the plurality of fault features, the noise distribution diagram of the rotor head and the corresponding flight abnormal event, thereby improving the accuracy of the plurality of key fault features of the rotor head. Determining a fault path of the rotor head in the rotation process according to the service life of the rotor head, the rotation display diagram of the rotor head and the plurality of key fault features; the multiple fault nodes are determined according to detection of the fault paths, the emergency rotation event of the rotor head is determined according to the multiple fault nodes, the remaining flight paths of the autorotorcraft and the current damage state of the rotor head, and the accuracy of the emergency rotation event of the rotor head is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of fault detection methods of rotor heads, and particularly relates to a fault detection method and system of a rotor head under dynamic work. BACKGROUND

[0002] With the development of science and technology, the rotor head of a autorotating rotorcraft is a key structural component connecting rotor blades and a fuselage, which bears the important functions of transmitting lift, controlling flight attitude and ensuring normal operation of the rotor. The rotor head is usually composed of a hub, a bearing and a connecting piece, and can realize free rotation and tilting of the rotor blades to adapt to various operating requirements in flight. In the prior art, the rotor head is in a continuous rotating state, and the fault of the rotor head is determined by comparing the rotating parameters of the rotor head with the preset rotating parameters, the fault detection is performed along a single dimension, and the noise distribution map of the rotor head and the corresponding flight abnormal event are not considered, which affects the accuracy of multiple key fault features of the rotor head and ignores the accuracy of emergency rotation events of the rotor head. SUMMARY

[0003] The present application aims to overcome the shortcomings of the prior art, and provides a fault detection method and system of a rotor head under dynamic work.

[0004] The present application provides a fault detection method of a rotor head under dynamic work, which comprises the following steps:

[0005] When the autorotating rotorcraft is in a flight state, multiple working parameters of the rotor head are collected, and the working state of the rotor head is determined according to the multiple working parameters of the rotor head and the flight mode of the autorotating rotorcraft;

[0006] A multidimensional state map of the rotor head is determined based on the rotating mode of the rotor head, the multiple working parameters and the vibration parameters of the rotor head in the rotating process;

[0007] Multiple dynamic abnormal regions are determined according to the identification of the multidimensional state map of the rotor head, multiple fault features are predicted according to the region position and the corresponding region mode of each dynamic abnormal region, and multiple key fault features of the rotor head are determined based on the multiple fault features, the noise distribution map of the rotor head and the corresponding flight abnormal event;

[0008] The service life of the rotor head is collected, and a fault path of the rotor head in the rotating process is determined according to the service life of the rotor head, the rotating display map of the rotor head and the multiple key fault features;

[0009] Multiple fault nodes are determined according to the detection of the fault path, and an emergency rotation event of the rotor head is determined according to the multiple fault nodes, the remaining flight path of the autorotating rotorcraft and the current damaged state of the rotor head.

[0010] The embodiment of the present application provides a rotor head fault detection system under dynamic working, which is applied to the rotor head fault detection method under dynamic working, and comprises the following steps:

[0011] The working state module is used for collecting a plurality of working parameters of the rotor head when the autorotating rotorcraft is in a flight state, and determining the working state of the rotor head according to the plurality of working parameters of the rotor head and the flight mode of the autorotating rotorcraft.

[0012] The multi-dimensional state diagram module is used for determining the multi-dimensional state diagram of the rotor head based on the rotation mode of the rotor head, the plurality of working parameters and the vibration parameters of the rotor head in the rotation process.

[0013] The key fault feature module is used for determining a plurality of dynamic abnormal regions according to the identification of the multi-dimensional state diagram of the rotor head, predicting a plurality of fault features according to the region position and the corresponding region mode of each dynamic abnormal region, and determining a plurality of key fault features of the rotor head based on the plurality of fault features, the noise distribution diagram of the rotor head and the corresponding flight abnormal event.

[0014] The fault path module is used for collecting the service life of the rotor head, and determining the fault path of the rotor head in the rotation process according to the service life of the rotor head, the rotation display diagram of the rotor head and the plurality of key fault features.

[0015] The emergency rotation event module is used for determining a plurality of fault nodes according to the detection of the fault path, and determining the emergency rotation event of the rotor head according to the plurality of fault nodes, the remaining flight path of the autorotating rotorcraft and the current damaged state of the rotor head.

[0016] Compared with the prior art, the present application has the following advantages:

[0017] In the embodiment of the present application, the method in the embodiment of the present application is used to collect a plurality of working parameters of the rotor head when the autorotating rotorcraft is in a flight state, and determine the working state of the rotor head according to the plurality of working parameters of the rotor head and the flight mode of the autorotating rotorcraft. The multi-dimensional state diagram of the rotor head is determined based on the rotation mode of the rotor head, the plurality of working parameters and the vibration parameters of the rotor head in the rotation process. A plurality of dynamic abnormal regions are determined according to the identification of the multi-dimensional state diagram of the rotor head, a plurality of fault features are predicted according to the region position and the corresponding region mode of each dynamic abnormal region, and a plurality of key fault features of the rotor head are determined based on the plurality of fault features, the noise distribution diagram of the rotor head and the corresponding flight abnormal event. The multi-dimensional state diagram of the rotor head is introduced, the overall consideration of the plurality of fault features, the noise distribution diagram of the rotor head and the corresponding flight abnormal event is compatible, and the accuracy of the plurality of key fault features of the rotor head is improved.

[0018] Therefore, the failure path of the rotor head in the rotating process is determined according to the service life of the rotor head, the rotating display diagram of the rotor head and the plurality of key failure features, the plurality of failure nodes are determined according to the detection of the failure path, the emergency rotating event of the rotor head is determined according to the plurality of failure nodes, the remaining flight path of the autorotating rotorcraft and the current damaged state of the rotor head, the failure path of the rotor head in the rotating process is introduced, the overall consideration of the plurality of failure nodes, the remaining flight path of the autorotating rotorcraft and the current damaged state of the rotor head is realized, and the accuracy of the emergency rotating event of the rotor head is improved. BRIEF DESCRIPTION OF DRAWINGS

[0019] Figure 1 is a flowchart of the failure detection method of the rotor head in the dynamic working state in the embodiment of the application;

[0020] Figure 2 is a flowchart of step S11 in the failure detection method of the rotor head in the dynamic working state in the embodiment of the application;

[0021] Figure 3 is a flowchart of step S12 in the failure detection method of the rotor head in the dynamic working state in the embodiment of the application;

[0022] Figure 4 is a flowchart of step S13 in the failure detection method of the rotor head in the dynamic working state in the embodiment of the application;

[0023] Figure 5 is a flowchart of step S14 in the failure detection method of the rotor head in the dynamic working state in the embodiment of the application;

[0024] Figure 6 is a flowchart of step S15 in the failure detection method of the rotor head in the dynamic working state in the embodiment of the application;

[0025] Figure 7 is a structural composition diagram of the failure detection system of the rotor head in the dynamic working state in the embodiment of the application. DETAILED DESCRIPTION

[0026] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application.

[0027] Please refer to Figures 1 to 7 A failure detection method of a rotor head in a dynamic working state is applied to a failure detection scene of the rotor head.

[0028] Step S11: collecting a plurality of working parameters of the rotor head when the autorotation rotorcraft is in a flight state, determining a working state of the rotor head according to the plurality of working parameters of the rotor head and a flight mode of the autorotation rotorcraft;

[0029] Step S12: determining a multi-dimensional state diagram of the rotor head based on a rotating mode of the rotor head, the plurality of working parameters and a vibration parameter of the rotor head in a rotating process;

[0030] Step S13: determining a plurality of dynamic abnormal regions according to recognition of the multi-dimensional state diagram of the rotor head, predicting a plurality of fault features according to a region position and a corresponding region mode of each dynamic abnormal region, and determining a plurality of key fault features of the rotor head based on the plurality of fault features, a noise distribution diagram of the rotor head and a corresponding flight abnormal event;

[0031] Step S14: collecting a service life of the rotor head, and determining a fault path of the rotor head in the rotating process according to the service life of the rotor head, the rotating display diagram of the rotor head and the plurality of key fault features;

[0032] Step S15: determining a plurality of fault nodes according to detection of the fault path, and determining an emergency rotating event of the rotor head according to the plurality of fault nodes, a remaining flight path of the autorotation rotorcraft and a current damaged state of the rotor head.

[0033] Reference Figure 2 In step S11, the specific steps are as follows:

[0034] S111: real-time monitoring a flight state of the autorotation rotorcraft, the rotor head being in a continuous rotating state when the autorotation rotorcraft is in the flight state, and collecting a plurality of working parameters of the rotor head in the continuous rotating state, and determining a working mode diagram of the rotor head according to the plurality of working parameters of the rotor head and the continuous rotating state of the rotor head;

[0035] S112: determining a plurality of working mode combinations according to recognition of the working mode diagram of the rotor head, determining corresponding working features based on recognition of each working combination mode, collecting a plurality of working features of the rotor head, and determining a working state of the rotor head according to the plurality of working features of the rotor head, a current flight parameter combination of the autorotation rotorcraft and a flight mode of the autorotation rotorcraft.

[0036] In the embodiment of the present application, the flight state of the autorotation rotorcraft is real-time monitored, the rotor head is in a continuous rotating state when the autorotation rotorcraft is in the flight state, and a plurality of working parameters of the rotor head are collected in the continuous rotating state, and a working mode diagram of the rotor head is determined according to the plurality of working parameters of the rotor head and the continuous rotating state of the rotor head, which is compatible with the overall consideration of the plurality of working parameters of the rotor head and the continuous rotating state of the rotor head, and ensures the accuracy of the working mode diagram of the rotor head.

[0037] At this time, the flight state parameters of the autorotating rotorcraft are acquired in real time by a flight data acquisition system (such as a flight control computer, an inertial measurement unit IMU, a GPS, etc.), including flight altitude, speed, attitude angle (pitch, roll, yaw), acceleration, etc.; these parameters are collected and transmitted to the processing unit at a high frequency (such as 50-100 Hz); the rotor head speed is monitored by a rotor speed sensor (such as a Hall effect sensor or an optical encoder) to confirm that it is in a continuous rotation state (such as 300-400 RPM); only when the speed is stable within the normal working range, the subsequent parameter acquisition is carried out; a speed threshold is set, when the speed is lower than the minimum working speed (such as 200 RPM), the system determines that it is in an abnormal working state, and the data acquisition is suspended.

[0038] At the same time, under the condition of continuous rotation of the rotor head, a plurality of key working parameters are synchronously collected: vibration parameters: three-axis vibration data are collected by acceleration sensors installed at different positions of the rotor head; temperature parameters: the temperature of the key parts of the rotor head is collected by temperature sensors; stress parameters: stress data of the rotor blade root are collected by strain gauges; acoustic parameters: sound characteristics during the operation of the rotor head are collected by a microphone array; further, the collected multi-dimensional parameters are processed by a data fusion algorithm to generate a working mode diagram of the rotor head; this is a visual representation of a multi-dimensional feature space, which usually includes: parameter change curves on the time axis; correlation analysis diagram between parameters; frequency domain feature distribution diagram.

[0039] Further, a plurality of working mode combinations are determined according to the identification of the working mode diagram of the rotor head, corresponding working features are determined based on the identification of each working combination mode, the working state of the rotor head is determined according to the plurality of working features of the rotor head, the current flight parameter combination of the autorotating rotorcraft and the flight mode of the autorotating rotorcraft, which takes into account the overall consideration of the plurality of working features of the ABC rotor head, the current flight parameter combination of the autorotating rotorcraft and the flight mode of the autorotating rotorcraft, ensuring the accuracy of the working state of the rotor head.

[0040] At this time, the working mode diagram generated by S111 is subjected to image recognition and mode analysis, and is decomposed into a plurality of working mode combinations; each working mode combination represents the working mode of the rotor head under certain conditions, such as "high load stable mode", "low load transition mode", "vibration abnormal mode", etc.; optionally, an image recognition algorithm (such as convolutional neural network CNN) is used to extract features from the diagram, and a clustering algorithm (such as K-means) is used to classify similar working modes into the same combination.

[0041] Deep analysis is performed on each identified work form combination, key feature parameters thereof are extracted, and work features are formed; these features include vibration feature frequency, temperature change rate, stress distribution mode, acoustic features, etc., at the same time, feature extraction algorithms (such as wavelet transform and Fourier transform) are used to extract features from original data, and principal component analysis (PCA) is used for dimension reduction to retain the most representative features; a plurality of work features of the rotor head are systematically collected and stored, and a feature database is established; each feature includes feature values, time stamps, confidence levels and other metadata.

[0042] A current flight parameter combination of the autorotating rotorcraft is synchronously acquired, including airspeed, altitude, attitude angle, acceleration, engine power and the like; these parameters reflect the influence of the current flight state on the work of the rotor head; the flight parameters are acquired in real time through a flight control system interface, and data preprocessing and standardization are performed; at the same time, according to the flight parameter combination and historical flight data, the current flight mode is determined, such as "take-off mode", "cruise mode", "maneuver mode", "landing mode" and the like; under different flight modes, the work state evaluation criteria of the rotor head are different; a flight mode recognition algorithm (such as decision tree or support vector machine SVM) is used to classify the flight parameters to determine the current flight mode; further, a plurality of work features of the rotor head, the current flight parameter combination and the flight mode are comprehensively analyzed, and through multi-factor fusion analysis, the work state of the rotor head is finally determined, such as "normal state", "warning state", "abnormal state" and the like; a multi-sensor data fusion algorithm (such as D-S evidence theory or Bayesian network) is used for comprehensive judgment, and the work state and its confidence level are output.

[0043] Reference Figure 3 In step S12, the specific steps are as follows:

[0044] S121: The rotor head rotates along the work state, and rotates images of the rotor head at different time periods are collected; a plurality of sub-rotation features of the rotor head are determined according to the recognition of the rotation images of the rotor head at different time periods; and the rotation form of the rotor head is determined based on the plurality of sub-rotation features and the work state of the rotor head;

[0045] S122: A plurality of work parameters of the rotor head are collected; a first state diagram is determined according to the rotation form of the rotor head and the plurality of work parameters of the rotor head; a vibration parameter of the rotor head in the rotation process is collected, and a vibration position corresponding to the vibration parameter is marked; a second state diagram is determined according to the vibration parameter of the rotor head in the rotation process, the corresponding vibration position and the rotation form of the rotor head; and a multi-dimensional state diagram of the rotor head is constructed according to the first state diagram and the second state diagram.

[0046] In the embodiments of the present application, the rotor head rotates along the working state, and rotates images of the rotor head at different time periods are collected, a plurality of sub-rotation characteristics of the rotor head are determined according to recognition of the rotation images of the rotor head at different time periods, and the rotation mode of the rotor head is determined based on the plurality of sub-rotation characteristics and the working state of the rotor head, which comprehensively considers the plurality of sub-rotation characteristics and the working state of the rotor head, and ensures the accuracy of the rotation mode of the rotor head.

[0047] At this time, the current "working state" (such as "cruise stable state") of the rotor head has been determined in S112; in this step, the rotor head continuously rotates in this state, and the high-speed industrial camera (frame rate is usually above 500-1000 fps) is used to collect images of the rotor head from multiple angles (such as top view, side view, oblique view); at the same time, the subtle motion characteristics of the rotor head under real dynamic working conditions are captured, such as blade deformation, hub motion, relative displacement of connecting parts, etc.; usually 500-1000 Hz to ensure that high-frequency vibration or slight deformation can be captured; multi-angle synchronous shooting combined with stroboscopic lighting or laser assisted positioning to reduce motion blur; covering at least 2-3 complete working periods (such as 10-30 seconds) to capture behaviors under different loads and attitudes.

[0048] Image processing and feature extraction are performed on the collected image sequence to identify a plurality of "sub-rotation characteristics" of the rotor head during rotation; the sub-rotation characteristics include but are not limited to: blade angle change (periodic swing); blade bending / twisting deformation; hub axial / radial displacement; relative motion of connecting parts (such as bearings, hinges); blade tip trajectory deviation; optionally, extract the blade profile and calculate the deformation; calculate the pixel displacement for analyzing the blade motion trajectory; track specific marker points (such as reflective markers attached to the hub); such as using CNN or VisionTransformer to automatically extract motion features.

[0049] The plurality of extracted sub-rotation characteristics and the current working state (such as "cruise stable state") are data fused and pattern matched to finally determine the "rotation mode" of the rotor head; the rotation mode describes the comprehensive motion pattern of the rotor head under a specific working state, for example: "stable synchronous mode": all blades move consistently without abnormal deviation; "asymmetric swing mode": there is periodic abnormal swing of a blade; "blade lag mode": there is a delay in the response of the blade, which is caused by abnormal damping; optionally, use a time series clustering algorithm (such as DTW or K-means) to classify the sub-features; combine a state mapping model (such as SVM or random forest) to judge the current rotation mode type; the output is "mode label" + "confidence", for example: "asymmetric swing mode, confidence 92%".

[0050] Specifically, the model: a certain type of autorotation rotorcraft; working state: cruising steady state (height 1000 meters, speed 120 km / h); acquisition equipment: 3 high-speed industrial cameras (1000 fps), arranged above the rotor head, front side, and oblique back to collect for 20 seconds; in the cruising state, the rotor head rotates stably at 400 rpm; use three cameras to shoot synchronously, each frame of image has a time stamp; each camera collects 20,000 frames of images (1000 fps x 20 seconds); sub-rotation feature extraction: blade angle change: through edge detection + template matching, calculate the angle change of each blade relative to the hub, find that blade #2 has about 1.5° extra swing per cycle; blade bending deformation: through optical flow analysis of the displacement difference between the middle section and the tip of the blade, it is found that blade #2 has obvious bending (maximum deformation 3mm) when rotating at high speed; hub axial displacement: through marker point tracking, it is found that the hub has a 0.2mm periodic axial jump during rotation.

[0051] Input the extracted sub-features into the pre-trained morphological classification model (random forest model trained based on historical data); model output: morphological label: "asymmetric swing morphology"; confidence: 94%; main abnormal source: blade #2 has periodic swing and bending abnormalities; reason: damper of blade #2 is aging or connecting piece is loose; through S121 step, the "asymmetric swing morphology" of the rotor head in cruising state is successfully identified, and the abnormality of blade #2 is accurately located; the determination of this rotation morphology provides a key basis for subsequent vibration analysis, fault feature extraction and fault path prediction.

[0052] Further, a plurality of working parameters of the rotor head are collected, and a first state diagram is determined according to the rotation morphology of the rotor head and the plurality of working parameters of the rotor head; the vibration parameters of the rotor head in the rotation process are collected, and the vibration positions corresponding to the vibration parameters are marked, and a second state diagram is determined according to the vibration parameters of the rotor head in the rotation process, the corresponding vibration positions and the rotation morphology of the rotor head; a multi-dimensional state diagram of the rotor head is constructed according to the first state diagram and the second state diagram, which comprehensively considers the vibration parameters of the rotor head in the rotation process, the corresponding vibration positions and the rotation morphology of the rotor head, and ensures the accuracy of the second state diagram.

[0053] At this time, a plurality of working parameters of the rotor head are collected, and a first state diagram is determined according to the rotating mode of the rotor head and the plurality of working parameters of the rotor head; in S121, the rotating mode of the rotor head (such as the “asymmetric swing mode”) has been determined; on this basis, a plurality of working parameters of the rotor head are collected, such as rotating speed, torque, temperature, stress, etc., and the first state diagram is constructed in combination with the rotating mode; working parameter collection: rotating speed: the rotating speed (unit: rpm) of the rotor head is collected in real time through a rotating speed sensor; torque: the torque (unit: N·m) transmitted by the rotor head is measured through a torque sensor; temperature: the temperature (unit: ℃) of the key parts of the rotor head is monitored through a temperature sensor; stress: the stress (unit: MPa) of the key parts of the rotor head is measured through a strain gauge.

[0054] First state diagram construction: the collected working parameters are combined with the rotating mode to construct a state diagram of a multi-dimensional parameter space; the state diagram can be a multi-dimensional scatter plot, a heat map or an isogram, which shows the relationship and distribution between different parameters; for example, the rotating speed is taken as the X axis, the torque is taken as the Y axis, and the temperature is taken as the color gradient to draw the state diagram.

[0055] A second state diagram is determined according to the vibration parameters of the rotor head in the rotating process, the corresponding vibration positions and the rotating mode of the rotor head; in the rotating process of the rotor head, the vibration parameters (such as vibration acceleration, vibration frequency, vibration amplitude) are collected, and the specific positions (such as blades, hubs, connecting pieces, etc.) corresponding to the vibration parameters are marked; in combination with the rotating mode, the second state diagram is constructed; vibration parameter collection: vibration acceleration: the vibration acceleration (unit: g) of different positions of the rotor head is measured through an acceleration sensor; vibration frequency: the main frequency and harmonic components (unit: Hz) of the vibration are obtained through frequency spectrum analysis; vibration amplitude: the maximum amplitude (unit: mm) of the vibration is measured; a plurality of vibration sensors (such as the positions of the blade root, the hub center and the connecting piece) are arranged on the rotor head; the data of each sensor corresponds to a specific position, which is used for subsequent analysis.

[0056] Second state diagram construction: the vibration parameters are combined with the position information to construct a vibration distribution diagram; the state diagram can be a 3D vibration distribution diagram, a vibration spectrum diagram or a vibration heat map; for example, the position of the rotor head is taken as the X axis, the vibration frequency is taken as the Y axis, and the vibration amplitude is taken as the color gradient to draw the state diagram.

[0057] According to the first state diagram and the second state diagram, a multi-dimensional state diagram of the rotor head is constructed, the first state diagram (working parameter state diagram) and the second state diagram (vibration parameter state diagram) are fused to construct a multi-dimensional state diagram of the rotor head; at the same time, the working parameters and the vibration parameters are spliced or weighted fused at the data level; the characteristics of the working parameters and the vibration parameters are extracted to construct a comprehensive feature vector; the states of the working parameters and the vibration parameters are analyzed respectively, and then the overall state is comprehensively judged; the multi-dimensional state diagram can be a high-dimensional scatter plot, a parallel coordinate diagram or a dimensionality reduction visualization diagram (such as t-SNE, PCA); the state diagram shows the comprehensive state of the rotor head under the joint action of the working parameters and the vibration parameters; for example, the working parameters are taken as one group of dimensions, and the vibration parameters are taken as another group of dimensions, and a comprehensive state diagram is drawn.

[0058] Specifically, a certain autorotation rotorcraft in a cruising state (rotational speed 400 rpm, flight height 1000 m) determines that the rotor head is in an “asymmetric swing mode” through S121; for the first state diagram, the working parameters are collected: rotational speed: 400 rpm (stable); torque: 120 N·m (normal range 100-150 N·m); temperature: 65℃ (normal range ≤80℃); stress: blade root stress 150 MPa (normal range ≤200 MPa); the first state diagram: taking the rotational speed as the X-axis, the torque as the Y-axis, and the temperature as the color gradient, a state diagram is drawn; the state diagram shows that the current working point (400 rpm, 120 N·m, 65℃) is located at the center of the normal working area.

[0059] For the second state diagram, the vibration parameters are collected: blade #1 root vibration acceleration: 1.0 g (normal); blade #2 root vibration acceleration: 1.8 g (abnormal, normal ≤1.2 g); hub center vibration acceleration: 0.5 g (normal); vibration main frequency: 30 Hz (normal 25 Hz); vibration position marker: blade #2 root vibration is abnormal, other positions are normal; the second state diagram: taking the rotor head position as the X-axis, the vibration frequency as the Y-axis, and the vibration amplitude as the color gradient, a state diagram is drawn; the state diagram shows that the vibration amplitude of blade #2 root at 30 Hz is abnormal (red highlight).

[0060] For the multi-dimensional state diagram, the first and second state diagrams are fused, and the working parameters (rotational speed, torque, temperature, stress) and the vibration parameters (vibration acceleration, frequency, position) are fused; using a parallel coordinate diagram to show: rotational speed: 400 rpm; torque: 120 N·m; temperature: 65℃; stress: 150 MPa; blade #2 vibration: 1.8 g; vibration frequency: 30 Hz; the multi-dimensional state diagram: the parallel coordinate diagram shows that the vibration parameters of blade #2 deviate from the normal range, and other parameters are normal; comprehensive judgment: the overall state of the rotor head is “warning state”, and blade #2 has a potential fault.

[0061] Reference Figure 4In step S13, the specific steps are:

[0062] S131: Collecting a multi-dimension state graph of the rotor head, determining a plurality of high-frequency vibration regions based on recognition of the multi-dimension state graph of the rotor head, determining a plurality of to-be-detected regions associated with the plurality of high-frequency vibration regions according to tracing of the plurality of high-frequency vibration regions, and determining a plurality of dynamic abnormal regions according to region morphologies of the plurality of to-be-detected regions, region morphologies of each high-frequency vibration region, and a rotation morphology of the rotor head.

[0063] S133: In each dynamic abnormal region, a plurality of sub-dynamic abnormal features are determined according to recognition of the dynamic abnormal region, meanwhile, a damaged position of the rotor head is collected, a fault range is constructed according to the plurality of sub-dynamic abnormal features, the damaged position of the rotor head, and a plurality of vibration parameters, and a plurality of fault features are predicted according to the fault range, region positions of each dynamic abnormal region, and corresponding region morphologies.

[0064] S133: A plurality of noise features are determined based on sound detection of the rotor head, a noise distribution graph of the rotor head is constructed according to the plurality of noise features and the rotation morphology of the rotor head, and a flight abnormal event associated with the noise features is marked, and a plurality of key fault features of the rotor head are determined according to the plurality of fault features, the noise distribution graph of the rotor head, and the corresponding flight abnormal event.

[0065] In the embodiment of the present application, a multi-dimension state graph of the rotor head is collected, a plurality of high-frequency vibration regions are determined based on recognition of the multi-dimension state graph of the rotor head, a plurality of to-be-detected regions associated with the plurality of high-frequency vibration regions are determined according to tracing of the plurality of high-frequency vibration regions, and a plurality of dynamic abnormal regions are determined according to region morphologies of the plurality of to-be-detected regions, region morphologies of each high-frequency vibration region, and a rotation morphology of the rotor head, which is compatible with overall consideration of the region morphologies of the plurality of to-be-detected regions, the region morphologies of each high-frequency vibration region, and the rotation morphology of the rotor head, and ensures the accuracy of the plurality of dynamic abnormal regions.

[0066] At this time, a multi-dimension state graph of the rotor head is collected, the multi-dimension state graph is a comprehensive data graph constructed in the previous step (such as S122), which contains: rotation morphology information (such as asymmetric swing, periodic bending, etc.) of the rotor head; time-space distribution of vibration parameters (such as vibration amplitude, frequency, phase, etc.); working parameters (such as rotation speed, torque, temperature, etc.); vibration position markers (such as blade root, bearing seat, connecting piece, etc.).

[0067] High-frequency vibration region refers to the region with vibration frequency exceeding a certain threshold (e.g. 50Hz, 100Hz, etc.); these regions are often related to mechanical looseness, fatigue cracks, bearing damage, etc.; identification method: perform spectral analysis on the multi-dimensional state graph, extract the vibration frequency components of each position; set a high-frequency threshold (e.g. 50Hz), filter out regions exceeding the threshold; cluster the high-frequency regions to form multiple independent high-frequency vibration regions; output: the positions, frequency ranges, vibration amplitudes, etc. of multiple high-frequency vibration regions.

[0068] High-frequency vibration often does not exist in isolation, it can be transmitted to other regions through the mechanical structure; the region to be detected is the region affected by high-frequency vibration or the source of high-frequency vibration; tracing method: based on the mechanical structure model, analyze the transmission path of high-frequency vibration; identify the source of high-frequency vibration (such as loose connections, bearing damage, etc.); identify the affected areas (such as blade roots, bearing seats, transmission shafts, etc.); output: the positions of multiple regions to be detected, the correlation strength with high-frequency vibration regions, etc.

[0069] Determine multiple dynamic abnormal regions according to the region morphology of multiple regions to be detected, the region morphology of each high-frequency vibration region, and the rotation morphology of the rotor head; dynamic abnormal region refers to the region that exhibits abnormal behavior during the rotation of the rotor head; the determination of these regions needs to consider: the region morphology of the region to be detected (such as geometric shape, material properties, structural strength, etc.); the region morphology of the high-frequency vibration region (such as vibration distribution, frequency characteristics, phase relationship, etc.); the rotation morphology of the rotor head (such as asymmetric swing, periodic bending, etc.); determination method: perform morphology matching analysis on the region to be detected and the high-frequency vibration region; combined with the rotation morphology, identify the region that exhibits abnormal behavior during rotation; determine the dynamic abnormal region through pattern recognition algorithms (such as clustering, classification, etc.); output: the positions, abnormal characteristics, abnormal degree, etc. of multiple dynamic abnormal regions.

[0070] Specifically, a certain autorotating rotorcraft in cruising state, the rotor head speed is 300rpm (5Hz); through the S122 step, the multi-dimensional state graph of the rotor head has been constructed; read the multi-dimensional state graph, including: rotation morphology: asymmetric swing morphology, blade #2 exists periodic swing and bending anomaly; vibration parameters: blade #2 root vibration amplitude 1.3g, other blades 1.0g; working parameters: speed 300rpm, torque 120Nm, temperature 45℃; vibration position: blade #2 root (position 8).

[0071] Spectrum analysis of the multi-dimensional state diagram: the vibration frequency of the blade #2 root (position 8) is 30 Hz, and the amplitude is 1.3 g; the vibration frequencies of other positions are mainly 5 Hz (rotational frequency) and 10 Hz (2 times the rotational frequency); set the high frequency threshold to 20 Hz: the vibration frequency of the blade #2 root (position 8) is 30 Hz > 20 Hz, which is determined as a high frequency vibration area; high frequency vibration area: area A: blade #2 root (position 8), frequency 30 Hz, amplitude 1.3 g.

[0072] Based on the mechanical structure model, analyze the transmission path of high frequency vibration: high frequency vibration source: loose connection of blade #2 root; vibration transmission path: blade #2 root → bearing seat → transmission shaft; areas to be detected: area B: blade #2 root connector (loose); area C: bearing seat (affected by high frequency vibration); area D: transmission shaft (affected by high frequency vibration).

[0073] Analyze the area morphology of the areas to be detected: area B: the connector has normal geometry, but is loose; area C: the bearing seat structure is complete, but there is high frequency vibration transmission; area D: the transmission shaft has normal geometry, but there is high frequency vibration transmission; analyze the area morphology of the high frequency vibration area: area A: vibration frequency 30 Hz, amplitude 1.3 g, unrelated to rotational frequency (5 Hz); combined with the rotational morphology (asymmetric swing morphology): there is periodic swing and bending abnormality in the rotation of blade #2; high frequency vibration and swing abnormality appear synchronously; dynamic abnormal area determination: dynamic abnormal area 1: blade #2 root connector (area B); abnormal feature: loose connector causes high frequency vibration; abnormality degree: moderate abnormality; dynamic abnormal area 2: bearing seat (area C); abnormal feature: affected by high frequency vibration transmission; abnormality degree: mild abnormality.

[0074] Further, in each dynamic abnormal area, a plurality of sub-dynamic abnormal features are determined according to the identification of the dynamic abnormal area, and the damaged position of the rotor head is collected, a fault range is constructed according to the plurality of sub-dynamic abnormal features, the damaged position of the rotor head and the plurality of vibration parameters, and a plurality of fault features are predicted according to the fault range, the area position of each dynamic abnormal area and the corresponding area morphology, which is compatible with the overall consideration of the identification of the dynamic abnormal area and ensures the accuracy of the plurality of sub-dynamic abnormal features.

[0075] At this time, multiple dynamic abnormal regions have been identified in S131 (such as the root connector of blade #2 and the bearing seat); this step is to perform more detailed feature extraction on each dynamic abnormal region to identify multiple sub-dynamic abnormal features; sub-dynamic abnormal features include: vibration frequency characteristics (such as main frequency, harmonic, sideband, etc.); vibration amplitude characteristics (such as peak value, RMS value, kurtosis, etc.); vibration phase characteristics (such as phase difference, phase drift, etc.); vibration time domain characteristics (such as impact, periodic pulse, etc.); identification method: time-frequency analysis is performed on the vibration signal of each dynamic abnormal region; feature parameters in frequency domain, time domain and time-frequency domain are extracted; compare with the characteristics in normal state to identify abnormal features; output: multiple sub-dynamic abnormal feature lists of each dynamic abnormal region.

[0076] Damaged locations refer to locations in the rotor head structure where obvious damage or abnormalities have occurred, such as cracks, wear, deformation, etc.; these locations can be collected through the following methods: visual inspection: use high-definition cameras or endoscopes to take pictures of each part of the rotor head to identify surface damage; non-destructive testing: use ultrasonic, X-ray, infrared thermal imaging, etc. to detect internal damage; historical data: refer to past maintenance records and fault history to determine vulnerable locations; output: a list of damaged locations of the rotor head, including location coordinates, damage type, damage degree, etc.

[0077] A fault range is constructed based on multiple sub-dynamic abnormal features, damaged locations of the rotor head, and multiple vibration parameters; the fault range refers to the spatial range and impact degree of the fault; this step combines sub-dynamic abnormal features, damaged locations, and vibration parameters to construct the fault range; construction method: spatially correlate sub-dynamic abnormal features with damaged locations; analyze the propagation path and attenuation law of vibration parameters; determine the core area and impact area of the fault; output: fault range diagram, including core fault area, impact area, impact degree, etc.

[0078] Fault features refer to the specific manifestations and attributes of the fault, such as fault type, fault degree, fault development trend, etc.; this step predicts multiple fault features based on the fault range, regional location and regional morphology of the dynamic abnormal region; prediction method: analyze the correlation between vibration features in the fault range and damaged locations; predict fault type based on regional morphology (such as geometric shape, material properties, etc.); predict fault development trend based on the change trend of vibration parameters; output: multiple fault feature lists, including fault type, fault degree, fault development trend, etc.

[0079] Specifically, the rotor head type is a three-blade rotor head of a rotating rotorcraft; the working state is a cruising state (rotation speed 300 RPM); the dynamic abnormal region (from S131) is the root connector of blade #2 (moderate abnormality); the bearing seat (mild abnormality).

[0080] Dynamic anomaly region 1: Blade #2 root connector; Sub-dynamic anomaly feature 1: Vibration frequency 30Hz (non-rotational speed multiple); Sub-dynamic anomaly feature 2: Vibration amplitude 1.3g (exceeds normal value 1.0g); Sub-dynamic anomaly feature 3: Vibration phase synchronizes with Blade #2 rotational position; Sub-dynamic anomaly feature 4: Periodic impact pulses exist in vibration signal.

[0081] Dynamic anomaly region 2: Bearing seat; Sub-dynamic anomaly feature 1: Vibration frequency 30Hz (same frequency as Blade #2); Sub-dynamic anomaly feature 2: Vibration amplitude 0.8g (slightly higher than normal value 0.6g); Sub-dynamic anomaly feature 3: Vibration phase lags Blade #2 by about 10 degrees; Sub-dynamic anomaly feature 4: Vibration signal is relatively smooth without impact pulses.

[0082] Through high-definition camera and endoscope detection, the following damaged locations are found: Damaged location 1: Blade #2 root connector bolt loosening (visible gap); Damaged location 2: Blade #2 root connector surface slight wear; Damaged location 3: Bearing seat internal ball slight wear (found by endoscope).

[0083] Spatial correlation analysis: Sub-dynamic anomaly features 1-4 are highly correlated with damaged locations 1-2; Sub-dynamic anomaly features 1-3 are partially correlated with damaged location 3; Vibration propagation analysis: 30Hz vibration is transmitted from Blade #2 root to bearing seat; Vibration amplitude decays from Blade #2 root (1.3g) to bearing seat (0.8g); Vibration phase lags from Blade #2 root to bearing seat by about 10 degrees; Fault range construction: Core fault region: Blade #2 root connector; Mainly affected area: Bearing seat; Secondary affected area: Blade #2 body; Impact degree: Moderate anomaly in core region, slight anomaly in affected regions.

[0084] Based on fault range and region location analysis: Fault feature 1: Blade #2 root connector bolt loosening; Fault type: Mechanical loosening; Fault degree: Moderate; Development trend: If not handled in time, it will lead to further wear of the connector; Fault feature 2: Blade #2 root connector surface wear; Fault type: Surface wear; Fault degree: Slight; Development trend: Wear speed is slow, but needs regular monitoring; Fault feature 3: Bearing seat internal ball wear; Fault type: Rolling element wear; Fault degree: Slight; Development trend: Influenced by Blade #2 vibration, wear accelerates.

[0085] Based on regional morphology analysis: the blade #2 root connector is a metal-metal contact structure, and loosening will cause high-frequency vibration; the bearing seat is a rolling bearing structure, and vibration transmission will cause uneven wear of the ball bearings; comprehensive prediction: main fault features: blade #2 root connector bolt loosening (moderate); secondary fault features: connector surface wear (mild), bearing seat ball bearing wear (mild); fault development trend: if not treated, the loosening will intensify, leading to increased vibration and accelerated wear of other components.

[0086] Therefore, based on sound detection of the rotor head, a plurality of noise features are determined, a noise distribution map of the rotor head is constructed according to the plurality of noise features and the rotating morphology of the rotor head, and a flight abnormal event associated with the noise features is marked, a plurality of key fault features of the rotor head are determined according to the plurality of fault features, the noise distribution map of the rotor head and the corresponding flight abnormal event, and the overall consideration of the plurality of fault features, the noise distribution map of the rotor head and the corresponding flight abnormal event is compatible, and the accuracy of the plurality of key fault features of the rotor head is ensured.

[0087] At this time, based on sound detection of the rotor head, a plurality of noise features are determined, the noise features include: frequency features (such as fundamental frequency, harmonics, sidebands, narrowband / wideband noise); amplitude features (such as sound pressure level, peak value, RMS value); time domain features (such as impact sound, continuous noise, impulse noise); spatial distribution features (such as sound intensity difference in different directions); collection method: arranging a plurality of acoustic sensors (usually 4-8) around the rotor head; synchronously collecting sound signals, the sampling rate is usually ≥48kHz; filtering, denoising and time-frequency analysis (such as FFT, STFT, wavelet transform) are performed on the sound signals; output: a plurality of noise feature lists, each feature contains frequency, amplitude, time domain characteristics, etc.

[0088] The noise distribution map is a spatial distribution map describing the noise features of the rotor head under different rotating morphologies; the construction method is: associating the noise features with the rotating morphologies (such as asymmetric swing, periodic bending); calculating the sound source positioning (such as beamforming, acoustic holography) according to the sound signals of different sensor positions; drawing a noise distribution heat map on the three-dimensional model of the rotor head; the noise distribution map includes: noise source position (such as blade root, bearing seat, connector); noise intensity (represented by color or contour line); noise frequency distribution (such as fundamental frequency, harmonics); output: noise distribution map of the rotor head (three-dimensional visualization).

[0089] Flight anomaly event refers to abnormal phenomena caused by abnormal noise of the rotor head during flight, such as increased vibration, unstable flight posture, and sluggish control response, etc. The marking method is to analyze the time synchronization of noise characteristics and flight parameters (such as attitude angle, acceleration, and control input); identify the time correlation between noise abnormalities and flight anomaly events; mark the noise characteristics associated with flight anomaly events on the noise distribution map; output: flight anomaly events marked on the noise distribution map (such as "blade #2 noise anomaly causes roll angle fluctuation").

[0090] According to the multiple fault characteristics, the noise distribution map of the rotor head, and the corresponding flight anomaly event, determine the multiple key fault characteristics of the rotor head, which are the core faults of the rotor head finally determined after comprehensive analysis of the fault characteristics (from S132), the noise distribution map, and the flight anomaly event; cross-verify the fault characteristics of S132 and the noise characteristics; confirm the fault source location in combination with the noise distribution map; evaluate the impact of the fault on flight safety according to the flight anomaly event; determine the key fault characteristics after comprehensive analysis; output a list of multiple key fault characteristics, each containing: fault location; fault type (such as loose, wear, and crack); fault degree (mild, moderate, and severe); impact on flight (such as increased vibration and sluggish control); and recommended measures (such as immediate maintenance and regular monitoring).

[0091] Specifically, based on the results of S131 and S132: dynamic abnormal area: ; blade #2 root connector (moderate abnormality); bearing seat (mild abnormality); fault characteristics: blade #2 root connector bolt loose (moderate); blade #2 root connector surface wear (mild); bearing seat internal ball wear (mild).

[0092] Six high-sensitivity microphones are arranged around the rotor head; the sampling rate is 96kHz, and 10 seconds of sound data are collected; the microphone near the blade #2 root connector detects 30Hz high-frequency noise with a sound pressure level of 85dB, and there is a periodic impact pulse; the microphone near the bearing seat detects 30Hz noise with a sound pressure level of 78dB and a phase lag; noise characteristics: feature 1: 30Hz high-frequency impact noise (blade #2 root); feature 2: 30Hz continuous noise (bearing seat).

[0093] The noise of the blade #2 root connector is most obvious during asymmetric swinging; the noise of the bearing seat is enhanced when the blade #2 swings; the noise source is concentrated in the blade #2 root connector and the bearing seat; noise distribution map: blade #2 root connector area: red (high noise); bearing seat area: yellow (medium noise); other areas: green (normal noise).

[0094] 30Hz impact noise occurs, roll angle fluctuation ± 2° (normal ± 0.5°); steering input response delay 0.3 seconds (normal 0.1 seconds); noise distribution icon: "blade #2 root 30Hz impact noise causes roll angle fluctuation"; "bearing seat 30Hz noise causes steering response delay".

[0095] Cross-validation: 30Hz noise is consistent with the 30Hz vibration characteristics of S132; the noise source location is consistent with the fault feature location; comprehensive analysis: loose bolts at the root of blade #2 cause impact noise and vibration, affecting flight attitude; bearing seat ball wear causes continuous noise, affecting steering response; Key fault features: loose bolts at the root of blade #2 (moderate); Location: root of blade #2; Type: mechanical looseness; Degree: moderate; Impact: roll angle fluctuation; Recommendation: immediate maintenance; bearing seat internal ball wear (mild); Location: bearing seat; Type: wear; Degree: mild; Impact: steering response delay; Recommendation: regular monitoring.

[0096] Reference Figure 5 In step S14, the specific steps are:

[0097] S141: Collect the past use events of the rotor head, determine the service life of the rotor head according to the identification of the past use events of the rotor head, and determine a plurality of rotating nodes of the rotor head based on the dynamic detection of the rotor head, and construct a rotating display diagram of the rotor head according to the node position, node state and rotating form of the plurality of rotating nodes of the rotor head;

[0098] S142: Determine the first fault influence coefficient according to the service life of the rotor head and the rotating display diagram of the rotor head, determine the second fault influence coefficient according to the service life of the rotor head and a plurality of key fault features, and determine the fault path of the rotor head in the rotating process based on the first fault influence coefficient, the second fault influence coefficient and the fault path mapping relationship.

[0099] In the embodiment of the present application, the past use events of the rotor head are collected, the service life of the rotor head is determined according to the identification of the past use events of the rotor head, and a plurality of rotating nodes of the rotor head are determined based on the dynamic detection of the rotor head, and a rotating display diagram of the rotor head is constructed according to the node position, node state and rotating form of the plurality of rotating nodes of the rotor head, which is compatible with the overall consideration of the dynamic detection of the rotor head, and ensures the accuracy of the plurality of rotating nodes of the rotor head.

[0100] At this time, through the channels of checking maintenance logs, flight logbooks, digital maintenance systems, etc., all usage events of the rotor head since it was put into use are collected, including: flight hours (such as 1,200 hours accumulated); take-off and landing times (such as 2,400 times accumulated); maintenance records (such as replacing bearings, tightening bolts, lubrication maintenance, etc.); fault history (such as blade loosening, bearing abnormal noise, etc.); environmental conditions (such as high humidity, salt spray environment, sand and dust environment, etc.); these events are classified and sorted out to identify the key factors affecting the service life of the rotor head.

[0101] According to the flight hours and maintenance records, the equivalent service life of the rotor head is calculated; for example: the manufacturer's design life is 10 years or 2,000 flight hours, whichever comes first; if the rotor head has been used for 1,200 hours and has no major failure, the equivalent service life is: 1,200÷2,000×10=6 years.

[0102] Use a high-speed camera system (such as more than 1,000 frames per second) to shoot the rotor head rotating process from multiple angles; use laser displacement sensors, acceleration sensors, etc. to measure the displacement and vibration of key parts; through image recognition and signal processing technology, the motion characteristics of the rotor head in the rotating process are extracted; the rotating node refers to the key position of the rotor head in the rotating process with specific motion characteristics, such as: blade root connection point; bearing seat; hub center; control pull rod connection point; for each node, record: node position (three-dimensional coordinates); node state (displacement, velocity, acceleration); relative motion relationship between nodes.

[0103] Integrate the position and state information of all rotating nodes into a three-dimensional coordinate system; use graphical technology (such as three-dimensional modeling software) to draw a rotating display diagram of the rotor head; mark in the diagram: the position of each node; the motion trajectory of each node; the state parameters of each node (such as vibration amplitude, displacement change); the connection relationship between nodes; mark abnormal nodes (such as vibration out of limit, displacement anomaly) by color coding or icon; rotating form integration: integrate the rotating form (such as asymmetric swing) determined in S121 into the rotating display diagram; show the overall motion state of the rotor head in the rotating process through animation or dynamic chart.

[0104] Specifically, an autogyro appears abnormal vibration in cruise flight; collect data: flight hours: 1,200 hours; take-off and landing times: 2,400 times; maintenance records: replace bearings in the second year; tighten blade connection bolts in the fourth year; fault history: blade #2 loosening in the third year; environmental conditions: coastal area, high humidity, salt spray environment; determine the service life: design life: 10 years or 2,000 hours; equivalent service life: 1,200÷2,000×10=6 years; consider environmental factors (salt spray corrosion), correct the service life: 6×1.2=7.2 years.

[0105] Three high-speed cameras are used to shoot the rotor head from different angles; acceleration sensors are installed at the blade root, bearing seat and hub center; a rotating node schematic table is collected, as shown in Table 1:

[0106] Table 1 Rotating node schematic table

[0107] Node name Position coordinates (x, y, z) State parameters Blade #1 root (0.5,0,0) Vibration 0.8 g Blade #2 root (-0.5,0,0) Vibration 1.5 g Bearing seat (0,0,0.2) Vibration 1.2 g Blade hub center (0,0,0) Vibration 0.5 g

[0108] Three-dimensional modeling software is used to draw the rotor head structure; the position and state parameters of each node are marked on the diagram; the vibration amplitude is marked with color: green: normal (<1.0g); yellow: warning (1.0-1.5g); red: danger (>1.5g); the root of blade #2 is displayed in yellow, and the bearing seat is displayed in yellow; the asymmetric swing form determined in S121 is displayed in animation; the swing amplitude of blade #2 in the rotating process is significantly larger than that of other blades.

[0109] From the rotating display diagram, it can be seen intuitively that: the vibration of the root of blade #2 is high (1.5g), and there is a risk of loosening; the vibration of the bearing seat is high (1.2g), and there is a risk of bearing wear; the whole presents an asymmetric swing form, which is consistent with the analysis result of S121; this rotating display diagram provides an intuitive basis for subsequent fault influence coefficient calculation (S142).

[0110] Further, the first fault influence coefficient is determined according to the service life of the rotor head and the rotating display diagram of the rotor head, the second fault influence coefficient is determined according to the service life of the rotor head and a plurality of key fault characteristics, and the fault path of the rotor head in the rotating process is determined based on the first fault influence coefficient, the second fault influence coefficient and the fault path mapping relationship, which is compatible with the overall consideration of the first fault influence coefficient, the second fault influence coefficient and the fault path mapping relationship, and ensures the accuracy of the fault path of the rotor head in the rotating process.

[0111] At this time, the first failure influence coefficient is determined according to the service life of the rotor head and the rotation display diagram of the rotor head, the first failure influence coefficient (FIC1) is used to evaluate the failure risk degree of the rotor head due to the service life and the current structure state (reflected by the rotation display diagram); the calculation method is: the aging factor is calculated according to the service life (such as 6 years, equivalent to 1,200 flight hours): aging factor = service life / design life = 6 / 10 = 0.6; the structure factor is calculated according to the number and severity of abnormal nodes in the rotation display diagram: each abnormal node is given a weight according to the vibration amplitude: normal (<1.0g): weight 0; warning (1.0-1.5g): weight 1; dangerous (>1.5g): weight 2; structure factor = sum of weights of all nodes / total number of nodes; for example: in 4 nodes, 2 are warning (weight 1) and 2 are normal (weight 0); structure factor = (1+1+0+0) / 4 = 0.5; FIC1 = aging factor x structure factor = 0.6 x 0.5 = 0.3.

[0112] The second failure influence coefficient is determined according to the service life of the rotor head and a plurality of key failure characteristics, the second failure influence coefficient (FIC2) is used to evaluate the failure risk degree of the rotor head due to the key failure characteristics (such as cracks, wear, looseness, etc.); the key failure characteristic identification result from S13, such as: blade root connector looseness; bearing seat internal ball wear; blade surface micro-cracks; the calculation method is: the aging factor is calculated according to the service life (same as above): 0.6; the characteristic factor is calculated according to the severity of the key failure characteristics: each failure characteristic is given a weight according to the severity: mild: weight 1; moderate: weight 2; severe: weight 3; characteristic factor = sum of weights of all characteristics / total number of characteristics; for example: in 3 characteristics, 1 is moderate (weight 2) and 2 are mild (weight 1); characteristic factor = (2+1+1) / 3 = 1.33; FIC2 = aging factor x characteristic factor = 0.6 x 1.33 = 0.8.

[0113] The failure path of the rotor head in the rotation process is determined based on the first failure influence coefficient, the second failure influence coefficient and the failure path mapping relationship, the failure path mapping relationship is a pre-established rule library, which is used to determine the type and development trend of the failure path according to the combined value of FIC1 and FIC2.

[0114] A schematic table of the failure path mapping relationship is collected, and the schematic table of the failure path mapping relationship is shown in Table 2:

[0115] Table 2 Schematic table of failure path mapping relationship

[0116] FIC1 FIC2 Fault path type Development trend <0.3 <0.5 A Stable 0.3-0.6 0.5-1.0 B Slow development >0.6 >1.0 C Rapid development

[0117] According to the calculation results: FIC1=0.3; FIC2=0.8; look up table: fault path type: B; development trend: slow development.

[0118] Specifically, the rotor head has been used for 6 years (equivalent to 1,200 flight hours); the rotating display diagram shows that the blade #2 root vibration is 1.5g (warning); the bearing seat vibration is 1.2g (warning); other nodes are normal; the key fault features are: blade root connector loosening (moderate); bearing seat internal ball wear (mild); blade surface micro-cracks (mild).

[0119] FIC1 calculation: aging factor = 6 / 10 = 0.6; structure factor = (1+1+0+0) / 4 = 0.5; FIC1 = 0.6x0.5 = 0.3; FIC2 calculation: aging factor = 0.6; feature factor = (2+1+1) / 3 = 1.33; FIC2 = 0.6x1.33 = 0.8; fault path determination: FIC1 = 0.3, FIC2 = 0.8; look up table fault path type: B (slow development); result analysis: the rotor head is currently in fault path B, indicating that the fault is developing slowly; main risk points: blade #2 root connector loosening (moderate); bearing seat internal ball wear (mild); recommended measures: focus on checking the blade connector and bearing during the next maintenance; strengthen the monitoring of vibration parameters, and if the deterioration trend is found, maintenance should be carried out in advance.

[0120] Reference Figure 6 In step S15, the specific steps are:

[0121] S151: Collecting fault paths, determining a plurality of sub-fault areas according to the detection of the fault paths, determining corresponding rotating fault combinations based on the identification of the plurality of sub-fault areas, and determining corresponding fault nodes according to the detection of each rotating fault combination, so as to collect a plurality of fault nodes;

[0122] S152: Collecting the flight path of the autorotating rotorcraft and the current position of the autorotating rotorcraft, determining the remaining flight path of the autorotating rotorcraft according to the matching of the flight path of the autorotating rotorcraft and the current position of the autorotating rotorcraft, and simultaneously collecting a plurality of damaged positions of the rotor head, determining the current damaged state of the rotor head according to the plurality of damaged positions of the rotor head, the corresponding damaged parts and the fault path of the rotor head;

[0123] S153: Determining a first emergency coefficient according to the plurality of fault nodes and the remaining flight path of the autorotating rotorcraft, determining a second emergency coefficient according to the plurality of fault nodes and the current damaged state of the rotor head, and determining an emergency rotating event of the rotor head based on the first emergency coefficient, the second emergency coefficient and an emergency rotating event mapping relationship.

[0124] In the embodiments of the present application, the fault path is collected, a plurality of sub-fault areas are determined according to the detection of the fault path, corresponding rotating fault combinations are determined based on the identification of the plurality of sub-fault areas, and corresponding fault nodes are determined according to the detection of each rotating fault combination to collect a plurality of fault nodes, which is compatible with the overall consideration of the detection of each rotating fault combination and ensures the accuracy of the corresponding fault nodes.

[0125] At this time, first, the fault path determined in the previous step (such as S14) is obtained; for example, the fault path is: path A: rapid development, high risk; path B: slow development, medium risk; path C: stable state, low risk; according to the characteristics of the fault path, a plurality of sub-areas affected on the rotor head are identified; these sub-areas are usually key structural parts of the rotor head, for example: blade root connection area; bearing seat area; blade surface area; hub area; each sub-area corresponds to different fault types and risk levels.

[0126] Each sub-fault area corresponds to one or more rotating fault combinations; a rotating fault combination refers to a combination of a plurality of fault phenomena occurring simultaneously in the area; for example: blade root connection area: fault combination: loose connection + abnormal vibration; bearing seat area: fault combination: bearing wear + abnormal noise; blade surface area: fault combination: surface crack + aerodynamic imbalance; these combinations reflect the relevance between different fault phenomena during the rotation of the rotor head.

[0127] According to the detection of each rotating fault combination, corresponding fault nodes are determined to collect a plurality of fault nodes, and for each rotating fault combination, the specific fault nodes are further refined; a fault node refers to a specific component or position on the rotor head, for example: blade #2 root connector; bearing seat internal ball; blade #2 surface middle section; these nodes are the specific targets of subsequent fault analysis and maintenance operations.

[0128] Specifically, the autorotation rotorcraft is in cruise flight, the flight height is 1000 meters, and the flight speed is 120 kilometers per hour; through the previous step (S14), it is determined that the fault path of the rotor head is: path B: slow development, medium risk; fault path: path B (slow development, medium risk); sub-failure area identification: blade root connection area; bearing seat area; blade surface area; determine the rotating failure combination based on the sub-failure area: blade root connection area: rotating failure combination: connection loose + vibration anomaly; bearing seat area: rotating failure combination: bearing wear + noise anomaly; blade surface area: rotating failure combination: surface crack + aerodynamic imbalance; blade root connection area: fault node: blade #2 root connector; bearing seat area: fault node: bearing seat internal ball; blade surface area: fault node: blade #2 surface middle section; result: successfully collected multiple fault nodes: blade #2 root connector; bearing seat internal ball; blade #2 surface middle section.

[0129] Further, the flight path of the autorotation rotorcraft and the current position of the autorotation rotorcraft are collected, and the remaining flight path of the autorotation rotorcraft is determined according to the matching of the flight path of the autorotation rotorcraft and the current position of the autorotation rotorcraft, and meanwhile, multiple damaged positions of the rotor head are collected, and the current damaged state of the rotor head is determined according to the multiple damaged positions of the rotor head, the corresponding damaged parts and the fault path of the rotor head, which is compatible with the overall consideration of the multiple damaged positions of the rotor head, the corresponding damaged parts and the fault path of the rotor head, and guarantees the accuracy of the current damaged state of the rotor head.

[0130] At this time, the flight plan of the autorotation rotorcraft is obtained, including: starting point, ending point, en route waypoint; expected flight time, height, speed and other parameters; for example: the flight path is A→B→C→D, and the total flight time is 2 hours; the current position of the rotorcraft is obtained in real time through the GPS and inertial navigation system; for example: currently located at point B, and has flown for 40 minutes; match the current position with the preset flight path; calculate the remaining flight segment: B→C→D; the expected remaining flight time is 1 hour and 20 minutes; emergency landing point analysis: whether there is an emergency landing point on the path; the location and distance of the nearest airport.

[0131] Obtaining the identified fault nodes from S151; for example: blade #2 root connector; bearing seat inner ball; blade #2 surface middle section; detailed analysis for each damaged location: blade #2 root connector: damage type: connector loosening; damage degree: moderate (vibration value 1.5g); bearing seat inner ball: damage type: bearing wear; damage degree: slight (noise value 85dB); blade #2 surface middle section: damage type: surface crack; damage degree: slight (crack length 2cm); associating the damage with the fault path (such as path B: slow development); analyzing the development trend of each damaged part: connector loosening intensifies vibration; bearing wear gradually worsens; surface crack expands.

[0132] Specifically, the autorotation rotorcraft is performing a flight task of A→B→C→D; total flight time: 2 hours; current state: has arrived at point B, and has flown for 40 minutes; rotor head fault path: path B (slow development, moderate risk).

[0133] Pre-set flight path: A (starting point)→B (waypoint)→C (waypoint)→D (end point); current position: located at point B, and has flown for 40 minutes; remaining flight path determination: remaining flight section: B→C→D; remaining flight time: 1 hour and 20 minutes; nearby emergency landing airport: emergency landing point E is 20 kilometers away from point B; airport F is 15 kilometers away from point C.

[0134] Damaged location collection (from S151): blade #2 root connector; bearing seat inner ball; blade #2 surface middle section; detailed analysis of damaged parts: blade #2 root connector: damage type: connector loosening; damage degree: moderate; specific performance: vibration value 1.5g (normal value should be <0.8g); development trend: if continue to fly, vibration intensifies to 2.0g; bearing seat inner ball: damage type: bearing wear; damage degree: slight; specific performance: noise value 85dB (normal value should be <75dB); development trend: wear gradually increases, noise increases to 90dB; blade #2 surface middle section: damage type: surface crack; damage degree: slight; specific performance: crack length 2cm; development trend: crack expands to 3-4cm.

[0135] Current impairment status comprehensive assessment: Overall condition: Moderate risk; Main risk points: Blade #2 root joint looseness leading to increased vibration; Bearing wear affecting rotor head overall stability; Surface crack propagation affecting aerodynamic performance; Development prediction: Following current failure path (slow development), within remaining 1 hour 20 minutes flight time: Vibration increases by 25%; Noise increases by 5dB; Crack propagation by 50%; Remaining flight path risk assessment: B→C segment (40 minutes): Risk controllable; C→D segment (40 minutes): Higher risk; Current impairment status summary: Rotor head is in a moderate risk state; Three main impaired parts interact with each other, forming a vicious cycle; It is suggested to perform a landing check at point C, or consider a precautionary landing at the nearest airport F.

[0136] Therefore, the first emergency coefficient is determined according to the plurality of fault nodes and the remaining flight path of the autorotating rotorcraft, the second emergency coefficient is determined according to the plurality of fault nodes and the current impairment status of the rotor head, and the emergency rotation event of the rotor head is determined based on the first emergency coefficient, the second emergency coefficient and the emergency rotation event mapping relationship, which comprehensively considers the first emergency coefficient, the second emergency coefficient and the emergency rotation event mapping relationship, and ensures the accuracy of the emergency rotation event of the rotor head.

[0137] At this time, the first emergency coefficient is determined according to the plurality of fault nodes and the remaining flight path of the autorotating rotorcraft, the fault node analysis: the identified fault nodes are obtained from S151, for example: blade #2 root joint looseness (moderate); bearing seat internal ball wear (mild); surface micro crack (mild); remaining flight path analysis: obtain the remaining flight path information determined in S152: remaining flight segment: BCD; remaining flight time: 1 hour 20 minutes; road segment characteristics: ; BC segment: level flight, relatively stable; CD segment: contains climbing and turning, heavy load; first emergency coefficient calculation: establish an evaluation model: fault severity x road segment risk weight x time factor.

[0138] Specific calculation: BC segment (40 minutes): blade looseness: 0.6x0.8x0.5=0.24; bearing wear: 0.4x0.8x0.5=0.16; surface crack: 0.3x0.8x0.5=0.12; BC segment comprehensive coefficient: 0.52; CD segment (40 minutes): blade looseness: 0.6x1.2x0.5=0.36; bearing wear: 0.4x1.2x0.5=0.24; surface crack: 0.3x1.2x0.5=0.18; CD segment comprehensive coefficient: 0.78; first emergency coefficient=max(0.52,0.78)=0.78.

[0139] A second emergency coefficient is determined according to the current damaged state of the multiple fault nodes and the rotor head, and the current damaged state in S152 is obtained: blade #2 root: vibration 1.5g, loose risk; bearing seat: vibration 1.2g, noise 85dB; surface: crack length 2cm, expansion; damaged state quantitative evaluation: a state evaluation matrix is established: vibration level: 1.5g (reference value 1.0g) -> risk coefficient 1.5; noise level: 85dB (reference value 80dB) -> risk coefficient 1.25; crack length: 2cm (allowable value 3cm) -> risk coefficient 0.67; second emergency coefficient calculation: comprehensive evaluation model: risk coefficient of each part x weight x development trend factor; specific calculation: blade root: 1.5x0.5x1.2=0.9; bearing seat: 1.25x0.3x1.1=0.4125; surface crack: 0.67x0.2x1.3=0.1742; second emergency coefficient=0.9+0.4125+0.1742=1.4867.

[0140] An emergency rotation event of the rotor head is determined based on the first emergency coefficient, the second emergency coefficient, and an emergency rotation event mapping relationship, emergency coefficient comprehensive analysis: first emergency coefficient: 0.78 (path related); second emergency coefficient: 1.4867 (state related); comprehensive emergency coefficient=(0.78+1.4867) / 2=1.1334; emergency rotation event mapping relationship: preset emergency event determination standard: 0.0-0.5: normal flight, continue to monitor; 0.5-1.0: reduce flight height, slow down flight; 1.0-1.5: find the nearest airport, prepare for emergency landing; 1.5-2.0: immediately emergency landing; 2.0: highest level of emergency, start all emergency procedures; emergency rotation event determination: comprehensive emergency coefficient 1.1334 falls in the interval of 1.0-1.5; corresponding emergency rotation event: find the nearest airport, prepare for emergency landing.

[0141] Specifically, the autorotating rotorcraft is performing a flight task from A to D; current position: B point, having flown for 40 minutes; remaining path: BCD, estimated to be 1 hour and 20 minutes; identified fault nodes: blade #2 root connection loose; bearing seat internal ball wear; surface small crack.

[0142] First emergency coefficient calculation: BC segment evaluation (40 minutes of level flight): blade loose risk: moderate (0.6); road risk: low (0.8); time factor: 0.5; calculation result: 0.24; CD segment evaluation (40 minutes including climb and turn): blade loose risk: moderate (0.6); road risk: high (1.2); time factor: 0.5; calculation result: 0.36; first emergency coefficient=0.78 (take the maximum value).

[0143] Second emergency coefficient calculation: current damaged state: blade root vibration: 1.5g (over limit 50%); bearing noise: 85dB (over limit 5dB); surface crack: 2cm (within limit); risk assessment of each part: blade root: risk coefficient 0.9; bearing seat: risk coefficient 0.4125; surface crack: risk coefficient 0.1742; second emergency coefficient = 1.4867.

[0144] Integrated emergency coefficient = (0.78 + 1.4867) / 2 = 1.1334; query mapping table: 1.1334 is located in the interval of 1.0-1.5; corresponding event: find the nearest airport and prepare for emergency landing"; final decision: emergency rotation event: find the nearest airport and prepare for emergency landing"; specific implementation scheme: immediately contact air traffic control and report the emergency; query the nearest available airport (find F airport, distance 15 minutes flight); reduce flight height to safe height; reduce speed to economic cruising speed; closely monitor the state changes of each fault node; prepare emergency landing procedure.

[0145] Please refer to Figure 7 , Figure 7 is a structural composition schematic diagram of a fault detection system of a rotor head in dynamic work in an embodiment of the application; the fault detection system of the rotor head in dynamic work comprises:

[0146] The working state module 21 is configured to collect a plurality of working parameters of the rotor head when the autorotating rotorcraft is in a flight state, and determine a working state of the rotor head according to the plurality of working parameters of the rotor head and a flight mode of the autorotating rotorcraft.

[0147] The multi-dimensional state diagram module 22 is configured to determine a multi-dimensional state diagram of the rotor head based on a rotation form of the rotor head, the plurality of working parameters and a vibration parameter of the rotor head in the rotation process.

[0148] The key fault feature module 23 is configured to determine a plurality of dynamic abnormal regions according to recognition of the multi-dimensional state diagram of the rotor head, predict a plurality of fault features according to region positions and corresponding region forms of the dynamic abnormal regions, and determine a plurality of key fault features of the rotor head based on the plurality of fault features, a noise distribution diagram of the rotor head and a corresponding flight abnormal event.

[0149] The fault path module 24 is configured to collect a service life of the rotor head, and determine a fault path of the rotor head in the rotation process according to the service life of the rotor head, a rotation display diagram of the rotor head and the plurality of key fault features.

[0150] The emergency rotation event module 25 is configured to determine a plurality of fault nodes according to detection of the fault path, and determine an emergency rotation event of the rotor head according to the plurality of fault nodes, a remaining flight path of the autorotating rotorcraft and a current damaged state of the rotor head.

[0151] Any combination of the technical features of the above embodiments is possible, and for the sake of brevity, not all possible combinations are described. However, any combination of the technical features of the above embodiments is deemed to be within the scope of the present disclosure.

Claims

1. A method of fault detection of a rotor head under dynamic operation, characterized in that, The method comprises the following steps: collecting a plurality of working parameters of the rotor head when the autorotating rotorcraft is in a flight state, and determining a working state of the rotor head according to the plurality of working parameters of the rotor head and a flight mode of the autorotating rotorcraft; determining a multi-dimensional state diagram of the rotor head based on a rotating mode of the rotor head, the plurality of working parameters, and vibration parameters of the rotor head in the rotating process; determining a plurality of dynamic abnormal regions according to the identification of the multi-dimensional state diagram of the rotor head, predicting a plurality of fault features according to the region position and corresponding region mode of each dynamic abnormal region, and determining a plurality of key fault features of the rotor head based on the plurality of fault features, a noise distribution diagram of the rotor head, and corresponding flight abnormal events; collecting the service life of the rotor head, and determining a fault path of the rotor head in the rotating process according to the service life of the rotor head, the rotating display diagram of the rotor head, and the plurality of key fault features; determining a plurality of fault nodes according to the detection of the fault path, and determining an emergency rotating event of the rotor head according to the plurality of fault nodes, the remaining flight path of the autorotating rotorcraft, and the current damaged state of the rotor head.

2. The method of fault detection of a rotor head under dynamic operation according to claim 1, characterized in that The method comprises the following steps: monitoring the flight state of the autorotating rotorcraft in real time, the rotor head being in a continuous rotating state when the autorotating rotorcraft is in the flight state, collecting a plurality of working parameters of the rotor head in the continuous rotating state, and determining a working mode diagram of the rotor head according to the plurality of working parameters of the rotor head and the continuous rotating state of the rotor head; determining a plurality of working mode combinations according to the identification of the working mode diagram of the rotor head, determining corresponding working features based on the identification of each working combination mode, collecting a plurality of working features of the rotor head, and determining the working state of the rotor head according to the plurality of working features of the rotor head, the current flight parameter combination of the autorotating rotorcraft, and the flight mode of the autorotating rotorcraft.

3. The method of fault detection of a rotor head under dynamic operation according to claim 1, wherein, The method comprises the following steps: rotating the rotor head along the working state, collecting rotating images of the rotor head at different time periods, determining a plurality of sub-rotating features of the rotor head according to the identification of the rotating images of the rotor head at different time periods, and determining the rotating mode of the rotor head based on the plurality of sub-rotating features and the working state of the rotor head; collecting a plurality of working parameters of the rotor head, determining a first state diagram according to the rotating mode of the rotor head and the plurality of working parameters of the rotor head, collecting vibration parameters of the rotor head in the rotating process, marking the vibration positions corresponding to the vibration parameters, determining a second state diagram according to the vibration parameters of the rotor head in the rotating process, the corresponding vibration positions, and the rotating mode of the rotor head, and constructing a multi-dimensional state diagram of the rotor head according to the first state diagram and the second state diagram.

4. The method of fault detection of a rotor head under dynamic operation according to claim 1, wherein, The method comprises the following steps: collecting a multi-dimensional state diagram of the rotor head; determining a plurality of dynamic abnormal regions according to recognition of the multi-dimensional state diagram of the rotor head; predicting a plurality of fault features according to region positions and corresponding region morphologies of the dynamic abnormal regions; and determining a plurality of key fault features of the rotor head based on the plurality of fault features, a noise distribution diagram of the rotor head, and corresponding flight abnormal events. The method comprises the following steps: collecting a multi-dimensional state diagram of the rotor head; determining a plurality of high-frequency vibration regions based on recognition of the multi-dimensional state diagram of the rotor head; determining a plurality of to-be-detected regions associated with the high-frequency vibration regions; and determining a plurality of dynamic abnormal regions according to region morphologies of the to-be-detected regions, region morphologies of the high-frequency vibration regions, and a rotation morphology of the rotor head.

5. The method of fault detection of a rotor head under dynamic operation according to claim 4, characterized in that The method comprises the following steps: collecting a multi-dimensional state diagram of the rotor head; determining a plurality of dynamic abnormal regions according to recognition of the multi-dimensional state diagram of the rotor head; predicting a plurality of fault features according to region positions and corresponding region morphologies of the dynamic abnormal regions; and determining a plurality of key fault features of the rotor head based on the plurality of fault features, a noise distribution diagram of the rotor head, and corresponding flight abnormal events. In each dynamic abnormal region, a plurality of sub-dynamic abnormal features are determined according to recognition of the dynamic abnormal region, meanwhile, a damaged position of the rotor head is collected, a fault range is constructed according to the plurality of sub-dynamic abnormal features, the damaged position of the rotor head, and a plurality of vibration parameters, and a plurality of fault features are predicted according to the fault range, region positions of the dynamic abnormal regions, and corresponding region morphologies; A plurality of noise features are determined based on sound detection of the rotor head, a noise distribution diagram of the rotor head is constructed according to the plurality of noise features and a rotation morphology of the rotor head, and a flight abnormal event associated with the noise features is marked, and a plurality of key fault features of the rotor head are determined according to the plurality of fault features, the noise distribution diagram of the rotor head, and corresponding flight abnormal events.

6. The method of fault detection of a rotor head under dynamic operation according to claim 1, wherein, The method comprises the following steps: collecting a used life of the rotor head; determining a fault path of the rotor head in a rotation process according to the used life of the rotor head, a rotation display diagram of the rotor head, and the plurality of key fault features. The method comprises the following steps: collecting a past use event of the rotor head; determining a used life of the rotor head according to recognition of the past use event of the rotor head, meanwhile, a plurality of rotation nodes of the rotor head are determined based on dynamic detection of the rotor head, and a rotation display diagram of the rotor head is constructed according to node positions, node state after, and a rotation morphology of the plurality of rotation nodes of the rotor head.

7. The method of fault detection of a rotor head under dynamic operation according to claim 6, characterized in that The method comprises the following steps: collecting a used life of the rotor head; determining a fault path of the rotor head in a rotation process according to the used life of the rotor head, a rotation display diagram of the rotor head, and the plurality of key fault features. The method comprises the following steps: determining a first fault influence coefficient according to the used life of the rotor head and the rotation display diagram of the rotor head; determining a second fault influence coefficient according to the used life of the rotor head and the plurality of key fault features; and determining the fault path of the rotor head in the rotation process based on the first fault influence coefficient, the second fault influence coefficient, and a fault path mapping relationship.

8. The method of fault detection of a rotor head under dynamic operation according to claim 1, wherein, The method comprises the following steps: determining a plurality of fault nodes according to detection of the fault path; and determining an emergency rotation event of the rotor head according to the plurality of fault nodes, a remaining flight path of the rotor, and a current damaged state of the rotor head. The fault path is collected, a plurality of sub-fault areas are determined according to detection of the fault path, a corresponding rotating fault combination is determined based on identification of the plurality of sub-fault areas, a corresponding fault node is determined according to detection of each rotating fault combination, and a plurality of fault nodes are collected.

9. The method of fault detection of a rotor head under dynamic operation according to claim 8, characterized in that The plurality of fault nodes are determined according to detection of the fault path, the emergency rotating event of the rotor head is determined according to the plurality of fault nodes, the remaining flight path of the autorotating rotorcraft, and the current damaged state of the rotor head, and further includes: The flight path of the autorotating rotorcraft and the current position of the autorotating rotorcraft are collected, the remaining flight path of the autorotating rotorcraft is determined according to matching of the flight path of the autorotating rotorcraft and the current position of the autorotating rotorcraft, and meanwhile, a plurality of damaged positions of the rotor head are collected, the current damaged state of the rotor head is determined according to the plurality of damaged positions of the rotor head, corresponding damaged parts, and the fault path of the rotor head; A first emergency coefficient is determined according to the plurality of fault nodes and the remaining flight path of the autorotating rotorcraft, a second emergency coefficient is determined according to the plurality of fault nodes and the current damaged state of the rotor head, and the emergency rotating event of the rotor head is determined based on the first emergency coefficient, the second emergency coefficient, and an emergency rotating event mapping relationship.

10. A fault detection system for a rotor head under dynamic operation, characterized by, The fault detection system of the rotor head under dynamic working is applied to the fault detection method of the rotor head under dynamic working as claimed in any one of claims 1-9, and the fault detection system of the rotor head under dynamic working includes: The working state module is used to collect a plurality of working parameters of the rotor head when the autorotating rotorcraft is in a flight state, and determine the working state of the rotor head according to the plurality of working parameters of the rotor head and the flight mode of the autorotating rotorcraft; The multi-dimensional state diagram module is used to determine the multi-dimensional state diagram of the rotor head based on the rotating form of the rotor head, the plurality of working parameters, and the vibration parameters of the rotor head in the rotating process; The key fault feature module is used to determine a plurality of dynamic abnormal areas according to identification of the multi-dimensional state diagram of the rotor head, predict a plurality of fault features according to the area position and corresponding area form of each dynamic abnormal area, and determine a plurality of key fault features of the rotor head based on the plurality of fault features, the noise distribution diagram of the rotor head, and corresponding flight abnormal events; The fault path module is used to collect the service life of the rotor head, and determine the fault path of the rotor head in the rotating process according to the service life of the rotor head, the rotating display diagram of the rotor head, and the plurality of key fault features; The emergency rotating event module is used to determine a plurality of fault nodes according to detection of the fault path, and determine the emergency rotating event of the rotor head according to the plurality of fault nodes, the remaining flight path of the autorotating rotorcraft, and the current damaged state of the rotor head.

Images