Methods, equipment and media for bolt shaft fatigue testing under multimodal sensing

Abstract

This invention discloses a method, equipment, and medium for multimodal sensing-based fatigue testing of bolt shafts, relating to the field of mechanical structure fatigue testing. The method includes: deploying a multimodal sensing layer on the bolt shaft, performing multimodal sensing monitoring, and establishing a multimodal sensing monitoring dataset; extracting periodic ultrasonic sensing signals, performing signal feature extraction, sending the signal feature extraction results and acquisition cycle to the fatigue identification main channel, and establishing a fatigue damage evolution baseline; extracting strain sensing time-series signals and acceleration sensing time-series signals; performing additional damage analysis and establishing a time-series additional damage factor; performing time-series superposition fitting of the fatigue damage evolution baseline using the time-series additional damage factor, and updating the fatigue damage state of the bolt shaft using the time-series superposition fitting results. This method solves the technical problem of low damage identification sensitivity in existing bolt shaft fatigue monitoring, achieving the technical effect of improving the sensitivity of bolt shaft fatigue damage identification.

Description

Technical Field

[0001] This application relates to the field of mechanical structure fatigue testing, and in particular to bolt shaft fatigue testing methods, equipment and media under multimodal sensing. Background Technology

[0002] As a critical mechanical connection component, the accurate monitoring of fatigue damage in bolted shafts is crucial for ensuring the safe operation of major equipment. Failure to identify early fatigue damage in a timely manner can easily lead to catastrophic structural failures. Currently, the mainstream method for fatigue monitoring of bolted shafts involves acquiring signals using a single modal sensor (such as an ultrasonic probe or strain gauge) and identifying damage based on time-domain / frequency-domain feature analysis. However, because single modal sensors can only acquire localized damage information and are susceptible to interference from environmental noise, load fluctuations, and other factors, existing methods suffer from incomplete damage feature extraction, inaccurate characterization of evolution patterns, and low sensitivity in early damage identification. These limitations make it difficult to meet the dynamic assessment requirements of bolted shaft fatigue states under complex operating conditions.

[0003] Currently, bolt shaft fatigue monitoring suffers from low sensitivity in damage identification. Summary of the Invention

[0004] This application provides a method, equipment, and medium for bolt shaft fatigue testing under multimodal sensing. It employs a multimodal sensing layer constructed by arranging a coupled probe, strain-sensing washer, and acceleration sensor on the bolt shaft. Monitoring is conducted, and a multimodal sensing monitoring dataset is established. Periodic ultrasonic sensing signals are extracted from the dataset, signal features are extracted, and the results, along with the acquisition cycle, are transmitted to the fatigue identification main channel to establish a fatigue damage evolution baseline. Strain-sensing and acceleration-sensing time-series signals are extracted from the dataset, and additional damage analysis is performed using these two time-series signals to construct a time-series additional damage factor. This time-series additional damage factor is then used to time-series superimpose and fit the fatigue damage evolution baseline to update the bolt shaft fatigue damage state. These techniques solve the technical problem of low damage identification sensitivity in existing bolt shaft fatigue monitoring, achieving the technical effect of improving the sensitivity of bolt shaft fatigue damage identification.

[0005] This application provides a method for fatigue testing of bolt shafts under multimodal sensing, comprising: deploying a multimodal sensing layer on the bolt shaft, performing multimodal sensing monitoring, and establishing a multimodal sensing monitoring dataset, wherein the multimodal sensing layer includes a coupling probe, a strain sensing washer, and an acceleration sensor; extracting periodic ultrasonic sensing signals from the multimodal sensing monitoring dataset, performing signal feature extraction on the periodic ultrasonic sensing signals, sending the signal feature extraction results and acquisition period to the fatigue identification main channel, and establishing a fatigue damage evolution baseline; extracting strain sensing time-series signals and acceleration sensing time-series signals from the multimodal sensing monitoring dataset; performing additional damage analysis using the strain sensing time-series signals and acceleration sensing time-series signals to establish a time-series additional damage factor; performing time-series superposition fitting of the fatigue damage evolution baseline using the time-series additional damage factor, and updating the fatigue damage state of the bolt shaft using the time-series superposition fitting result.

[0006] In a possible implementation, the signal feature extraction of the periodic ultrasonic sensing signal is performed, and the signal feature extraction results and acquisition cycle are sent to the fatigue identification main channel to establish a fatigue damage evolution baseline. The following processing is performed: signal feature extraction includes echo delay increment extraction, echo amplitude attenuation rate extraction, and echo waveform offset extraction; the independent identification channel and cumulative damage channel within the fatigue identification main channel are initialized using the acquisition cycle; after receiving the signal feature extraction results using the independent identification channel, transient fatigue state analysis is performed on each period node to establish a periodic fatigue state with period node identifiers; after receiving the signal feature extraction results and the periodic fatigue state using the cumulative damage channel, the periodic signal change rate is calculated, and the fatigue damage evolution baseline is generated using the periodic signal change rate calculation results and the periodic fatigue state.

[0007] In a possible implementation, after receiving the signal feature extraction result and the periodic fatigue state using the cumulative damage channel, the periodic signal change rate is calculated, and the following processing is performed: establishing the calibration fatigue cycle of the bolt shaft; using the periodic fatigue state to perform node segmentation of the calibration fatigue cycle, establishing node segmentation results, and configuring dynamic weighting coefficients with the node segmentation results; during the periodic signal change rate calculation, multi-feature damage weighted calculation is performed through the dynamic weighting coefficients, and the periodic fatigue state is corrected according to the multi-feature damage weighted calculation results to generate a fatigue damage evolution baseline.

[0008] In a possible implementation, the step of performing a time-series superposition fitting of the fatigue damage evolution baseline using the time-series additional damage factor, and updating the fatigue damage state of the bolt shaft using the time-series superposition fitting result, involves the following processing: configuring a backtracking time window and setting a time-series decreasing weight based on the window length of the backtracking time window; performing impact transition identification based on the time-series additional damage factor and configuring the impact transition identification result; and performing a time-series superposition fitting of the fatigue damage evolution baseline using the time-series additional damage factor and the time-series decreasing weight.

[0009] In a possible implementation, the step of using the impact transition identification result and time-decreasing weights to perform time-series superposition fitting of the time-series additional damage factors to the fatigue damage evolution baseline involves the following processing: when any time-series additional damage factor is an impact transition identification result, time node compensation corresponding to the fatigue damage is performed according to the corresponding time-series additional damage factor; when any time-series additional damage factor is a non-impact transition identification result, multiple time-series additional damage factors are weighted and superimposed based on the time-decreasing weights; and time-series superposition fitting is completed by time node compensation and / or weighted superposition.

[0010] In a possible implementation, the step of updating the fatigue damage state of the bolt shaft using the time-series superposition fitting results involves the following processing: performing fatigue early warning trigger analysis using the updated fatigue damage state of the bolt shaft to establish a first early warning result; identifying fatigue abnormal impact based on strain sensing time-series signals and acceleration sensing time-series signals to establish a second early warning result; and issuing an early warning based on the first early warning result and the second early warning result.

[0011] In a possible implementation, the multimodal sensing monitoring is performed, a multimodal sensing monitoring dataset is established, and the following processing is performed: acquiring historical fatigue test data, setting the acquisition cycle of the coupling probe according to the historical fatigue test data; under the acquisition cycle, emitting an ultrasonic signal through the coupling probe and receiving the signal echo to establish a periodic ultrasonic sensing signal; activating the strain sensing gasket and acceleration sensor to perform time-series data acquisition to establish strain sensing time-series signals and acceleration sensing time-series signals; and establishing a multimodal sensing monitoring dataset based on the periodic ultrasonic sensing signal, strain sensing time-series signal, and acceleration sensing time-series signal.

[0012] In a possible implementation, the step of sending the signal feature extraction results and acquisition cycle to the fatigue identification main channel to establish a fatigue damage evolution baseline also includes the following processing: establishing an aging fit for the coupled probe, configuring a drift correction factor using the aging fit; and using the drift correction factor to perform periodic parameter drift correction for the fatigue identification main channel before establishing a fatigue damage evolution baseline.

[0013] This application also provides an electronic device, including: a memory for storing executable instructions; and a processor for executing the executable instructions stored in the memory to implement a bolt shaft fatigue testing method under multimodal sensing.

[0014] This application also provides a computer-readable storage medium, including: a computer program stored thereon, which, when executed by a processor, implements a bolt shaft fatigue testing method under multimodal sensing.

[0015] The proposed method, equipment, and medium for bolt shaft fatigue testing under multimodal sensing, as described in this application, firstly deploys a multimodal sensing layer on the bolt shaft to perform multimodal sensing monitoring and establish a multimodal sensing monitoring dataset. The multimodal sensing layer includes a coupling probe, a strain-sensing washer, and an acceleration sensor. Next, periodic ultrasonic sensing signals are extracted from the multimodal sensing monitoring dataset, and signal feature extraction is performed on these signals. The extracted signal feature results and acquisition period are sent to the fatigue identification main channel to establish a fatigue damage evolution baseline. Then, strain-sensing time-series signals and acceleration-sensing time-series signals are extracted from the multimodal sensing monitoring dataset. Additional damage analysis is performed using these signals to establish a time-series additional damage factor. Finally, the fatigue damage evolution baseline is time-series superimposed and fitted using the time-series superimposed and fitted results to update the bolt shaft fatigue damage state. This achieves the technical effect of improving the sensitivity of bolt shaft fatigue damage identification. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings of the embodiments of the present invention will be briefly described below. Flowcharts are used in this application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed precisely in sequence. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from these processes.

[0017] Figure 1 This is a flowchart illustrating the bolt shaft fatigue testing method under multimodal sensing provided in an embodiment of this application.

[0018] Figure 2 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0019] Explanation of reference numerals in the attached drawings: Input device 201, processor 202, memory 203, output device 204. Detailed Implementation

[0020] The above description is merely an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, specific embodiments of this application are given below.

[0021] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description of this application will be provided in conjunction with the accompanying drawings. The described embodiments should not be considered as limitations on this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0022] In the following description, references to "some embodiments" describe a subset of all possible embodiments. However, it is understood that "some embodiments" can be the same or different subsets of all possible embodiments and can be combined with each other without conflict. The terms "first" and "second" are used merely to distinguish similar objects and do not represent a specific ordering of objects. The terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules not explicitly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only.

[0023] This application provides a bolt shaft fatigue testing method under multimodal sensing, such as... Figure 1 As shown, the method includes:

[0024] Step S100: Deploy a multimodal sensing layer on the bolt shaft, perform multimodal sensing monitoring, and establish a multimodal sensing monitoring dataset. The multimodal sensing layer includes a coupling probe, a strain sensing washer, and an acceleration sensor.

[0025] Specifically, the multimodal sensing layer refers to a monitoring system that integrates multiple sensors (such as ultrasonic probes, strain sensors, and accelerometers) to simultaneously acquire signals of multiple physical quantities. Specifically, the bolt shaft surface is cleaned to ensure the installation location is free of impurities. According to design requirements, an ultrasonic coupling probe is used and fixed at the designated position on the bolt shaft. Through ultrasonic transmitting and receiving devices, non-destructive testing of the internal structure of the bolt shaft is achieved. A strain-sensing washer is installed between the bolt and the connected component. The washer contains a strain sensor that can monitor the axial and circumferential strain of the bolt shaft in real time. An accelerometer is installed at a suitable position on the bolt shaft to monitor the vibration acceleration signal of the bolt shaft during operation. The sensors are connected via a data acquisition card (such as an NI data acquisition card), and the sampling frequency (e.g., 10kHz) and sampling time (e.g., 10 seconds) are set. The data acquisition system is started to collect multimodal sensing data of the bolt shaft under different operating conditions and store it as a multimodal sensing monitoring dataset.

[0026] For example, when installing a bolt shaft for a bridge connection, a special ultrasonic coupling agent is used to ensure good contact between the probe and the surface of the bolt shaft when installing the coupling probe; when installing the strain sensing washer, the washer is ensured to fit tightly against the bolt and the connected parts to avoid signal interference; when installing the accelerometer, the sensor is fixed with screws to ensure that it does not loosen during vibration.

[0027] In one possible implementation, the step of performing multimodal sensing monitoring and establishing a multimodal sensing monitoring dataset, step S100, further includes step S110: acquiring historical fatigue test data and setting the acquisition cycle of the coupling probe based on the historical fatigue test data. Specifically, fatigue test data similar to the current bolt shaft is acquired from a database or historical test records, including ultrasonic signal characteristics, strain data, and acceleration data. The fatigue damage evolution law in the historical fatigue test data is analyzed, and the acquisition cycle of the coupling probe is set based on the analysis results. For example, assuming that historical fatigue test data shows that the fatigue damage of the bolt shaft grows slowly in the first 100 hours, with a damage increase of approximately 10% per 100 hours, while after 100 hours, the damage increase is approximately 5% per 10 hours. Therefore, the acquisition cycle of the coupling probe can be set as follows: data is acquired every 20 hours in the first 100 hours; and data is acquired every 10 hours after 100 hours.

[0028] In step S120, during the acquisition cycle, an ultrasonic signal is emitted through the coupling probe, and the signal echo is received to establish a periodic ultrasonic sensing signal. Specifically, the emission frequency (e.g., 1MHz) and pulse width (e.g., 10μs) of the ultrasonic signal are set in the coupling probe. When the acquisition cycle arrives, the coupling probe is triggered to emit an ultrasonic signal, receive the reflected echo signal, and extract the periodic signal features through filtering and amplification.

[0029] Step S130: Activate the strain-sensing washer and accelerometer to perform timing data acquisition, establishing strain-sensing timing signals and accelerometer timing signals. Specifically, activate the strain-sensing washer and accelerometer. Acquire strain data from the strain-sensing washer and acceleration data from the accelerometer. Perform low-pass filtering on the acquired timing signals to remove high-frequency noise. Normalize the filtered signals, adjusting the signal amplitude to between 0 and 1.

[0030] Step S140 involves establishing a multimodal sensing and monitoring dataset based on periodic ultrasonic sensing signals, strain sensing time-series signals, and acceleration sensing time-series signals. Specifically, the signal data acquired in steps S120 and S130 are integrated into a single dataset. This dataset is stored in a database and includes timestamps, ultrasonic signal characteristics, strain data, and acceleration data. This approach, by analyzing historical fatigue test data and appropriately setting the acquisition cycle of the coupled probe, ensures more frequent data acquisition during the rapid evolution phase of fatigue damage, thereby improving the richness and effectiveness of the data.

[0031] Step S200: Extract periodic ultrasonic sensing signals from the multimodal sensing monitoring dataset, perform signal feature extraction of the periodic ultrasonic sensing signals, send the signal feature extraction results and acquisition cycle to the fatigue identification main channel, and establish a fatigue damage evolution baseline.

[0032] Specifically, periodic ultrasonic sensing signals, acquired via coupled probes and reflecting the ultrasonic echo characteristics of the internal structure of the bolt shaft, are selected from the multimodal sensing and monitoring dataset. Digital signal processing techniques (such as bandpass filters) are used to remove noise and other non-periodic components, extracting the pure periodic ultrasonic sensing signals. A Fast Fourier Transform (FFT) is performed on the extracted periodic ultrasonic sensing signals to convert the time-domain signal to a frequency-domain signal. Characteristic parameters such as amplitude, frequency, and phase of the frequency-domain signal are extracted. For example, the dominant frequency, maximum amplitude, and phase shift of the signal are calculated. The extracted signal features (such as amplitude, frequency, and phase) and the acquisition period are sent to the fatigue identification main channel via a communication interface (such as Ethernet or USB). In the fatigue identification main channel, a fatigue damage evolution baseline is established based on the initially acquired signal features. This baseline serves as a reference curve to describe the fatigue damage evolution process. For example, the amplitude, frequency, and phase of the initially acquired ultrasonic signal are used as baseline parameters.

[0033] In one possible implementation, the signal feature extraction of the periodic ultrasonic sensing signal, sending the signal feature extraction results and acquisition cycle to the fatigue identification main channel, and establishing a fatigue damage evolution baseline, step S200 further includes step S210. The signal feature extraction includes echo delay increment extraction, echo amplitude attenuation rate extraction, and echo waveform offset extraction. Specifically, time delay analysis is performed on the ultrasonic signal for each acquisition cycle to calculate the echo delay increment, which is obtained by measuring the time delay change of the ultrasonic signal echo. Amplitude analysis is performed on the ultrasonic signal for each acquisition cycle to calculate the echo amplitude attenuation rate, which is obtained by analyzing the amplitude change of the ultrasonic signal echo. Waveform comparison is performed on the ultrasonic signal for each acquisition cycle to calculate the echo waveform offset, which is obtained by comparing the difference between the standard waveform and the actual waveform.

[0034] Step S220: Initialize the independent identification channel and cumulative damage channel within the fatigue identification main channel using the acquisition cycle. Specifically, according to the acquisition cycle, an independent identification channel is set for each cycle node to analyze transient fatigue state. According to the acquisition cycle, a cumulative damage channel is set to calculate the cycle signal change rate and generate a fatigue damage evolution baseline. The initial values ​​of the independent identification channel and cumulative damage channel are set to 0.

[0035] Step S230: After receiving the signal feature extraction results using the independent identification channel, analyze the transient fatigue state of each cycle node and establish a cycle fatigue state with a cycle node identifier. Specifically, in each independent identification channel, receive the signal feature extraction results (echo delay increment, echo amplitude attenuation rate, echo waveform offset). Analyze the transient fatigue state based on changes in signal characteristics. Generate an identified cycle fatigue state for each cycle node. Specifically, threshold ranges can be set for echo delay increment, echo amplitude attenuation rate, and echo waveform offset based on historical data or experimental results. These thresholds are used to determine the severity of the transient fatigue state. For example, set the delay increment threshold to 2μs, the amplitude attenuation rate threshold to 0.02V, and the waveform offset threshold to 2°. Based on the range of signal feature changes, classify the transient fatigue state into "normal state," "minor damage," "moderate damage," and "severe damage." For example: if the delay increment is <2μs, the amplitude attenuation rate is <0.02V, and the waveform offset is <2°, it is considered "normal state"; if the delay increment is between 2-4μs, the amplitude attenuation rate is between 0.02-0.04V, and the waveform offset is between 2-4°, it is considered "slight damage"; if the delay increment is between 4-6μs, the amplitude attenuation rate is between 0.04-0.06V, and the waveform offset is between 4-6°, it is considered "moderate damage"; if the delay increment is >6μs, the amplitude attenuation rate is >0.06V, and the waveform offset is >6°, it is considered "severe damage". Transient fatigue state is determined using logical judgment or rule-based algorithms based on signal feature extraction results.

[0036] Step S240: After receiving the signal feature extraction result and the periodic fatigue state using the cumulative damage channel, the periodic signal change rate is calculated. The fatigue damage evolution baseline is generated using the periodic signal change rate calculation result and the periodic fatigue state. Specifically, the periodic signal change rate refers to the rate of change of signal features (such as echo delay increment, echo amplitude attenuation rate, and echo waveform offset) between adjacent acquisition cycles. For each signal feature, the change between adjacent cycles is calculated and standardized as a change rate. For example, for the echo delay increment, the change rate can be expressed as the ratio of the delay increment change between adjacent cycles to the acquisition cycle. The fatigue damage evolution baseline is a reference curve used to describe the evolution of fatigue damage over time. Based on the periodic signal change rate and the periodic fatigue state, the fatigue damage evolution baseline is generated using a cumulative damage model (such as a linear cumulative damage model). This implementation method, through independent identification channels and cumulative damage channels, can simultaneously achieve short-term and long-term fatigue monitoring, capturing both transient changes and assessing cumulative damage, thus providing more comprehensive and accurate data support for fatigue monitoring of bolted shafts.

[0037] In one possible implementation, after receiving the signal feature extraction result and the periodic fatigue state using the cumulative damage channel, the periodic signal change rate is calculated. Step S240 further includes step S241, establishing the calibration fatigue cycle of the bolt shaft. Specifically, historical fatigue test data is analyzed to determine the fatigue characteristics of the bolt shaft at different stages. Based on the evolution law of fatigue damage, the entire fatigue test process is divided into three stages: initial, middle, and final. A calibration fatigue cycle is set for each stage, for example: initial stage: 0-50 hours (amplitude attenuation rate is more sensitive, time delay increment and waveform offset change are small, but begin to gradually increase); middle stage: 50-100 hours (amplitude attenuation rate and time delay increment are both sensitive, waveform offset begins to change significantly, fatigue damage gradually accumulates); final stage: 100-150 hours (time delay increment and waveform offset are more sensitive, amplitude attenuation rate changes tend to stabilize, fatigue damage accumulates rapidly).

[0038] Step S242 involves performing node segmentation to calibrate the fatigue cycle using the cyclic fatigue state, establishing node segmentation results, and configuring dynamic weighting coefficients based on these results. Specifically, according to the calibrated fatigue cycle, the collected cyclic fatigue state data is segmented into different nodes, each corresponding to a specific time range and fatigue stage. Based on the fatigue stage characteristics of each node, different weighting coefficients are assigned to different signal characteristics (amplitude attenuation rate, time delay increment, waveform offset). The allocation of weighting coefficients is based on the sensitive characteristics of each stage; for example, in the initial stage, the amplitude attenuation rate has a higher weight, while the time delay increment and waveform offset have lower weights; in the middle stage, the amplitude attenuation rate and time delay increment have similar weights, while the waveform offset has a moderate weight; and in the final stage, the time delay increment and waveform offset have higher weights, while the amplitude attenuation rate has a lower weight.

[0039] Step S243: During the calculation of the periodic signal change rate, multi-feature damage weighted calculation is performed using the dynamic weighting coefficients. The periodic fatigue state is corrected based on the multi-feature damage weighted calculation results to generate a fatigue damage evolution baseline. Specifically, the change rate of each signal feature is weighted according to the dynamic weighting coefficients, with the total weighted value = (amplitude attenuation rate change rate × amplitude attenuation rate weight) + (delay increment change rate × delay increment weight) + (waveform offset change rate × waveform offset weight). Based on the multi-feature damage weighted calculation results, the fatigue state of each cycle is corrected, and the corrected fatigue state more accurately reflects the actual degree of fatigue damage. The corrected fatigue states of each cycle are accumulated to form a complete fatigue damage evolution baseline. This implementation method, through dynamic weighting coefficients, adjusts the weights of signal features according to the fatigue characteristics of different stages, enabling a more accurate reflection of the fatigue damage evolution process. By using multi-feature damage weighted calculation to correct the fatigue state of each cycle, the limitations of a single feature are avoided, improving the accuracy of fatigue state assessment.

[0040] In one possible implementation, the step of sending the signal feature extraction results and acquisition cycle to the fatigue identification main channel to establish a fatigue damage evolution baseline, further includes step S250, which involves establishing an aging fit for the coupling probe and configuring a drift correction factor using the aging fit. Specifically, signal feature data of the coupling probe at different usage times (e.g., 0 hours, 50 hours, 100 hours, 150 hours, etc.), including amplitude, time delay, waveform, etc., are collected to analyze the aging trend of the coupling probe. Mathematical modeling methods (e.g., linear regression, polynomial fitting, etc.) are used to fit the aging data of the coupling probe to establish an aging model, which describes the change law of the coupling probe performance over time. Based on the aging model, a drift correction factor for each acquisition cycle is calculated to correct the drift of the coupling probe signal and ensure the accuracy of the signal features.

[0041] Step S260: After performing periodic parameter drift correction on the fatigue identification main channel using the drift correction factor, a fatigue damage evolution baseline is established. Specifically, the signal characteristics of each acquisition cycle are corrected using the drift correction factor. The corrected signal characteristics more accurately reflect the actual fatigue state. Based on the corrected signal characteristics, the fatigue state of each cycle is calculated, and the fatigue states of each cycle are accumulated to generate a complete fatigue damage evolution baseline. This implementation method, by establishing an aging fitting model for the coupled probe and configuring the drift correction factor, can effectively correct signal drift and ensure the accuracy of signal characteristics. The corrected signal characteristics can more accurately reflect the actual fatigue state, thereby improving the reliability of the fatigue damage evolution baseline.

[0042] Step S300: Extract the strain sensing time-series signal and acceleration sensing time-series signal from the multimodal sensing monitoring dataset.

[0043] Specifically, data filtering algorithms are used to extract preprocessed strain sensing time-series signals and acceleration sensing time-series signals from multimodal sensing and monitoring datasets.

[0044] Step S400: Perform additional damage analysis using the strain sensing time-series signal and the acceleration sensing time-series signal to establish a time-series additional damage factor.

[0045] Specifically, the temporal additional damage factor refers to a factor obtained by analyzing time-series signals to quantify the impact of additional damage. Machine learning algorithms (such as Support Vector Machines, SVM) are used to analyze preprocessed strain-sensing and acceleration-sensing time-series signals to identify additional damage features. For example, an SVM model is trained using labeled training data, and the preprocessed strain-sensing and acceleration-sensing time-series signals are input into the SVM model to identify additional damage features. Based on the identification results, a temporal additional damage factor is established to quantify the impact of additional damage on fatigue damage. For example, the additional damage factor can be defined as the ratio of the additional damage feature to the normal state feature. For instance, assuming the SVM model identifies an additional damage feature in the strain-sensing time-series signal with an amplitude of 0.2, while the amplitude under normal conditions is 0.1, then the additional damage factor can be defined as 0.2 / 0.1 = 2.

[0046] Step S500: Perform time-series superposition fitting of the fatigue damage evolution baseline using the time-series additional damage factor, and update the fatigue damage state of the bolt shaft using the time-series superposition fitting result.

[0047] Specifically, time-series superposition fitting refers to mathematically fitting an additional damage factor to a time-series fatigue damage evolution baseline to update the fatigue damage state. Mathematical modeling methods (such as linear or nonlinear regression) are used to superimpose and fit the time-series additional damage factor to the time-series fatigue damage evolution baseline. The fatigue damage state of the bolt shaft is updated based on the fitting results, and the updated fatigue damage state is displayed using visualization tools (such as plotting libraries in Matlab or Python).

[0048] For example, assuming the fatigue damage evolution baseline is a linearly increasing curve with a slope of 0.1, and the additional damage factor is 2, then the slope of the updated fatigue damage state curve is 0.1 × 2 = 0.2. The fatigue damage state curve plotted using Matlab can visually demonstrate the fatigue damage evolution process of the bolt shaft.

[0049] In one possible implementation, the step of performing a time-series superposition fitting of the fatigue damage evolution baseline using the time-series additional damage factor, and updating the fatigue damage state of the bolt shaft using the time-series superposition fitting result, further includes step S510: configuring a backtracking time window and setting a time-series decreasing weight according to the window length of the backtracking time window. Specifically, a backtracking time window is defined to consider the impact of additional damage over a past period on the current fatigue state. The length of the backtracking time window can be set according to actual needs, such as 30 hours, 50 hours, or 100 hours. A decreasing weight is set for each time point according to the length of the backtracking time window to reflect the characteristic that the more distant the additional damage, the smaller its impact on the current state. The decreasing weight can be linearly decreasing, exponentially decreasing, or other suitable decreasing functions.

[0050] Step S520: Impact transition identification is performed based on the time-series additional damage factors, and the impact transition identification results are configured. Specifically, the time-series additional damage factors are analyzed to identify damage that increases significantly within a short period (i.e., impact transitions). Impact transitions represent the rapid accumulation or sudden occurrence of fatigue damage. For example, the change in additional damage factors between adjacent time points is calculated, and time points where the change exceeds a threshold are identified and marked as impact transitions. The identified impact transitions are marked, and their occurrence time and magnitude are recorded.

[0051] Step S530 involves performing a time-series superposition fitting of the time-series additional damage factor onto the fatigue damage evolution baseline using the impact transition identification results and time-series decreasing weights. Specifically, during the time-series superposition fitting process, special attention is paid to the impact transition identification results. Identified impact transition points are given higher weights or special processing to reflect their significant impact on the fatigue damage state. Impact transition points represent the acceleration stage of fatigue damage, and their influence needs to be highlighted during the fitting process. For non-impact transition points, time-series decreasing weights are applied to reflect the characteristic that additional damage from more recent times has a smaller impact on the current state. By combining the impact transition identification results and time-series decreasing weights, the time-series additional damage factor is superimposed and fitted onto the fatigue damage evolution baseline. The fitting result is used to update the fatigue damage state of the bolt shaft. This implementation, by combining time-series decreasing weights and impact transition identification results, can dynamically adjust the fatigue damage state, ensuring the accuracy and real-time performance of the fatigue damage evolution baseline.

[0052] In one possible implementation, step S530 further includes step S531, whereby, when any time-series additional damage factor is an impact transition identification result, time-series compensation for the fatigue damage is performed based on the corresponding time-series additional damage factor. Specifically, based on the impact transition identification result, it is determined which time points are impact transition points. For each impact transition point, time-series compensation is performed on the fatigue damage evolution baseline based on the magnitude of its additional damage factor, directly reflecting the impact of the impact transition on the fatigue damage evolution baseline. For example, if the additional damage factor of the impact transition point is 0.2, this value is directly added to the fatigue damage evolution baseline at the current time point.

[0053] Step S532: When any temporal additional damage factor is a non-impact transition identification result, multiple temporal additional damage factors are weighted and superimposed based on the temporal decreasing weight. Specifically, for non-impact transition points, the additional damage factors are weighted and superimposed according to the temporal decreasing weight. The superposition strategy can be adjusted according to actual needs, such as linear superposition or nonlinear superposition.

[0054] Step S533 involves completing the time-series superposition fitting using time node compensation and / or weighted superposition. Specifically, the results of time node compensation and / or weighted superposition are combined to complete the time-series superposition fitting. Based on the fitting results, the fatigue damage state of the bolt shaft is dynamically updated. This implementation method, by integrating additional damage factors, time-series decreasing weights, and impact transition identification results, can more comprehensively reflect the evolution process of fatigue damage and improve the reliability of fatigue monitoring. In particular, time node compensation can highlight the rapid accumulation or sudden damage of fatigue damage, making the update of the fatigue damage state more sensitive and accurate.

[0055] In one possible implementation, the step S500, which updates the fatigue damage state of the bolt shaft using the time-series superposition fitting result, further includes step S540, which uses the updated fatigue damage state of the bolt shaft to perform fatigue early warning trigger analysis and establish a first early warning result. Specifically, a fatigue damage early warning threshold is set based on historical data, experimental results, or industry standards. For example, an early warning is triggered when the fatigue damage level reaches 0.8. Based on the updated fatigue damage state, it is analyzed whether the preset fatigue damage threshold has been reached. If the fatigue damage state exceeds the preset threshold, a fatigue early warning is triggered, and the early warning information is recorded. The early warning information includes the time point, fatigue damage level, and early warning level.

[0056] Step S550: Based on the strain sensing time-series signal and the acceleration sensing time-series signal, fatigue abnormal impact identification is performed, and a second early warning result is established. Specifically, the strain sensing time-series signal and the acceleration sensing time-series signal are analyzed to identify whether abnormal impact exists. Abnormal impact can be identified through signal abrupt changes, peak values, or frequency variations. For example, an abnormal impact is considered to exist when the strain signal exceeds a certain threshold or the peak value of the acceleration signal exceeds a certain threshold. If an abnormal impact is identified, a fatigue abnormal impact early warning is triggered, and the early warning information is recorded. The early warning information includes the time point, the magnitude of the abnormal impact, and the early warning level.

[0057] Step S560: Issue a warning based on the first and second warning results. Specifically, a comprehensive warning is issued by combining the first warning result (fatigue damage state) and the second warning result (fatigue abnormal impact). An appropriate alarm method, such as audible and visual alarm, SMS notification, or system prompt, is selected according to the warning level and type. All warning information is recorded in the system log for subsequent analysis and tracing. This implementation method, by combining fatigue damage state and fatigue abnormal impact, achieves multi-dimensional warning analysis, improving the accuracy and reliability of the warnings.

[0058] This application employs a multimodal sensing layer constructed by arranging a coupled probe, strain-sensing washer, and acceleration sensor on the bolt shaft. Monitoring is conducted, and a multimodal sensing monitoring dataset is established. Periodic ultrasonic sensing signals are extracted from the dataset, signal features are extracted, and the results, along with the acquisition cycle, are transmitted to the fatigue identification main channel to establish a fatigue damage evolution baseline. Strain-sensing time-series signals and acceleration-sensing time-series signals are extracted from the dataset, and additional damage analysis is performed using these two time-series signals to construct a time-series additional damage factor. This time-series additional damage factor is then used to perform time-series superposition fitting of the fatigue damage evolution baseline to update the fatigue damage state of the bolt shaft. These techniques solve the technical problem of low damage identification sensitivity in existing bolt shaft fatigue monitoring, achieving the technical effect of improving the sensitivity of bolt shaft fatigue damage identification.

[0059] Based on the foregoing embodiments, this application also provides an electronic device and a computer-readable storage medium storing a computer program. When the computer program is executed by the processor of the electronic device, it can implement the methods described in any of the preceding embodiments.

[0060] Figure 2 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention, showing a block diagram of an exemplary electronic device suitable for implementing the embodiments of the present invention. Figure 2The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments of the present invention. This electronic device is in the form of a general-purpose computing device, and its components may include, but are not limited to, an input device 201, a processor 202, a memory 203, and an output device 204. The processor 202 may be one or more; the memory 203 may include a computer-readable medium and at least one program product having a set (at least one) of program modules configured to perform the functions of the embodiments of this application.

[0061] The memory 203 shown in this embodiment of the invention can be any combination of one or more computer-readable media. The computer-readable storage medium can be, but is not limited to, infrared, semiconductor systems, devices or components, or any combination thereof, used to store software programs, computer-executable programs and modules, such as the program instructions / modules corresponding to the bolt shaft fatigue testing method under multimodal sensing in this embodiment of the invention. The processor 202 executes various functional applications and data processing of the computer device by running the software programs, instructions and modules stored in the memory 203, thereby realizing the above-mentioned bolt shaft fatigue testing method under multimodal sensing.

[0062] The specific embodiments described above do not constitute a limitation on the scope of protection of this application. Those skilled in the art should understand that various modifications, combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application. In some cases, the actions or steps described in this application can be performed in a different order than that shown in the embodiments and still achieve the desired results. Furthermore, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A method for bolt shaft fatigue testing under multi-modal perception, characterized in that, The method comprises: Deploying a multi-modal sensing layer on the bolt shaft, performing multi-modal sensing monitoring, establishing a multi-modal sensing monitoring dataset, the multi-modal sensing layer comprising a coupling probe, a strain sensing gasket, and an acceleration sensor; Extracting periodic ultrasonic sensing signals in the multi-modal sensing monitoring dataset, performing signal feature extraction of the periodic ultrasonic sensing signals, sending the signal feature extraction results and the collection period to a fatigue identification main channel, and establishing a fatigue damage evolution baseline; Extracting strain sensing time series signals and acceleration sensing time series signals in the multi-modal sensing monitoring dataset; Performing additional damage analysis using the strain sensing time series signals and the acceleration sensing time series signals, and establishing a time series additional damage factor; Performing time series superposition fitting of the fatigue damage evolution baseline using the time series additional damage factor, and updating the fatigue damage state of the bolt shaft using the time series superposition fitting result; The signal feature extraction comprises echo time delay increment extraction, echo amplitude decay rate extraction, and echo waveform offset extraction. Initializing independent identification channels and cumulative damage channels in the fatigue identification main channel using the collection period; After receiving the signal feature extraction results using the independent identification channels, analyzing the transient fatigue state of each period node, and establishing a period fatigue state with period node identification; After receiving the signal feature extraction results and the period fatigue state using the cumulative damage channels, performing period signal change rate calculation, generating a fatigue damage evolution baseline using the period signal change rate calculation result and the period fatigue state; The time series superposition fitting of the fatigue damage evolution baseline using the time series additional damage factor comprises: Configuring a backtracking time window and setting a time series decreasing weight according to the window length of the backtracking time window; Performing impact transition identification according to the time series additional damage factor, and configuring the impact transition identification result; Performing time series superposition fitting of the fatigue damage evolution baseline using the time series additional damage factor according to the impact transition identification result and the time series decreasing weight. The period signal change rate calculation after receiving the signal feature extraction results and the period fatigue state using the cumulative damage channels comprises:

2. The bolt shaft fatigue testing method under multi-modal perception as claimed in claim 1, wherein, Establishing a calibrated fatigue period of the bolt shaft; Performing node segmentation of the calibrated fatigue period using the period fatigue state, establishing a node segmentation result, and configuring a dynamic weight coefficient using the node segmentation result; During the period signal change rate calculation, performing multi-feature damage weighted calculation through the dynamic weight coefficient, correcting the period fatigue state according to the multi-feature damage weighted calculation result, and generating a fatigue damage evolution baseline. The time series superposition fitting of the fatigue damage evolution baseline using the time series additional damage factor comprises:

3. The bolt shaft fatigue testing method under multi-modal perception as claimed in claim 1, wherein, ​ When any time sequence additional damage factor is an impact jump identification result, time node compensation corresponding to the fatigue damage is performed according to the corresponding time sequence additional damage factor; When any time sequence additional damage factor is a non-impact jump identification result, the weighted superposition of multiple time sequence additional damage factors is performed based on time sequence decreasing weights; The time sequence superposition fitting is completed by time node compensation and / or weighted superposition.

4. The bolt shaft fatigue testing method under multi-modal perception as claimed in claim 1, wherein, The bolt shaft fatigue damage state is updated by using the time sequence superposition fitting result, including: Fatigue early warning trigger analysis is performed by using the updated bolt shaft fatigue damage state, and a first early warning result is established; Fatigue abnormal impact identification is performed according to the strain sensing time sequence signal and the acceleration sensing time sequence signal, and a second early warning result is established; The first early warning result and the second early warning result are used for early warning.

5. The bolt shaft fatigue testing method under multi-modal perception as claimed in claim 1, wherein, The multi-modal sensing monitoring is performed, and a multi-modal sensing monitoring data set is established, including: Historical fatigue test data is acquired, and a collection period of the coupled probe is set according to the historical fatigue test data; Under the collection period, the coupled probe emits an ultrasonic signal and receives a signal echo, and a periodic ultrasonic sensing signal is established; The strain sensing gasket and the acceleration sensor are activated to perform time sequence data acquisition, and strain sensing time sequence signals and acceleration sensing time sequence signals are established; A multi-modal sensing monitoring data set is established according to the periodic ultrasonic sensing signal, the strain sensing time sequence signal, and the acceleration sensing time sequence signal.

6. The bolt shaft fatigue testing method under multi-modal sensing as claimed in claim 1, wherein, The signal feature extraction result and the collection period are sent to a fatigue identification main channel to establish a fatigue damage evolution baseline, and the method further includes: An aging fitting of the coupled probe is established, and a drift correction factor is configured by using the aging fitting; After periodic parameter drift correction of the fatigue identification main channel is performed by using the drift correction factor, the fatigue damage evolution baseline is established.

7. An electronic device, comprising: The electronic device includes: A memory for storing executable instructions; A processor for executing the executable instructions stored in the memory, realizing the bolt shaft fatigue test method under multi-modal sensing according to any one of claims 1 to 6.

8. A computer-readable storage medium having stored thereon a computer program, characterized in that, The program is executed by the processor to realize the bolt shaft fatigue test method under multi-modal sensing according to any one of claims 1 to 6.

Images