On-line monitoring and early warning method and system for service state of ear-free check ring

Abstract

The invention discloses a service state online monitoring and early warning method and system for an ear-free check ring. The method comprises the following steps: acquiring an acoustic signal acquired by a monitoring module arranged at a mounting part of the ear-free check ring, wherein the signal comprises a passive acoustic emission signal generated by the ear-free check ring in a stress state; processing the acoustic signal to extract characteristic parameters related to an interface contact state and / or a structural integrity state; determining a current health state index based on the characteristic parameters; and generating and outputting state evaluation or early warning information according to the index and a preset evaluation rule. The system comprises a monitoring module, a signal processing and analyzing module, an output module and an active excitation module. According to the method, real-time and online monitoring and early warning of the service state of the ear-free check ring are achieved, early degradation such as fretting wear and microcracks can be recognized, sudden failure is effectively prevented, and the reliability and safety of equipment operation are improved.

Description

Technical Field

[0001] This invention relates to the field of lugless retainer condition monitoring technology, specifically to a method and system for online monitoring and early warning of the service status of lugless retainers. Background Technology

[0002] In mechanical transmission and equipment systems, the lugless retaining ring, as a core axial positioning and locking element, is widely used in rotary or reciprocating motion mechanisms in key fields such as aerospace, rail transportation and precision machinery due to its compact structure and convenient installation.

[0003] These components are typically concealed within space-constrained shaft grooves or bores, enduring complex alternating loads and vibrations over extended periods. Their service condition directly impacts the reliability and safety of the entire transmission chain. The failure mode of lugless retaining rings is progressive and insidious, usually beginning with preload relaxation due to interfacial fretting wear or the initiation of micro-cracks within the material. If effective identification and early warning are not provided in the early stages of performance degradation, sudden breakage or detachment can trigger axial movement of the fixed components, leading to excessive equipment vibration, loss of transmission accuracy, and potentially even severe cascading failures such as bearing sintering and gear damage, resulting in significant safety risks and economic losses. Therefore, real-time, accurate condition monitoring of lugless retaining rings during service, and subsequent predictive maintenance, is of paramount engineering importance. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for online monitoring and early warning of the service status of equipment without ear loops, so as to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for online monitoring and early warning of the service status of vehicles without ear loops, comprising: Acoustic signals are acquired by a monitoring module located at the installation location of the target earless retainer ring, wherein the acoustic signals include passive acoustic emission signals generated by the earless retainer ring under stress. The acoustic signal is processed to extract at least one feature parameter associated with the interface contact state and / or structural integrity state of the ear ring-less structure. Based on at least one of the feature parameters, determine the current health status indicators of the earless ring; Based on the current health status indicators and preset evaluation rules, generate and output status evaluation information or early warning information related to the earless ring.

[0006] Preferably, the step of acquiring the acoustic signal further includes: A preset excitation signal is applied to the structure carrying the earless retainer ring through the active excitation module; Acquire acoustic signals collected by the monitoring module, including the response signal of the earless retainer ring to the excitation signal; The processing steps include: processing the response signal to extract feature parameters associated with the characteristics of the earless retainer structure.

[0007] Preferably, prior to the processing step, a step of constructing a baseline model is also included: When the ear-free retainer is in a known healthy state, its baseline acoustic signal under various preset working conditions is acquired; Process the baseline acoustic signal and extract the corresponding baseline feature parameter set; The step of determining the current health status index includes: comparing and analyzing the at least one feature parameter extracted in real time with the parameters under the corresponding working conditions in the baseline feature parameter set, and calculating the current health status index based on the degree of deviation.

[0008] Preferably, the at least one characteristic parameter includes at least one of the following extracted from the passive acoustic emission signal: event count rate, signal amplitude distribution statistics, signal energy rate, and frequency domain characteristic parameters obtained by time-frequency transformation.

[0009] Preferably, processing the response signal to extract feature parameters includes extracting at least one of the following: the time difference of arrival of the response signal, the signal attenuation coefficient, or the change in the frequency response function relative to the baseline state.

[0010] Preferably, the step of determining the current health status indicators includes: The extracted feature parameters are normalized separately. Assign a weight coefficient to each normalized feature parameter based on its correlation with the degradation state of the earless retainer; The weighted feature parameters are fused and calculated to generate a comprehensive scalar value as the current health status indicator.

[0011] Preferably, the generation of early warning information based on current health status indicators and preset assessment rules includes: Based on the historical operating data of the ear-proof ring and the historical trend of the current health status indicators, the thresholds for triggering different levels of warnings are dynamically adjusted. The current health status indicators are compared with dynamically adjusted thresholds. If the conditions are met, a warning message of the corresponding level is triggered. The warning levels include at least performance degradation warning and failure risk warning.

[0012] Preferably, it also includes a prediction step: The sequence data of the current health status indicators over time and the current operating parameters are input into the trained prediction model. The predictive model outputs predictive maintenance information for the earless retainer ring, which includes an estimated remaining service life or a suggested maintenance time. The status assessment information or early warning information includes the predictive maintenance information.

[0013] This application also proposes an online monitoring and early warning system for the service status of systems without ear loops, for implementing the above method, the system comprising: The monitoring module includes at least one acoustic sensor arranged near the mounting location of the earless retainer ring for acquiring acoustic signals; The signal processing and analysis module is communicatively connected to the monitoring module and is configured to perform the above-described processing steps, determination steps, and generation steps. The output module is used to output the status assessment information or early warning information.

[0014] Preferably, it also includes an active excitation module, which is communicatively connected to the signal processing and analysis module. The active excitation module is configured to apply a preset excitation signal to the structure carrying the earless retainer ring according to the instructions of the signal processing and analysis module. The monitoring module is also used to acquire acoustic signals including the response signal of the earless retainer ring to the excitation signal.

[0015] Compared with the prior art, the beneficial effects of the present invention are: by acquiring acoustic signals, extracting feature parameters, determining health status indicators and generating early warning information, it is possible to monitor the status changes of ear rings in real time, realize early identification and timely warning, and achieve real-time online monitoring and early warning of the service status of ear rings, solving the problems of difficulty in early identification and untimely warning. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the online monitoring and early warning method for service status according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the connection of the online monitoring and early warning system for service status according to an embodiment of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Please see Figure 1-2 A method for online monitoring and early warning of the service status of a lugless retainer ring includes: acquiring acoustic signals collected by a monitoring module disposed at the installation location of the target lugless retainer ring, wherein the acoustic signals include passive acoustic emission signals generated by the lugless retainer ring under stress; processing the acoustic signals to extract at least one feature parameter associated with the interface contact state and / or structural integrity state of the lugless retainer ring; determining the current health status index of the lugless retainer ring based on at least one feature parameter; and generating and outputting status assessment information or early warning information related to the lugless retainer ring based on the current health status index and preset assessment rules.

[0019] For ease of understanding, the following explains some key terms in this embodiment: Earless retaining rings, as an axial positioning and locking element, are characterized by the absence of the protruding ears found in traditional retaining rings. They are typically installed inside shaft grooves or bore grooves to restrict the axial movement of components. Their reliability in service directly affects the operational safety of the entire mechanical system.

[0020] The monitoring module, typically consisting of one or more acoustic sensors, is designed to collect acoustic signals generated by the earless retainer during service. These sensors, such as piezoelectric sensors, fiber optic sensors, or microelectromechanical systems (MEMS) microphones, are positioned near the earless retainer's mounting location to ensure effective capture of relevant acoustic information.

[0021] Acoustic signals refer to elastic wave signals that propagate in air or solid media and are acquired by a monitoring module. In monitoring scenarios without ear rings, acoustic signals include passive acoustic emission signals generated by the ear rings under stress, as well as potential environmental noise.

[0022] Passive acoustic emission signals specifically refer to elastic waves generated by the instantaneous release of stress waves during processes such as deformation, crack initiation and propagation, friction, or phase transformation of materials. These signals are characterized by wide bandwidth and transient nature, and their generation is directly related to changes in the internal or interface state of the material. They are important indicators for evaluating the integrity of earless retaining ring structures and the interface contact state.

[0023] Characteristic parameters refer to physical or statistical quantities extracted from the original acoustic signal that can quantify the service status of the earless retainer. These parameters may include the signal amplitude, energy, frequency distribution, etc., and their changing trends correspond to the performance degradation or failure mode of the earless retainer.

[0024] A health status index is a comprehensive numerical value or grade used to quantify the current service condition of an earless ring. This index is typically calculated based on one or more characteristic parameters and can intuitively reflect the health level of the earless ring, such as different levels from normal to severely degraded.

[0025] Assessment rules are logical criteria used to determine the health status of ear loops and trigger corresponding information outputs. These rules can be established based on empirical thresholds, statistical models, or machine learning models, and are used to map health status indicators into specific status assessment information or early warning information.

[0026] Status assessment information or early warning information is the result output by the system based on the health status indicators and assessment rules of the earless retainer. Status assessment information typically provides a description of the current operating status of the earless retainer, while early warning information is issued when a potential risk is detected, prompting that maintenance measures need to be taken.

[0027] This application provides an embodiment of a method for online monitoring and early warning of the service status of a lugless retainer. First, acoustic signals are acquired by a monitoring module located at the mounting site of the target lugless retainer. The monitoring module can consist of one or more acoustic sensors. These can be piezoelectric sensors directly coupled to the surface of the structure supporting the lugless retainer, or an air-coupled microphone array positioned in the vicinity of the mounting site. Acoustic signal acquisition can be set to a continuous mode, i.e., continuously during the service life of the lugless retainer, or a periodic mode, acquiring signals at preset time intervals. The acquired acoustic signals include passive acoustic emission signals generated by the lugless retainer under stress. These signals are elastic waves generated due to microscopic changes in the material's interior or interface during service, such as the initiation and propagation of microcracks or fretting wear of the interface.

[0028] Furthermore, the acquired acoustic signals are processed to extract at least one characteristic parameter associated with the interface contact state and / or structural integrity state of the earless retainer. Acoustic signal processing may include preprocessing steps such as filtering, amplification, and analog-to-digital conversion to remove noise and prepare the data. Characteristic parameter extraction may be based on time-domain analysis, such as calculating the root mean square value, peak value, or energy integral of the acoustic signal. Alternatively, the energy distribution or dominant frequency component within a specific frequency band can be obtained as characteristic parameters by performing a Fourier transform on the acoustic signal. Changes in these characteristic parameters are considered to be associated with changes in the earless retainer's preload, wear, or crack propagation.

[0029] Based on this, at least one extracted feature parameter is used to determine the current health status index of the earmuff-less device. The determination of the current health status index can be based on one or more extracted feature parameters. The real-time value of a key feature parameter can be compared with a preset reference value, and the health status can be directly judged based on the degree of deviation. Alternatively, a simple arithmetic average of multiple feature parameters can be performed to obtain a comprehensive value as the health status index. This index aims to quantify the health level of the earmuff-less device; for example, a higher value indicates a worse health condition.

[0030] Finally, based on the current health status indicators and preset evaluation rules, status evaluation information or early warning information related to the ear loop removal is generated and output. The generation of status evaluation information or early warning information can be based on preset evaluation rules. A series of fixed thresholds can be set; when the current health status indicators exceed a certain threshold, an early warning of the corresponding level is triggered. The output information can be displayed in text form on the operation interface or through an audible and visual alarm device. The status evaluation information can include the current health level of the ear loop removal, while the early warning information can indicate potential failure risks and suggest inspection or maintenance.

[0031] This application achieves early, online monitoring and early warning of the service status of earless retaining rings by acquiring passive acoustic emission signals online in real time and processing them to extract characteristic parameters related to interface contact status and / or structural integrity status. This allows for the determination of health status indicators and the generation of early warning information. Consequently, it effectively identifies potential failure risks of earless retaining rings, preventing equipment malfunctions and safety accidents caused by their hidden failures, and improving the operational reliability and safety of mechanical systems.

[0032] In some of the embodiments described above in this application, a method for acquiring passive acoustic emission signals to monitor the state of the earless retainer ring is proposed. However, during its implementation, the passive signal may be greatly affected by external loads and environmental factors, resulting in unstable or incomplete extraction of feature parameters, which cannot reliably reflect the structural characteristics of the earless retainer ring, thereby affecting the accuracy of the health status assessment.

[0033] In this regard, this application further proposes that the step of acquiring acoustic signals also includes applying a preset excitation signal to the structure carrying the earless retainer through an active excitation module, and acquiring an acoustic signal collected by the monitoring module that includes the response signal of the earless retainer to the excitation signal; the processing step includes processing the response signal to extract feature parameters associated with the characteristics of the earless retainer structure.

[0034] The active excitation module is used to apply a controllable, preset energy input to the lugless retainer and its supporting structure to induce a measurable response from the lugless retainer. This actively applied excitation signal overcomes the drawback of passive acoustic emission signals being greatly affected by environmental and load uncertainties, providing a stable and repeatable input condition, thus making subsequent response signal analysis more reliable and comparable. The active excitation module can be a piezoelectric ceramic exciter, which generates ultrasonic or vibration waves by applying electrical pulses or sweep frequency signals and couples them to the shaft or bore wall where the lugless retainer is located. Alternatively, it can be an electromagnetic exciter, which drives a small hammer device or vibrator through electromagnetic force to subject the lugless retainer's supporting structure to periodic or transient mechanical impact or vibration excitation.

[0035] After the active excitation signal is applied, the monitoring module is used to collect the dynamic response of the earless retaining ring and its supporting structure to the excitation signal. This response signal contains information about the current structural characteristics of the earless retaining ring, such as its stiffness, damping, and whether there are cracks or loosening. By analyzing these response signals, the structural integrity of the earless retaining ring can be directly and comprehensively assessed. The monitoring module can use a piezoelectric acoustic emission sensor or an accelerometer, which can be directly attached or fixed near the installation location of the earless retaining ring to capture the structural vibration or sound wave propagation signal caused by the excitation with high sensitivity. Alternatively, the monitoring module can also be a laser vibrometer, which measures the vibration displacement, velocity, or acceleration of the surface of the earless retaining ring or its supporting structure in a non-contact manner to obtain the response signal.

[0036] The subsequent processing steps aim to quantify parameters reflecting changes in the internal structural state of the earless retainer from the acquired response signal. These parameters are crucial for assessing the health of the earless retainer, as they are directly related to physical properties such as the material's elastic modulus, geometric dimensions, and connection tightness. Time-domain analysis can be performed on the response signal to extract parameters such as peak value, root mean square value, energy, and attenuation rate; or frequency-domain analysis can be performed to obtain the spectrum through Fast Fourier Transform (FFT) and extract parameters such as resonant frequency, harmonic components, and frequency band energy distribution. Furthermore, modal analysis methods can be employed to identify the structure's modal parameters, such as modal frequencies, modal damping ratios, and mode shapes, through excitation-response data. These parameters are highly sensitive to structural damage.

[0037] Through the above technical solution, this application overcomes the problem of unstable or incomplete feature parameter extraction caused by changes in external loads and environmental factors when relying solely on passive acoustic emission signals. By applying a preset excitation signal to the structure carrying the earless retainer using the active excitation module, a stable and controllable input condition can be provided, reducing environmental interference and ensuring the reliability of the acquired response signals. The response signals acquired by the monitoring module directly reflect the dynamic response of the earless retainer to a specific excitation, including its current structural characteristic information, such as stiffness, damping, or the presence of microscopic damage. Subsequently, these response signals are specially processed to extract feature parameters associated with the structural characteristics of the earless retainer, such as resonant frequency, modal parameters, or signal attenuation characteristics. These parameters can more directly and comprehensively quantify the internal state changes of the earless retainer, compensating for the shortcomings of passive acoustic emission signals in reflecting structural characteristics. Therefore, this application can more accurately and reliably assess the health status of the earless retainer, improving the accuracy and stability of monitoring, thus providing a more solid data foundation for subsequent state assessment and early warning.

[0038] In some of the embodiments described above in this application, a method is proposed to process the response signal to extract feature parameters for evaluating the health status of the ear ring-less device. However, in its implementation, there is a lack of comparison with benchmark data under the health status, resulting in insufficient accuracy and reliability of the health status assessment. It is also impossible to effectively quantify the deviation between real-time data and the ideal state, thereby affecting the accuracy of early warning.

[0039] In response, this application further proposes a step of constructing a baseline model before the above processing steps: when the ear-proof ring is in a known healthy state, acquire its baseline acoustic signals under various preset working conditions; process the baseline acoustic signals and extract the corresponding baseline feature parameter set; the above step of determining the current health status index includes: comparing and analyzing the at least one feature parameter extracted in real time with the parameters under the corresponding working conditions in the baseline feature parameter set, and calculating the current health status index based on its deviation.

[0040] The baseline modeling step aims to provide a reliable reference benchmark for assessing the health status of the earless retainer. This step is typically performed after the earless retainer's initial installation, major overhaul, or after rigorous quality testing confirms it is in optimal working condition. Baseline data can be collected before the equipment leaves the factory, or initially after non-destructive testing or performance testing confirms the earless retainer and its mounting location are in good condition before baseline data collection.

[0041] Acquiring baseline acoustic signals of the earless retainer under various preset operating conditions while it is in a known healthy state is to comprehensively capture its acoustic characteristics under different operating conditions. A known healthy state refers to the earless retainer's condition when it is structurally intact, functionally normal, and shows no signs of damage or degradation. The various preset operating conditions can cover a wide range of typical operating conditions that the earless retainer may encounter during actual service, including its operation at different speeds, loads, temperatures, or vibration frequencies. By collecting acoustic signals under these conditions, a benchmark database covering its normal operating range can be established. Low-speed, medium-speed, and high-speed operating conditions, as well as no-load, half-load, and full-load operating conditions can be set, and acoustic signals can be collected under each of these nine combined operating conditions.

[0042] Processing the baseline acoustic signal and extracting the corresponding baseline feature parameter set involves transforming the original acoustic signal into quantifiable and representative numerical features. This processing may include filtering, noise reduction, time-domain analysis, frequency-domain analysis, or time-frequency-domain analysis of the acoustic signal. Time-domain statistical features such as root mean square value, peak value, kurtosis, and skewness can be extracted from the acoustic signal, or frequency-domain features such as power spectral density and specific frequency band energy can be obtained through Fast Fourier Transform (FFT). Time-frequency features can also be extracted through wavelet transform or short-time Fourier transform. These extracted feature parameters constitute the baseline feature parameter set corresponding to different operating conditions under a healthy state without an earmuff.

[0043] Comparing the real-time extracted feature parameters with the parameters under the corresponding operating conditions in the baseline feature parameter set is a crucial step in assessing the current health status of the earless retainer. During real-time monitoring, the system retrieves the baseline parameter that best matches the current operating conditions from the pre-established baseline feature parameter set. Subsequently, the real-time collected and extracted feature parameters are compared with the matching baseline parameter. The comparative analysis methods may include calculating the absolute difference, relative difference, percentage deviation, or using more complex statistical distance metrics such as Euclidean distance or Mahalanobis distance.

[0044] The current health status index is calculated based on the degree of deviation, quantifying the results of comparative analysis into a comprehensive health indicator. A greater degree of deviation generally indicates a worse health condition for those without earmuffs. The calculation of the health status index can be a simple weighted sum, where the deviations of multiple feature parameters are multiplied by their respective weights and then summed. Alternatively, it can be the output of a machine learning model, taking the degree of deviation as input and outputting a health score from 0 to 100 through a pre-trained classifier or regressor.

[0045] Through the above technical solution, this application provides a reliable benchmark reference for online monitoring and early warning of the service status of earless retaining rings. Based on the acoustic signals collected by the monitoring module and the feature parameters extracted by the signal processing and analysis module, a baseline model is introduced, enabling the real-time extracted feature parameters to be accurately compared with baseline data of the earless retaining ring in a known healthy state. This comparative analysis effectively quantifies the deviation between the current state and the ideal healthy state, thus overcoming the problem of insufficient assessment accuracy due to the lack of a comparison standard. Especially in scenarios where the active excitation module applies an excitation signal and obtains a response signal, comparison with the baseline response signal characteristics can more sensitively capture minute changes in the structural characteristics of the earless retaining ring, such as preload relaxation or microcrack initiation. The current health status index calculated based on the degree of deviation can more accurately and reliably reflect the actual health status of the earless retaining ring, improving the accuracy of early warning, helping to promptly detect potential failure risks, and avoiding serious consequences caused by sudden failure of the earless retaining ring.

[0046] In some embodiments described above in this application, characteristic parameters are proposed to extract information associated with the interface contact state and / or structural integrity state of the ear ring-less device, in order to determine health status indicators. However, in its implementation, the specific types and ranges of the characteristic parameters are not explicitly defined, which may result in insufficiently comprehensive or representative extracted parameters, affecting the accuracy and reliability of the monitoring results.

[0047] In this regard, this application further proposes that the at least one characteristic parameter includes at least one of the following extracted from the passive acoustic emission signal: event count rate, signal amplitude distribution statistics, signal energy rate, and frequency domain characteristic parameters obtained by time-frequency transformation.

[0048] The event count rate refers to the number of acoustic emission events detected per unit time. This parameter directly reflects the activity level of transient elastic wave events generated by microscopic damage mechanisms such as fretting wear, microcrack initiation or propagation, etc., when the lugless retainer is under stress. A fixed signal amplitude threshold can be set, counting signal pulses exceeding this threshold as one event, and calculating the total number of events and their rate within a preset time window. Alternatively, an adaptive threshold method can be used to dynamically adjust the threshold based on the background noise level to more accurately capture valid acoustic emission events.

[0049] The signal amplitude distribution statistics refer to parameters obtained by statistically analyzing the amplitude of acoustic emission signals. These statistics, such as peak amplitude, root mean square (RMS) value, average amplitude, and skewness and kurtosis of the amplitude distribution, can characterize the intensity and energy of acoustic emission events, thereby reflecting the severity or evolution of damage. Peak amplitudes of all acoustic emission events over a period of time can be collected, and the mean and standard deviation of these peak amplitudes can be calculated. Alternatively, a probability density function of the acoustic emission signal amplitude can be constructed, and characteristic parameters, such as the proportion of events within a specific amplitude range, can be extracted from it.

[0050] The signal energy rate refers to the energy carried by the acoustic emission signal per unit time. This parameter quantifies the overall energy release level of the acoustic emission source and is closely related to the cumulative effects of structural changes, stress concentration, or material damage without an earpiece. The energy rate can be calculated by integrating the instantaneous power of the acoustic emission signal, accumulating it within a certain time window, and then dividing by the time window length. Alternatively, a more refined algorithm can be used to calculate the energy of each independent acoustic emission event, then accumulate and time-normalize the energies of these events.

[0051] The frequency domain characteristic parameters obtained through time-frequency transformation refer to the features extracted from the spectrum or time-frequency plot of the acoustic signal after time-frequency analysis. These parameters, such as dominant frequency, bandwidth, spectral centroid, and energy proportion of a specific frequency band, can reveal subtle changes in the internal structure, interface contact state, or damage mode of the earless retainer material, because different damage mechanisms often correspond to specific frequency responses. A short-time Fourier transform can be performed on the acquired acoustic signal to obtain its spectrum, and then the dominant frequency components or energy concentration within a specific frequency range can be extracted from the spectrum. Alternatively, wavelet transform can be used to decompose the signal into different frequency scales, and the statistical characteristics of the wavelet coefficients at each scale can be analyzed.

[0052] Through the above technical solution, this application solves the problems of incomplete and insufficient representativeness of feature parameter selection by specifically defining the types of key feature parameters extracted from passive acoustic emission signals, thereby improving the accuracy and reliability of health status monitoring. The event count rate directly reflects the activity level of fretting wear or microcrack initiation in the earless retainer under stress, helping to identify early damage trends. The signal amplitude distribution statistics reveal the severity of wear or structural damage by analyzing the statistical characteristics of the signal amplitude, providing a more objective quantitative basis. The signal energy rate measures the energy level of the acoustic signal, correlated with the intensity changes of structural changes or stress concentration, enhancing sensitivity to overall state changes. The frequency domain feature parameters obtained through time-frequency transformation capture the evolution of frequency components using time-frequency analysis methods, identifying hidden fault modes such as material fatigue or loosening, compensating for the lack of time domain information. The comprehensive application of these parameters ensures the comprehensiveness and specificity of feature extraction, making the health status assessment of the earless retainer more accurate and reliable. This multi-dimensional and multi-angle feature extraction method can more comprehensively capture various signs of degradation that may occur during the service of the earless retainer ring, thereby achieving early warning of potential faults, effectively avoiding misjudgment or missed judgment due to the limitations of a single feature parameter, and improving the robustness and diagnostic capability of the monitoring system.

[0053] In some of the embodiments described above in this application, a method is proposed to process the response signal to extract feature parameters for monitoring the structural characteristics of the earless retainer. However, in its implementation, the selection and extraction of feature parameters may lack specificity and fail to accurately capture subtle changes in structural characteristics, resulting in insufficient accuracy and reliability of the state assessment.

[0054] In this regard, this application further proposes to process the response signal to extract feature parameters, including extracting at least one of the following: the arrival time difference of the response signal, the signal attenuation coefficient, or the change in the frequency response function relative to the baseline state.

[0055] Processing the response signal to extract feature parameters refers to identifying and quantifying key information related to the characteristics of the earless ring structure from the acoustic signal collected by the monitoring module after a preset excitation signal is applied from the active excitation module to the structure carrying the earless ring. This is done using specific algorithms and techniques. Digital signal processing techniques, such as filtering, denoising, and framing, can be used for preprocessing, followed by feature extraction algorithms based on wavelet transform, Fourier transform, or empirical mode decomposition to extract signal features from the time or frequency domain. Alternatively, machine learning or deep learning models can be used, taking the original response signal as input, to train the model to automatically learn and extract potential features related to the health status of the earless ring.

[0056] The arrival time difference of the response signal is extracted as the difference between the time it takes for the response signal to arrive at the monitoring module after the excitation signal is emitted and the expected or baseline time. This difference can reflect changes in the propagation path of sound waves in the earless retainer and its supporting structure, such as changes in sound velocity or path elongation due to cracks, loosening, or material damage. The arrival time difference can be obtained by calculating the time delay between the excitation signal and the response signal using the cross-correlation function method and comparing it with the baseline time delay under healthy conditions. Alternatively, the arrival time of the initial or main energy of the signal can be determined by analyzing the peak value or specific threshold point of the signal envelope, and then the time difference can be calculated.

[0057] The signal attenuation coefficient measures the degree of energy loss during sound wave propagation in a shielded structure. Impaired structural integrity, such as material fatigue, microcrack propagation, or poor interfacial contact, leads to increased sound wave energy dissipation, thus exacerbating signal attenuation. The attenuation coefficient can be calculated by comparing the energy or amplitude of the excitation signal to that of the response signal. Alternatively, it can be obtained by calculating the root mean square value or peak amplitude ratio of the signal, or by analyzing the amplitude variation of the signal over different propagation distances and fitting an attenuation model.

[0058] Extracting the change in the frequency response function relative to the baseline state refers to the fact that the frequency response function (FRF) describes the system's response characteristics to excitation at different frequencies. When the structural characteristics of a system without lugs change, such as stiffness, mass, and damping, its dynamic characteristics, such as natural frequencies, mode shapes, and damping ratios, will also change, resulting in a shift in the FRF. These changes can be quantified by comparing it with the baseline FRF under healthy conditions. The FRF is obtained by performing a Fourier transform on the excitation and response signals and calculating their ratio. Then, the current FRF is compared point-by-point or region-by-region with the pre-established baseline FRF to calculate its amplitude or phase difference. Modal analysis techniques can also be used to extract modal parameters, such as natural frequencies and damping ratios, from the FRF and monitor the drift of these modal parameters relative to the baseline state.

[0059] Through the above technical solution, this application can specifically extract key parameters closely related to the structural characteristics of the earless retainer ring from the response signal generated by active excitation. By extracting the arrival time difference of the response signal, it can sensitively capture minute changes in the sound wave propagation path, thereby effectively indicating damage or loosening inside the structure. By extracting the signal attenuation coefficient, it can quantify the degree of sound wave energy dissipation, directly reflecting material loss or degradation of the interface contact state. By extracting the change in the frequency response function relative to the baseline state, it can accurately identify the shift in the dynamic characteristics of the earless retainer ring, thereby revealing the gradual changes in its stiffness, damping, and other structural parameters. The selection of these parameters is highly targeted and physically significant, overcoming the lack of accuracy in feature parameter extraction in traditional methods, improving the accuracy and reliability of monitoring the structural characteristics of the earless retainer ring, and providing a more solid data foundation for subsequent health status assessment.

[0060] In some of the embodiments described above in this application, a current health status index is proposed to assess the health status of patients without ear loops. In its implementation, multiple feature parameters may have different dimensions and importance. Without normalization and weight allocation, the index calculation will be inaccurate and multi-parameter information cannot be effectively integrated, thereby affecting the accuracy and reliability of the assessment.

[0061] In this regard, this application further proposes a step for determining the current health status indicator, which includes: normalizing the extracted multiple feature parameters respectively, assigning a weight coefficient to each normalized feature parameter based on its correlation with the degradation of the ear ring state, and merging the weighted multiple feature parameters to generate a comprehensive scalar value as the current health status indicator.

[0062] The extracted feature parameters are normalized to eliminate the influence of differences in units and numerical ranges between different feature parameters, ensuring the comparability of all parameters in subsequent calculations. Min-max normalization can be used to linearly scale each parameter value to a preset interval, such as [0, 1]. Alternatively, Z-score normalization can be used to convert each parameter value into a standard normal distribution with a mean of 0 and a standard deviation of 1, to adapt to the characteristics of different data distributions.

[0063] Based on this, a weight coefficient is assigned to each normalized feature parameter, determined by its correlation with the degradation of the lug retainer. The key to this step is identifying and quantifying the indicative ability of different feature parameters to the degradation of the lug retainer's health. This can be achieved by combining domain expert experience with historical failure data analysis to determine which parameters have a stronger correlation with early fatigue damage or preload relaxation, thus assigning them higher weights. Alternatively, statistical methods, such as correlation analysis or principal component analysis, can be used to quantify the correlation strength between each parameter and the degradation trend of the lug retainer, thereby automatically generating weight coefficients.

[0064] Subsequently, the weighted feature parameters are fused to generate a comprehensive scalar value as the current health status indicator. This fusion calculation integrates standardized and weighted feature information from different dimensions into a single numerical value, thus providing a comprehensive and easy-to-understand assessment result of the health status without an earmuff. Typically, the fusion calculation can use a weighted summation method, that is, the sum of the products of each normalized parameter and its corresponding weight coefficient, or it can use more complex fusion algorithms, such as fuzzy logic-based fusion, to handle nonlinear relationships or uncertainties between parameters.

[0065] Through the above technical solution, this application effectively solves the problem of inaccurate calculation of health status indicators caused by the different dimensions and importance of multiple characteristic parameters. Normalization eliminates the dimensional differences between parameters, enabling the comparison and fusion of characteristic information from different sources on a unified scale. The allocation of weighting coefficients highlights the characteristic parameters that are more sensitive and critical to the degradation of the ear-ring-less condition, thereby improving the sensitivity and accuracy of the health status indicators in responding to the actual degradation process. Finally, through fusion calculation, these processed parameters are integrated into a comprehensive scalar value, which not only simplifies the health status assessment process but also makes the assessment results more comprehensive, objective, and reliable. This enables more accurate identification of early degradation signs in the service status monitoring of ear-ring-less vehicles, providing a solid data foundation for timely early warning and predictive maintenance.

[0066] In some of the embodiments described above in this application, a method is proposed to generate early warning information based on current health status indicators and preset evaluation rules to monitor the status of earless rings. However, in the implementation process, the preset evaluation rules may use fixed thresholds, which cannot adapt to the historical changes and trends of earless rings in actual operation. This may cause the early warning information to be triggered too early or too late, resulting in false alarms or missed alarms, and affecting the accuracy and timeliness of the monitoring system.

[0067] In response, this application further proposes to generate early warning information based on current health status indicators and preset evaluation rules, including: dynamically adjusting the thresholds for triggering different levels of early warnings based on the historical operating data of the ear-blocking ring and the historical trend of the current health status indicators; comparing the current health status indicators with the dynamically adjusted thresholds, and triggering the corresponding level of early warning information if the conditions are met, with the early warning levels including at least performance degradation early warning and failure risk early warning.

[0068] When generating early warning information, the thresholds for triggering different levels of warnings are dynamically adjusted based on historical operating data of the earless ring and the historical trends of current health status indicators. This feature aims to address the problem that fixed thresholds cannot adapt to changes in the actual operating status of the earless ring. By introducing historical data and trend analysis, the warning thresholds can reflect the actual degradation degree and risk level of the earless ring in real time and adaptively, thereby improving the accuracy and timeliness of warnings. Machine learning algorithms, such as regression analysis or time series prediction models, can be used to model the historical trends of historical operating data and health status indicators. The model learns the changing patterns of health status indicators at different degradation stages of the earless ring and predicts the evolution of future health status indicators accordingly, thereby dynamically calculating warning thresholds that match the current health status and future trends. Alternatively, methods based on expert experience and statistical analysis can be used. These methods, such as moving averages, exponential smoothing, or Kalman filtering, can be used to analyze the historical trends of health status indicators, identify their rate of change, volatility, and other characteristics, and then dynamically adjust the warning thresholds in conjunction with a pre-set expert rule base.

[0069] Subsequently, the current health status indicators are compared with dynamically adjusted thresholds. This comparison is the direct basis for determining whether the absence of an earplug has met the warning conditions. By comparing the real-time acquired current health status indicators with dynamically adjusted thresholds that better reflect actual operating conditions, false alarms or missed alarms caused by inappropriate thresholds can be avoided, ensuring the effectiveness of the warning. If the health status indicator is a numerical value representing the degree of health, then a warning is triggered when the current health status indicator falls below a certain dynamically adjusted threshold. The comparison can also be based on interval judgment, setting a dynamically adjusted threshold interval for each warning level. When the current health status indicator falls into a specific interval, a warning for that interval is triggered.

[0070] If the conditions are met, a corresponding level of early warning information will be triggered. This feature is the execution part of the early warning mechanism, ensuring that when an anomaly is detected, an alert can be issued to the user or maintenance personnel in a timely and accurate manner so that appropriate intervention measures can be taken. The system can send SMS, email, or application push notifications to preset recipients. The early warning information should include the identification of the missing ear ring, current health status indicators, the triggered early warning level, and possible causes or suggestions. Early warning information can also be displayed through a visual interface, such as changing the color of the missing ear ring status indicator light on the monitoring system's dashboard, or popping up a warning window accompanied by an audible and visual alarm.

[0071] The warning levels include at least performance degradation warnings and failure risk warnings. This feature defines the hierarchy of warning information, enabling maintenance personnel to take different levels of response measures based on the severity of the warning. Performance degradation warnings typically indicate that the performance of the lugless retainer ring has begun to decline but has not yet reached a critical state, allowing for planned maintenance; while failure risk warnings indicate that the lugless retainer ring is approaching or has reached the failure threshold, requiring immediate emergency measures to avoid equipment failure. Different warning levels can be implemented by setting two or more dynamically adjustable thresholds, with a lower threshold triggering a performance degradation warning and a higher, more stringent threshold triggering a failure risk warning. The classification of warning levels can also be combined with the rate or trend of change of health status indicators. Even if the health status indicator has not yet reached the absolute threshold for failure risk, if it deteriorates rapidly in a short period of time, and its rate of change exceeds the preset critical value, a failure risk warning can be directly triggered to deal with sudden failures.

[0072] Through the above technical solution, this application effectively solves the problems of false alarms and missed alarms in the monitoring of the service status of earless retainers using traditional fixed threshold early warning methods. By dynamically adjusting the early warning threshold based on the historical operating data of the earless retainer and the historical trend of the current health status indicators, the system can fully consider the gradual performance degradation process of the earless retainer due to wear, fatigue, and other factors during actual operation, enabling the early warning threshold to adaptively reflect its true health status and potential risks. This dynamic adjustment mechanism avoids the limitations of fixed thresholds in adapting to equipment aging and changes in operating conditions, thus ensuring the accuracy of the early warning. Furthermore, by comparing the real-time acquired current health status indicators with these dynamically adjusted thresholds, it is possible to more accurately determine whether the earless retainer has met the early warning conditions. When conditions are met, corresponding warning information is triggered, especially distinguishing between performance degradation warnings and failure risk warnings. This allows maintenance personnel to take targeted maintenance strategies in a timely manner based on the severity of the warning. For example, planned maintenance can be carried out in the early stages of performance degradation, or emergency measures can be taken immediately when failure risk is imminent. This effectively avoids equipment damage and safety accidents caused by sudden failure of the lug retainer, improves the accuracy, timeliness and reliability of lug retainer service status monitoring, extends equipment service life and reduces maintenance costs.

[0073] In some of the solutions mentioned above in this application, the current health status indicators are proposed to assess the condition of the ear ring without ear protection. However, in the process of implementation, only real-time status assessment is considered, and the ability to predict historical trends and future degradation patterns is lacking. It is impossible to estimate the remaining service life or recommend maintenance time points, thus failing to achieve true predictive maintenance, resulting in delayed maintenance decisions and an inability to effectively prevent the risk of sudden failure.

[0074] In response, this application further proposes an online monitoring and early warning method for the service status of earless retainers, which includes a prediction step: inputting the sequence data of the current health status indicators formed over time and the current operating parameters into a trained prediction model; outputting predictive maintenance information of the earless retainers through the prediction model, wherein the predictive maintenance information includes an estimated remaining service life or a suggested maintenance time point; and the status assessment information or early warning information includes the predictive maintenance information.

[0075] The prediction step aims to predict the future service status and potential failure risks of the lug retainer by analyzing its historical health status and current operating conditions, thereby achieving a shift from passive monitoring to proactive predictive maintenance. This step can exist as a standalone computing module or be integrated into the signal processing and analysis module, responsible for executing the prediction algorithm. The sequence data of the current health status indicators over time, along with the current operating condition parameters, are input into the trained prediction model. Specifically, the sequence data of the current health status indicators over time refers to the set of health status indicators of the lug retainer acquired and recorded at different time points. This sequence data reflects the trend and pattern of the lug retainer's performance changes over time and is key to understanding its degradation process. It can be a historical database containing health indicator values ​​from the past few hours, days, or months, or a continuous data segment extracted through a sliding time window. The current operating condition parameters refer to the operating environment and load conditions of the lug retainer at the monitoring time. These parameters have a significant impact on the degradation rate and pattern of the lug retainer. They may include ambient temperature, bearing speed, applied load magnitude, vibration level, etc. These parameters can be acquired in real time through additional sensors or obtained from the equipment control system. The trained predictive model refers to a mathematical model that learns and optimizes from historical data, enabling it to identify deterioration patterns in the earless retainer and predict its future state. This model can be implemented using various techniques, including machine learning methods such as support vector regression, random forest regression, artificial neural networks (especially recurrent neural networks or long short-term memory networks, which are advantageous for time-series data), statistical methods such as autoregressive moving average models or Kalman filtering, or degradation models based on physical failure mechanisms. The model is trained to accurately capture the complex nonlinear relationship between health status indicators and operating parameters, and to predict future health status or directly predict remaining service life. The predictive model outputs predictive maintenance information for the earless retainer. This step refers to the predictive model, after receiving input data, performing internal calculations and inferences to generate an assessment result regarding the future state of the earless retainer. These results are organized into easily understandable and operable maintenance information. The predictive maintenance information includes estimated remaining service life (RUL) or recommended maintenance time points. The estimated remaining service life (RUL) refers to the length of time, from the current moment, the ear-less retainer ring is expected to continue operating normally under current conditions until it reaches a preset failure threshold. RUL can be calculated directly from a predictive model or by predicting a sequence of future health status indicators and combining this with the preset failure threshold. Recommended maintenance time points refer to future times recommended for preventative maintenance or replacement operations based on a combination of factors such as estimated remaining service life, historical maintenance experience, maintenance costs, and downtime. This can be a specific date or a time window.The status assessment information or early warning information includes the predictive maintenance information. This step aims to output the predictive maintenance information generated by the predictive model as part of the status assessment report or early warning notification. This means that when users receive an assessment of the current health status or an early warning of potential risks related to ear loops, they can also simultaneously receive guidance on future trends and maintenance recommendations, thus providing a more comprehensive and forward-looking basis for decision-making.

[0076] By introducing a predictive step, this application achieves a shift from passive condition assessment to proactive predictive maintenance, effectively addressing the shortcomings of existing solutions in predictive capabilities. Specifically, by inputting the time-varying sequence data of current health status indicators and current operating parameters into a trained predictive model, it can fully utilize historical degradation trends and current operating environment information, avoiding the one-sidedness of relying solely on instantaneous indicators. This allows the predictive model to more accurately capture the potential failure modes and degradation patterns of the lugless retainer ring. The predictive maintenance information output by the predictive model, including estimated remaining service life or recommended maintenance time points, directly provides a quantitative assessment of future risks and specific maintenance guidance, compensating for the deficiency of simple condition assessment in providing forward-looking decision-making basis. Furthermore, integrating predictive maintenance information into condition assessment or early warning information enhances the comprehensiveness and practicality of the output information, enabling maintenance personnel to obtain the current status, future trends, and corresponding maintenance recommendations of the lugless retainer ring in a timely manner, thereby optimizing maintenance plans, effectively preventing sudden failure risks, and improving the reliability and safety of the lugless retainer ring in service.

[0077] In mechanical transmission and equipment systems, lugless retaining rings, as core axial positioning and locking components, are widely used in rotary or reciprocating motion mechanisms in key fields such as aerospace, rail transportation, and precision machinery due to their compact structure and convenient installation. These components are usually concealed in space-constrained shaft grooves or bores, and are subjected to complex alternating loads and vibration impacts over long periods of time. Their service condition directly affects the reliability and safety of the entire transmission chain. The failure mode of lugless retaining rings is progressive and insidious, usually beginning with the relaxation of preload caused by interfacial fretting wear, or the initiation of micro-cracks within the material. If effective identification and early warning are not provided in the early stages of performance degradation, sudden breakage or detachment will trigger axial movement of the fixed components, leading to excessive equipment vibration, loss of transmission accuracy, and even potentially causing serious cascading failures such as bearing sintering and gear damage, resulting in significant safety risks and economic losses.

[0078] To address this, this application proposes an online monitoring and early warning system for the service status of earless retaining rings, used to implement the aforementioned method. The system includes a monitoring module comprising at least one acoustic sensor positioned near the earless retaining ring mounting location to collect acoustic signals; a signal processing and analysis module, communicatively connected to the monitoring module, configured to execute the aforementioned processing, determination, and generation steps; and an output module for outputting status assessment information or early warning information. The core innovation of this embodiment lies in the fact that by arranging acoustic sensors near the earless retaining ring mounting location to collect passive acoustic emission signals, and combining this with real-time analysis of characteristic parameters by the signal processing and analysis module, online monitoring of the interface contact state and structural integrity state of the earless retaining ring is achieved, enabling early identification of fretting wear and micro-cracks and preventing sudden failures.

[0079] In practical implementation, the acoustic sensors of the monitoring module are fixedly installed on the structural surface near the shaft groove or hole groove of the earless retainer ring to ensure effective capture of the passive acoustic emission signals generated by the earless retainer ring under stress. The acoustic sensors can be piezoelectric sensors, fiber optic sensors, or microelectromechanical system (MEMS) microphones, etc., and their placement must meet the sensitivity requirements for signal acquisition, typically not exceeding 50 mm from the earless retainer ring installation location to minimize environmental noise interference. The signal processing and analysis module is connected to the monitoring module via wired or wireless communication. First, the acquired acoustic signals undergo bandpass filtering and amplification preprocessing to eliminate low-frequency vibrations and high-frequency noise. Then, characteristic parameters related to the interface contact state and structural integrity are extracted, including the root mean square value, energy integral value, and dominant frequency components within a specific frequency band of the acoustic signal. Based on these characteristic parameters, a preset algorithm calculates the current health status index of the earless retainer ring, which quantifies the continuous change process from normal state to severe degradation in numerical form. Finally, according to preset evaluation rules, when the health status index exceeds a threshold, a corresponding level of warning information is generated. The output module outputs status assessment information or early warning information through a local display screen, audible and visual alarm device, or remote network interface, so that operators can take timely predictive maintenance measures.

[0080] Through the above technical solution, this application achieves real-time and accurate monitoring of the service status of earless retaining rings. It can identify fretting wear and the initiation of microcracks at the interface in the early stages of performance degradation, effectively avoiding equipment failures and safety risks caused by hidden failures, and improving the reliability and safety of the mechanical system. Compared with traditional maintenance methods that rely on periodic disassembly and inspection, this system does not require interruption of equipment operation, reducing maintenance costs. Furthermore, through the direct acquisition and analysis of passive acoustic emission signals, it ensures the objectivity and accuracy of the monitoring results, providing reliable technical support for predictive maintenance of critical equipment.

[0081] In some of the embodiments described above in this application, a signal processing and analysis module is proposed to execute the monitoring method. However, when implementing a monitoring method that requires active excitation, the system lacks the necessary active excitation module, which results in the inability to effectively collect response signals and limits the accuracy of state assessment.

[0082] In response, this application further proposes an online monitoring and early warning system for the service status of a ringless device, which also includes an active excitation module, communicatively connected to a signal processing and analysis module. The active excitation module is configured to apply a preset excitation signal to the structure carrying the ringless device according to the instructions of the signal processing and analysis module. The monitoring module is also used to collect acoustic signals including the response signal of the ringless device to the excitation signal.

[0083] The active excitation module is a device capable of applying controllable mechanical or acoustic excitation to a structure supporting a lugless retainer. Its function is to actively induce a measurable structural response in the lugless retainer, thereby obtaining structural characteristic information that is difficult to obtain through passive monitoring. The active excitation module can be implemented in various ways. It can be a piezoelectric exciter, generating mechanical vibration by applying an electrical signal and transferring energy to the lugless retainer and its supporting structure; it can be an electromagnetic exciter, driving a vibrating component through electromagnetic force to excite the structure; or it can be an ultrasonic transducer, emitting high-frequency ultrasonic pulses to detect the internal structure or interface state of the lugless retainer.

[0084] The active excitation module establishes a communication connection with the signal processing and analysis module to enable the transmission of data and control commands between them. This communication connection ensures that the signal processing and analysis module can precisely control and synchronize the active excitation module. The specific communication method can be a wired connection, such as data transmission and command exchange via Ethernet, USB, or RS-485 interface; or a wireless connection, such as communication via Wi-Fi, Bluetooth, or Zigbee wireless protocols.

[0085] The active excitation module is configured to apply a preset excitation signal to the structure carrying the earless retainer ring according to the instructions of the signal processing and analysis module. This means that the operation of the active excitation module is centrally controlled by the signal processing and analysis module, which can accurately select the type, frequency, amplitude, and duration of the excitation signal according to monitoring requirements. The signal processing and analysis module can send instructions to the active excitation module to apply a frequency sweep signal to detect the response characteristics of the earless retainer ring at different frequencies, or to apply a short-time pulse signal to analyze the propagation time or attenuation of sound waves in the structure. This controlled excitation method enables the system to acquire dynamic response data of the earless retainer ring under specific excitation.

[0086] The monitoring module is also used to acquire acoustic signals including the response signal of the earless retaining ring to the excitation signal. This means that the function of the monitoring module is expanded, enabling it not only to acquire the passive acoustic emission signal generated by the earless retaining ring under normal service conditions, but also to simultaneously acquire the response of the earless retaining ring and its supporting structure to these excitation signals when the active excitation module applies excitation. When the active excitation module applies a vibration excitation, the acoustic sensors in the monitoring module record the vibration response generated by the earless retaining ring. These response signals contain key information such as the structural integrity of the earless retaining ring, the preload state, or the interface contact condition.

[0087] Through the aforementioned technical solution, the system can actively apply a preset excitation signal to the structure supporting the earless retainer, thereby inducing a measurable response from the earless retainer. The monitoring module expands its acquisition range, capturing not only passive acoustic signals but also these response signals simultaneously. This combination of active excitation and response acquisition overcomes the limitation of insufficient information in certain situations when relying solely on passive acoustic signals, enabling the system to acquire richer and more controllable acoustic data. By analyzing the propagation characteristics, attenuation, or frequency changes of the response signal, the structural characteristics, preload state, or interface contact condition of the earless retainer can be more accurately assessed, especially in the early stages of damage or minor changes, improving the sensitivity and accuracy of condition assessment. Furthermore, active excitation allows for repeatable testing under specific operating conditions, facilitating the establishment of more reliable baseline models and trend analyses, thereby enhancing the comprehensiveness and reliability of the entire monitoring and early warning method.

[0088] The following example will provide a more detailed explanation of the above technical solution: The monitoring object was the earless retaining ring of a high-pressure turbine bearing for a certain type of aero-engine. The retaining ring was made of bearing steel and installed in a rectangular groove on the outer ring of the bearing. The monitoring hardware included two piezoelectric acoustic emission sensors with a resonant frequency of 150kHz and a piezoelectric ceramic exciter with a center frequency of 250kHz. The sensor installation positions were determined through simulation optimization: the first sensor was placed on the outer wall of the bearing housing, 15 mm axially from the center of the earless retaining ring mounting groove; the second sensor was at a 90-degree angle to the first sensor on the same circumferential surface, at the same distance. During installation, a high-temperature coupling agent was used to ensure good acoustic coupling between the sensor and the polished bearing housing surface, and the sensor was fixed with magnetic clamps. The exciter was bonded to the other side of the outer wall of the bearing housing, 20 mm from the center of the mounting groove, using high-temperature resistant epoxy resin adhesive.

[0089] Signal acquisition is conducted in two modes: passive and active. During engine operation, the system continuously acquires acoustic signals from two sensors at a sampling rate of 5MHz to capture the passive acoustic emission signals generated by the earless retainer ring under stress. In addition, the system automatically performs an active test every hour: the signal processing and analysis module controls the exciter to emit a 5-cycle, Hanning-window modulated sine wave pulse with a center frequency of 250kHz and a peak voltage of 80V. Simultaneously with the excitation, the two sensors are triggered to acquire the response signals within a subsequent 2-millisecond time window at a sampling rate of 10MHz.

[0090] Feature parameter extraction is based on a defined algorithm and parameters. For passive signals, a bandpass filter of 150-350kHz is first applied. The event count rate is calculated as follows: the event detection threshold is set to three times the moving average of 100 consecutive sampling points, and the number of events exceeding this threshold within each 10-second time window is counted, with the result expressed as events per second. The signal energy rate is calculated as the average of the sum of squares of the amplitudes of all sample points of the filtered signal within the same time window, in V². Furthermore, for each detected acoustic emission event waveform, 512 sampling points are taken before and after the trigger point for wavelet packet transform. The db4 mother wavelet is used for decomposition to the fourth level, and the proportion of wavelet coefficient energy at the third node to the total energy is extracted as a frequency domain feature parameter. For the response signal of active excitation, the signal envelope is extracted using Hilbert transform, and the time difference between the rising edges of the signal envelopes received by the two sensors reaching 50% of their maximum amplitude is calculated as the arrival time difference, with a measurement accuracy of 0.1 microseconds. Meanwhile, for the response signal of the first sensor, the waveform of the last 10 cycles after the main wave packet is extracted, and the exponential decay curve A(t)=A0·exp(-αt) is fitted. The resulting decay coefficient α is the characteristic parameter characterizing energy loss.

[0091] The baseline model was established when the lug retainer was in a known healthy condition, i.e., after an engine overhaul, the installation torque was confirmed to be 8.5 N·m and there was no visual damage. Multiple sets of acoustic signals were collected under three steady-state speeds of 5000, 9000, and 13000 rpm and two simulated load combinations. The average value and standard deviation of the five characteristic parameters were calculated for each condition, thereby constructing a baseline database storing the reference values ​​of the parameters and their normal fluctuation ranges under each condition.

[0092] The health status index is calculated through multi-feature fusion. In real-time monitoring, the system uses the baseline operating condition that best matches the current speed and load. The five extracted feature parameter values ​​are compared with the mean and standard deviation of this baseline operating condition, and normalized using the Z-score method: subtracting the baseline mean from the measured value and then dividing by the baseline standard deviation. Each normalized feature parameter is assigned a fixed weight based on its correlation with historical degradation data: event count rate weight 0.25, signal energy rate weight 0.20, wavelet energy percentage weight 0.15, arrival time difference weight 0.25, and attenuation coefficient weight 0.15. The final health status index HI is obtained by calculating the weighted Euclidean distance. This indicator is a dimensionless scalar value, and an increase in its value clearly indicates that the state deviates from the health benchmark.

[0093] The early warning and prediction functions operate dynamically based on this indicator. The system initially sets the performance degradation early warning threshold to HI>3.0 and the failure risk early warning threshold to HI>5.0. These thresholds are not fixed; every 100 hours of system operation, the two thresholds are dynamically adjusted based on the moving average and trend of the health status indicators over the past 100 hours, determined by the slope of a linear regression, with an adjustment range of ±0.5. Furthermore, to achieve predictive maintenance, the system includes a pre-trained prediction model. This model is a two-layer long short-term memory neural network, with 64 neurons in each layer. It takes the health status indicator sequence of the past 24 hours and current operating parameters as input, and outputs a predicted sequence of indicators for the next 24 hours. The model is trained using accelerated life test data from the same model of engine bench. When the predicted sequence indicates that an indicator will exceed the dynamic failure risk threshold within the next 8 hours, the system's output warning message will include specific maintenance recommendations for checking within the next available maintenance window.

[0094] Through the specific implementation methods described above, in bench tests, this system was able to trigger a performance degradation warning when the preload of the lug-less retaining ring decreased by 15% due to fretting wear, and to trigger a failure risk warning approximately 50 hours before fatigue crack propagation led to complete failure. This verifies the effectiveness, early warning capability, and reliability of the method of this invention. It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A method for online monitoring and early warning of service status of equipment without ear retainers, characterized in that, include: Acoustic signals are acquired by a monitoring module located at the installation location of the target earless retainer ring, wherein the acoustic signals include passive acoustic emission signals generated by the earless retainer ring under stress. The acoustic signal is processed to extract at least one feature parameter associated with the interface contact state and / or structural integrity state of the ear ring-less structure. Based on at least one of the feature parameters, determine the current health status indicators of the earless ring; Based on the current health status indicators and preset evaluation rules, generate and output status evaluation information or early warning information related to the earless ring.

2. The method for online monitoring and early warning of service status of earless retaining rings according to claim 1, characterized in that, The step of acquiring acoustic signals also includes: A preset excitation signal is applied to the structure carrying the earless retainer ring through the active excitation module; Acquire acoustic signals collected by the monitoring module, including the response signal of the earless retainer ring to the excitation signal; The processing steps include: processing the response signal to extract feature parameters associated with the characteristics of the earless retainer structure.

3. The method for online monitoring and early warning of service status of earless retaining rings according to claim 2, characterized in that, Prior to the processing steps, a step of building a baseline model is also included: When the ear-free retainer is in a known healthy state, its baseline acoustic signal under various preset working conditions is acquired; Process the baseline acoustic signal and extract the corresponding baseline feature parameter set; The step of determining the current health status index includes: comparing and analyzing the at least one feature parameter extracted in real time with the parameters under the corresponding working conditions in the baseline feature parameter set, and calculating the current health status index based on the degree of deviation.

4. The method for online monitoring and early warning of service status of a device without an ear retainer ring according to any one of claims 1-3, characterized in that, The at least one characteristic parameter includes at least one of the following extracted from the passive acoustic emission signal: event count rate, signal amplitude distribution statistics, signal energy rate, and frequency domain characteristic parameters obtained by time-frequency transformation.

5. The method for online monitoring and early warning of service status of earless retaining rings according to claim 2 or 3, characterized in that, The process of processing the response signal to extract feature parameters includes extracting at least one of the following: the time difference of arrival of the response signal, the signal attenuation coefficient, or the change in the frequency response function relative to the baseline state.

6. The method for online monitoring and early warning of service status of earless retaining rings according to claim 1, characterized in that, The steps for determining the current health status indicators include: The extracted feature parameters are normalized separately. Assign a weight coefficient to each normalized feature parameter based on its correlation with the degradation state of the earless retainer; The weighted feature parameters are fused and calculated to generate a comprehensive scalar value as the current health status indicator.

7. The method for online monitoring and early warning of service status of earless retaining rings according to claim 1, characterized in that, The generation of early warning information based on current health status indicators and preset assessment rules includes: Based on the historical operating data of the ear-proof ring and the historical trend of the current health status indicators, the thresholds for triggering different levels of warnings are dynamically adjusted. The current health status indicators are compared with dynamically adjusted thresholds. If the conditions are met, a warning message of the corresponding level is triggered. The warning levels include at least performance degradation warning and failure risk warning.

8. The method for online monitoring and early warning of service status of earless retaining rings according to claim 1, characterized in that, It also includes a prediction step: The sequence data of the current health status indicators over time and the current operating parameters are input into the trained prediction model. The predictive model outputs predictive maintenance information for the earless retainer ring, which includes an estimated remaining service life or a suggested maintenance time. The status assessment information or early warning information includes the predictive maintenance information.

9. An online monitoring and early warning system for the service status of vehicles without ear retainers, characterized in that, The system for implementing the method as described in any one of claims 1-8 comprises: The monitoring module includes at least one acoustic sensor arranged near the mounting location of the earless retainer ring for acquiring acoustic signals; The signal processing and analysis module is communicatively connected to the monitoring module and is configured to perform the processing step, the determination step, and the generation step as described in any one of claims 1-8; The output module is used to output the status assessment information or early warning information.

10. The online monitoring and early warning system for the service status of earless retaining rings according to claim 9, characterized in that, It also includes an active excitation module, which is communicatively connected to the signal processing and analysis module. The active excitation module is configured to apply a preset excitation signal to the structure carrying the earless retainer ring according to the instructions of the signal processing and analysis module. The monitoring module is also used to acquire acoustic signals including the response signal of the earless retainer ring to the excitation signal.

Images