Gear fault diagnosis method and system based on piezoelectric stress wave signal active sensing

Abstract

The invention discloses a gear fault diagnosis method and system based on piezoelectric stress wave signal active sensing, and relates to the technical field of fault diagnos.A piezoelectric sensing unit is arranged on a tested gear, meshing impact stress waves are captured in real time in a single-to-periodic mode, and a fault diagnosis result is obtained. A high-resolution time-angle combined waveform is obtained through high-pass-envelope-angle domain interpolation, and transient energy imaging is achieved; rapid wave crest counting is executed, emergency faults such as tooth missing and tooth breakage are found instantly, comprehensive criteria are constructed according to absolute peak values, relative deviation and adjacent tooth comparison three-dimensional features, tooth-by-tooth mapping is carried out, and anomalies are marked; and fusing the abnormal tooth ratio and the average severity to generate a damage index, continuously tracking the evolution trend of the abnormal tooth ratio, automatically distinguishing early warning and progressive degradation, completing full life cycle monitoring from micro-crack to peeling under a non-stop condition, outputting emergency, early warning and degradation three-level prompts, and realizing early warning. And the operation and maintenance safety and economy of high-reliability scenes such as wind power and rail transit are improved.

Description

Technical Field

[0001] This invention belongs to the field of fault diagnosis technology, specifically relating to a gear fault diagnosis method and system based on active sensing of piezoelectric stress wave signals. Background Technology

[0002] As a core component of power transmission, the safe operation of gears directly affects the efficiency of the entire production line and the safety of personnel.

[0003] Existing gear fault diagnosis technologies largely rely on vibration acceleration sensors to collect macroscopic vibration signals from the gearbox or bearing housing, or on acoustic emission technology to capture high-frequency elastic waves. These methods have several significant drawbacks. First, in terms of signal acquisition, the macroscopic vibration signals acquired by existing vibration monitoring methods have relatively low frequencies, making them highly susceptible to interference from vibrations of other components in the transmission chain and environmental noise. This results in a low signal-to-noise ratio and insufficient sensitivity to early, minor localized damage to gears. Often, the damage is only effectively identified when it has progressed to a certain extent and caused significant vibration modulation, thus missing the optimal early warning window. Second, regarding fault feature extraction and localization, even if existing methods detect anomalies, they often struggle to accurately correlate fault features with specific individual gear teeth. Traditional spectral analysis or time-frequency analysis can indicate a fault in the gear pair, but it cannot achieve precise identification at the tooth level. The inability to distinguish which gear tooth is malfunctioning hinders accurate maintenance and root cause analysis. Furthermore, existing diagnostic methods often rely on simplistic or black-box logic, such as single threshold alarms or complex intelligent models trained with extensive historical fault data. The former is prone to false alarms or missed alarms when operating conditions change, while the latter suffers from high generalization requirements and poor interpretability, and its application is limited for equipment without a large number of fault samples. Finally, in terms of condition assessment and trend prediction, existing technologies tend to focus on binary judgments of fault presence or absence, or provide overall statistical indicators such as kurtosis and RMS, lacking a comprehensive index that can integrate multi-dimensional features, intuitively quantify the severity of damage, and reflect its deterioration trend. This makes it difficult for operators to accurately assess the overall health and remaining life of gears, and lacks detailed data support when formulating maintenance strategies.

[0004] To address the aforementioned problems, this invention proposes a gear fault diagnosis method and system based on active sensing of piezoelectric stress wave signals. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a gear fault diagnosis method and system based on active sensing of piezoelectric stress wave signals, solving the problem of difficulty in high signal-to-noise ratio sensing, accurate positioning, and quantitative evaluation of early local damage to gears in existing technologies.

[0006] The objective of this invention can be achieved through the following technical solutions: A gear fault diagnosis method based on active sensing of piezoelectric stress wave signals, the method comprising: Step 1: Based on the piezoelectric module pre-installed on the target gear, the piezoelectric stress wave of the target gear during operation is extracted in real time with the single rotation time of the target gear as the monitoring period, and the time-related envelope waveform of the piezoelectric stress wave of the target gear during the monitoring period is constructed. Step 2: Extract the number of teeth of the target gear and perform peak detection on the piezoelectric stress wave envelope waveform. If the peak detection fails, an emergency fault alert will be output. If the peak detection passes, proceed to step three and perform a second analysis; Step 3: Based on the piezoelectric stress wave envelope waveform diagram within the monitoring period, perform peak mapping between gear teeth, construct a peak sequence combined with the gear tooth segment sequence, determine the comprehensive characteristics of each gear tooth in the target gear for secondary judgment, and mark abnormal gear teeth and normal gear teeth; Step 4: Based on the ratio of abnormal teeth to normal teeth and the comprehensive characteristics of abnormal faulty teeth, assess the damage index of the target gear during the monitoring period, conduct continuous monitoring, and output early warning fault reminders or deterioration fault reminders.

[0007] As a further embodiment of the present invention, the piezoelectric module in step one includes a piezoelectric sensing unit and a signal transmission unit, which are electrically connected. The piezoelectric sensing unit is fixedly installed on the target gear at a position preset by the operator. The piezoelectric stress wave collected by the piezoelectric sensing unit is transmitted as data via the signal transmission unit.

[0008] As a further aspect of the present invention, the specific method for constructing the piezoelectric stress wave envelope waveform diagram of the target gear in relation to time during the monitoring period in step one is as follows: A magnet is attached to the non-working surface of a single tooth of the target gear. A Hall effect sensor is used to monitor the tooth of the target gear. The time interval between two consecutive pulse signals output by the Hall effect sensor is recorded as the single rotation time of the target gear and marked as a monitoring period T_mc. The duration of the monitoring period T_mc is inversely proportional to the rotational speed of the target gear. Within the monitoring period T_mc, the piezoelectric stress wave of the target gear during operation is extracted in real time based on the piezoelectric module; High-pass filtering of the piezoelectric stress wave is performed at a cutoff frequency of 1Hz to eliminate temperature drift of the piezoelectric sensing unit. Then, a 6th-order Butterworth high-pass filter with a cutoff frequency of 1-3Hz is used to remove low-frequency macroscopic vibrations of the gears while retaining transient stress wave energy. After performing a Hilbert transform and taking the absolute value, a 20kHz low-pass filter is applied to obtain a pulse envelope signal that reflects the impact profile. Based on the single-rotation synchronous pulse provided by the Hall effect sensor, the pulse envelope signal is interpolated in the angular domain every 0.1°. A two-dimensional coordinate system is constructed with the time line as the horizontal axis and the interpolated angular domain values ​​as the vertical axis. The pulse envelope signal is plotted in the two-dimensional coordinate system to obtain the piezoelectric stress wave envelope waveform diagram reflecting the correlation between the target gear and time within the monitoring period T_mc, denoted as WED.

[0009] As a further aspect of the present invention, in step two, the specific method for extracting the number of teeth of the target gear and performing peak detection on the piezoelectric stress wave envelope waveform is as follows: Extract the number of teeth of the target gear, denoted as K; Using the single-rotation synchronous pulse output by the Hall effect sensor as the zero angle, the mechanical angle of the target gear from 0° to 360° corresponding to the monitoring period T_mc is divided into K tooth segments, which are recorded as tooth segment sequences TS1, TS2, ..., TSk according to the order of the division time. The angular width of each gear tooth section is defined as Δθ = 360° / K; Based on the piezoelectric stress wave envelope waveform diagram WED and the angle width Δθ of each tooth segment, the envelope waveform is segmented to obtain the envelope waveform segments associated with each of the K tooth segments, which are arranged as the envelope waveform segment sequence B1, B2, ..., BK. Among them, the envelope waveform segment B1 corresponds to the tooth segment of the first tooth, representing the complete engagement-disengagement process of the first tooth, and the rest are similar. Extract wave peaks from the envelope waveform segments of each gear tooth, and compare the total number of wave peaks F with the number of gear teeth K; If F≠K, the target gear is determined to have failed the peak detection, and an emergency fault alert is output to the operator.

[0010] As a further embodiment of the present invention, in step two, if F=K, it is determined that the target gear has passed the peak detection, and then proceeds to step three to perform secondary analysis.

[0011] As a further aspect of the present invention, the specific method for performing the secondary analysis in step three is as follows: Sort the K gear teeth according to the tooth segment sequence TS1, TS2, ..., TSk, and denote it as the tooth sequence L1, L2, ..., LK; Extract the envelope waveform segment sequence B1, B2, ..., BK corresponding to K tooth segments, and extract the peak values ​​corresponding to the peaks on the vertical axis of the two-dimensional coordinate system from the K envelope waveform segments, denoted as the peak sequence Z1, Z2, ..., ZK; Establish a mapping relationship (LN, ZN) between any peak ZN in the peak sequence Z1, Z2, ..., ZK and the corresponding tooth LN in the tooth sequence, where N is the counting index, and the value range is from 1 to K; Repeat the above steps to construct the mapping set {(L1,Z1),(L2,Z2),...,(LK,ZK)}; Calculate the global statistical characteristics of the peak sequence, including the peak mean μ_Z and the peak standard deviation σ_Z; Extract any mapping relationship (LN, ZN), and determine the absolute peak feature R_abs_N=ZN, the relative peak feature R_rel_N=(ZN-μ_Z) / σ_Z, and the adjacent tooth comparison feature R_nei_N=|ZN-(ZN-1+ZN+1) / 2| / ((ZN-1+ZN+1) / 2) of the associated gear tooth LN. Among them, the adjacent tooth comparison feature of the first and last gear teeth is calculated by cyclic adjacent method. The absolute peak feature R_abs_N, the relative peak feature R_rel_N, and the adjacent tooth comparison feature R_nei_N are combined to define the comprehensive feature CR_N={R_abs_N,R_rel_N,R_nei_N} of the gear tooth LN; Similarly, by determining the comprehensive features associated with each of the K gear teeth, we obtain the comprehensive feature sequence CR_1, CR_2, ..., CR_K; A secondary judgment is made based on the comprehensive characteristics of each tooth, and abnormal teeth are marked as normal teeth.

[0012] As a further aspect of the present invention, in step three, the specific method for performing a secondary judgment based on the comprehensive characteristics corresponding to each gear tooth and marking abnormal gear teeth as normal gear teeth is as follows: Obtain the comprehensive feature CR_N of any gear tooth LN, and extract the absolute peak feature R_abs_N, the relative peak feature R_rel_N, and the adjacent tooth comparison feature R_nei_N; Extract the operator's preset absolute threshold TH_abs, relative threshold TH_rel, and adjacent tooth comparison threshold TH_nei; The comprehensive feature CR_N of the gear tooth LN is compared with the corresponding threshold. If there exist R_abs_N≤TH_abs, R_rel_N≤TH_rel, and R_nei_N≤TH_nei, then mark tooth LN as a normal tooth; Conversely, it is marked as an abnormal tooth; Similarly, mark all abnormal teeth and normal teeth in the target gear.

[0013] As a further aspect of the present invention, the specific method for evaluating the damage index of the target gear during the monitoring period in step four is as follows: The total number of abnormal teeth in the target gear is counted and denoted as X. Then the total number of normal teeth is KX. The ratio Y of the number of abnormal teeth to the number of normal teeth is calculated using Y=X / (KX); For any tooth LN marked as an abnormal tooth, calculate the severity component of its comprehensive feature CR_N that exceeds a preset threshold, including: The absolute peak severity component S_abs_N = max(0, (R_abs_N - TH_abs) / TH_abs); The relative peak severity component S_rel_N = max(0, (|R_rel_N|-TH_rel) / TH_rel); The severity component of adjacent tooth contrast is S_nei_N=max(0,(R_nei_N-TH_nei) / TH_nei); The individual damage index S_N of the gear tooth LN is calculated using S_N=δ*S_abs_N+β*S_rel_N+γ*S_nei_N, where δ,β,γ are preset weighting coefficients, and δ+β+γ=1; Similarly, the mean of the individual damage index of all abnormal teeth is calculated as the average severity of the abnormal teeth, denoted as Avg_S. If the total number of abnormal teeth is 0, Avg_S is defined as 0. Based on the ratio Y and the average severity Avg_S, the damage index DI of the target gear within the monitoring period T_mc is calculated by DI=ω1×Y+ω2×Avg_S, where ω1 and ω2 are fusion weights, and ω1+ω2=1.

[0014] As a further aspect of the present invention, in step four, after determining the damage index DI of the target gear within the monitoring period T_mc, the damage index DI is compared with the preset damage index threshold DI_TH. If DI≥DI_TH, output a target gear early warning fault reminder; Conversely, continuous monitoring will be conducted. If the damage index DI of the target gear increases for Q consecutive monitoring cycles, a target gear deterioration fault alert will be output, where Q is a preset integer.

[0015] A gear fault diagnosis system based on active sensing of piezoelectric stress wave signals, the system comprising: The active envelope construction module, based on the piezoelectric module pre-deployed on the target gear, uses the single rotation time of the target gear as the monitoring period to extract the piezoelectric stress wave of the target gear in real time during operation and constructs the time-related piezoelectric stress wave envelope waveform of the target gear within the monitoring period. The peak warning and initial inspection module extracts the number of teeth of the target gear and performs peak detection on the piezoelectric stress wave envelope waveform. If the peak detection fails, an emergency fault reminder is output. If the peak detection passes, proceed to the multi-feature mapping diagnostic module for secondary analysis; The multi-feature mapping diagnostic module performs peak mapping between gear teeth based on the piezoelectric stress wave envelope waveform diagram within the monitoring period, constructs a peak sequence combined with a gear tooth segment sequence, determines the comprehensive features corresponding to each tooth in the target gear for secondary judgment, and marks abnormal teeth and normal teeth. The damage assessment module evaluates the damage index of the target gear within the monitoring period based on the ratio of abnormal to normal gear teeth and the comprehensive characteristics of abnormal faulty gear teeth, and performs continuous monitoring, outputting early warning fault alerts or deterioration fault alerts.

[0016] The beneficial effects of this invention are: This invention utilizes piezoelectric stress wave active sensing technology. A Hall effect sensor combined with a pre-built magnet determines the single-rotation time of a gear and constructs a piezoelectric stress wave envelope waveform in real time for continuous monitoring. The advantage lies in the rapid identification of emergency faults through peak detection to ensure immediate response. Combined with secondary analysis, peak mapping and comprehensive feature evaluation between gear teeth are performed to accurately mark abnormal gear teeth and quantify the damage index. This invention can comprehensively cover the diagnostic needs from sudden failures to gradual degradation, support early warning and trend analysis, thereby promoting preventive maintenance and reducing the risk of unexpected downtime. This invention achieves high-precision synchronous monitoring of gear operating status through the collaborative design of piezoelectric modules and Hall effect sensors. Multi-stage filtering effectively eliminates temperature drift and low-frequency vibration interference, thus preserving the key stress wave energy reflecting transient gear impacts. Furthermore, Hilbert transform is used to extract the envelope signal, and angle domain interpolation is combined to construct a high-resolution waveform, clearly revealing the gear's microscopic defects and impact characteristics. This improves the signal-to-noise ratio and monitoring sensitivity, enhances resistance to environmental interference, and provides a basis for accurate diagnosis and preventative maintenance of early gear faults. This invention uses Hall effect sensor synchronous pulses to precisely divide the gear rotation angle into segments corresponding to the number of teeth, achieving accurate correlation between the piezoelectric stress wave envelope waveform and the meshing process of each tooth, thus ensuring high accuracy and reliability of detection. It can monitor the health status of gears in real time and automatically. By quickly comparing the number of peaks with the number of teeth, it can immediately identify abnormalities such as broken teeth or severe wear and trigger emergency fault alerts, effectively preventing equipment damage and safety accidents. At the same time, it seamlessly transfers to secondary analysis after detection, further improving the comprehensiveness of fault diagnosis and maintenance efficiency. This invention constructs a mapping relationship between gear teeth and envelope peak values, and integrates three types of features—absolute, relative, and adjacent tooth comparison—to form a comprehensive criterion, achieving a multi-dimensional and accurate assessment of gear condition. Its advantages lie in comprehensively considering both overall statistical regularities and local adjacent tooth correlations, effectively avoiding the limitations of a single threshold criterion, and improving the robustness and accuracy of anomaly detection. It can not only sensitively identify isolated defects but also discover overall degradation trends based on relative features, thus enabling more reliable and comprehensive gear health status diagnosis and early warning in engineering practice. This invention achieves multi-dimensional and precise quantification of gear damage by comprehensively considering the proportion of abnormal teeth and the severity of each individual tooth. It not only reflects the breadth of the damage but also deeply characterizes the severity of the damage by integrating multiple features such as absolute peak value, relative peak value, and comparison with adjacent teeth, making the assessment results more comprehensive and objective. The final early warning strategy is both sensitive and reliable. It sets thresholds for immediate warning and continuously monitors the increasing trend of the damage index to identify potential deterioration processes in the early stage. This enables full-cycle health management from sudden fault alarms to gradual deterioration warnings, improving the predictability of condition monitoring and the timeliness of maintenance decisions. Attached Figure Description

[0017] The invention will now be further described with reference to the accompanying drawings.

[0018] Figure 1 This is a flowchart illustrating the method described in this invention. Detailed Implementation

[0019] 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.

[0020] like Figure 1 As shown, this application provides a gear fault diagnosis method and system based on active sensing of piezoelectric stress wave signals; As an embodiment 1 of this application, it specifically includes: Step 1: Based on the piezoelectric module pre-installed on the target gear, the piezoelectric stress wave of the target gear during operation is extracted in real time with the single rotation time of the target gear as the monitoring period, and the time-related envelope waveform of the piezoelectric stress wave of the target gear during the monitoring period is constructed. Step 2: Extract the number of teeth of the target gear and perform peak detection on the piezoelectric stress wave envelope waveform. If the peak detection fails, an emergency fault alert will be output. If the peak detection passes, proceed to step three and perform a second analysis; Step 3: Based on the piezoelectric stress wave envelope waveform diagram within the monitoring period, perform peak mapping between gear teeth, construct a peak sequence combined with the gear tooth segment sequence, determine the comprehensive characteristics of each gear tooth in the target gear for secondary judgment, and mark abnormal gear teeth and normal gear teeth; Step 4: Based on the ratio of abnormal teeth to normal teeth and the comprehensive characteristics of abnormal faulty teeth, assess the damage index of the target gear during the monitoring period, conduct continuous monitoring, and output early warning fault reminders or deterioration fault reminders.

[0021] Example 2 This embodiment further discloses the detailed steps involved in the active envelope construction module based on Embodiment 1, including the following: This module, the active envelope construction module, is based on a piezoelectric module pre-deployed on the target gear. Using the single-rotation time of the target gear as the monitoring period, it extracts the piezoelectric stress wave of the target gear in real time during operation and constructs a time-correlated envelope waveform diagram of the piezoelectric stress wave of the target gear within the monitoring period. The specific implementation is as follows: First, the operator needs to identify the target gear. It should be noted that the target gear selected in this solution refers to large gears used in fields such as wind power. During normal operation, the gear speed needs to be limited to a certain range. Therefore, high-speed gears such as those the size of engines are not covered in this solution.

[0022] The target gear mainly acquires data based on a piezoelectric module, which includes a piezoelectric sensing unit and a signal transmission unit. The piezoelectric sensing unit and the signal transmission unit transmit data in an electrically connected manner. The piezoelectric sensing unit and the signal transmission unit form a set of signal reception and signal excitation relationships. The piezoelectric sensing unit is mainly used to collect piezoelectric stress wave signals (i.e., the function of signal excitation), while the signal transmission unit is used to receive piezoelectric stress wave signals and transmit them to the terminal for subsequent processing (the function of signal reception).

[0023] The piezoelectric sensing unit is pre-fixed on the target gear at a position predetermined by the operator. After the piezoelectric sensing unit collects the piezoelectric stress wave signal, it performs data transmission through the signal transmission unit for subsequent data processing.

[0024] The target gear also needs to integrate a Hall effect sensor for collaborative data acquisition. A magnet prepared in advance by the operator needs to be attached to the non-working surface of a single tooth of the target gear (or a tooth specified by the operator). Then, the Hall effect sensor is used to monitor the tooth of the target gear in operation. Thus, the Hall effect sensor and the pre-prepared magnet form a signal receiver (the function of receiving signals) and a signal transmitter (the function of generating signals). That is, the magnet is essentially a signal transmitter, and the Hall effect sensor is a signal receiver. Specifically: The time interval between two consecutive pulse signals output by the Hall effect sensor is recorded as the single rotation time of the target gear, that is, the time it takes for the target gear to rotate one revolution.

[0025] The duration of a single rotation of the target gear is marked as a monitoring period T_mc. The duration of the monitoring period T_mc is inversely proportional to the rotational speed of the target gear. The faster the rotational speed of the target gear, the shorter the duration of the monitoring period T_mc. Meanwhile, the rotational speed (instantaneous speed) of the target gear is determined in real time by the feedback signal of the variable frequency motor driver of the target gear.

[0026] Next, within the monitoring period T_mc, the piezoelectric stress wave of the target gear during operation is extracted in real time based on the piezoelectric module; Then, a high-pass filter is applied to the piezoelectric stress wave at a cutoff frequency of 1Hz to eliminate the temperature drift of the piezoelectric sensing unit. A 6th-order Butterworth high-pass filter is then used at a cutoff frequency of 1-3Hz to remove the low-frequency macroscopic vibration of the gear while retaining the energy of the transient stress wave. Finally, after performing a Hilbert transform and taking the absolute value, a 20kHz low-pass filter is applied to obtain the pulse envelope signal that reflects the impact profile. Based on the single-rotation synchronization pulse provided by the Hall effect sensor, the pulse envelope signal is interpolated in the angle domain every 0.1°. The interpolated angle domain values ​​and the corresponding time lines are plotted in a two-dimensional coordinate system with the time line as the horizontal axis and the interpolated angle domain values ​​as the vertical axis. This yields the piezoelectric stress wave envelope waveform diagram of the target gear in the monitoring period T_mc, which is related to time and the rotation angle of the target gear itself. This waveform is denoted as WED.

[0027] Example 3 This embodiment further discloses the detailed steps involved in the peak warning preliminary detection module based on embodiment 2, specifically including the following: This module, the Peak Warning Preliminary Inspection Module, primarily aims to extract the number of teeth on the target gear and perform peak detection on the piezoelectric stress wave envelope waveform. If the peak detection fails, an emergency fault alert is output. If the peak detection passes, relevant data is transmitted to the Multi-Feature Mapping Diagnostic Module for secondary analysis. It should be noted that the physical principle used in this embodiment is based on the characteristic that each tooth of a normal gear generates a typical stress wave pulse during normal meshing. By detecting whether the number of peaks in the stress wave envelope waveform matches the number of teeth, it is possible to quickly determine whether the gear has experienced a serious structural fault that could lead to a major accident. The specific details are as follows: First, extract the number of teeth of the target gear and denote it as K; Next, using a Hall effect sensor, when a specific tooth on the gear (the tooth with the magnet installed) rotates, a pulse signal is output. The position corresponding to this pulse signal is defined as the 0° mechanical angle of the gear rotation, which is used as the detection starting point. Then, when the specific tooth (the tooth with the magnet installed) rotates again, another pulse signal is output (corresponding to the 360° mechanical angle). At this time, the two pulse signals and the interval between them represent a monitoring period T_mc. The mechanical angle from 0° to 360° of the target gear corresponding to the monitoring period T_mc is divided into K tooth segments, and they are recorded as tooth segment sequences TS1, TS2, ..., TSk according to the time sequence of the division. The angular width of each tooth segment is defined as Δθ = 360° / K. Based on the predefined K angle segments, the piezoelectric stress wave envelope waveform WED collected within the same monitoring period T_mc is also synchronously divided into K segments, resulting in the envelope waveform segment sequence B1, B2, ..., BK. Among them, the envelope waveform segment B1 corresponds to the tooth segment TS1 of the first gear tooth, representing the complete meshing-out process of the first gear tooth. The rest are treated similarly to ensure the accurate mapping between the physical position of the gear and the collected waveform data.

[0028] Next, in each envelope waveform segment, the peaks are extracted. Under normal circumstances, a healthy tooth will generate a significant stress peak during meshing. The total number of peaks detected in all envelope waveform segments is counted and recorded as F. The total number of peaks F is compared with the number of teeth K. If F=K, it means that a valid stress wave peak value was detected in each tooth segment, indicating that all teeth participated in meshing basically completely and no serious fault that would cause the signal to be completely lost was detected. The target gear passed the peak detection. If F≠K, the target gear is determined to have failed the peak detection, and an emergency fault alert is output to the operator. When F≠K, there are mainly two cases, with F<K being the most common. This means that no wave crest was detected in a tooth section, indicating that the corresponding tooth may be completely broken, severely damaged, or have significant spalling, causing it to fail to generate normal stress impact during meshing. If F > K, it means that multiple abnormal peaks have appeared in a single tooth section, which may be caused by severe impact, foreign object intervention or other abnormal impact events. Therefore, if F≠K, the machine needs to be stopped immediately for inspection to prevent the fault from escalating.

[0029] Example 4 This embodiment further discloses the detailed steps involved in the multi-feature mapping diagnostic module based on embodiment 3, specifically including the following: This module, the multi-feature mapping diagnostic module, primarily performs secondary analysis. Based on the piezoelectric stress wave envelope waveform within the monitoring period, it performs peak mapping between gear teeth, constructs a peak sequence combined with a gear tooth segment sequence, determines the comprehensive characteristics corresponding to each tooth in the target gear, performs secondary judgment, and marks abnormal and normal gear teeth, as detailed below: First, define the tooth sequence. Sort the K teeth according to the marked tooth segment sequence TS1, TS2, ..., TSk to obtain the tooth sequence L1, L2, ..., LK. Next, extract the envelope waveform segment sequence B1, B2, ..., BK corresponding to K tooth segments, and extract the peak values ​​corresponding to the peaks on the vertical axis of the two-dimensional coordinate system from the K envelope waveform segments, and record them as the peak sequence Z1, Z2, ..., ZK in chronological order; Any gear tooth LN is mapped to the peak stress wave ZN generated within the monitoring period T_mc, forming a mapping relationship (LN, ZN), where N is the counting index, with a value range from 1 to K; Perform the same operation on K gear teeth to obtain K mapping relationships corresponding to K gear teeth. Combine them to obtain the mapping set {(L1,Z1),(L2,Z2),...,(LK,ZK)}. Calculate the mean of the K peak values ​​corresponding to the K gear teeth, and denote it as the peak mean μ_Z; Next, the standard deviation of the K peak values ​​corresponding to the K gear teeth is calculated from the peak mean μ_Z, and denoted as the peak standard deviation σ_Z; The mean peak value μ_Z and the standard deviation of peak value σ_Z are the global statistical characteristics of the peak sequence Z1, Z2, ..., ZK, representing the average level and dispersion of the overall meshing impact intensity of the target gear under the current working condition, providing a benchmark for subsequent relative comparisons.

[0030] Then, extract the mapping relationship (LN,ZN) corresponding to any tooth LN from the mapping set {(L1,Z1),(L2,Z2),...,(LK,ZK)}, extract the peak value ZN of tooth LN, and determine the absolute peak feature R_abs_N=ZN, the relative peak feature R_rel_N=(ZN-μ_Z) / σ_Z, and the adjacent tooth comparison feature R_nei_N=|ZN-(ZN-1+ZN+1) / 2| / ((ZN-1+ZN+1) / 2). Among them, the adjacent tooth comparison feature of the first and last teeth is calculated by cyclic adjacent method.

[0031] Thus, the absolute peak feature R_abs_N, relative peak feature R_rel_N, and adjacent tooth comparison feature R_nei_N of the gear tooth LN were obtained; The absolute peak characteristic R_abs_N refers to the original peak value ZN of the gear tooth LN, which reflects the absolute intensity of the stress impact generated by the gear tooth LN during meshing. If the intensity is too low, it may indicate severe wear of the tooth surface, material fatigue, or the presence of hidden cracks; if it is abnormally high, it may indicate abnormal impact caused by local hard spots, foreign object intervention, or severe misalignment.

[0032] The relative peak characteristic R_rel_N is the deviation of the peak value of the gear tooth LN from the overall average value, normalized by the standard deviation. It is used to measure the degree to which the impact strength of the gear tooth LN deviates significantly from the average level of the gear as a whole. For example, the absolute value of the gear tooth LN decreases due to the overall load decrease, but it becomes weaker relative to other teeth, which suggests that the tooth degrades faster.

[0033] The adjacent tooth comparison feature R_nei_N is the relative deviation between the peak value of tooth LN and the average peak value of its two adjacent teeth. It is used to measure the degree of abnormality of the impact intensity of tooth LN in its directly adjacent local microenvironment. During continuous meshing, the instantaneous working conditions of several adjacent teeth are the most similar. Therefore, their peak values ​​should be highly correlated. If the peak value of a tooth deviates significantly from the average level of its left and right neighbors, it indicates that there is a very local abnormality in the tooth itself, such as pitting, small-scale peeling, micro-cracks, etc., which makes it easier for operators to detect faults in advance.

[0034] The absolute peak feature R_abs_N, the relative peak feature R_rel_N, and the adjacent tooth comparison feature R_nei_N are combined into the comprehensive feature CR_N={R_abs_N,R_rel_N,R_nei_N} of the gear tooth LN.

[0035] Repeat the above steps to determine the comprehensive features associated with each of the K gear teeth, resulting in the comprehensive feature sequence CR_1, CR_2, ..., CR_K; Finally, obtain the comprehensive feature CR_N of any gear tooth LN, and extract the absolute peak feature R_abs_N, the relative peak feature R_rel_N, and the adjacent tooth comparison feature R_nei_N; Extract the operator's preset absolute threshold TH_abs, relative threshold TH_rel, and adjacent tooth comparison threshold TH_nei, and compare the comprehensive feature CR_N of the tooth LN with the corresponding threshold; That is, comparing the absolute peak feature R_abs_N with the absolute threshold TH_abs, comparing the relative peak feature R_rel_N with the relative peak feature R_rel_N, and comparing the adjacent tooth comparison feature R_nei_N with the adjacent tooth comparison threshold TH_nei. If there exists R_abs_N≤TH_abs, R_rel_N≤TH_rel, and R_nei_N≤TH_nei, that is, when the corresponding tooth LN performs normally in all three dimensions, then the tooth LN is marked as a normal tooth; otherwise, the tooth LN is marked as an abnormal tooth. By repeating the above steps, you can mark all abnormal teeth and normal teeth in the target gear.

[0036] Example 5 This embodiment further discloses the detailed steps involved in the damage status assessment module based on embodiment 4, specifically including the following: This module, the Damage Status Assessment module, primarily aims to assess the damage index of the target gear within the monitoring period by analyzing the ratio of abnormal to normal teeth and the comprehensive characteristics of abnormal faulty teeth. It then continuously monitors the gear and outputs early warning or deterioration fault alerts, as detailed below: First, based on the method described in Example 4, all abnormal teeth and normal gears in the target gear are determined. It should be clarified that abnormal teeth are not faulty teeth, but teeth that produce abnormal data due to factors such as aging, and are not equivalent to faulty teeth. The total number of abnormal teeth in the target gear is counted and denoted as X. Then the total number of normal teeth is KX. Next, the ratio Y of the number of abnormal teeth to the number of normal teeth can be calculated by using Y=X / (KX); For any tooth LN in the target gear that is marked as an abnormal tooth, calculate the severity component of its comprehensive feature CR_N that exceeds a preset threshold, including the absolute peak severity component, the relative peak severity component, and the adjacent tooth comparative severity component, specifically defined as: The absolute peak severity component S_abs_N = max(0, (R_abs_N - TH_abs) / TH_abs) represents the percentage of absolute peak severity exceeding the limit; The relative peak severity component S_rel_N=max(0,(|R_rel_N|-TH_rel) / TH_rel) represents the percentage of relative peak exceeding the limit; The severity component of adjacent tooth contrast, S_nei_N=max(0,(R_nei_N-TH_nei) / TH_nei), represents the percentage of adjacent tooth contrast exceeding the standard. Since the absolute peak severity component S_abs_N, the relative peak severity component S_rel_N, and the adjacent tooth contrast severity component S_nei_N are all percentages and dimensionless, they can be calculated directly. The individual damage index S_N of gear LN is calculated by using S_N=δ*S_abs_N+β*S_rel_N+γ*S_nei_N, where δ,β,γ are preset weighting coefficients, and δ+β+γ=1; Next, by repeating the above steps, the mean of the individual damage index of all abnormal teeth is calculated and used as the average severity of the abnormal teeth, denoted as Avg_S. If the total number of abnormal teeth is 0, then Avg_S=0 is defined. Based on the ratio Y and the average severity Avg_S, the damage index DI of the target gear in the monitoring period T_mc is calculated by using DI=ω1×Y+ω2×Avg_S, where ω1 and ω2 are the fusion weights, ω1+ω2=1, and the damage index DI is a continuous value between 0 and positive infinity, which represents a quantitative score of the overall health status of the target gear in this monitoring period T_mc. DI=0 represents the ideal state, with no abnormal gear teeth. The damage index DI is compared with the preset damage index threshold DI_TH. The damage index threshold DI_TH is a preset high-risk threshold. Once DI reaches or exceeds this value, it means that the overall damage of the gear has exceeded the safety boundary, and there is a high risk of functional failure or rapid deterioration. Maintenance personnel must immediately arrange shutdown inspection or intervention. That is, if DI≥DI_TH, the target gear early warning fault reminder is directly output. Conversely, it indicates that the abnormal teeth of the target gear are within a controllable range, requiring only continuous monitoring and no replacement or repair is needed; In subsequent continuous monitoring, if the damage index DI of the target gear increases within Q consecutive monitoring cycles (if the damage index DI≥DI_TH, then the target gear early warning fault reminder will be output first), then the target gear deterioration fault reminder will be output, where Q is an integer preset by the operator.

[0037] All data in the formulas described above have been calculated with dimensions removed. Furthermore, any content not described in detail in this specification is existing technology known to those skilled in the art.

[0038] The above description is merely an example and illustration of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the invention or exceed the scope defined in the claims, all of which should fall within the protection scope of the present invention.

[0039] It should be stated that all user data collected in this application was collected with the user's consent and authorization. Furthermore, the uses of user data are legal and compliant, and the use and processing of user data comply with the relevant laws, regulations, and standards of the relevant regions.

Claims

1. A gear fault diagnosis method based on active sensing of piezoelectric stress wave signals, characterized in that, The method includes: Step 1: Based on the piezoelectric module pre-installed on the target gear, the piezoelectric stress wave of the target gear during operation is extracted in real time with the single rotation time of the target gear as the monitoring period, and the time-related envelope waveform of the piezoelectric stress wave of the target gear during the monitoring period is constructed. Step 2: Extract the number of teeth of the target gear and perform peak detection on the piezoelectric stress wave envelope waveform. If the peak detection fails, an emergency fault alert will be output. If the peak detection passes, proceed to step three and perform a second analysis; Step 3: Based on the piezoelectric stress wave envelope waveform diagram within the monitoring period, perform peak mapping between gear teeth, construct a peak sequence combined with the gear tooth segment sequence, determine the comprehensive characteristics of each gear tooth in the target gear, perform secondary judgment, and mark abnormal gear teeth and normal gear teeth; Step 4: Based on the ratio of abnormal teeth to normal teeth and the comprehensive characteristics of abnormal faulty teeth, assess the damage index of the target gear during the monitoring period, conduct continuous monitoring, and output early warning fault reminders or deterioration fault reminders.

2. The method according to claim 1, characterized in that, The piezoelectric module in step one includes a piezoelectric sensing unit and a signal transmission unit, which are electrically connected. The piezoelectric sensing unit is fixedly installed on the target gear at a position preset by the operator. The piezoelectric stress wave collected by the piezoelectric sensing unit is transmitted as data via the signal transmission unit.

3. The method according to claim 1, characterized in that, In step one, the specific method for constructing the time-correlated piezoelectric stress wave envelope waveform of the target gear within the monitoring period is as follows: A magnet is attached to the non-working surface of a single tooth of the target gear. A Hall effect sensor is used to monitor the tooth of the target gear. The time interval between two consecutive pulse signals output by the Hall effect sensor is recorded as the single rotation time of the target gear and marked as a monitoring period T_mc. The duration of the monitoring period T_mc is inversely proportional to the rotational speed of the target gear. Within the monitoring period T_mc, the piezoelectric stress wave of the target gear during operation is extracted in real time based on the piezoelectric module; High-pass filtering of the piezoelectric stress wave is performed at a cutoff frequency of 1Hz to eliminate temperature drift of the piezoelectric sensing unit. Then, a 6th-order Butterworth high-pass filter with a cutoff frequency of 1-3Hz is used to remove low-frequency macroscopic vibrations of the gears while retaining transient stress wave energy. After performing a Hilbert transform and taking the absolute value, a 20kHz low-pass filter is applied to obtain a pulse envelope signal that reflects the impact profile. Based on the single-rotation synchronous pulse provided by the Hall effect sensor, the pulse envelope signal is interpolated in the angular domain every 0.1°. A two-dimensional coordinate system is constructed with the time line as the horizontal axis and the interpolated angular domain values ​​as the vertical axis. The pulse envelope signal is plotted in the two-dimensional coordinate system to obtain the piezoelectric stress wave envelope waveform diagram reflecting the correlation between the target gear and time within the monitoring period T_mc, denoted as WED.

4. The method according to claim 3, characterized in that, In step two, the specific method for extracting the number of teeth of the target gear and detecting the peak of the piezoelectric stress wave envelope waveform is as follows: Extract the number of teeth of the target gear, denoted as K; Using the single-rotation synchronous pulse output by the Hall effect sensor as the zero angle, the mechanical angle of the target gear from 0° to 360° corresponding to the monitoring period T_mc is divided into K tooth segments, which are recorded as tooth segment sequences TS1, TS2, ..., TSk according to the order of the division time. The angular width of each gear tooth section is defined as Δθ = 360° / K; Based on the piezoelectric stress wave envelope waveform diagram WED and the angle width Δθ of each tooth segment, the envelope waveform is segmented to obtain the envelope waveform segments associated with each of the K tooth segments, which are arranged as the envelope waveform segment sequence B1, B2, ..., BK. Among them, the envelope waveform segment B1 corresponds to the tooth segment of the first tooth, representing the complete engagement-disengagement process of the first tooth, and the rest are similar. Extract wave peaks from the envelope waveform segments of each gear tooth, and compare the total number of wave peaks F with the number of gear teeth K; If F≠K, the target gear is determined to have failed the peak detection, and an emergency fault alert is issued to the operator.

5. The method according to claim 4, characterized in that, In step two, if F=K, the target gear is determined to have passed the peak detection and proceeds to step three to perform secondary analysis.

6. The method according to claim 5, characterized in that, In step three, the specific method for performing the secondary analysis is as follows: Sort the K gear teeth according to the tooth segment sequence TS1, TS2, ..., TSk, and denote it as the tooth sequence L1, L2, ..., LK; Extract the envelope waveform segment sequence B1, B2, ..., BK corresponding to K tooth segments, and extract the peak values ​​corresponding to the peaks on the vertical axis of the two-dimensional coordinate system from the K envelope waveform segments, denoted as the peak sequence Z1, Z2, ..., ZK; Establish a mapping relationship (LN, ZN) between any peak ZN in the peak sequence Z1, Z2, ..., ZK and the corresponding tooth LN in the tooth sequence, where N is the counting index, and the value range is from 1 to K; Repeat the above steps to construct the mapping set {(L1,Z1),(L2,Z2),...,(LK,ZK)}; Calculate the global statistical characteristics of the peak sequence, including the peak mean μ_Z and the peak standard deviation σ_Z; Extract any mapping relationship (LN, ZN), and determine the absolute peak feature R_abs_N=ZN, the relative peak feature R_rel_N=(ZN-μ_Z) / σ_Z, and the adjacent tooth comparison feature R_nei_N=|ZN-(ZN-1+ZN+1) / 2| / ((ZN-1+ZN+1) / 2) of the associated gear tooth LN. Among them, the adjacent tooth comparison feature of the first and last gear teeth is calculated by cyclic adjacent method. The absolute peak feature R_abs_N, the relative peak feature R_rel_N, and the adjacent tooth comparison feature R_nei_N are combined to define the comprehensive feature CR_N={R_abs_N,R_rel_N,R_nei_N} of the gear tooth LN; Similarly, by determining the comprehensive features associated with each of the K gear teeth, we obtain the comprehensive feature sequence CR_1, CR_2, ..., CR_K; A secondary judgment is made based on the comprehensive characteristics of each tooth, and abnormal teeth are marked as normal teeth.

7. The method according to claim 6, characterized in that, In step three, a secondary judgment is made based on the comprehensive characteristics of each gear tooth. The specific method for marking abnormal gear teeth and normal gear teeth is as follows: Obtain the comprehensive feature CR_N of any gear tooth LN, and extract the absolute peak feature R_abs_N, the relative peak feature R_rel_N, and the adjacent tooth comparison feature R_nei_N; Extract the operator's preset absolute threshold TH_abs, relative threshold TH_rel, and adjacent tooth comparison threshold TH_nei; The comprehensive feature CR_N of the gear tooth LN is compared with the corresponding threshold. If there exist R_abs_N≤TH_abs, R_rel_N≤TH_rel, and R_nei_N≤TH_nei, then mark tooth LN as a normal tooth; Conversely, it is marked as an abnormal tooth; Similarly, mark all abnormal teeth and normal teeth in the target gear.

8. The method according to claim 7, characterized in that, In step four, the specific method for evaluating the damage index of the target gear during the monitoring period is as follows: The total number of abnormal teeth in the target gear is counted and denoted as X. Then the total number of normal teeth is KX. The ratio Y of the number of abnormal teeth to the number of normal teeth is calculated using Y=X / (KX); For any tooth LN marked as an abnormal tooth, calculate the severity component of its comprehensive feature CR_N that exceeds a preset threshold, including: The absolute peak severity component S_abs_N = max(0, (R_abs_N - TH_abs) / TH_abs); The relative peak severity component S_rel_N = max(0, (|R_rel_N|-TH_rel) / TH_rel); The severity component of adjacent tooth contrast is S_nei_N=max(0,(R_nei_N-TH_nei) / TH_nei); The individual damage index S_N of the gear tooth LN is calculated using S_N=δ*S_abs_N+β*S_rel_N+γ*S_nei_N, where δ,β,γ are preset weighting coefficients, and δ+β+γ=1; Similarly, the mean of the individual damage index of all abnormal teeth is calculated as the average severity of the abnormal teeth, denoted as Avg_S. If the total number of abnormal teeth is 0, Avg_S is defined as 0. Based on the ratio Y and the average severity Avg_S, the damage index DI of the target gear within the monitoring period T_mc is calculated by DI=ω1×Y+ω2×Avg_S, where ω1 and ω2 are fusion weights, and ω1+ω2=1.

9. The method according to claim 8, characterized in that, In step four, after determining the damage index DI of the target gear within the monitoring period T_mc, the damage index DI is compared with the preset damage index threshold DI_TH. If DI≥DI_TH, output a target gear early warning fault reminder; Conversely, continuous monitoring will continue. If the damage index DI of the target gear increases for Q consecutive monitoring cycles, a target gear deterioration fault alert will be output, where Q is a preset integer.

10. A gear fault diagnosis system based on active sensing of piezoelectric stress wave signals, characterized in that, The system includes: The active envelope construction module, based on the piezoelectric module pre-deployed on the target gear, uses the single rotation time of the target gear as the monitoring period to extract the piezoelectric stress wave of the target gear in real time during operation and constructs the time-related piezoelectric stress wave envelope waveform of the target gear within the monitoring period. The peak warning and initial inspection module extracts the number of teeth of the target gear and performs peak detection on the piezoelectric stress wave envelope waveform. If the peak detection fails, an emergency fault reminder is output. If the peak detection passes, proceed to the multi-feature mapping diagnostic module for secondary analysis; The multi-feature mapping diagnostic module performs peak mapping between gear teeth based on the piezoelectric stress wave envelope waveform diagram within the monitoring period, constructs a peak sequence combined with a gear tooth segment sequence, determines the comprehensive features corresponding to each gear tooth in the target gear for secondary judgment, and marks abnormal gear teeth and normal gear teeth. The damage assessment module evaluates the damage index of the target gear within the monitoring period based on the ratio of abnormal to normal gear teeth and the comprehensive characteristics of abnormal faulty gear teeth, and performs continuous monitoring, outputting early warning fault alerts or deterioration fault alerts.

Images