Gear breakage identification method and device, equipment and medium

Abstract

The invention discloses a gear breakage identification method and device, equipment and a medium, and relates to the technical field of fault diagnosis, and the method comprises the steps: carrying out the single-sample fault diagnosis of impact single-sample data of each gear measurement point of a current gear, and obtaining a single-sample fault diagnosis result; performing trend analysis of gear impact and meshing spectrum impact based on the impact single sample data to obtain a gear impact trend result and a meshing spectrum impact trend result respectively, and obtaining gear impact continuity statistics of the impact single sample data; and according to the current health level of the current gear, the single sample fault diagnosis result, the gear impact trend result, the meshing spectrum impact trend result and the gear impact continuity statistical magnitude, determining whether the current gear has a gear breakage fault, and judging whether gear breakage early warning is needed. Multi-angle features are extracted through single sample analysis and trend analysis. And comprehensive gear breaking fault identification is carried out in combination with health levels and multi-angle features, so that misjudgment of a single index is avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of fault diagnosis, and in particular to a gear breakage identification method, device, equipment and medium. Background Art

[0002] Gears often need to bear large torque and radial loads during operation, and are affected by a variety of factors such as manufacturing errors, improper assembly, poor lubrication, overload, and poor operation and maintenance, resulting in a relatively high failure rate. Once a gear fails, it will not only affect the normal operation of the equipment and reduce production efficiency, but may also cause major safety accidents and bring serious economic losses.

[0003] In the early stages of gear failure, the fault characteristics are often not obvious and the signal changes are relatively weak, making it difficult to be captured in time by conventional diagnostic methods. For example, in the early stages of tooth surface fatigue crack development, when the cracks have just started to germinate, there may only be microscopic cracks in a very small local area of ​​the tooth surface. At this time, the impact on the overall vibration and temperature of the gear is very small.

[0004] At present, gear fault detection utilizes machine learning and artificial intelligence algorithms, such as artificial neural networks, support vector machines, decision trees, etc., to learn a large amount of gear fault sample data, explore hidden patterns and rules, and realize automatic identification of fault types. In order to achieve early fault diagnosis, advanced signal processing algorithms, such as wavelet analysis, are often used to reduce noise and extract weak features from the collected signals. By continuously optimizing the analysis parameters, the early fault signal characteristics are amplified to enhance their identifiability. However. In actual complex working conditions, the complex working environment and operating conditions will introduce a large number of interference signals. Therefore, the gear signal characteristics are difficult to extract accurately. In the diagnostic process, a single diagnostic method can often only reflect the gear fault condition from a specific angle. When facing complex gear system failures.

[0005] In summary, how to realize real-time and comprehensive collection and analysis of various data information of gears during operation, accurately capture abnormal impact signals, and then realize early warning and accurate diagnosis of gear failures is a technical problem to be solved in this field. Summary of the invention

[0006] In view of this, the purpose of the present invention is to provide a gear breakage identification method, device, equipment and medium, which can collect and analyze various data information of the gear in the operation process in real time and comprehensively, accurately capture abnormal impact signals, and thus achieve early warning and accurate diagnosis of gear failure. The specific scheme is as follows:

[0007] In a first aspect, the present application discloses a gear breakage identification method, comprising:

[0008] Performing single sample fault diagnosis on the impact single sample data of each gear measuring point of the current gear to obtain a single sample fault diagnosis result;

[0009] Performing trend analysis of gear impact and meshing spectrum impact based on each of the impact single sample data to obtain gear impact trend results and meshing spectrum impact trend results respectively, and obtaining gear impact continuity statistics of each of the impact single sample data;

[0010] Determine whether the current gear has a gear break fault according to the current health level of the current gear, the single sample fault diagnosis result, the gear impact trend result, the meshing spectrum impact trend result and the gear impact continuity statistic, so as to obtain a corresponding gear break identification result;

[0011] If the gear breakage identification result is that there is no gear breakage fault, a gear breakage warning judgment process is performed to obtain a corresponding gear breakage warning result.

[0012] Optionally, performing single sample fault diagnosis on the impact single sample data of each gear measuring point of the current gear to obtain a single sample fault diagnosis result includes:

[0013] Performing spectrum analysis on the impact single sample data of each gear measuring point of the current gear, and obtaining the target frequency components after the preset number of spectrum analyses in order of amplitude from large to small, and determining the maximum amplitude, so as to set the amplitude threshold based on the maximum amplitude;

[0014] The number of frequency components in the target frequency components that are greater than the amplitude threshold is counted, and a single-sample fault diagnosis result is determined based on the number of frequency components and a fault judgment condition.

[0015] Optionally, determining a single sample fault diagnosis result based on the number of frequency components and a fault determination condition includes:

[0016] Determining whether the number of frequency components meets a fault determination condition established based on a preset fault feature number threshold;

[0017] If the number of the frequency components is greater than or equal to the preset fault feature number threshold, it is determined that the impact single sample data has a single tooth broken tooth diagnosis feature or a tooth root crack feature, so as to determine the corresponding single sample fault diagnosis result;

[0018] If the number of the frequency components is less than the preset fault feature number threshold, it is determined that the impact single sample data has no fault diagnosis feature to determine the corresponding single sample fault diagnosis result.

[0019] Optionally, the trend analysis of gear impact and meshing spectrum impact is performed based on each of the impact single sample data to obtain a gear impact trend result and a meshing spectrum impact trend result, respectively, including:

[0020] Respectively calculating the strength index of the gear impact signal strength in a first single time period and the strength index of the meshing spectrum impact signal strength in a second single time period of each of the impact single sample data, so as to respectively obtain the first time series trend data of the gear impact signal strength and the second time series trend data of the meshing spectrum impact signal strength;

[0021] By using a multi-time scale dynamic smoothing method, and based on the first time series trend data and the second time series trend data, a trend evolution analysis is performed on the impact trend data within a preset time period to obtain gear impact trend results and meshing spectrum impact trend results, respectively.

[0022] Optionally, the multi-time scale dynamic smoothing method is used to perform trend evolution analysis on the impact trend data within a preset time period based on the first time series trend data and the second time series trend data to obtain a gear impact trend result and a meshing spectrum impact trend result, respectively, including:

[0023] Using a first preset time window, a sliding average process is performed on the first time series trend data within the preset time period to obtain a first short-term moving average;

[0024] Using a second preset time window to perform sliding average processing on the first time series trend data within the preset time period to obtain a first long-term moving average;

[0025] Using the first preset time window, a sliding average process is performed on the second time series trend data within the preset time period to obtain a second short-term moving average;

[0026] Using a second preset time window, a sliding average process is performed on the second time series trend data within the preset time period to obtain a second long-term moving average;

[0027] Respectively extracting the corresponding data mean points on the first short-term moving average and the first long-term moving average in the same single time period, and performing difference processing to obtain a first short-term variation;

[0028] Respectively extract the corresponding data mean points on the second short-term moving average and the second long-term moving average in the same single time period, and perform difference processing to obtain the second short-term variation;

[0029] Performing difference processing on two data mean points on the first long-term moving average that are located in adjacent single time periods to obtain a first long-term variation;

[0030] Performing difference processing on two data mean points on the second long-term moving average that are located in adjacent single time periods to obtain a second long-term variation;

[0031] If the first short-time variation satisfies the short-time variation threshold condition, and the first long-time variation satisfies the long-time variation threshold condition, then determining that the gear impact trend result is a trend rising result;

[0032] If the second short-time variation satisfies the short-time variation threshold condition, and the second long-time variation satisfies the long-time variation threshold condition, it is determined that the meshing spectrum impact trend result is a trend rising result.

[0033] Optionally, the multi-time scale dynamic smoothing method is used to perform trend evolution analysis on the impact single sample data within a preset time period based on the first time series trend data and the second time series trend data to obtain gear impact trend results and meshing spectrum impact trend results, respectively, including:

[0034] Using a first preset time window, a sliding average process is performed on the first time series trend data within the preset time period to obtain a first short-term moving average;

[0035] Using a second preset time window to perform sliding average processing on the first time series trend data within the preset time period to obtain a first long-term moving average;

[0036] Using the first preset time window, a sliding average process is performed on the second time series trend data within the preset time period to obtain a second short-term moving average;

[0037] Using a second preset time window, a sliding average process is performed on the second time series trend data within the preset time period to obtain a second long-term moving average;

[0038] Performing a trend evolution analysis on the gear impact signal strength of the impact trend data based on the relative position of the first short-term moving average and the first long-term moving average and the size of the slope of the moving average to obtain a gear impact trend result;

[0039] Based on the relative position of the second short-term moving average and the second long-term moving average and the size of the moving average slope, a trend evolution analysis is performed on the meshing spectrum impact signal intensity of the impact trend data to obtain a meshing spectrum impact trend result.

[0040] Optionally, the step of obtaining the gear impact continuity statistics of each impact single sample data includes:

[0041] Based on the first time series trend data of each impact single sample data and the actual operating mileage information of the equipment where the current gear is located, the gear impact continuity statistics of the impact trend data within every 100 kilometers are calculated.

[0042] Optionally, the calculation of the gear impact continuity statistics of the impact trend data within every 100 kilometers based on the first time series trend data of each impact single sample data and the actual operating mileage information of the equipment where the current gear is located includes:

[0043] Counting the target frequency of the intensity index in the first single time period of the first time series trend data of each of the impact single sample data being greater than the preset intensity index threshold in the single time period;

[0044] According to the target frequency, the actual operating mileage of the equipment where the current gear is located, and by using a preset gear impact continuity statistics equation, the gear impact continuity statistics of the impact trend data within every 100 kilometers are calculated;

[0045] The preset gear impact continuity statistics equation is:

[0046] ;

[0047] in, represents the target frequency, Indicates the actual operating mileage. Represents the gear impact continuity statistic.

[0048] Optionally, determining whether the current gear has a gear breakage fault according to the current health level of the current gear, the single sample fault diagnosis result, the gear impact trend result, the meshing spectrum impact trend result and the gear impact continuity statistic includes:

[0049] If the current health level of the current gear is at any level of the unhealthy level, the feature ratio of the number of single-tooth broken tooth diagnosis features or tooth root crack diagnosis features in the single-sample fault diagnosis result to the total number of results of the single-sample fault diagnosis result is greater than a first preset fault feature ratio, and the gear impact trend result is a trend rising result, it is determined that the current gear has a gear breakage fault, and a gear breakage fault alarm is issued;

[0050] If the current health level of the current gear is at any level of the unhealthy levels, the feature ratio of the number of single-tooth broken tooth diagnosis features or tooth root crack diagnosis features in the single-sample fault diagnosis result to the total number of results of the single-sample fault diagnosis result is greater than a first preset fault feature ratio, and the meshing spectrum impact trend result is a trend decline result and / or the gear impact continuity statistic is greater than a preset statistic threshold, it is determined that the current gear has a gear breakage fault, and a gear breakage fault alarm is issued;

[0051] Optionally, if the current health level of the current gear is before any level of the unhealthy levels, the method further includes:

[0052] According to the health assessment of key train components and the intelligent operation and maintenance model output of the current health level of the current gear, it is determined whether the current health level is at any unhealthy level to obtain a corresponding level judgment result; wherein the unhealthy level includes a preset sub-health level, a preset minor fault level, a preset medium fault level, and a preset severe fault level.

[0053] Optionally, the gear breakage warning judgment process to obtain a corresponding gear breakage warning result includes:

[0054] If an external alarm message is received, and the feature quantity of the single-tooth broken tooth diagnosis feature or the tooth root crack diagnosis feature in the single-sample fault diagnosis result accounts for a feature quantity ratio of the total number of results of the single-sample fault diagnosis result that is greater than a second preset fault feature ratio, a gear breakage warning is issued; wherein the second preset fault feature ratio is less than the first preset fault feature ratio;

[0055] Or, if no external alarm information is received, and if it is detected that the meshing spectrum impact trend result is an upward trend result and the gear impact continuity statistic is greater than a preset statistic threshold, a gear breakage warning is issued.

[0056] In a second aspect, the present application discloses a gear breakage identification device, comprising:

[0057] A single sample diagnosis module is used to perform single sample fault diagnosis on the impact single sample data of each gear measuring point of the current gear to obtain a single sample fault diagnosis result;

[0058] A trend analysis module, used for performing trend analysis of gear impact and meshing spectrum impact based on each of the impact single sample data, so as to obtain a gear impact trend result and a meshing spectrum impact trend result respectively, and to obtain a gear impact continuity statistic of each of the impact single sample data;

[0059] a fault identification module, for determining whether the current gear has a gear breakage fault according to the current health level of the current gear, the single sample fault diagnosis result, the gear impact trend result, the meshing spectrum impact trend result and the gear impact continuity statistic, so as to obtain a corresponding gear breakage identification result;

[0060] The fault warning module is used to perform a gear breakage warning judgment process to obtain a corresponding gear breakage warning result if the gear breakage identification result is that there is no gear breakage fault.

[0061] In a third aspect, the present application discloses an electronic device, comprising:

[0062] Memory, used to store computer programs;

[0063] The processor is used to execute the computer program to implement the steps of the gear breakage identification method disclosed above.

[0064] In a fourth aspect, the present application discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, the steps of the gear breakage identification method disclosed above are implemented.

[0065] It can be seen that the present application discloses a gear breakage identification method, including: performing single sample fault diagnosis on the impact single sample data of each gear measuring point of the current gear respectively to obtain a single sample fault diagnosis result; performing trend analysis of gear impact and meshing spectrum impact based on each of the impact single sample data to obtain gear impact trend results and meshing spectrum impact trend results respectively, and obtaining the gear impact continuity statistics of each of the impact single sample data; determining whether the current gear has a gear breakage fault according to the current health level of the current gear, the single sample fault diagnosis result, the gear impact trend result, the meshing spectrum impact trend result and the gear impact continuity statistic to obtain the corresponding gear breakage identification result; if the gear breakage identification result is that there is no gear breakage fault, then performing a gear breakage warning judgment process to obtain the corresponding gear breakage warning result. It can be seen that multi-angle features are extracted through single sample analysis and trend analysis (gear impact, meshing spectrum trend). Then, comprehensive gear breakage fault identification is performed in combination with health level and multi-angle features to avoid misjudgment of a single indicator. The false alarm rate is reduced through triple verification of single sample features, trends, and statistics. Further, early warning of gear breakage faults (including early gear crack faults) is achieved through single sample analysis and trend analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0067] Figure 1 A flow chart of a gear breakage identification method disclosed in the present application;

[0068] Figure 2 A flow chart of a gear fracture trend analysis method disclosed in this application;

[0069] Figure 3 A flowchart of a specific gear breakage identification method disclosed in this application;

[0070] Figure 4 This is a schematic diagram of the structure of a gear breakage identification device disclosed in this application;

[0071] Figure 5 This is a structural diagram of an electronic device disclosed in this application. DETAILED DESCRIPTION

[0072] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0073] As a key transmission component in many mechanical equipment, gears are widely used in many fields, including transportation, industrial machinery, aerospace, medical equipment, etc. For example, they play an indispensable role in automobile transmissions, machine tool spindle transmissions, aircraft engine accessory transmissions, etc. However, since gears often need to withstand large torque and radial loads during operation, and are affected by a variety of factors such as manufacturing errors, improper assembly, poor lubrication, overload, and poor operation and maintenance, their failure rate is relatively high.

[0074] Once a gear fails, it will not only affect the normal operation of the equipment and reduce production efficiency, but may also cause major safety accidents and bring serious economic losses. Gear failures manifest themselves in various forms, and different types of failures may have similarities in certain characteristic manifestations. This makes it difficult to accurately distinguish which specific type of failure it belongs to by relying on only a single diagnostic method or partial diagnostic means. When the gear failure is in the early stages, since the fault characteristics are often not obvious and the signal changes are relatively weak, it is difficult to be captured in time by conventional diagnostic methods. For example, in the early stage of tooth surface fatigue crack development, when the cracks have just begun to germinate, there may only be microscopic cracks in a very small local area of ​​the tooth surface. At this time, the impact on the overall vibration, temperature, etc. of the gear is very small.

[0075] At present, machine learning and artificial intelligence algorithms, such as artificial neural networks, support vector machines, decision trees, etc., are used to learn a large number of gear fault sample data, explore hidden patterns and rules, and realize automatic identification of fault types. In order to realize early fault diagnosis, advanced signal processing algorithms, such as wavelet analysis, are often used to reduce noise and extract weak features from the collected signals. By continuously optimizing analysis parameters, the characteristics of early fault signals are amplified to enhance their identifiability. However, under normal conditions, gear impact signals have specific rules and characteristics. However, once a fault occurs, such as tooth surface wear and cracks, the impact signal will change accordingly. In actual working conditions, complex working environments and operating conditions will introduce a large number of interference signals. For example, in large-scale industrial production sites, many mechanical equipment are running at the same time, which will generate vibrations and noises of various frequencies. For signal acquisition during gear fault diagnosis, it is like distinguishing a specific "melody" in a noisy "background sound". Take a multi-stage gear reducer as an example. The gears interact with each other, and there are vibration interferences from motors, couplings and other components around them. When one of the gears fails, it is very difficult to separate the key characteristic signals that can accurately reflect the gear failure from these numerous interference signals. Traditional technical means have the following disadvantages:

[0076] Strong data dependence: A large amount of representative fault sample data is required for training. Otherwise, the model has poor generalization ability and it is difficult to accurately identify fault types that have not appeared in the training data or faults under different working conditions.

[0077] Poor model interpretability: The internal decision-making process and feature extraction mechanism of some complex intelligent algorithms, such as deep neural networks, are difficult to understand intuitively, which is not conducive to trust in diagnostic results and further fault analysis.

[0078] Poor adaptability to working conditions: In practical applications, gears often need to operate under a variety of working conditions, but the adaptability of existing diagnostic methods under different working conditions is obviously limited.

[0079] Limited comprehensive diagnostic capabilities: A single diagnostic method can often only reflect the gear failure from a specific angle. When faced with complex gear system failures, it is necessary to combine multiple diagnostic methods for collaborative diagnosis, but there are currently many limitations in this regard.

[0080] To this end, the present invention provides a gear breakage identification solution, which can realize real-time and comprehensive collection and analysis of various data information of the gear during operation, accurately capture abnormal impact signals, and thus achieve early warning and accurate diagnosis of gear failures.

[0081] Reference Figure 1 As shown, an embodiment of the present invention discloses a gear breakage identification method, comprising:

[0082] Step S11: performing single-sample fault diagnosis on the impact single-sample data of each gear measuring point of the current gear to obtain a single-sample fault diagnosis result.

[0083] In this embodiment, before the single sample fault diagnosis is performed on the impact single sample data of each gear measuring point of the current gear, it also includes: collecting impact signals from each gear measuring point of the current gear within each preset time period to obtain impact signals as impact single sample data. It can be understood that the impact signals of each gear measuring point when the current gear is running are first collected by sensors, wherein the gear measuring point is a sensor position installed on the gear or transmission equipment, which is used to monitor the vibration, noise and other signals of specific parts in real time. Different gear measuring points focus on the status of different areas of the gear (such as the root, top of the tooth, and meshing point) to accurately locate the fault location. By fusing the impact data of multiple gear measuring points, the limitations of a single position signal are reduced. It should be noted that the preset time period is 24 hours. That is, the impact signal of the preset time length is collected every day as the impact single sample data, and the impact signal contains the impact component. The preset time length can be 10 seconds, or other specific time lengths, which are not specifically limited.

[0084] In this embodiment, the impact single sample data of each gear measuring point of the current gear is subjected to spectrum analysis, and the target frequency components after the first preset number of spectrum analyses are obtained in order of amplitude from large to small, and the maximum amplitude is determined to set the amplitude threshold based on the maximum amplitude; the number of frequency components in the target frequency components that are greater than the amplitude threshold is counted, and the single sample fault diagnosis result is determined based on the number of frequency components and the fault judgment condition. It can be understood that the spectrum analysis processing of the above-obtained impact single sample data is performed to obtain the first 20 target frequency components sorted by amplitude size. The target frequency component refers to the first 20 frequency points with the highest amplitude in the spectrum analysis results within the preset time length. The 20 target frequency components are selected to cover the main key fault characteristic frequencies, because gear faults (such as broken teeth, cracks) will significantly change the meshing frequency and its harmonics or sideband amplitudes, and when the gear is operating normally, the meshing frequency and its harmonics are the main vibration sources, and broken teeth or cracks will cause the meshing frequency amplitude to increase, or generate new sidebands. The first 20 orders cover these key frequency bands to ensure that the fault signal is not missed.

[0085] Specifically, search for the first 20 order spectra in the spectrum analysis results of the current gear, and obtain the maximum amplitude among the first 20 order fault spectra ; Set the amplitude threshold to 50% of the maximum amplitude, so the amplitude threshold is ; Count the top 20 fault spectra and the "top 20 maximum values ​​and amplitudes reaching or exceeding " is the order n (number of frequency components).

[0086] Further, the determination of the single sample fault diagnosis result based on the number of frequency components and the fault judgment condition includes: judging whether the number of frequency components meets the fault judgment condition constructed based on the preset fault feature number threshold; if the number of frequency components is greater than or equal to the preset fault feature number threshold, judging that the impact single sample data has a single tooth broken tooth diagnosis feature or a tooth root crack feature, so as to determine the corresponding single sample fault diagnosis result; if the number of frequency components is less than the preset fault feature number threshold, judging that the impact single sample data has no fault diagnosis feature, so as to determine the corresponding single sample fault diagnosis result. It can be understood that the fault judgment condition is set as the size relationship between the number of frequency components and the number of 8 fault features. When the number of frequency components is greater than or equal to 8, it is judged that there is a single tooth broken tooth or tooth root crack fault of the gear. Therefore, if n≥8, it is considered that there is a single tooth broken tooth diagnosis feature or a tooth root crack diagnosis feature of the gear.

[0087] For example: when the meshing frequency and its first few harmonic amplitudes of the 20 target frequency components are high and the amplitudes of other frequency components are low, the impact single sample data is judged to have no fault diagnosis results; when the amplitude of the meshing frequency of the 20 target frequency components increases significantly and new sidebands appear, the impact single sample data is judged to have the diagnostic features of a single gear tooth breakage; when the amplitude of the high-order harmonics of the meshing frequency of the 20 target frequency components increases abnormally, the impact single sample data is judged to have the diagnostic features of tooth root cracks.

[0088] In this way, by performing spectrum analysis on the impact single sample data of the current gear every day, it is ensured that the fault characteristics of the day can be discovered in time. Among them, by screening the first 20 target frequency components to capture the key characteristics of the gear fault, the calculation amount is reduced, the calculation efficiency is optimized, and the actual engineering needs are adapted. The impact single sample data is preliminarily diagnosed every day to screen out potential fault signals and reduce the amount of data that needs to be tracked for a long time. However, the impact single sample data may be affected by accidental factors such as instantaneous noise and temporary equipment abnormalities, resulting in misjudgment. For example, a high-amplitude spectrum line that appears in a certain measurement may be caused by temporary vibration rather than a real fault; in the early stage of the fault, the spectrum characteristics of cracks or wear may be very weak (such as the amplitude is lower than the threshold), and single sample analysis is prone to miss these early signals. Single sample analysis only reflects the current state and cannot determine whether the fault continues to develop or worsen. Therefore, after performing fault diagnosis of the impact single sample, further analysis from other angles is performed to obtain comprehensive data.

[0089] Step S12: performing trend analysis of gear impact and meshing spectrum impact based on each of the impact single sample data to obtain gear impact trend results and meshing spectrum impact trend results respectively, and obtaining gear impact continuity statistics of each of the impact single sample data.

[0090] In this embodiment, the strength index of the gear impact signal strength within the first single time period and the strength index of the meshing spectrum impact signal strength within the second single time period of each of the impact single sample data are calculated respectively to obtain the first time series trend data of the gear impact signal strength and the second time series trend data of the meshing spectrum impact signal strength respectively; through a multi-time scale dynamic smoothing method, and based on the first time series trend data and the second time series trend data, a trend evolution analysis is performed on the impact trend data within a preset time period to obtain the gear impact trend result and the meshing spectrum impact trend result respectively. It can be understood that a single time period can be set to 24 hours, as can be seen from the specific embodiments in the above steps. Therefore, the root mean square value of the gear impact dB (Decibel) of the impact single sample data is calculated on a daily basis to obtain the first single-day strength index; the root mean square value of the meshing spectrum impact dB of the impact single sample data is calculated on a daily basis to obtain the second single-day strength index; then, based on the first single-day strength index of the day and the first single-day strength index of the history, the gear impact dB root mean square trend data is formed, that is, the first time series trend data; based on the second single-day strength index of the day and the second single-day strength index of the history, the meshing spectrum impact dB root mean square trend data is formed, that is, the second time series trend data; in this way, the impact trend data is obtained, and the impact trend data specifically includes the first time series trend data and the second time series trend data, and then according to the long and short moving average sliding average method, the gear impact dB root mean square trend data and the meshing spectrum impact dB root mean square trend data are respectively trend analyzed to analyze whether there is an obvious upward trend, and a trend upward result is obtained. Among them, the trend analysis can be performed separately in two different implementation modes, or in combination with the two implementation modes. The specific trend analysis process is as follows:

[0091] In one embodiment, a first preset time window is used to perform sliding average processing on the first time series trend data within the preset time period to obtain a first short-time moving average; a second preset time window is used to perform sliding average processing on the first time series trend data within the preset time period to obtain a first long-time moving average; a first preset time window is used to perform sliding average processing on the second time series trend data within the preset time period to obtain a second short-time moving average; a second preset time window is used to perform sliding average processing on the second time series trend data within the preset time period to obtain a second long-time moving average; the corresponding data mean points on the first short-time moving average and the first long-time moving average in the same single time period are respectively extracted, and difference processing is performed to obtain a first short-time change; the same The corresponding data mean points on the second short-time moving average and the second long-time moving average in a single time period are processed for difference to obtain the second short-time variation; the two data mean points on the first long-time moving average located in adjacent single time periods are processed for difference to obtain the first long-time variation; the two data mean points on the second long-time moving average located in adjacent single time periods are processed for difference to obtain the second long-time variation; if the first short-time variation satisfies the short-time variation threshold condition, and the first long-time variation satisfies the long-time variation threshold condition, then the gear impact trend result is determined to be a trend-up result; if the second short-time variation satisfies the short-time variation threshold condition, and the second long-time variation satisfies the long-time variation threshold condition, then the meshing spectrum impact trend result is determined to be a trend-up result. It can be understood that if Figure 2As shown, the preset time period is set to 30 days including the current day, and then a smaller time window, such as 7 days, is used to perform sliding processing on the first time series trend data or the second time series trend data (RMS trend data z), and then the weekly change (7-day change) is calculated to obtain the first short-term moving average or the second short-term moving average, respectively. A larger time window, such as 28 days, is further used to perform sliding processing on the first time series trend data or the second time series trend data (RMS trend data z), and then the 28-day change is calculated to obtain the first long-term moving average or the second long-term moving average, respectively. Furthermore, if the analysis requirement is for a fault diagnosis requirement on a certain day, the data mean point (RMS mean) corresponding to the current day is obtained from the first short-term moving average, and the data mean point corresponding to the current day is obtained from the first long-term moving average. The data mean point on the first short-term moving average is subtracted from the numerical mean point on the first long-term moving average to obtain the first short-term change, so as to measure the magnitude of the short-term deviation from the long-term trend (the first short-term change), wherein if the first short-term change is greater than the threshold A, it indicates that the recent impact intensity is higher than the long-term average level. At the same time, the difference between the long-term sliding average of the day (the data mean point of the day on the first long-term moving average) and the long-term sliding average of the previous day (the data mean point of the previous day on the first long-term moving average) is calculated to obtain the first long-term change, which is used to reflect the evolution rate of the long-term trend. If the first long-term change is greater than the threshold value B, it indicates that the long-term trend is generally on the rise. Therefore, if the first short-term change is greater than the threshold value A, and the first long-term change is greater than the threshold value B, it indicates that the gear impact trend result is an upward trend result. The trend analysis process of the meshing general impact trend is the same as the analysis process of the gear impact trend, which will not be repeated here. In this way, by directly comparing the mean changes in different time windows, we can directly focus on the acceleration or slowdown of the trend.

[0092] In another embodiment, a first preset time window is used to perform sliding average processing on the first time series trend data within the preset time period to obtain a first short-time moving average; a second preset time window is used to perform sliding average processing on the first time series trend data within the preset time period to obtain a first long-time moving average; a first preset time window is used to perform sliding average processing on the second time series trend data within the preset time period to obtain a second short-time moving average; a second preset time window is used to perform sliding average processing on the second time series trend data within the preset time period to obtain a second long-time moving average; based on the relative position of the first short-time moving average and the first long-time moving average, and the size of the slope of the moving average, a trend evolution analysis is performed on the gear impact signal strength of the impact trend data to obtain a gear impact trend result; based on the relative position of the second short-time moving average and the second long-time moving average, and the size of the slope of the moving average, a trend evolution analysis is performed on the meshing spectrum impact signal strength of the impact trend data to obtain a meshing spectrum impact trend result. It can be understood that the preset time period is set to 30 days including the current day, and then a smaller time window, such as 7 days, is used to perform sliding processing on the first time series trend data or the second time series trend data (RMS impact trend data z), and then the weekly change (7-day change) is calculated to obtain the first short-term moving average or the second short-term moving average, and a larger time window, such as 28 days, is further used to perform sliding processing on the first time series trend data or the second time series trend data (RMS trend data z), and then After that, calculate the 28-day change to get the first long-term moving average or the second long-term moving average respectively; then, perform a relative position analysis of the slope of the short-term moving average and the long-term moving average. For example, if the short-term moving average is higher than the long-term moving average for K consecutive days (such as 3 days) and exceeds the threshold A, it is marked as "short-term trend rising". Perform a moving average slope analysis on the short-term moving average and the long-term moving average. For example, calculate the slopes of the short-term moving average and the long-term moving average (such as linear regression fitting). If the slope of the short-term moving average > the slope of the long-term moving average, and the slope difference exceeds the threshold B, it is marked as "accelerating upward trend". Finally, a comprehensive judgment process is performed, and the "trend rising result" is generated if any of the following conditions are met:

[0093] Condition 1: The short-term trend is upward (the moving average position continues to cross upward).

[0094] Condition 2: Accelerating upward trend (the slope of the moving average increases significantly).

[0095] According to the above-mentioned slope relative position analysis and moving average slope analysis methods, the trend directions of the first short-time moving average and the first long-time moving average, and the trend directions of the second short-time moving average and the second long-time moving average are analyzed respectively to obtain the gear impact trend results and the meshing spectrum impact trend results respectively. Among them, if the trend evolution analysis of the gear impact signal intensity meets the rising condition, it is marked as "gear impact trend rising"; if the trend evolution analysis of the meshing spectrum impact signal intensity meets the rising condition, it is marked as "meshing spectrum trend rising".

[0096] In this way, by converting the original signal into a single-day strength index (calculating the root mean square value), eliminating high-frequency noise and instantaneous interference, the calculated single-day strength index is smoothed on multiple time scales to separate short-term fluctuations from long-term trends. The relative position of the short-term moving average and the long-term moving average can be used to verify whether the fault is persistent, and the long-term moving average verifies the long-term health status, forming a closed loop of "instant detection + trend tracking". The combination of the two realizes the upgrade from "single-day snapshot" to "full-cycle trend", providing a reliable technical path for the accurate identification and early warning of gear breakage faults.

[0097] In this embodiment, based on the first time series trend data of each impact single sample data and the actual operating mileage information of the equipment where the current gear is located, the gear impact continuity statistics of the impact trend data within every 100 kilometers are calculated. Specifically, the target frequency of the intensity index in the first time series trend data of each impact single sample data being greater than the preset intensity index threshold within a single time period is counted; the gear impact continuity statistics of the impact trend data within every 100 kilometers are calculated according to the target frequency, the actual operating mileage of the equipment where the current gear is located, and the preset gear impact continuity statistics equation; the preset gear impact continuity statistics equation is:

[0098] ;

[0099] in, Indicates the target frequency of the daily strength index being greater than the preset single time period strength index threshold. Indicates the actual operating mileage of a single day. It represents the statistics of gear impact continuity. The single time period is a single day.

[0100] It can be understood that the impact continuity statistic (the number of gear dB per 100 kilometers) is calculated based on the first time series trend data. Specifically, the target frequency of the first single-day intensity index greater than 0 in the first time series trend data (impact trend data) of the impact single sample data of a certain day and the actual single-day operating history of the equipment where the current gear is located are counted. The abnormal frequency of the gear impact signal per 100 kilometers is calculated through the above-mentioned preset gear impact continuity statistic equation to obtain the gear impact continuity statistic.

[0101] In this way, if the gear impact dB value (first time series trend data) rises occasionally on a certain day, but the gear impact continuity statistics do not increase significantly, it may be a short-term interference rather than a continuous fault, which can rule out occasional interference, and if the dB number per 100 kilometers continues to rise, it indicates that the fault impact event becomes more frequent with the increase in mileage, which further verifies the fault deterioration. Moreover, by normalizing the gear impact continuity statistics of the impact single sample data within every 100 kilometers, the problem of misjudgment caused by directly counting the number of impacts per day due to large differences in daily mileage of different equipment (such as 200 kilometers on one day and only 50 kilometers on another day) is avoided, and the data under different working conditions can be uniformly judged, for example: Equipment A: 200 kilometers per day, 10 dB>0 detected → dB number per 100 kilometers = 5. Equipment B: 50 kilometers per day, 3 dB>0 detected → dB number per 100 kilometers = 6; Conclusion: Equipment B has a higher impact frequency and needs to be paid special attention.

[0102] Step S13: Determine whether the current gear has a gear breakage fault based on the current health level of the current gear, the single sample fault diagnosis result, the gear impact trend result, the meshing spectrum impact trend result and the gear impact continuity statistic to obtain a corresponding gear breakage identification result.

[0103] In this embodiment, the gear health level is pre-set to five levels: normal, sub-health level, slight fault level, medium fault level, and severe fault level. The current health level of the current gear state is obtained. Specifically, the current health level of the current gear is output according to the health assessment of key components of the train and the intelligent operation and maintenance model, and it is determined whether the current health level is in any level of the unhealthy level to obtain the corresponding level judgment result; wherein, the unhealthy level includes a preset sub-health level, a preset slight fault level, a preset medium fault level, and a preset severe fault level. Among them, after obtaining the impact single sample data, the obtained impact single sample data is transmitted to the external train key component health assessment and intelligent operation and maintenance model, through which the train key component health assessment and intelligent operation and maintenance model can output the health level of the current gear, wherein the current health level can be determined by the train key component health assessment and intelligent operation and maintenance model based on historical single sample analysis results (daily single sample failure ratio (such as broken teeth, crack feature ratio)), trend data, statistics (dB per hundred kilometers, impact event frequency, etc.), and external information (load intensity, lubrication status, operating time, etc.). The train key component health assessment and intelligent operation and maintenance model obtains the health score of the current gear through weighted calculation of the above-mentioned indicators, and determines the current health level based on the score division rules, wherein the score division rules are as follows: normal: health score <0.2, sub-health: 0.2≤health score <0.5, minor fault: 0.5≤health score <0.7, moderate fault: 0.7≤health score <0.9, serious fault: health score ≥0.9. It should be noted that the above health level determination process is only an example for illustration, and the health level determination and output process of the train key component health assessment and intelligent operation and maintenance model are not specifically limited.

[0104] In this embodiment, if the current health level of the current gear is at any level in the unhealthy level, the feature ratio of the single tooth broken tooth diagnosis feature or the tooth root crack diagnosis feature in the single sample fault diagnosis result to the total number of results of the single sample fault diagnosis result is greater than the first preset fault feature ratio, and the gear impact trend result is a trend rising result, then it is determined that the current gear has a gear breakage fault, and a gear breakage fault alarm is performed; it can be understood that if the current gear is at any level in the unhealthy level, and the gear impact trend result is a trend rising result (the short-term moving average is continuously higher than the long-term moving average). The single sample fault ratio> the first preset fault feature ratio (0.5), it is determined that the current gear has a gear breakage fault. That is, the gear impact trend worsens and instantaneous faults occur frequently, and the breakage is confirmed to occur, and a gear breakage fault alarm is directly performed.

[0105] For example: Scenario 1: Health level: minor fault; gear impact trend result: rising for 5 consecutive days (short moving average > long moving average); single sample failure ratio: 0.6, judgment result: trigger condition, output "gear breakage fault, maintenance recommended".

[0106] In this embodiment, if the current health level of the current gear is at any level in the unhealthy level, the feature number of the single tooth broken tooth diagnosis feature or the tooth root crack diagnosis feature in the single sample fault diagnosis result accounts for a feature ratio of the total number of results of the single sample fault diagnosis result that is greater than the first preset fault feature ratio, and the meshing spectrum impact trend result is a trend decline result and / or the gear impact continuity statistic is greater than the preset statistic threshold, then it is determined that the current gear has a gear break fault, and a gear break fault alarm is performed. It can be understood that if the current health level of the current gear is at any level in the unhealthy level, the gear impact trend result does not rise (the trend is stable or fluctuating). The single sample fault ratio> the first preset fault feature ratio, at this time, if the following additional conditions (one of them is sufficient) are met: the meshing spectrum impact trend result decreases, or the gear impact continuity statistic ratio> the preset statistic threshold (the surface impact event density increases), it is determined that the current gear has a gear break fault, that is, the gear impact trend has not deteriorated, but the meshing state is abnormal or the impact frequency increases, and it is confirmed that the break occurs, and the gear break fault alarm is directly performed.

[0107] For example: Scenario 2: Health level: sub-healthy; gear impact trend result: stable; single sample failure ratio: 0.6; meshing spectrum impact trend result: decreasing (tooth surface wear); judgment result: trigger condition, output alarm.

[0108] Furthermore, if the historical fault identification result of the current gear has been determined to be a broken fault but has not been repaired, the conclusion of the gear broken fault is maintained and a gear broken fault alarm is directly issued, that is, ensuring that unhandled faults are not missed and are forced to be tracked until repaired.

[0109] Step S14: If the gear breakage identification result is that there is no gear breakage fault, a gear breakage warning judgment process is performed to obtain a corresponding gear breakage warning result.

[0110] In this embodiment, early signs of gear breakage faults are captured, and an early warning is issued before the fault reaches the alarm threshold, supporting preventive maintenance. Specifically, if the gear breakage identification result is that there is no gear breakage fault, a gear breakage early warning judgment process is performed to obtain the corresponding gear breakage early warning result, including: if an external alarm message is received, and the number of features of the single-tooth broken tooth diagnosis feature or the tooth root crack diagnosis feature in the single-sample fault diagnosis result accounts for a feature ratio of the total number of results of the single-sample fault diagnosis result that is greater than a second preset fault feature ratio, a gear breakage early warning is performed; wherein the second preset fault feature ratio is less than the first preset fault feature ratio; or, if no external alarm message is received, and if it is detected that the meshing spectrum impact trend result is a trend rising result, and the gear impact continuity statistic is greater than the preset statistic threshold, a gear breakage early warning is performed. It can be understood that the breakage early warning does not require a health level threshold, and the warning can still be triggered even if the health level is normal. The triggering conditions for the breakage early warning (including early crack fault of the gear) are:

[0111] Condition ①: When there is external warning / alarm information (in the field of rail transit, it refers to the real-time alarm information output by the onboard fault monitoring device of the train running gear for the current gear in this cycle), the single sample fault ratio> the second preset fault feature ratio (0.2), for example, ≥2 times in 10 analyses show fault features. It can be seen that the combination of external alarm and weak fault signal can provide early warning.

[0112] Condition ②: In the absence of external warning / alarm information, the meshing spectrum impact trend increases, and the gear impact continuity statistics > the preset statistical threshold (0.5), triggering an early warning. It can be seen that early faults can be identified independently through abnormal meshing state and changes in impact frequency.

[0113] Scenario 1: Health level: normal; lubrication system alarm (pre-alarm information); single sample failure ratio: 0.3 (exceeding threshold D=0.2); judgment result: trigger condition ①, output "early fracture characteristics exist, inspection is recommended".

[0114] Scenario 2: Health level: normal; meshing spectrum trend: rising for 3 consecutive days (early stage of crack extension); gear impact continuity statistics: rising from 5 to 8 (increase ratio 60%>0.5); judgment result: trigger condition ②, output warning.

[0115] like Figure 3 As shown in the figure, the overall process of the gear breakage identification method is shown, including six main steps: data collection, single sample analysis, trend analysis, statistical calculation, fault judgment and early warning analysis. Specifically:

[0116] First, perform data collection. Specifically, input: impact signal during gear operation (collected by sensor). Output: single sample data of impact at each gear measurement point (time domain waveform).

[0117] Specific operation: Install sensors at key positions of gears (such as tooth roots and near bearings), collect impact signals in real time, and store them as single sample data on a daily basis.

[0118] Secondly, single sample fault diagnosis analysis is performed. Specifically, input: single sample data of impact measured once. Output: single sample fault diagnosis result (broken tooth, crack or no fault).

[0119] Tool operation: Perform spectrum analysis (FFT) on single sample data and extract the first 20 spectral lines.

[0120] The order n with a statistical amplitude ≥ 50% of the maximum amplitude, if n ≥ 8, it is determined that there is a broken tooth or crack fault; otherwise, it is determined that there is no fault.

[0121] Example: In a certain measurement, 10 of the first 20 spectral lines have amplitudes ≥ 50% of the maximum amplitude → it is determined to be a broken tooth fault.

[0122] Further trend analysis is performed. Specifically, input: daily root mean square (RMS) value of gear impact dB and meshing spectrum impact dB. Output: gear impact trend result, meshing spectrum impact trend result (up / down mark).

[0123] Specific operation: Calculate the root mean square value of gear impact dB and meshing spectrum impact dB on a daily basis to form trend data. Use the long and short moving average sliding average method (such as 7-day short moving average and 30-day long moving average) to smooth the trend data. If the short moving average is continuously higher than the long moving average, it is marked as "trend rising"; otherwise it is marked as "trend stable or falling".

[0124] Example: Gear Shock dB trend is rising for 5 consecutive days (short moving average > long moving average) → marked as “Trending Up”.

[0125] Then, based on the trend analysis, trend statistics are further performed. Specifically, input: gear impact dB trend data. Output: gear dB number per 100 kilometers (impact continuity statistics).

[0126] Specific operation: Count the number of times the gear impact dB>0 in a single day, and calculate the dB per 100 kilometers.

[0127] Example: On a certain day, the number of gear impact dB>0 is 10 times, and the operating mileage is 200 kilometers → dB per 100 kilometers = 5.

[0128] Execute gear breakage fault determination. Specifically, input: health level, single sample fault ratio, trend indicator, dB per 100 kilometers. Output: gear breakage fault alarm (yes / no).

[0129] Precondition: Health level ≥ sub-health. Judgment conditions (alarm will be triggered if any of the following conditions are met): Condition ①: Gear impact trend increases + single sample failure ratio > threshold A (0.5); Condition ②: Gear impact trend is stable + single sample failure ratio > threshold B (0.5) + (meshing spectrum trend decreases or dB increase ratio per 100 km > C (0.5)). Condition ③: Historical failures have not been repaired.

[0130] Example: Health level: minor fault, gear impact trend increases, single sample fault ratio: 0.6, judgment result: trigger alarm.

[0131] Finally, the gear breakage warning analysis is performed. Specifically, the input is: pre-warning / alarm information, single sample failure ratio, meshing spectrum trend, dB per 100 km. Output: early warning of gear breakage (yes / no). Judgment conditions (early warning can be obtained if any of the following conditions are met): Condition ①: Pre-warning / alarm information exists + single sample failure ratio > threshold D (0.2). Condition ②: No pre-warning / alarm information + meshing spectrum trend increases + dB per 100 km increase ratio > E (0.5).

[0132] Example: The meshing spectrum trend increases for three consecutive days, and the dB per 100 kilometers increases from 5 to 8 (the increase ratio is 60%). Judgment result: trigger an early warning.

[0133] It can be seen that the present application discloses a gear breakage identification method, including: performing single sample fault diagnosis on the impact single sample data of each gear measuring point of the current gear respectively to obtain a single sample fault diagnosis result; performing trend analysis of gear impact and meshing spectrum impact based on each of the impact single sample data to obtain gear impact trend results and meshing spectrum impact trend results respectively, and obtaining the gear impact continuity statistics of each of the impact single sample data; determining whether the current gear has a gear breakage fault according to the current health level of the current gear, the single sample fault diagnosis result, the gear impact trend result, the meshing spectrum impact trend result and the gear impact continuity statistic to obtain the corresponding gear breakage identification result; if the gear breakage identification result is that there is no gear breakage fault, then performing a gear breakage warning judgment process to obtain the corresponding gear breakage warning result. It can be seen that multi-angle features are extracted through single sample analysis and trend analysis (gear impact, meshing spectrum trend). Then, comprehensive gear breakage fault identification is performed in combination with health level and multi-angle features to avoid misjudgment of a single indicator. The false alarm rate is reduced through triple verification of single sample features, trends, and statistics. Further, early warning of gear breakage faults (including early gear crack faults) is achieved through single sample analysis and trend analysis.

[0134] Reference Figure 4As shown, the present invention provides a gear breakage identification device, comprising:

[0135] The single sample diagnosis module 11 is used to perform single sample fault diagnosis on the impact single sample data of each gear measuring point of the current gear to obtain a single sample fault diagnosis result;

[0136] A trend analysis module 12, for performing trend analysis of gear impact and meshing spectrum impact based on each of the impact single sample data, so as to obtain a gear impact trend result and a meshing spectrum impact trend result respectively, and to obtain a gear impact continuity statistic of each of the impact single sample data;

[0137] A fault identification module 13 is used to determine whether the current gear has a gear breakage fault according to the current health level of the current gear, the single sample fault diagnosis result, the gear impact trend result, the meshing spectrum impact trend result and the gear impact continuity statistic, so as to obtain a corresponding gear breakage identification result;

[0138] The fault warning module 14 is used to perform a gear breakage warning judgment process to obtain a corresponding gear breakage warning result if the gear breakage identification result is that there is no gear breakage fault.

[0139] It can be seen that the present application discloses performing single sample fault diagnosis on the impact single sample data of each gear measuring point of the current gear respectively to obtain a single sample fault diagnosis result; performing trend analysis of gear impact and meshing spectrum impact based on each of the impact single sample data to obtain gear impact trend results and meshing spectrum impact trend results respectively, and obtaining the gear impact continuity statistics of each of the impact single sample data; determining whether the current gear has a gear break fault according to the current health level of the current gear, the single sample fault diagnosis result, the gear impact trend result, the meshing spectrum impact trend result and the gear impact continuity statistics to obtain the corresponding gear break identification result; if the gear break identification result is that there is no gear break fault, a gear break warning judgment process is performed to obtain the corresponding gear break warning result. It can be seen that multi-angle features are extracted through single sample analysis and trend analysis (gear impact, meshing spectrum trend). Then, comprehensive gear breakage fault identification is performed in combination with health level and multi-angle features to avoid misjudgment of a single indicator. The false alarm rate is reduced through triple verification of single sample features, trends, and statistics. Further, early warning of gear breakage faults (including early gear crack faults) is achieved through single sample analysis and trend analysis.

[0140] Furthermore, the present application also discloses an electronic device. Figure 5This is a structural diagram of an electronic device 20 according to an exemplary embodiment. The content in the diagram cannot be regarded as any limitation on the scope of use of the present application.

[0141] Figure 5 The present invention provides a schematic diagram of the structure of an electronic device 20 provided in an embodiment of the present application. The electronic device 20 may specifically include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. The memory 22 is used to store a computer program, which is loaded and executed by the processor 21 to implement the relevant steps in the gear breakage identification method disclosed in any of the aforementioned embodiments. In addition, the electronic device 20 in this embodiment may specifically be an electronic computer.

[0142] In this embodiment, the power supply 23 is used to provide working voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and the external device, and the communication protocol it follows is any communication protocol that can be applied to the technical solution of the present application, and is not specifically limited here; the input and output interface 25 is used to obtain external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs and is not specifically limited here.

[0143] Among them, the processor 21 may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor 21 can be implemented in at least one hardware form of DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array). The processor 21 may also include a main processor and a coprocessor. The main processor is a processor for processing data in the awake state, also known as a CPU (Central Processing Unit); the coprocessor is a low-power processor for processing data in the standby state. In some embodiments, the processor 21 may be integrated with a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the display screen. In some embodiments, the processor 21 may also include an AI (Artificial Intelligence) processor, which is used to process computing operations related to machine learning.

[0144] In addition, the memory 22, as a carrier for storing resources, can be a read-only memory, a random access memory, a disk or an optical disk, etc. The resources stored thereon can include an operating system 221, a computer program 222, etc., and the storage method can be temporary storage or permanent storage.

[0145] The operating system 221 is used to manage and control the hardware devices and computer programs 222 on the electronic device 20, so as to realize the operation and processing of the massive data 223 in the memory 22 by the processor 21, which can be Windows Server, Netware, Unix, Linux, etc. In addition to including a computer program that can be used to complete the gear breakage identification method performed by the electronic device 20 disclosed in any of the aforementioned embodiments, the computer program 222 can further include a computer program that can be used to complete other specific tasks. In addition to data received by the electronic device and transmitted from an external device, the data 223 can also include data collected by its own input and output interface 25.

[0146] Furthermore, the present application also discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, the gear breakage identification method disclosed above is implemented. The specific steps of the method can refer to the corresponding contents disclosed in the above embodiments, and will not be repeated here.

[0147] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0148] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented with electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application. The steps of the method or algorithm described in conjunction with the embodiments disclosed herein can be implemented directly with hardware, a software module executed by a processor, or a combination of the two. The software module can be placed in a random access memory RAM (Random Access Memory), memory, read-only memory ROM (Read Only Memory), electrically programmable EPROM (Electrically Programmable Read Only Memory), electrically erasable programmable EEPROM (ElectricErasable Programmable Read Only Memory), register, hard disk, removable disk, CD-ROM (CompactDisc-Read Only Memory), or any other form of storage medium known in the technical field.

[0149] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0150] The scheme provided by the present invention is introduced in detail above. Specific examples are used in this article to illustrate the principle and implementation mode of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation mode and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A gear breakage identification method, characterized in that: include: Performing single sample fault diagnosis on the impact single sample data of each gear measuring point of the current gear to obtain a single sample fault diagnosis result; Performing trend analysis of gear impact and meshing spectrum impact based on each of the impact single sample data to obtain gear impact trend results and meshing spectrum impact trend results respectively, and obtaining gear impact continuity statistics of each of the impact single sample data; Determine whether the current gear has a gear break fault according to the current health level of the current gear, the single sample fault diagnosis result, the gear impact trend result, the meshing spectrum impact trend result and the gear impact continuity statistic, so as to obtain a corresponding gear break identification result; If the gear breakage identification result is that there is no gear breakage fault, a gear breakage warning judgment process is performed to obtain a corresponding gear breakage warning result.

2. The gear breakage identification method according to claim 1, characterized in that: The single sample fault diagnosis is performed on the impact single sample data of each gear measuring point of the current gear to obtain a single sample fault diagnosis result, including: Performing spectrum analysis on the impact single sample data of each gear measuring point of the current gear, and obtaining the target frequency components after the preset number of spectrum analyses in order of amplitude from large to small, and determining the maximum amplitude, so as to set the amplitude threshold based on the maximum amplitude; The number of frequency components in the target frequency components that are greater than the amplitude threshold is counted, and a single-sample fault diagnosis result is determined based on the number of frequency components and a fault determination condition.

3. The gear breakage identification method according to claim 2, characterized in that: The determining of a single sample fault diagnosis result based on the number of frequency components and the fault judgment condition includes: Determining whether the number of frequency components meets a fault determination condition established based on a preset fault feature number threshold; If the number of the frequency components is greater than or equal to the preset fault feature number threshold, it is determined that the impact single sample data has a single tooth broken tooth diagnosis feature or a tooth root crack feature, so as to determine the corresponding single sample fault diagnosis result; If the number of the frequency components is less than the preset fault feature number threshold, it is determined that the impact single sample data has no fault diagnosis feature to determine the corresponding single sample fault diagnosis result.

4. The gear breakage identification method according to claim 1, characterized in that: The trend analysis of gear impact and meshing spectrum impact is performed based on each of the impact single sample data to obtain the gear impact trend result and the meshing spectrum impact trend result respectively, including: Respectively calculating the strength index of the gear impact signal strength in a first single time period and the strength index of the meshing spectrum impact signal strength in a second single time period of each of the impact single sample data, so as to respectively obtain the first time series trend data of the gear impact signal strength and the second time series trend data of the meshing spectrum impact signal strength; By using a multi-time scale dynamic smoothing method, and based on the first time series trend data and the second time series trend data, a trend evolution analysis is performed on the impact trend data within a preset time period to obtain gear impact trend results and meshing spectrum impact trend results, respectively.

5. The gear breakage identification method according to claim 4, characterized in that: The method uses a multi-time scale dynamic smoothing method and performs trend evolution analysis on the impact trend data within a preset time period based on the first time series trend data and the second time series trend data to obtain a gear impact trend result and a meshing spectrum impact trend result, respectively, including: Using a first preset time window, a sliding average process is performed on the first time series trend data within the preset time period to obtain a first short-term moving average; Using a second preset time window to perform sliding average processing on the first time series trend data within the preset time period to obtain a first long-term moving average; Using the first preset time window, a sliding average process is performed on the second time series trend data within the preset time period to obtain a second short-term moving average; Using a second preset time window, a sliding average process is performed on the second time series trend data within the preset time period to obtain a second long-term moving average; Respectively extracting the corresponding data mean points on the first short-term moving average and the first long-term moving average in the same single time period, and performing difference processing to obtain a first short-term variation; Respectively extract the corresponding data mean points on the second short-term moving average and the second long-term moving average in the same single time period, and perform difference processing to obtain the second short-term variation; Performing difference processing on two data mean points on the first long-term moving average that are located in adjacent single time periods to obtain a first long-term variation; Performing difference processing on two data mean points on the second long-term moving average that are located in adjacent single time periods to obtain a second long-term variation; If the first short-time variation satisfies the short-time variation threshold condition, and the first long-time variation satisfies the long-time variation threshold condition, then determining that the gear impact trend result is a trend rising result; If the second short-time variation satisfies the short-time variation threshold condition, and the second long-time variation satisfies the long-time variation threshold condition, it is determined that the meshing spectrum impact trend result is a trend rising result.

6. The gear breakage identification method according to claim 4, characterized in that: The method uses a multi-time scale dynamic smoothing method and performs trend evolution analysis on the impact trend data within a preset time period based on the first time series trend data and the second time series trend data to obtain a gear impact trend result and a meshing spectrum impact trend result, respectively, including: Using a first preset time window, a sliding average process is performed on the first time series trend data within the preset time period to obtain a first short-term moving average; Using a second preset time window to perform sliding average processing on the first time series trend data within the preset time period to obtain a first long-term moving average; Using the first preset time window, a sliding average process is performed on the second time series trend data within the preset time period to obtain a second short-term moving average; Using a second preset time window, a sliding average process is performed on the second time series trend data within the preset time period to obtain a second long-term moving average; Performing a trend evolution analysis on the gear impact signal strength of the impact trend data based on the relative position of the first short-term moving average and the first long-term moving average and the size of the slope of the moving average to obtain a gear impact trend result; Based on the relative position of the second short-term moving average and the second long-term moving average and the size of the moving average slope, a trend evolution analysis is performed on the meshing spectrum impact signal intensity of the impact trend data to obtain a meshing spectrum impact trend result.

7. The gear breakage identification method according to claim 4, characterized in that: The step of obtaining the gear impact continuity statistics of each impact single sample data comprises: Based on the first time series trend data of each impact single sample data and the actual operating mileage information of the equipment where the current gear is located, the gear impact continuity statistics of the impact trend data within every 100 kilometers are calculated.

8. The gear breakage identification method according to claim 7, characterized in that: The calculation of the gear impact continuity statistics of the impact trend data within every 100 kilometers based on the first time series trend data of each impact single sample data and the actual operating mileage information of the equipment where the current gear is located includes: Counting the target frequency of the intensity index in the first single time period of the first time series trend data of each of the impact single sample data being greater than the preset intensity index threshold in the single time period; According to the target frequency, the actual operating mileage of the equipment where the current gear is located, and by using a preset gear impact continuity statistics equation, the gear impact continuity statistics of the impact trend data within every 100 kilometers are calculated; The preset gear impact continuity statistics equation is: ； in, represents the target frequency, Indicates the actual operating mileage. Represents the gear impact continuity statistic.

9. The gear breakage identification method according to claim 1, characterized in that: The determining whether the current gear has a gear breakage fault according to the current health level of the current gear, the single sample fault diagnosis result, the gear impact trend result, the meshing spectrum impact trend result and the gear impact continuity statistic includes: If the current health level of the current gear is at any level of the unhealthy level, the feature ratio of the number of single-tooth broken tooth diagnosis features or tooth root crack diagnosis features in the single-sample fault diagnosis result to the total number of results of the single-sample fault diagnosis result is greater than a first preset fault feature ratio, and the gear impact trend result is a trend rising result, it is determined that the current gear has a gear breakage fault, and a gear breakage fault alarm is issued; If the current health level of the current gear is at any level among the unhealthy levels, the feature ratio of the feature number of the single-tooth broken tooth diagnosis feature or the tooth root crack diagnosis feature in the single-sample fault diagnosis result to the total result number of the single-sample fault diagnosis result is greater than the first preset fault feature ratio, and the meshing spectrum impact trend result is a trend downward result and / or the gear impact continuity statistic is greater than the preset statistic threshold, then it is determined that the current gear has a gear breakage fault, and a gear breakage fault alarm is issued.

10. The gear breakage identification method according to claim 9, characterized in that: If the current health level of the current gear is higher than any level of the unhealthy levels, the method further includes: According to the health assessment of key train components and the intelligent operation and maintenance model output of the current health level of the current gear, it is determined whether the current health level is at any unhealthy level to obtain a corresponding level judgment result; wherein the unhealthy level includes a preset sub-health level, a preset minor fault level, a preset medium fault level, and a preset severe fault level.

11. The gear breakage identification method according to claim 1, characterized in that: The gear breakage warning judgment process is performed to obtain the corresponding gear breakage warning result, including: If an external alarm message is received, and the feature quantity of the single-tooth broken tooth diagnosis feature or the tooth root crack diagnosis feature in the single-sample fault diagnosis result accounts for a feature quantity ratio of the total number of results of the single-sample fault diagnosis result that is greater than a second preset fault feature ratio, a gear breakage warning is issued; wherein the second preset fault feature ratio is less than the first preset fault feature ratio; Or, if no external alarm information is received, and it is detected that the meshing spectrum impact trend result is an upward trend result and the gear impact continuity statistic is greater than a preset statistic threshold, a gear breakage warning is issued.

12. A gear breakage identification device, characterized in that: include: A single sample diagnosis module is used to perform single sample fault diagnosis on the impact single sample data of each gear measuring point of the current gear to obtain a single sample fault diagnosis result; A trend analysis module, used for performing trend analysis of gear impact and meshing spectrum impact based on each of the impact single sample data, so as to obtain a gear impact trend result and a meshing spectrum impact trend result respectively, and to obtain a gear impact continuity statistic of each of the impact single sample data; a fault identification module, for determining whether the current gear has a gear breakage fault according to the current health level of the current gear, the single sample fault diagnosis result, the gear impact trend result, the meshing spectrum impact trend result and the gear impact continuity statistic, so as to obtain a corresponding gear breakage identification result; The fault warning module is used to perform a gear breakage warning judgment process to obtain a corresponding gear breakage warning result if the gear breakage identification result is that there is no gear breakage fault.

13. An electronic device, characterized in that: include: Memory, used to store computer programs; A processor is used to execute the computer program to implement the steps of the gear breakage identification method according to any one of claims 1 to 11.

14. A computer-readable storage medium, characterized in that: Used to store computer programs; wherein, when the computer program is executed by a processor, the steps of the gear breakage identification method according to any one of claims 1 to 11 are implemented.

Images