A method for intelligent diagnosis of gear failure of a deceleration system

By collecting and processing speed, load, and vibration data in the deceleration system, meshing frequency drift values ​​and load impact energy correction values ​​are generated to correct the transmission stage order data. This solves the problem of the influence of speed fluctuations and load impacts on gear fault diagnosis and improves the accuracy of fault type identification.

CN123111685APending Publication Date: 2026-11-17DONGGUAN FUTENG MOTOR CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611261636.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-19
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

In a deceleration system, speed fluctuations and load impacts cause shifts in the meshing order position and changes in order vibration caused by load impacts in gear fault diagnosis, affecting the accuracy of fault type identification.

Method used

By collecting speed, load, and vibration data, meshing frequency drift values ​​and load impact energy correction values ​​are generated to correct the transmission stage order data. Combined with the order analysis section and impact influence range, gear fault type identification is performed.

Benefits of technology

Maintaining continuous alignment and impact correction of gear fault order characteristics under speed fluctuations and load impacts improves the accuracy of gear fault type identification.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123111685A_ABST
    Figure CN123111685A_ABST
Patent Text Reader

Abstract

The application discloses a kind of gear fault intelligent diagnosis methods of deceleration system, it is related to the technical field of fault diagnosis, comprising: collection rotation speed data, load data, vibration data and transmission structure data, the transmission structure data includes transmission stage number, gear pair number and tooth number data;Based on tooth number data and rotation speed data, the meshing frequency drift value corresponding to each transmission stage number is generated, and the vibration data is aligned by the meshing frequency drift value Order, generate transmission stage order data;From load data, extract load impact time, and the energy difference of vibration data before and after load impact time is generated, and the load impact energy correction value is generated;According to load impact energy correction value, correct transmission stage order data, generate gear fault stable order data;The application keeps the continuous alignment and impact correction of gear fault order characteristics under the joint action of rotation speed fluctuation and load impact, improves the accuracy of gear fault type identification result.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of fault diagnosis technology, and specifically to an intelligent diagnostic method for gear faults in a reduction system. Background Technology

[0002] Speed ​​reduction systems are widely used in transmission equipment, production lines and electromechanical actuators. Gear fault diagnosis usually relies on the analysis of operating signals such as vibration response, speed and load to determine the meshing order and sideband changes. It can reflect the operating status of the gear pair without stopping the machine and provide data basis for judging faults such as wear, cracks and broken teeth.

[0003] Currently, during the variable speed operation of the deceleration system and the diagnosis of multi-stage transmission linkage, continuous fluctuations in rotational speed will change the gear meshing frequency, causing the order position under the same transmission stage number to drift. If it is still treated as a fixed speed or discrete frequency segment, it is easy to cause misalignment of the meshing order position and the sideband order position, thereby affecting the comparability of vibration characteristics under different transmission stage numbers.

[0004] Secondly, during sudden load changes, impact loading, or operating condition switching, load impacts can increase local order vibration values ​​and superimpose them with the order vibration changes caused by gear faults. If the impact range is not distinguished, the energy changes caused by load impacts can easily be misjudged as gear fault characteristics, thus affecting the accuracy of gear fault type identification results. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides an intelligent diagnostic method for gear faults in a reduction system.

[0006] A method for intelligent diagnosis of gear faults in a reduction system, comprising:

[0007] Collect rotational speed data, load data, vibration data, and transmission structure data. The transmission structure data includes transmission stage number, gear pair number, and number of teeth data.

[0008] Based on tooth count data and rotational speed data, the meshing frequency drift value corresponding to each transmission stage number is generated, and the vibration data is aligned by order using the meshing frequency drift value to generate transmission stage order data.

[0009] Extract the load impact time from the load data, and perform energy difference analysis on the vibration data before and after the load impact time to generate a load impact energy correction value.

[0010] The transmission stage order data is corrected based on the load impact energy correction value to generate stable order data for gear faults. The stable order data for gear faults is then matched with the gear fault sample data to generate gear fault type identification results.

[0011] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0012] This invention uses effective meshing frequency data covering all speed sampling times to enable the meshing frequency drift value to continuously reflect the gear meshing frequency changes under speed fluctuations, and to maintain a continuous correspondence between the cumulative meshing phase value and the order resampled vibration data. This makes the transmission stage order data comparable under different transmission stage numbers and gear pair numbers, thereby reducing the interference of speed changes on the identification of meshing order position and sideband order position.

[0013] Furthermore, this invention corrects the changes in order vibration values ​​caused by load impact by using load impact energy correction values ​​and load impact energy positive correction values. It also limits the correction range by combining the order analysis section, impact order position, and impact influence order section, so that the stable order data of gear faults retains fault characteristics that match the sample fault order position and sample order vibration values, thereby improving the accuracy of gear fault type identification results.

[0014] In summary, this invention can maintain continuous alignment and impact correction of gear fault order characteristics under the combined effects of speed fluctuations and load impacts, thereby improving the accuracy of gear fault type identification results. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0016] Figure 1 The flowchart of an intelligent diagnostic method for gear faults in a deceleration system provided by the present invention is shown. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Please see Figure 1 As shown in the figure, this embodiment discloses an intelligent diagnostic method for gear faults in a reduction system, the method comprising:

[0019] S11: Collect rotational speed data, load data, vibration data, and transmission structure data. The transmission structure data includes transmission stage number, gear pair number, and number of teeth data.

[0020] Specifically, the rotational speed data includes the rotational speed sampling time, the rotational speed measurement point number, the gear shaft number corresponding to the rotational speed measurement point, and the rotational speed sampling value; the load data includes the load sampling time, the load measurement point number, and the load sampling value; the vibration data includes the vibration sampling time, the vibration measurement point number, and the vibration acceleration value; and the tooth count data includes the number of teeth of the driving gear and the number of teeth of the driven gear.

[0021] The reduction system is a multi-stage gear transmission reducer, and the speed data is collected by a speed sensor installed on the input shaft, intermediate shaft or output shaft;

[0022] The speed sampling time is the time point at which the speed sensor outputs the speed sampling value. The speed measurement point number is the installation measurement point number of the speed sensor. The gear shaft number corresponding to the speed measurement point is the gear shaft number measured by the speed sensor. The speed sampling value is the rotational speed of the gear shaft corresponding to the speed measurement point at the speed sampling time. The load data is collected by the torque sensor, support force sensor, or motor current conversion unit. The load sampling time is the time point at which the load sampling value is collected. The load measurement point number is the collection position number corresponding to the load sampling value. The load sampling value is the torque value, support reaction force value, or equivalent load value.

[0023] Vibration data is collected by accelerometers installed in the gearbox housing, bearing housing, or near the transmission stage. The vibration sampling time is the time point at which the vibration acceleration value is collected, the vibration measurement point number is the installation measurement point number of the accelerometer, and the vibration acceleration value is the acceleration response value at the corresponding vibration sampling time.

[0024] Transmission structure data is obtained from the deceleration system assembly file, equipment ledger or manual configuration file. The transmission stage number is used to distinguish different transmission stages in the deceleration system, and the gear pair number is used to distinguish different meshing gear pairs in the same transmission stage.

[0025] The tooth count data records the number of teeth of the driving gear and the driven gear participating in the meshing under the gear pair number. The gear pair number also records the driving gear, driven gear, and the gear shaft where the driving gear and driven gear are located. The gear shaft number corresponding to the speed measurement point and the gear pair number together determine the gear participating in the meshing corresponding to the speed sampling value. The load measurement point number and vibration measurement point number are assigned to the corresponding transmission stage number according to the installation position.

[0026] S12: Based on tooth count data and rotational speed data, generate meshing frequency drift values ​​corresponding to each transmission stage number, and align the vibration data by order using the meshing frequency drift values ​​to generate transmission stage order data;

[0027] Specifically, the steps for generating the meshing frequency drift value corresponding to each transmission stage number are as follows:

[0028] S121: Determine the number of gear teeth involved in meshing on the gear shaft corresponding to the speed measuring point according to the transmission stage number and gear pair number, and use the number of gear teeth as the meshing tooth value;

[0029] In one specific embodiment, the transmission stage number is represented as follows: ,in, The serial number is the number of the transmission stage; the gear pair number is represented as... ,in, This is the serial number for the gear pair, specifically for the transmission stage number. and gear pair number Read the gear pair number The corresponding number of teeth on the driving gear and driven gear is recorded. Simultaneously, the gear shaft number corresponding to the speed measuring point is read. When the gear shaft number corresponding to the speed measuring point matches the gear shaft of the driving gear, the number of teeth on the driving gear is used as the meshing tooth value. When the gear shaft number corresponding to the speed measuring point matches the gear shaft of the driven gear, the number of teeth on the driven gear is used as the meshing tooth value. The meshing tooth value is expressed as follows: ,in, Numbering of transmission stages and gear pair number The corresponding meshing tooth values, the number of teeth of the driving gear and the number of teeth of the driven gear are all obtained from the tooth count data.

[0030] S122: Based on the meshing tooth values ​​and the speed sampling values ​​under the same transmission stage number, calculate the gear meshing frequency values ​​at each speed sampling time and generate real-time meshing frequency data;

[0031] Specifically, the steps for generating real-time meshing frequency data are as follows:

[0032] S122.1: Convert the speed sampling value into the gear shaft rotation frequency value, and multiply the gear shaft rotation frequency value by the meshing tooth value to obtain the gear meshing frequency value;

[0033] In a specific embodiment, the rotational speed sampling time is represented as: ,in, Let be the sequence number of the rotational speed sampling time, and let represent the rotational speed sampling value as follows: ,in, Numbering of transmission stages At the speed sampling time The speed sampling value, when the speed sampling value When the unit is revolutions per minute, the gear shaft rotation frequency is expressed as:

[0034]

[0035] in, Numbering of transmission stages At the speed sampling time The gear shaft rotation frequency value is 60, which is a time conversion constant between minutes and seconds.

[0036] The gear meshing frequency value is expressed as:

[0037]

[0038] in, Numbering of transmission stages Gear pair number At the speed sampling time The gear meshing frequency value is as follows. Numbering of transmission stages At the speed sampling time The rotational frequency of the gear shaft below, Numbering of transmission stages and gear pair number The corresponding meshing tooth value.

[0039] S122.2: Arrange gear meshing frequency values ​​according to transmission stage number, gear pair number and speed sampling time to generate real-time meshing frequency data.

[0040] In one specific embodiment, the gear meshing frequency value According to the transmission stage number Gear pair number and rotational speed sampling time The gears are arranged in a series of units, each including the transmission stage number, gear pair number, speed sampling time, and gear meshing frequency value. The arranged gear meshing frequency values ​​generate real-time meshing frequency data.

[0041] S123: The gear meshing frequency value in the real-time meshing frequency data is subtracted from the reference meshing frequency value of the corresponding transmission stage number and gear pair number, and the absolute value is taken to obtain the meshing frequency deviation value. The continuous speed sampling time with the meshing frequency deviation value greater than or equal to the preset meshing frequency deviation judgment value is divided into candidate meshing frequency drift segments. Candidate meshing frequency drift segments with a duration less than the preset minimum duration are eliminated. Adjacent interpolation is performed on the gear meshing frequency value corresponding to the eliminated speed sampling time to generate effective meshing frequency data covering all speed sampling times.

[0042] In one specific embodiment, the reference engagement frequency value is expressed as: ,in, Numbering of transmission stages and gear pair number The corresponding reference meshing frequency value is obtained by averaging the gear meshing frequency values ​​collected by the reduction system under stable operating conditions, or by converting the rated speed and number of teeth data. The deviation value of the meshing frequency is expressed as:

[0043]

[0044] in, Numbering of transmission stages Gear pair number At the speed sampling time The deviation of the meshing frequency is as follows. This refers to the gear meshing frequency value in the real-time meshing frequency data. Numbering of transmission stages and gear pair number The corresponding reference meshing frequency value.

[0045] The preset meshing frequency deviation judgment value is expressed as: ,in, Based on historical operational experimental data, when the meshing frequency deviates from the value... Greater than or equal to the preset meshing frequency deviation judgment value When the corresponding rotational speed sampling time is taken as the deviation sampling time, the consecutive deviation sampling times are divided into candidate meshing frequency drift segments.

[0046] The duration of the candidate meshing frequency drift segment is expressed as:

[0047]

[0048] in, The duration of the candidate meshing frequency drift segment. The sampling time of the first rotational speed in the candidate meshing frequency drift segment. The minimum duration of the sampling time for the last rotational speed in the candidate meshing frequency drift segment is expressed as follows: ,in, The settings are based on historical experimental data.

[0049] Remove duration Less than the preset minimum duration The candidate meshing frequency drift segment is identified, and the rotational speed sampling time within the eliminated candidate meshing frequency drift segment is taken as the eliminated rotational speed sampling time. Adjacent interpolation is performed on the gear meshing frequency value corresponding to the eliminated rotational speed sampling time. The gear meshing frequency value after adjacent interpolation is expressed as:

[0050]

[0051] in, The sampling time of the speed that was rejected The corresponding interpolated gear meshing frequency value, For the time of rotational speed sampling The most recent rotational speed sampling time that was not previously removed. For the time of rotational speed sampling The most recent rotational speed sampling time that has not been removed. The time for sampling rotational speed The corresponding gear meshing frequency value, The time for sampling rotational speed The corresponding gear meshing frequency value.

[0052] When there is no non-rejected speed sampling time before the rejected speed sampling time, the gear meshing frequency value corresponding to the nearest non-rejected speed sampling time after the rejected speed sampling time is used as the interpolated gear meshing frequency value. When there is no non-rejected speed sampling time after the rejected speed sampling time, the gear meshing frequency value corresponding to the nearest non-rejected speed sampling time before the rejected speed sampling time is used as the interpolated gear meshing frequency value. When there are no non-rejected speed sampling times before or after the rejected speed sampling time, the reference meshing frequency value corresponding to the transmission stage number and gear pair number is used as the interpolated gear meshing frequency value.

[0053] The gear meshing frequency values ​​corresponding to the speed sampling times that were not rejected and the interpolated gear meshing frequency values ​​corresponding to the speed sampling times that were rejected are written into the transmission stage number, gear pair number and speed sampling time to generate effective meshing frequency data covering all speed sampling times.

[0054] S124: Subtract the gear meshing frequency value in the effective meshing frequency data from the reference meshing frequency value of the corresponding transmission stage number and gear pair number to obtain the meshing frequency drift value corresponding to each transmission stage number.

[0055] In a specific embodiment, the meshing frequency drift value is expressed as:

[0056]

[0057] in, Numbering of transmission stages Gear pair number At the speed sampling time The meshing frequency drift value is below. This refers to the gear meshing frequency value in the effective meshing frequency data. Numbering of transmission stages and gear pair number The corresponding reference meshing frequency value.

[0058] Specifically, the steps for generating transmission stage order data are as follows:

[0059] S125: According to the vibration sampling time and the rotational speed sampling time, the vibration acceleration value and the rotational speed sampling value are registered at the same time, and the vibration acceleration value with a sampling time difference exceeding the preset synchronization error time is removed to generate vibration and rotational speed synchronization data.

[0060] In a specific embodiment, the vibration sampling time is represented as: ,in, Let be the sequence number of the vibration sampling time, and let represent the vibration acceleration value as follows: ,in, Vibration sampling time The vibration acceleration value at each vibration sampling time. From the time of rotational speed sampling Screening and vibration sampling time The rotational speed sampling time with the smallest difference is represented by the sampling time difference as follows:

[0061]

[0062] in, Vibration sampling time With rotational speed sampling time The difference in sampling time between them, with the preset synchronization error duration expressed as: ,in, Based on the sampling frequency settings of the rotational speed data and vibration data, the difference between sampling times is... Less than or equal to the preset synchronization error duration At that time, the vibration sampling time will be... Vibration acceleration value With rotational speed sampling time Speed ​​sampling value Perform simultaneous registration, when the sampling time difference Greater than the preset synchronization error duration At that time, vibration sampling times were excluded. Vibration acceleration value The retained vibration acceleration values ​​and corresponding rotational speed sampling values ​​are used to generate vibration rotational speed synchronization data.

[0063] S126: Input the vibration speed synchronization data and meshing frequency drift value into the order resampling function, correct the meshing phase cumulative value according to the meshing frequency drift value, and resample the vibration acceleration value according to the corrected meshing phase cumulative value to generate order resampled vibration data.

[0064] In one specific embodiment, the input to the order resampling function is the vibration speed synchronization data and the meshing frequency drift value, and the output of the order resampling function is the order resampled vibration data. First, the corrected gear meshing frequency value is calculated based on the reference meshing frequency value and the meshing frequency drift value. The corrected gear meshing frequency value is expressed as:

[0065]

[0066] in, Numbering of transmission stages Gear pair number At the speed sampling time The corrected gear meshing frequency value. The reference meshing frequency value, This represents the meshing frequency drift value.

[0067] The cumulative value of the meshing phase is calculated based on the corrected gear meshing frequency value. The cumulative value of the meshing phase is expressed as follows:

[0068]

[0069] in, Numbering of transmission stages Gear pair number Up to the speed sampling time The cumulative value of the meshing phase, The previous rotational speed sampling time The corresponding cumulative value of the meshing phase, This is the phase constant corresponding to one meshing cycle. This is the corrected gear meshing frequency value. The time interval between adjacent rotational speed sampling times.

[0070] The preset number of resampling cycles is expressed as: ,in, Based on the order resolution setting, the meshing phase sampling interval is expressed as:

[0071]

[0072] in, The sampling interval for the engagement phase. This is the preset number of resampling cycles.

[0073] According to the sampling interval of the meshing phase Cumulative value during meshing phase The sampling positions of the equal meshing phase are generated, and the vibration acceleration values ​​in the vibration speed synchronization data are linearly interpolated to obtain the order resampled vibration data. The order resampled vibration data includes the transmission stage number, gear pair number, order resampled point, vibration sampling time, and vibration acceleration value. The correspondence between the order resampled point and the vibration sampling time is written into the order resampled vibration data according to the time position before and after the linear interpolation.

[0074] S127: Divide the vibration data into order resampling points according to the preset order analysis points, perform order spectrum conversion on the vibration data in each order analysis section, and calibrate the meshing order position and sideband order position according to the transmission stage number, gear pair number and tooth count data to obtain the transmission stage order calibration data.

[0075] In a specific embodiment, the preset order analysis point number is represented as: ,in, Based on the required order resolution and vibration data sampling frequency for gear fault diagnosis, and according to the preset number of analysis points... The order resampling vibration data is divided into multiple order analysis segments. Each order analysis segment includes a start time and an end time. The start time of the order analysis segment is the vibration sampling time corresponding to the first order resampling point in the order analysis segment, and the end time of the order analysis segment is the vibration sampling time corresponding to the last order resampling point in the order analysis segment.

[0076] Numbering the same transmission stage Same gear pair number The order spectrum transformation was performed on the resampled vibration data within the same order analysis segment. The order spectrum transformation used discrete Fourier transform, and the order vibration values ​​were expressed as follows:

[0077]

[0078] in, Numbering of transmission stages Gear pair number Order analysis section and order position The corresponding order vibration value, For the first Each order of analysis segment, Numbering of transmission stages Gear pair number and order analysis section Inner The vibration acceleration values ​​corresponding to each order of resampling point. To preset the number of analysis points for each order, Numbering of the order spectrum, It is the imaginary unit.

[0079] The order position is represented as:

[0080]

[0081] in, For the first Each order position To preset the number of resampling cycles, The number of analysis points for the preset order.

[0082] The meshing order position is represented as:

[0083]

[0084] in, Numbering of transmission stages and gear pair number The corresponding meshing order position, the value 1 represents a meshing cycle in the meshing phase domain.

[0085] The order position of the sideband is represented as:

[0086]

[0087] in, Numbering of transmission stages and gear pair number The corresponding sideband order position, This is the sideband offset number, which is a non-zero integer. The range of values ​​for the meshing teeth and the sideband offset number is set according to the number of sidebands to be detected.

[0088] The order vibration values ​​corresponding to each order position after order spectrum conversion are retained, and the meshing order position and sideband order position are marked. The transmission stage number, gear pair number, order analysis section, start time of the order analysis section, end time of the order analysis section, order position, order vibration value and corresponding order position mark are written into the data to obtain the transmission stage order calibration data.

[0089] S128: Collect the order vibration values ​​in the transmission stage order calibration data according to the order analysis section and order position, and write the start time and end time of the order analysis section into the collection result to generate transmission stage order data.

[0090] In a specific embodiment, the order vibration values ​​in the transmission stage order calibration data are collected according to the transmission stage number, gear pair number, order analysis section, and order position. The transmission stage number, gear pair number, order analysis section, start time of the order analysis section, end time of the order analysis section, order position, and order vibration value are written in the corresponding data to generate transmission stage order data.

[0091] S13: Extract the load impact time from the load data, and perform energy difference analysis on the vibration data before and after the load impact time to generate the load impact energy correction value.

[0092] Specifically, the steps for generating the load impact energy correction value are as follows:

[0093] S131: Arrange the load sampling values ​​according to the load sampling time, calculate the difference between adjacent load sampling values, and divide the difference result by the time interval between adjacent load sampling times to obtain the load difference per unit time.

[0094] In a specific embodiment, the load sampling time is represented as ,in, Let be the sequence number of the load sampling time, and let the load sample value be expressed as . ,in, For the load sampling time The load sampled values ​​are based on the load sampling time. Load sampling values ​​arranged from morning to night The difference between adjacent load samples is calculated, and the result is divided by the time interval between adjacent load sampling times to obtain the load difference per unit time. The load difference per unit time is expressed as:

[0095]

[0096] in, For the load sampling time Compared to the previous load sampling time The load difference per unit time, For the load sampling time The load sampling value under the following conditions The previous load sampling time The load sampling value under the following conditions This represents the time interval between adjacent load sampling times.

[0097] S132: Compare the load difference per unit time with the preset load impact judgment value to obtain the candidate load impact time. Based on the load sampling values ​​before and after the candidate load impact time, obtain the load increase value before impact and the load decrease value after impact, and filter to obtain the load impact time.

[0098] In a specific embodiment, the preset load impact judgment value is expressed as: ,in, The load change judgment value is set based on historical operational experimental data, and the load difference per unit time is... Impact judgment value of preset load Compare the load differences per unit time. Greater than or equal to the preset load impact judgment value At that time, the load sampling time will be As candidate load impact moments, load sampling values ​​before and after the candidate load impact moments are read to calculate the load increase before impact and the load decrease after impact. The load impact moments are then selected by filtering the load increase before impact and the load decrease after impact.

[0099] Specifically, the steps for selecting the impact time of the load are as follows:

[0100] S132.1: The load sampling time when the load difference per unit time is greater than or equal to the preset load impact judgment value is taken as the candidate load impact time;

[0101] In a specific embodiment, the candidate load impact time is represented as: ,in, This refers to the sequence number of the candidate load impact time, when the load sampling time is... Corresponding load difference per unit time satisfy At that time, the load sampling time will be As the impact moment of the candidate load ,in, The preset load impact judgment value is set based on historical operational test data.

[0102] S132.2: Read the load sampling values ​​of multiple consecutive load sampling times before the impact time of the candidate load, and subtract the smallest load sampling value read from the load sampling value corresponding to the impact time of the candidate load to obtain the load increase value before impact;

[0103] In a specific embodiment, the preset number of pre-impact reads is expressed as: ,in, Based on the load data sampling frequency and the impact duration of the deceleration system, the impact time of the candidate load is read. Previous consecutive The load sample values ​​at each load sampling time are used to select the smallest load sample value among the read load sample values, which is then used as the candidate load impact time. Subtracting the minimum load sample value from the corresponding load sample value yields the pre-impact load increase, which is expressed as:

[0104]

[0105] in, Impact time of candidate load The corresponding increase in pre-impact load, Impact time of candidate load The load sampling value under the following conditions to Impact time of candidate load Previous consecutive Load sample values ​​at each load sampling time. This is the preset number of reads before the impact.

[0106] S132.3: Read the load sampling values ​​of multiple consecutive load sampling times after the candidate load impact time, subtract the minimum load sampling value read from the load sampling value corresponding to the candidate load impact time to obtain the load fall value after impact, filter the candidate load impact time where the load increase value before impact is greater than or equal to the preset load increase judgment value and the load fall value after impact is greater than or equal to the preset load recovery judgment value, and obtain the load impact time.

[0107] In a specific embodiment, the preset number of reads after the impact is expressed as: ,in, Based on the load data sampling frequency and the deceleration system impact recovery time settings, the impact time of the candidate load is read. Then continuous The load sample values ​​at each load sampling time are used to select the smallest load sample value among the read load sample values, which is then used as the candidate load impact time. Subtracting the minimum load sample value from the corresponding load sample value yields the load drop value after impact, which is expressed as:

[0108]

[0109] in, Impact time of candidate load The corresponding post-impact load drop value, Impact time of candidate load The load sampling value under the following conditions to Impact time of candidate load Then continuous Load sample values ​​at each load sampling time. The preset number of readings after impact is represented by the preset load increase judgment value. The preset load recovery judgment value is expressed as: The preset load increase judgment value and preset load recovery judgment value are set based on historical operational experimental data. At the moment of candidate load impact... Simultaneously satisfy and At that time, the impact time of the candidate load will be... As the moment of load impact.

[0110] S133: According to the preset sampling duration, the vibration acceleration values ​​before and after the load impact time are extracted respectively to obtain the vibration data before impact and the vibration data after impact.

[0111] In a specific embodiment, the moment of load impact is expressed as: ,in, The sequence number of the load impact moment is denoted as , and the vibration sampling moment is represented as . ,in, Let be the sequence number of the vibration sampling time, and let represent the vibration acceleration value as follows: ,in, Vibration sampling time The vibration acceleration value is expressed as the preset sampling time. ,in, Based on the vibration data sampling frequency and gear meshing cycle settings, the reading meets the requirements. Vibration sampling time The corresponding vibration acceleration value is used to obtain the vibration intercept data before the impact, and the data is read to meet the requirements. Vibration sampling time The corresponding vibration acceleration value is used to obtain the vibration data after the impact.

[0112] S134: The vibration acceleration values ​​in the pre-impact vibration data and post-impact vibration data are squared and summed to obtain the pre-impact vibration energy value and the post-impact vibration energy value. The post-impact vibration energy value under the same transmission stage number is subtracted from the pre-impact vibration energy value to generate the load impact energy correction value.

[0113] In a specific embodiment, the vibration acceleration value in the pre-impact vibration intercept data is represented as: ,in, The sequence number is the vibration acceleration value in the pre-impact vibration data segment, and the vibration acceleration value in the post-impact vibration data segment is represented as... ,in, To determine the sequence number of the vibration acceleration value in the post-impact vibration data, the vibration acceleration values ​​in the pre-impact and post-impact vibration data are squared and summed to obtain the pre-impact and post-impact vibration energy values, which are expressed as follows:

[0114]

[0115]

[0116] in, At the moment of load impact The corresponding pre-impact vibration energy value, At the moment of load impact The corresponding post-impact vibration energy value, This represents the number of vibration acceleration values ​​extracted from the pre-impact vibration data. To determine the number of vibration acceleration values ​​in the post-impact vibration data, the load measurement point number and vibration measurement point number are assigned to the corresponding transmission stage number according to their installation location. The post-impact vibration energy value under the same transmission stage number is subtracted from the pre-impact vibration energy value to generate the load impact energy correction value, which is expressed as:

[0117]

[0118] in, At the moment of load impact The corresponding load impact energy correction value.

[0119] The positive correction value for load impact energy is expressed as:

[0120]

[0121] in, At the moment of load impact The corresponding positive correction value for load impact energy, At the moment of load impact The corresponding load impact energy correction value.

[0122] S14: Correct the transmission stage order data according to the load impact energy correction value, generate stable order data of gear faults, and match the stable order data of gear faults with the gear fault sample data to generate gear fault type identification results.

[0123] Specifically, the gear fault sample data includes the sample transmission stage number, sample gear pair number, sample fault order location, sample fault order vibration value, and sample fault type. The steps for generating gear fault type identification results are as follows:

[0124] S141: Based on the load impact time, filter the order analysis segments where the start time is less than or equal to the load impact time and the end time is greater than the load impact time. From the meshing order position and sideband order position corresponding to the selected order analysis segments, select the order position with the largest order vibration value to obtain the impact order position, and write the load impact energy correction value into the impact order position.

[0125] In a specific embodiment, the moment of load impact is expressed as: The start time of the order analysis segment is represented as The end time of the order analysis segment is represented as ,in, The sequence number of the analysis segment is used to filter those that meet the requirements. The selected order analysis segments are used as the corresponding order analysis segments for impact. The order vibration values ​​corresponding to the meshing order position and the sideband order position are read from these segments. The order position with the largest order vibration value is selected as the impact order position. The impact order position is represented as... Correction value of load impact energy Write the impact order position The impact order position carrying the impact energy correction value is obtained.

[0126] S142: Using the impact order position as the center, select continuous order positions that meet the preset impact judgment range from the transmission stage order data to obtain the impact influence order segment.

[0127] In one specific embodiment, the impact order position Centered on the transmission stage order data, the position located at the impact order is read from the transmission stage order data. Before and after the order position, the read order position is used as the candidate impact order position. The order impact judgment value corresponding to each candidate impact order position is obtained through the impact extension judgment function. The candidate impact order positions whose order impact judgment values ​​fall within the preset impact judgment range are selected as impact-affected order positions. The segment with consecutive impact-affected order positions is divided into impact-affected order segments.

[0128] Specifically, the steps to obtain the impact order segment are as follows:

[0129] S142.1: Using the impact order position as the center, read the adjacent order positions before and after the impact order position to obtain the candidate impact order positions;

[0130] In a specific embodiment, the candidate impact order position is represented as: ,in, The sequence number is the position of the candidate impact order, and the preset order reading width value is represented as follows. ,in, Based on the order resolution of the transmission stage order data and the duration of the load impact, the impact order position is set. Centered on the impact, within the corresponding order analysis section, read the values ​​that satisfy the following conditions. The order position is used to obtain the candidate impact order position. .

[0131] S142.2: Input the order vibration value and load impact energy positive correction value corresponding to the candidate impact order position into the impact expansion judgment function to obtain the order impact judgment value corresponding to each candidate impact order position;

[0132] In a specific embodiment, candidate impact order position The corresponding order vibration value is expressed as The positive correction value for load impact energy is expressed as: Since the order vibration value and the positive correction value of the load impact energy have different values, before inputting the order vibration value and the positive correction value of the load impact energy into the impact expansion judgment function, the order vibration value and the positive correction value of the load impact energy are normalized respectively to obtain the normalized value of the order vibration and the normalized value of the load impact energy.

[0133] The normalized value of the order vibration is expressed as:

[0134]

[0135] in, Candidate impact order position The corresponding normalized value of the order vibration, Candidate impact order position The corresponding order vibration value, This represents the lower bound of the normalized order vibration. The upper limit of the normalized order vibration value is defined as the upper limit of the normalized order vibration value. The lower limit and upper limit of the normalized order vibration value are set based on historical experimental data. Greater than The range of normalized values ​​for order vibrations is: .

[0136] The normalized value of the load impact energy is expressed as:

[0137]

[0138] in, This is the normalized value of the load impact energy. This is a positive correction value for the load impact energy. This represents the normalized lower limit of the load impact energy. The normalized upper limit of the load impact energy is defined as the normalized lower limit and upper limit of the load impact energy, which are set based on historical operational experimental data. Greater than The normalized value of the load impact energy ranges from 100 to 100. .

[0139] The impact propagation determination function is expressed as:

[0140]

[0141] in, Candidate impact order position The corresponding order of impact determination value, Candidate impact order position The corresponding normalized value of the order vibration, This is the normalized value of the load impact energy. Position of candidate impact order For the impact order position, The preset order read width value is a value greater than 0.

[0142] S142.3: Compare the order impact judgment value with the preset impact judgment threshold, filter out candidate impact order positions where the order impact judgment value is greater than or equal to the preset impact judgment threshold, obtain the impact-affected order positions, and divide the segment with continuous impact-affected order positions into impact-affected order segments.

[0143] In a specific embodiment, the preset impact judgment threshold is expressed as: , Based on historical experimental data, the setting is 0 < <1;

[0144] When the candidate impact order position Corresponding order impact judgment value satisfy At that time, the candidate impact order position As the impact-affected order positions, the impact-affected order positions are arranged in ascending order, and the segments with consecutive impact-affected order positions are divided into impact-affected order segments.

[0145] It should be noted that: the order position of continuous impact is the order interval between adjacent impact positions that is less than or equal to the preset continuous order interval value, which is set according to the order resolution of the transmission stage order data.

[0146] S143: Input the order vibration value and the positive correction value of the load impact energy in the impact influence order section into the impact energy correction function to obtain the corrected order vibration value. Write the corrected order vibration value into the impact influence order section and retain the transmission stage order data outside the impact influence order section to generate gear fault stability order data.

[0147] In a specific embodiment, the order vibration value in the impact-affected order segment is expressed as: ,in, The order position number in the impact influence zone is represented by the positive correction value of the load impact energy. The input to the impact energy correction function is the order vibration value in the impact influence order segment and the positive correction value of the load impact energy. The output of the impact energy correction function is the corrected order vibration value, which is the normalized value of the load impact energy obtained in S142.2. The order vibration values ​​in the impact-affected order segment are proportionally corrected, and the corrected order vibration values ​​are expressed as follows:

[0148]

[0149] in, To correct for the order vibration value, The order vibration values ​​in the impact-affected segment, This is the normalized value of the load impact energy. The impact correction weight value is set based on historical experimental data, and the impact correction weight value is greater than or equal to 0 and less than or equal to 1.

[0150] When the load impact energy correction value When the value is less than or equal to 0, the order vibration value in the impact influence order section is used as the corrected order vibration value. The corrected order vibration value is written into the impact influence order section, and the transmission stage order data outside the impact influence order section is retained to generate gear fault stability order data.

[0151] S144: Filter gear fault sample data according to transmission stage number and gear pair number, match the order position and order vibration value in the stable order data of gear fault with the order position and order vibration value of sample fault respectively, filter sample fault types, and generate gear fault type identification results.

[0152] In one specific embodiment, gear fault sample data with matching transmission stage number and gear pair number are selected from the gear fault sample data. The selected gear fault sample data is then used as the gear fault sample data for matching. For each gear fault sample data to be matched, the order position and vibration value of the sample fault are read. The order position closest to the order position of the sample fault is selected from the stable order data of the gear fault. The selected order position is used as the matching order position, and the vibration value corresponding to the matching order position is read. The order position difference is obtained by subtracting the order position from the sample fault order position and taking the absolute value. The vibration value corresponding to the matching order position is obtained by subtracting the vibration value of the sample order position and taking the absolute value. The order position difference and the vibration difference are normalized respectively to obtain the normalized order position difference and the normalized order vibration difference.

[0153] The order position normalized difference is expressed as:

[0154]

[0155] in, For the first The order-normalized difference of the position corresponding to each gear fault sample data participating in the matching. This represents the matching order position in the stable order data of gear faults that is closest to the fault order position in the sample. The fault order position of the sample. This refers to the normalized range value for the order position. The normalized range value is set based on the range of order position values ​​for sample faults in the gear fault sample data, and the normalized range value is greater than 0. The range of the normalized difference for the order position is... .

[0156] The normalized difference of order vibrations is expressed as:

[0157]

[0158] in, For the first The normalized difference of the order vibration corresponding to the gear fault sample data participating in the matching. To match the order vibration value corresponding to the order position, The sample vibration order value, This refers to the normalized range value for order vibration. The normalized range value is set based on the range of order vibration values ​​in the gear fault sample data, and it must be greater than 0. The range of the normalized difference for order vibration is... .

[0159] The sample matching difference is expressed as:

[0160]

[0161] in, For the first The sample matching difference corresponding to each gear fault sample data participating in the matching. For the first The order-normalized difference of the position corresponding to each gear fault sample data participating in the matching. For the first The normalized difference of the order vibration corresponding to the gear fault sample data participating in the matching. Match weight values ​​to order positions. The order vibration matching weight value, the order position matching weight value, and the order vibration matching weight value are set based on historical experimental data. ,and and Both are greater than 0 and less than 1;

[0162] The preset matching judgment value is set based on historical experimental data. Gear fault sample data with the smallest sample matching difference and a sample matching difference less than or equal to the preset matching judgment value are selected. The sample fault type, transmission stage number and gear pair number in the selected gear fault sample data are written to generate gear fault type identification results.

[0163] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A method for intelligent diagnosis of gear faults in a reduction system, characterized in that, The method includes: Collect rotational speed data, load data, vibration data, and transmission structure data. The transmission structure data includes transmission stage number, gear pair number, and number of teeth data. Based on tooth count data and rotational speed data, the meshing frequency drift value corresponding to each transmission stage number is generated, and the vibration data is aligned by order using the meshing frequency drift value to generate transmission stage order data. Extract the load impact time from the load data, and perform energy difference analysis on the vibration data before and after the load impact time to generate a load impact energy correction value. The transmission stage order data is corrected based on the load impact energy correction value to generate stable order data for gear faults. The stable order data for gear faults is then matched with the gear fault sample data to generate gear fault type identification results.

2. The intelligent diagnostic method for gear faults in a reduction system according to claim 1, characterized in that, The rotational speed data includes the rotational speed sampling time, rotational speed measuring point number, the gear shaft number corresponding to the rotational speed measuring point, and the rotational speed sampling value. The load data includes the load sampling time, load measuring point number, and load sampling value. The vibration data includes the vibration sampling time, vibration measuring point number, and vibration acceleration value. The transmission structure data also includes the installation position of the load measuring point number and the vibration measuring point number under the transmission stage number. The tooth count data includes the number of teeth of the driving gear and the number of teeth of the driven gear.

3. The intelligent diagnostic method for gear faults in a reduction system according to claim 2, characterized in that, The steps for generating the meshing frequency drift value corresponding to each transmission stage number are as follows: According to the transmission stage number and gear pair number, determine the number of gear teeth that participate in meshing on the gear shaft corresponding to the speed measuring point, and use the number of gear teeth as the meshing tooth value; Based on the meshing tooth values ​​and the speed sampling values ​​under the same transmission stage number, the gear meshing frequency values ​​at each speed sampling time are calculated to generate real-time meshing frequency data. The gear meshing frequency value in the real-time meshing frequency data is subtracted from the reference meshing frequency value of the corresponding transmission stage number and gear pair number, and the absolute value is taken to obtain the meshing frequency deviation value. The continuous speed sampling time with the meshing frequency deviation value greater than or equal to the preset meshing frequency deviation judgment value is divided into candidate meshing frequency drift segments. Candidate meshing frequency drift segments with a duration less than the preset minimum duration are eliminated. Adjacent interpolation is performed on the gear meshing frequency value corresponding to the eliminated speed sampling time to generate effective meshing frequency data covering all speed sampling times. The difference between the gear meshing frequency value in the effective meshing frequency data and the reference meshing frequency value of the corresponding transmission stage number and gear pair number is used to obtain the meshing frequency drift value corresponding to each transmission stage number.

4. The intelligent diagnostic method for gear faults in a reduction system according to claim 3, characterized in that, The steps for generating real-time meshing frequency data are as follows: The speed sampling value is converted into the gear shaft rotation frequency value, and the gear shaft rotation frequency value is multiplied by the meshing tooth value to obtain the gear meshing frequency value. The gear meshing frequency values ​​are arranged according to the transmission stage number, gear pair number, and speed sampling time to generate real-time meshing frequency data.

5. The intelligent diagnostic method for gear faults in a reduction system according to claim 3, characterized in that, The steps for generating transmission stage order data are as follows: Based on the vibration sampling time and the rotational speed sampling time, the vibration acceleration value and the rotational speed sampling value are registered at the same time. Vibration acceleration values ​​whose sampling time difference exceeds the preset synchronization error time are removed to generate vibration and rotational speed synchronization data. The vibration speed synchronization data and meshing frequency drift value are input into the order resampling function. The meshing phase cumulative value is corrected according to the meshing frequency drift value. The vibration acceleration value is then resampled based on the corrected meshing phase cumulative value to generate order resampled vibration data. The vibration data is divided into order resampling points according to the preset order analysis points. The order spectrum of the vibration data in each order analysis section is converted. The meshing order position and the sideband order position are calibrated according to the transmission stage number, gear pair number and tooth number data to obtain the transmission stage order calibration data. The order vibration values ​​in the transmission stage order calibration data are collected according to the order analysis section and order position, and the start and end times of the order analysis section are written into the collection results to generate transmission stage order data.

6. The intelligent diagnostic method for gear faults in a reduction system according to claim 2, characterized in that, The steps for generating the load impact energy correction value are as follows: The load sampling values ​​are arranged according to the load sampling time. The difference between adjacent load sampling values ​​is calculated, and the difference is divided by the time interval between adjacent load sampling times to obtain the load difference per unit time. The load difference per unit time is compared with the preset load impact judgment value to obtain the candidate load impact time. Based on the load sampling values ​​before and after the candidate load impact time, the load increase value before impact and the load decrease value after impact are obtained, and the load impact time is selected. According to the preset sampling time, the vibration acceleration values ​​before and after the load impact time are extracted to obtain the vibration data before impact and the vibration data after impact. The vibration acceleration values ​​in the pre-impact and post-impact vibration data are squared and summed to obtain the pre-impact and post-impact vibration energy values. The post-impact vibration energy value under the same transmission stage number is subtracted from the pre-impact vibration energy value to generate the load impact energy correction value.

7. The intelligent diagnostic method for gear faults in a reduction system according to claim 6, characterized in that, The steps for selecting the impact time of the load are as follows: The load sampling time at which the load difference per unit time is greater than or equal to the preset load impact judgment value is taken as the candidate load impact time. Read the load sampling values ​​of multiple consecutive load sampling times before the impact time of the candidate load, and subtract the smallest load sampling value read from the load sampling value corresponding to the impact time of the candidate load to obtain the load increase value before impact; Read the load sampling values ​​of multiple consecutive load sampling times after the candidate load impact time, subtract the smallest load sampling value read from the load sampling value corresponding to the candidate load impact time to obtain the load drop value after impact, filter the candidate load impact time where the load increase value before impact is greater than or equal to the preset load increase judgment value and the load drop value after impact is greater than or equal to the preset load recovery judgment value, and obtain the load impact time.

8. The intelligent diagnostic method for gear faults in a reduction system according to claim 7, characterized in that, The gear fault sample data includes the sample transmission stage number, sample gear pair number, sample fault order location, sample fault order vibration value, and sample fault type. The steps to generate gear fault type identification results are as follows: Based on the load impact time, select the order analysis segment with a start time less than or equal to the load impact time and an end time greater than the load impact time. From the meshing order position and sideband order position corresponding to the selected order analysis segment, select the order position with the largest order vibration value to obtain the impact order position, and write the load impact energy correction value into the impact order position. Centered on the impact order position, continuous order positions that meet the preset impact judgment range are selected from the transmission stage order data to obtain the impact influence order segment; The order vibration value and load impact energy correction value in the impact influence order section are input into the impact energy correction function to obtain the corrected order vibration value. The corrected order vibration value is written into the impact influence order section, and the transmission stage order data outside the impact influence order section is retained to generate gear fault stability order data. Gear fault sample data are filtered according to transmission stage number and gear pair number. The order position and order vibration value in the stable order data of gear fault are matched with the order position and order vibration value of the sample fault, respectively, to filter the sample fault type and generate gear fault type identification results.

9. The intelligent diagnostic method for gear faults in a reduction system according to claim 8, characterized in that, The steps to obtain the impact order segment are as follows: Using the impact order position as the center, read the adjacent order positions before and after the impact order position to obtain the candidate impact order positions; Input the order vibration value and load impact energy correction value corresponding to the candidate impact order position into the impact expansion judgment function to obtain the order impact judgment value corresponding to each candidate impact order position. The order impact judgment value is compared with the preset impact judgment threshold, and candidate impact order positions with order impact judgment values ​​greater than or equal to the preset impact judgment threshold are selected to obtain the impact-affected order positions. The segment with continuous impact-affected order positions is divided into impact-affected order segments.