Bearing ring hardness detection method and application thereof

Through pulse eddy current detection and data mapping model, the problems of low efficiency and limitations in traditional detection methods are solved, and efficient and accurate bearing ring hardness detection is achieved, which is suitable for bearing production quality inspection.

CN119023740APending Publication Date: 2024-11-26XUZHOU UNIV OF TECH
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202411122842.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-15
Publication Date
2024-11-26

AI Technical Summary

Technical Problem

Traditional bearing ring hardness detection methods are inefficient and have limited applicability, making it difficult to achieve automation and online inspection, and there is a lack of effective data mapping models to evaluate hardness.

Method used

The conductivity of ferromagnetic materials is detected by pulse eddy current detection method, a reference database and data mapping model is constructed, and the relationship between conductivity and hardness is evaluated. A bearing ring hardness detection method based on eddy current detection is established.

Benefits of technology

It realizes efficient, accurate and flexible bearing ring hardness detection, eliminates interference factors during the detection process, improves the authenticity and quality of the detection signal, and is suitable for bearing production quality inspection.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119023740A_ABST
    Figure CN119023740A_ABST
Patent Text Reader

Abstract

The invention discloses a bearing ring hardness detection method and application thereof.According to the scheme, a reference database is ingeniously constructed, and then data with consistency meeting requirements are fitted to form a data mapping model between conductivity and hardness; according to the technical scheme, detection personnel can call different data mapping models by obtaining the workpiece basic information of the to-be-detected bearing ring, the environmental parameters during detection and the detection parameters, then the hardness of the to-be-detected bearing ring is calculated by combining the conductivity, and the scheme is flexible to implement, good in applicability, high in efficiency and high in practicability. And the production quality inspection of the bearing ring can be greatly facilitated.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the fields of nondestructive testing technology and bearing quality inspection technology, and in particular to a bearing ring hardness testing method and application thereof. Background Art

[0002] Bearings are precision mechanical components that convert the sliding friction between the rotating shaft and the shaft seat into rolling friction, thereby reducing friction losses. As the core component of the bearing, the bearing ring bears friction, radial pressure, axial pressure, complex alternating stress, etc., which determine the service life and performance of the bearing. In order to further improve the wear resistance, strength and service life of the bearing, different heat treatment processes are generally used to improve the hardness and strength of the bearing ring, such as carburizing, quenching, tempering, reheating, etc. After heat treatment, the bearing ring generally forms a multi-layer structure of hardened layer, transition layer and matrix layer, and the hardness of the hardened layer is expressed as the hardness of the bearing ring. During the heat treatment process, if the hardness is "too small", it will cause the problem of "soft" material, which will weaken the anti-friction performance and be easy to deform. If the hardness is "too large", it will cause the problem of "brittle" material, which will weaken the impact resistance and reduce the anti-fatigue ability. It can be seen that hardness is an important technical indicator for the manufacture of bearing rings and a technical parameter that is strictly required in production and manufacturing. Testing the hardness of bearing rings is an important technical method to control the manufacturing quality of bearing rings, and it is also an important technical means to sort out unqualified heat-treated bearing rings. Therefore, it is particularly important to test and evaluate the hardness of bearing rings.

[0003] In the traditional non-destructive testing method for bearing ring hardness, the hardness value is directly evaluated mainly through the calibration method. However, the calibration evaluation method is inefficient, limited in applicability, and difficult to automate. In order to improve the efficiency of hardness testing and be suitable for online hardness testing, there is an urgent need for an efficient, accurate, and highly applicable testing method. The eddy current non-destructive testing method is one of the key technologies to ensure the manufacturing quality of bearings and the safety of in-service operation. It has been applied in conductivity, hardness, and defect detection and is regarded as an effective and reliable hardness testing method. The inventor found through research that in the eddy current hardness test, the conductivity parameter of ferromagnetic materials is a variable corresponding to the hardness value, and there is a certain corresponding correlation between the two; and in the currently disclosed literature, there are few documents that reveal how to construct a reference database and form a refined data mapping model scheme to evaluate the bearing ring hardness according to other physical parameters such as conductivity. Based on this, the researchers used the conductivity parameter as the corresponding variable, tested the conductivity of the ferromagnetic material based on the eddy current detection method, and then evaluated the hardness of the ferromagnetic bearing rings based on the mapping relationship model between conductivity and hardness value. This scheme has positive practical significance for the non-destructive detection of the hardness of the bearing rings. Summary of the invention

[0004] In view of this, the purpose of the present invention is to provide a bearing ring hardness detection method and application thereof which is reliable in implementation, flexible in application and has good data reference.

[0005] In order to achieve the above technical objectives, the technical solution adopted by the present invention is:

[0006] A bearing ring hardness detection method is used for the hardness detection of a bearing ring made of ferromagnetic material, comprising:

[0007] S01. Select multiple bearing rings that meet the quality requirements as reference workpieces, perform hardness testing on them and conduct electrical conductivity testing through a pulsed eddy current testing device, generate test results, and then associate the test results with the corresponding reference workpiece information to form a reference data group, and use this to collect multiple reference data groups to build a reference database;

[0008] S02, extracting a plurality of reference data groups that meet preset requirements from a reference database, and establishing a data mapping model between the conductivity and hardness of the bearing ring;

[0009] S03, obtaining a bearing ring to be tested, and performing conductivity testing on the bearing ring through a probe of a pulsed eddy current testing device under preset environmental parameters and testing parameters to obtain conductivity data;

[0010] S04. Determine the hardness of the bearing ring according to the conductivity data and in combination with a data mapping model between the conductivity and hardness of the bearing ring.

[0011] As a possible implementation, further, this solution S01 includes:

[0012] S011. Select multiple bearing rings that meet the quality requirements as reference workpieces, record their corresponding specification information and processing parameter information, and obtain basic workpiece information;

[0013] S012. Under the conditions of preset environmental parameters and detection parameters, perform hardness detection on multiple bearing rings and perform conductivity detection through a pulsed eddy current detection device to generate detection results;

[0014] S013. According to the test results, the basic information of the workpiece corresponding to the bearing ring is obtained, and then the test results, the basic information of the workpiece, the environmental parameters during the test, and the test parameters are collected and associated to form a reference data group. Then, the reference data groups corresponding to multiple bearing rings are collected into a test data set to establish a corresponding reference database.

[0015] In this solution, the environmental parameters include at least one of the temperature of the bearing ring to be tested, the ambient temperature, and the ambient humidity, and the detection parameters include at least one of the pulse width, pulse amplitude, repetition frequency, probe size, type, distance between the probe and the sample, sampling rate, filter settings, and gain adjustment parameters.

[0016] As a preferred implementation option, preferably, in the present solution S012, the conductivity detection by the pulsed eddy current detection device includes:

[0017] A01. Initialize the pulsed eddy current testing device, set the pulse repetition frequency and pulse width to adjust the pulse energy and penetration depth;

[0018] A02. Take a bearing ring with known conductivity as a standard, test it with a pulsed eddy current testing device, and then calibrate the pulsed eddy current testing device based on the test results;

[0019] A03. Use a calibrated pulsed eddy current testing device to perform conductivity testing on multiple bearing rings used as reference workpieces. Collect multiple sets of conductivity test results for each test, and then calculate the average conductivity as the test result of the corresponding bearing ring.

[0020] As a preferred implementation option, preferably, the present solution A03 also includes performing Fourier transform on the detected signal to realize time domain signal analysis and frequency domain signal analysis of the signal, and using a filter to remove high-frequency noise and non-correlated frequency components in the signal;

[0021] Among them, the signal collected by the probe of the pulsed eddy current detection device is processed by the following Fourier formula:

[0022]

[0023] Among them, X(f) is the frequency domain signal, x(t) is the time domain signal, f is the operating frequency of the signal, j is the imaginary unit, and t is the time variable;

[0024] The filter adopts a low-pass filter, and its formula is as follows:

[0025]

[0026] Where H(f) is the frequency response function of the low-pass filter, f is the operating frequency of the signal, and f c is the cutoff frequency of the signal;

[0027] After using this low-pass filter, the frequency domain signal can be expressed as follows:

[0028] Y(f)=X(f)×H(f)

[0029] After inverse Fourier transform, the time domain signal of the acquired signal can be obtained, and the formula is as follows:

[0030]

[0031] y(t) is the filtered time domain signal, Y(f) is the filtered frequency domain signal, f is the operating frequency of the signal, j is the imaginary unit, and t is the time variable.

[0032] As a preferred implementation option, preferably, in the present scheme A03, when Fourier transform is performed on the detected signal to realize time domain signal analysis and frequency domain signal analysis of the signal, the time domain signal and the frequency domain signal are also differentiated and normalized to eliminate the interference of magnetic permeability on the time domain signal and the frequency domain signal, which includes:

[0033] (1) Differential processing

[0034] The difference ΔX(f) between adjacent frequency points or preset frequency intervals is calculated to weaken the widespread influence of magnetic permeability on the signal. The formula is as follows:

[0035] ΔX(f)=X(f)-X(f-δf)

[0036] Wherein, δf is the preset frequency interval, and X(f) is the frequency domain signal;

[0037] (2) Normalization

[0038] The differential signal is normalized to facilitate comparison of data from different samples or under different conditions. The formula is as follows:

[0039]

[0040] in, is the processed frequency domain signal;

[0041] The processed frequency domain signal is subjected to inverse Fourier transform to obtain the processed time domain signal, and the formula is as follows:

[0042]

[0043] Among them, y(t) is the processed time domain signal, is the processed frequency domain signal, f is the operating frequency of the signal, j is the imaginary unit, and t is the time variable.

[0044] As a preferred implementation option, preferably, this solution S02 includes:

[0045] S021. Extracting several reference data groups whose consistency of workpiece basic information, environmental parameters and detection parameters meets the requirements from the reference database;

[0046] S022. Perform data relationship analysis on several reference data groups to obtain parameters that affect conductivity and hardness, and then establish a mapping relationship between the conductivity and hardness of the bearing ring through linear regression, polynomial regression, decision tree, random forest, support vector machine SVM or neural network to generate a data mapping model corresponding to the basic information of the workpiece.

[0047] As a preferred implementation option, preferably, in this solution S021, the several reference data groups that meet the requirements are reference data groups with consistent basic information of the workpiece, and fluctuations of environmental parameters and detection parameters less than 5%, and the conductivity of the bearing rings corresponding to the several reference data groups are not completely the same;

[0048] S022 also includes:

[0049] The basic workpiece information of the bearing ring corresponding to several reference data groups is recorded, and after a data mapping model corresponding to the basic workpiece information is generated, information is also annotated to make the data mapping model correspond to the basic workpiece information.

[0050] As a preferred implementation option, preferably, in this solution S022, the environmental parameters include the temperature and humidity of the detection environment where the bearing ring is located;

[0051] The detection parameters include the amplitude and phase of the eddy current response of the corresponding bearing ring;

[0052] S022 also establishes a formula model for the conductivity and eddy current response of the bearing ring through a linear regression model, which includes:

[0053]

[0054] Among them, σ is the conductivity of the bearing ring, A, are the amplitude and phase of the eddy current response, respectively. a, b, and c are all model parameters, which are determined by substituting several reference data into the model.

[0055] The data mapping model between the conductivity and hardness of the bearing ring is expressed as follows:

[0056] H=f(σ)

[0057] Where H is the hardness, and f(σ) is the mapping function model between the conductivity and hardness of the bearing ring.

[0058] As a preferred implementation option, preferably, in the present solution S03, when the bearing ring to be detected is detected, basic workpiece information corresponding to the bearing ring to be detected is also obtained, which at least includes specification information and processing process parameter information corresponding to the bearing ring to be detected;

[0059] When the bearing ring to be inspected is inspected by the pulse eddy current detection device, the detection signal is also Fourier transformed, and then differential and normalized, and then inverse Fourier transformed to obtain the time domain signal and frequency domain signal after the magnetic permeability is eliminated, and then processed to obtain the conductivity data;

[0060] In S04, based on the basic workpiece information of the bearing ring to be tested, a data mapping model whose correlation meets the preset requirements is retrieved; then, based on the conductivity data, combined with the data mapping model between the conductivity and hardness of the bearing ring, the hardness of the bearing ring is determined.

[0061] Based on the above, the present invention also provides a bearing quality inspection method, which is applied with the bearing ring hardness inspection method described above.

[0062] By adopting the above technical scheme, the present invention has the following beneficial effects compared with the prior art: the present scheme cleverly constructs a reference database, and then fits the data that meets the consistency requirements to form a data mapping model between conductivity and hardness, so that the inspector can call different data mapping models by obtaining the basic information of the workpiece of the bearing ring to be inspected, the environmental parameters during the inspection, and the inspection parameters, and then substitute the conductivity obtained by the inspection into the data mapping model to calculate the hardness of the bearing ring to be inspected. In order to eliminate or suppress the interference factors in the inspection process, the present scheme calibrates the pulsed eddy current inspection device, denoises the data, and eliminates the magnetic permeability interference, so that the obtained signal data has high authenticity and quality, providing a guarantee for the inspection of conductivity, so that when the data mapping model is substituted later, a high-quality hardness evaluation result can be obtained. In the case of conductivity inspection, the present scheme can also directly perform conductivity inspection of the bearing ring to be inspected based on the commercially available pulsed eddy current conductivity meter. The scheme is not only flexible to implement, but also has good applicability and high efficiency, and can greatly facilitate the production quality inspection of the bearing ring. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] 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 some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0064] Figure 1 It is a brief implementation flow diagram of the detection method of this scheme;

[0065] Figure 2 This is a schematic diagram of an implementation process of step S01 of the detection method of this solution;

[0066] Figure 3 This is a schematic diagram of an implementation process of step S021 of the detection method of this solution;

[0067] Figure 4 This is a brief data flow diagram when testing the bearing ring to be tested after the conductivity and hardness data mapping relationship is established in this detection method. DETAILED DESCRIPTION

[0068] The present invention will be further described in detail below in conjunction with the accompanying drawings and examples. It is particularly noted that the following examples are only used to illustrate the present invention, but are not intended to limit the scope of the present invention. Similarly, the following examples are only partial embodiments of the present invention rather than all embodiments, and all other embodiments obtained by those of ordinary skill in the art without creative work are within the scope of protection of the present invention.

[0069] refer to Figure 1 As shown, this embodiment provides a bearing ring hardness detection method, which is used for hardness detection of a bearing ring made of ferromagnetic material, and includes:

[0070] S01. Select multiple bearing rings that meet the quality requirements as reference workpieces, perform hardness testing on them and conduct conductivity testing through a pulsed eddy current testing device to generate test results, and then associate the test results with the corresponding reference workpiece information (including specifications of the reference workpiece, process parameters, etc.) to form a reference data group, and use this to collect multiple reference data groups to build a reference database;

[0071] S02, extracting a plurality of reference data groups that meet preset requirements from a reference database, and establishing a data mapping model between the conductivity and hardness of the bearing ring;

[0072] S03, obtaining a bearing ring to be tested, and performing conductivity testing on the bearing ring through a probe of a pulsed eddy current testing device under preset environmental parameters and testing parameters to obtain conductivity data;

[0073] S04. Determine the hardness of the bearing ring according to the conductivity data and in combination with a data mapping model between the conductivity and hardness of the bearing ring.

[0074] In this scheme, by establishing a data mapping model of the conductivity and hardness of the reference workpiece, it is possible to evaluate the hardness of the bearing ring to be tested that is identical to the reference workpiece or whose consistency meets the preset requirements, that is, after the conductivity data of the bearing ring to be tested is detected by the probe of the existing pulsed eddy current detection device (eddy current conductivity meter), the hardness of the bearing ring to be tested is inferred in combination with the mapping model. Through the data mapping model of this scheme, the hardness detection efficiency of the bearing to be tested can be greatly improved, and at the same time, the bearing ring to be tested is non-destructive.

[0075] Specifically, in establishing the reference data, as a possible implementation method, further, this solution S01 includes:

[0076] S011. Select multiple bearing rings that meet the quality requirements as reference workpieces, record their corresponding specification information and processing parameter information, and obtain basic workpiece information;

[0077] S012. Under the conditions of preset environmental parameters and detection parameters, perform hardness detection on multiple bearing rings and perform conductivity detection through a pulsed eddy current detection device to generate detection results;

[0078] S013. According to the test results, the basic information of the workpiece corresponding to the bearing ring is obtained, and then the test results, the basic information of the workpiece, the environmental parameters during the test, and the test parameters are collected and associated to form a reference data group. Then, the reference data groups corresponding to multiple bearing rings are collected into a test data set to establish a corresponding reference database.

[0079] This solution helps to improve the reference precision of the benchmark workpiece by recording the basic workpiece information of the benchmark workpiece, and provides a high-quality data foundation for the subsequent establishment of a strongly relevant and reliable data mapping model.

[0080] As for data collection of the reference workpiece, as a preferred implementation option, preferably, in the present solution S012, the conductivity detection by the pulsed eddy current detection device includes:

[0081] A01. Initialize the pulsed eddy current testing device, set the pulse repetition frequency and pulse width to adjust the pulse energy and penetration depth;

[0082] A02. Take a bearing ring with known conductivity as a standard, test it with a pulsed eddy current testing device, and then calibrate the pulsed eddy current testing device based on the test results;

[0083] A03. Use a calibrated pulsed eddy current testing device to perform conductivity testing on multiple bearing rings used as reference workpieces. Collect multiple sets of conductivity test results for each test, and then calculate the average conductivity as the test result of the corresponding bearing ring.

[0084] In order to facilitate the removal of interference signals from the data collected by the detection to improve the accuracy of the data, as a preferred implementation option, preferably, this solution A03 also includes Fourier transforming the detected signal to realize time domain signal analysis and frequency domain signal analysis of the signal, and using a filter to remove high-frequency noise and non-correlated frequency components in the signal;

[0085] Among them, the signal collected by the probe of the pulsed eddy current detection device is processed by the following Fourier formula:

[0086]

[0087] Among them, X(f) is the frequency domain signal, x(t) is the time domain signal, f is the operating frequency of the signal, j is the imaginary unit, and t is the time variable;

[0088] The filter adopts a low-pass filter, and its formula is as follows:

[0089]

[0090] Where H(f) is the frequency response function of the low-pass filter, f is the operating frequency of the signal, and f c is the cutoff frequency of the signal;

[0091] After using this low-pass filter, the frequency domain signal can be expressed as follows:

[0092] Y(f)=X(f)×H(f)

[0093] After inverse Fourier transform, the time domain signal of the acquired signal can be obtained, and the formula is as follows:

[0094]

[0095] y(t) is the filtered time domain signal, Y(f) is the filtered frequency domain signal, f is the operating frequency of the signal, j is the imaginary unit, and t is the time variable.

[0096] In this scheme, after obtaining the conductivity data of the bearing ring to be tested, other models can be used for data verification. For example, as the theoretical basis of this scheme, during the pulsed eddy current testing process, when a time-varying magnetic field penetrates the bearing ring made of conductive material, eddy currents are induced inside the material. The properties of these eddy currents (such as distribution, density and phase) depend on the conductivity, magnetic permeability and geometric structure of the material.

[0097] Among them, the distribution of eddy current can be obtained by solving the boundary conditions in the Maxwell equation. In this scheme, the eddy current density J can be expressed as follows:

[0098] J=σE

[0099] Among them, σ is the conductivity of the bearing ring, and E is the induced electric field strength.

[0100] The induced electric field strength E can be derived from the time-varying magnetic field and can be calculated using Faraday's law of electromagnetic induction. The formula is as follows:

[0101]

[0102] in, is the electric potential, and A is the magnetic vector potential. During eddy current testing, the change of magnetic vector potential over time needs attention.

[0103] In the calculation of the magnetic vector potential A, the magnetic vector potential A can be calculated based on the relative position relationship between the current I(t) in the detection coil of the detection device and the bearing ring to be detected. When the geometric relationship of the coil is known, the Biot-Savart law can be used to calculate the magnetic field, and then the magnetic vector potential is obtained by integration. The formula is as follows:

[0104]

[0105] Among them, r, r` are the point in space and the position of the coil element respectively, and dI` is the tiny length segment of the coil element.

[0106] On this basis, the calculation of the induced electric field E can also be obtained by simplifying the time derivative of the magnetic vector potential A, which is as follows:

[0107]

[0108] The above calculation relies on knowing how the current I(t) changes with time and applying it to the above magnetic vector potential calculation formula. Once the induced electric field E is calculated, the eddy current density J and the conductivity of the bearing to be tested can be expressed by Ohm's law, that is:

[0109] J=σE

[0110] In the pulse eddy current detection process, the magnetic permeability and the thickness of the bearing ring to be detected will interfere with the detection structure of the conductivity. In terms of eliminating the interference of the magnetic permeability, as a better implementation option, preferably, in the present scheme A03, when Fourier transform is performed on the detected signal to realize the time domain signal analysis and frequency domain signal analysis of the signal, the time domain signal and the frequency domain signal are also differentiated and normalized to eliminate the interference of the magnetic permeability on the time domain signal and the frequency domain signal, which includes:

[0111] (1) Differential processing

[0112] The difference ΔX(f) between adjacent frequency points or preset frequency intervals is calculated to weaken the widespread influence of magnetic permeability on the signal. The formula is as follows:

[0113] ΔX(f)=X(f)-X(f-δf)

[0114] Wherein, δf is the preset frequency interval, and X(f) is the frequency domain signal;

[0115] (2) Normalization

[0116] The differential signal is normalized to facilitate comparison of data from different samples or under different conditions. The formula is as follows:

[0117]

[0118] in, is the processed frequency domain signal;

[0119] The processed frequency domain signal is subjected to inverse Fourier transform to obtain the processed time domain signal, and the formula is as follows:

[0120]

[0121] Among them, y(t) is the processed time domain signal, is the processed frequency domain signal, f is the operating frequency of the signal, j is the imaginary unit, and t is the time variable.

[0122] In terms of eliminating or suppressing the influence of thickness, this solution can obtain the basic information of the workpiece to be detected, and then select a data mapping model suitable for it for use, so as to reduce the interference of the processing parameters on the conductivity detection. For example, the parameters of the heat treatment process will affect the thickness of the hardened layer of the bearing ring. As a better implementation option, preferably, this solution S02 includes:

[0123] S021. Extracting several reference data groups whose consistency of workpiece basic information, environmental parameters and detection parameters meets the requirements from the reference database;

[0124] S022. Perform data relationship analysis on several reference data groups to obtain parameters that affect conductivity and hardness, and then establish a mapping relationship between the conductivity and hardness of the bearing ring through linear regression, polynomial regression, decision tree, random forest, support vector machine SVM or neural network to generate a data mapping model corresponding to the basic information of the workpiece.

[0125] As a preferred implementation option, preferably, in this scheme S021, several reference data groups that meet the requirements are reference data groups with consistent basic information of workpieces, and fluctuations in environmental parameters and detection parameters less than 5%, and the conductivity of bearing rings corresponding to several reference data groups is not exactly the same.

[0126] On the basis of the above, this plan S022 also includes:

[0127] The basic workpiece information of the bearing ring corresponding to several reference data groups is recorded, and after a data mapping model corresponding to the basic workpiece information is generated, information is also annotated to make the data mapping model correspond to the basic workpiece information.

[0128] In addition, as a preferred implementation option, preferably, in this solution S022, the environmental parameters include the temperature and humidity of the detection environment where the bearing ring is located;

[0129] The detection parameters include the amplitude and phase of the eddy current response of the corresponding bearing ring;

[0130] S022 also establishes a formula model for the conductivity and eddy current response of the bearing ring through a linear regression model, which includes:

[0131]

[0132] Among them, σ is the conductivity of the bearing ring, A, are the amplitude and phase of the eddy current response, respectively. a, b, and c are all model parameters, which are determined by substituting several reference data into the model.

[0133] The data mapping model between the conductivity and hardness of the bearing ring is expressed as follows:

[0134] H=f(σ)

[0135] Where H is the hardness, and f(σ) is the mapping function model between the conductivity and hardness of the bearing ring.

[0136] In addition to the linear regression model exemplified above, the present solution can also establish a mapping model between the conductivity and hardness of the bearing ring through a neural network. Since the neural network has a high applicability, during data training, it is not limited to using highly consistent data as training data, and it can also include greater difference data, for example, reference data groups with different processing parameters and specifications, as well as environmental parameters during detection and different detection parameters. The training method of the neural network includes the following:

[0137] B01. Retrieve a preset amount of reference data from the reference database, select 80% of the data as training data, and the remaining 20% ​​as verification data;

[0138] B02. The conductivity, basic information of the workpiece, environmental parameters during detection, and detection parameters in the reference data group are used as input items, and the hardness is used as the result item, and they are imported into the neural network for training to obtain a trained neural network, which is used to evaluate the hardness of the bearing ring to be detected;

[0139] B03. Import the verification data into the trained neural network for verification. When the accuracy meets the requirements, obtain the detection neural network for hardness detection. Otherwise, return to B02 to continue training until the model converges.

[0140] Since neural networks have high generalization capabilities, we can mark their test results and then store samples that meet the test accuracy requirements in the reference database to continuously improve the reference database. On this basis, we can regularly retrain the detection neural network to improve its detection capabilities.

[0141] By using the above-mentioned detection neural network method, when testing bearing rings of unknown hardness, if the conductivity, basic information of the workpiece, and environmental parameters and detection parameters during testing are known, hardness evaluation can be easily carried out; and with the continuous training and iteration of the detection neural network, this solution can also continuously improve its reliability.

[0142] In order to conveniently and accurately call the data mapping model adapted to the bearing ring to be inspected, as a better implementation option, preferably, in this scheme S03, when inspecting the bearing ring to be inspected, the basic information of the workpiece corresponding to the bearing ring to be inspected is also obtained, which at least includes the specification information and processing parameter information corresponding to the bearing ring to be inspected.

[0143] When the bearing ring to be inspected is inspected by the pulse eddy current detection device, the detection signal is also Fourier transformed, and then differential and normalized, and then inverse Fourier transformed to obtain the time domain signal and frequency domain signal after the magnetic permeability is eliminated, and then processed to obtain the conductivity data;

[0144] In S04, based on the basic workpiece information of the bearing ring to be tested, a data mapping model whose correlation meets the preset requirements is retrieved; then, based on the conductivity data, combined with the data mapping model between the conductivity and hardness of the bearing ring, the hardness of the bearing ring is determined.

[0145] In addition, when constructing a reference database, this solution can also enrich the quality and completeness of the data and improve its reuse value by making detailed records of the model or other physical indicators of the probe of the pulsed eddy current detection device.

[0146] Based on the above, the bearing hardness detection method of this scheme can be used in the production quality inspection of bearings, thereby providing a reliable guarantee for the stable quality of bearings.

[0147] The above descriptions are only some embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Any equivalent device or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A bearing ring hardness detection method, used for the hardness detection of a bearing ring made of ferromagnetic material, characterized in that: It includes: S01. Select multiple bearing rings that meet the quality requirements as reference workpieces, perform hardness testing on them and conduct electrical conductivity testing through a pulsed eddy current testing device to generate test results, then associate the test results with the corresponding reference workpiece information to form a reference data group, and use this to collect multiple reference data groups to build a reference database; S02, extracting a plurality of reference data groups that meet preset requirements from a reference database, and establishing a data mapping model between the conductivity and hardness of the bearing ring; S03, obtaining a bearing ring to be tested, and performing conductivity testing on the bearing ring through a probe of a pulsed eddy current testing device under preset environmental parameters and testing parameters to obtain conductivity data; S04. Determine the hardness of the bearing ring according to the conductivity data and in combination with a data mapping model between the conductivity and hardness of the bearing ring.

2. The bearing ring hardness detection method according to claim 1, characterized in that: S01 includes: S011. Select multiple bearing rings that meet the quality requirements as reference workpieces, record their corresponding specification information and processing parameter information, and obtain basic workpiece information; S012. Under the conditions of preset environmental parameters and detection parameters, perform hardness detection on multiple bearing rings and perform conductivity detection through a pulsed eddy current detection device to generate detection results; S013. According to the test results, the basic information of the workpiece corresponding to the bearing ring is obtained, and then the test results, the basic information of the workpiece, the environmental parameters during the test, and the test parameters are collected and associated to form a reference data group. Then, the reference data groups corresponding to multiple bearing rings are collected into a test data set to establish a corresponding reference database.

3. The bearing ring hardness detection method according to claim 2, characterized in that: In S012, the conductivity detection by the pulsed eddy current detection device includes: A01. Initialize the pulsed eddy current testing device, set the pulse repetition frequency and pulse width to adjust the pulse energy and penetration depth; A02. Take a bearing ring with known conductivity as a standard, test it with a pulsed eddy current testing device, and then calibrate the pulsed eddy current testing device based on the test results; A03. Use a calibrated pulsed eddy current testing device to perform conductivity testing on multiple bearing rings used as reference workpieces. Collect multiple sets of conductivity test results for each test, and then calculate the average conductivity as the test result of the corresponding bearing ring.

4. The bearing ring hardness detection method according to claim 3, characterized in that: A03 also includes performing Fourier transform on the detected signal to realize time domain signal analysis and frequency domain signal analysis of the signal, and using a filter to remove high frequency noise and non-correlated frequency components in the signal; Among them, the signal collected by the probe of the pulsed eddy current detection device is processed by the following Fourier formula: Among them, X(f) is the frequency domain signal, x(t) is the time domain signal, f is the operating frequency of the signal, j is the imaginary unit, and t is the time variable; The filter adopts a low-pass filter, and its formula is as follows: Where H(f) is the frequency response function of the low-pass filter, f is the operating frequency of the signal, and f c is the cutoff frequency of the signal; After using this low-pass filter, the frequency domain signal can be expressed as follows: Y(f)=X(f)×H(f) After inverse Fourier transform, the time domain signal of the acquired signal can be obtained, and the formula is as follows: y(t) is the filtered time domain signal, Y(f) is the filtered frequency domain signal, f is the operating frequency of the signal, j is the imaginary unit, and t is the time variable.

5. The bearing ring hardness detection method according to claim 4, characterized in that: In A03, when Fourier transform is performed on the detected signal to realize time domain signal analysis and frequency domain signal analysis of the signal, the time domain signal and the frequency domain signal are also differentiated and normalized to eliminate the interference of magnetic permeability on the time domain signal and the frequency domain signal, which includes: (1) Differential processing The difference ΔX(f) between adjacent frequency points or preset frequency intervals is calculated to weaken the widespread influence of magnetic permeability on the signal. The formula is as follows: ΔX(f)=X(f)-X(f-δf) Wherein, δf is the preset frequency interval, and X(f) is the frequency domain signal; (2) Normalization The differential signal is normalized to facilitate comparison of data from different samples or under different conditions. The formula is as follows: in, is the processed frequency domain signal; The processed frequency domain signal is subjected to inverse Fourier transform to obtain the processed time domain signal, and the formula is as follows: Among them, y(t) is the processed time domain signal, is the processed frequency domain signal, f is the operating frequency of the signal, j is the imaginary unit, and t is the time variable.

6. The bearing ring hardness detection method according to any one of claims 3 to 5, characterized in that: S02 includes: S021. Extracting several reference data groups whose consistency of workpiece basic information, environmental parameters and detection parameters meets the requirements from the reference database; S022. Perform data relationship analysis on several reference data groups to obtain parameters that affect conductivity and hardness, and then establish a mapping relationship between the conductivity and hardness of the bearing ring through linear regression, polynomial regression, decision tree, random forest, support vector machine SVM or neural network to generate a data mapping model corresponding to the basic information of the workpiece.

7. The bearing ring hardness detection method according to claim 6, characterized in that: In S021, several reference data groups that meet the requirements are reference data groups with consistent basic information of workpieces, and fluctuations of environmental parameters and detection parameters less than 5%. The conductivity of the corresponding bearing rings of several reference data groups is not completely the same; S022 also includes: The basic workpiece information of the bearing ring corresponding to several reference data groups is recorded, and after a data mapping model corresponding to the basic workpiece information is generated, information is also annotated to make the data mapping model correspond to the basic workpiece information.

8. The bearing ring hardness detection method according to claim 6, characterized in that: In S022, the environmental parameters include the temperature and humidity of the detection environment where the bearing ring is located; The detection parameters include the amplitude and phase of the eddy current response of the corresponding bearing ring; S022 also establishes a formula model for the conductivity and eddy current response of the bearing ring through a linear regression model, which includes: Among them, σ is the conductivity of the bearing ring, A, are the amplitude and phase of the eddy current response, respectively. a, b, and c are all model parameters, which are determined by substituting several reference data into the model. The data mapping model between the conductivity and hardness of the bearing ring is expressed as follows: H=f(σ) Where H is the hardness, and f(σ) is the mapping function model between the conductivity and hardness of the bearing ring.

9. The bearing ring hardness detection method according to claim 8, characterized in that: In S03, when the bearing ring to be inspected is inspected, basic workpiece information corresponding to the bearing ring to be inspected is also obtained, which at least includes specification information and processing process parameter information corresponding to the bearing ring to be inspected; When the bearing ring to be inspected is inspected by the pulse eddy current detection device, the detection signal is also Fourier transformed, and then differential and normalized, and then inverse Fourier transformed to obtain the time domain signal and frequency domain signal after the magnetic permeability is eliminated, and then processed to obtain the conductivity data; In S04, based on the basic workpiece information of the bearing ring to be tested, a data mapping model whose correlation meets the preset requirements is retrieved; then, based on the conductivity data, combined with the data mapping model between the conductivity and hardness of the bearing ring, the hardness of the bearing ring is determined.

10. A bearing quality inspection method, characterized in that: It includes a bearing ring hardness detection method as claimed in any one of claims 1 to 9.

Citation Information

Cited By

  • Method, equipment and system for detecting hardness of alloy forging part

    CN120831410A

  • Bearing ring polishing track self-adaptive control method and system

    CN121132415A