Ball screw abnormal state monitoring method based on current signals
By extracting the current signal of the motor in the ball screw drive system and diagnosing machine learning faults, the complex and hysteresis monitoring of the existing technology is solved, and the timely and accurate abnormal state monitoring of the ball screw is realized.
Patent Information
- Application Number
- CN202411932414.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-26
- Publication Date
- 2025-04-15
AI Technical Summary
The existing ball screw abnormal status monitoring methods are complex and cannot reflect the working status of the equipment in a timely and accurate manner, resulting in lag in fault detection.
By obtaining the current signal of the motor in the ball screw drive system, time domain feature extraction and frequency domain feature extraction are performed, feature collections are obtained, and fault diagnosis is performed using machine learning methods (such as global K neighborhood GKNN algorithm or support vector machine SVM algorithm) to output alarm information.
Timely and accurate abnormal status monitoring of ball screws is realized, avoiding the expansion of faults, and improving production efficiency and accuracy.
Smart Images
Figure CN119830116A_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to the technical field of mechanical equipment fault monitoring, and in particular to a method for monitoring abnormal state of a ball screw based on current signals. Background Art
[0002] Ball screws are widely used in CNC machine tools, automation equipment and precision machinery. Their main function is to convert rotational motion into precise linear motion, carrying the heavy responsibility of transmitting power and achieving positioning. However, in long-term operation, ball screws may experience abnormal operation due to factors such as wear, insufficient lubrication, overload, and manufacturing defects, such as vibration, noise, and temperature rise. In severe cases, they may cause equipment failure, affecting production efficiency and precision.
[0003] At present, the abnormal state monitoring methods of ball screws mostly use vibration sensors, temperature sensors or visual inspection. Although these methods can monitor some abnormal conditions, they often require more complex hardware equipment and installation space, and cannot reflect the working status of the ball screw in a timely and accurate manner. Summary of the invention
[0004] The present disclosure provides a method for monitoring the abnormal state of a ball screw based on a current signal, which can timely and accurately monitor the abnormal state of a ball screw. The technical solution at least includes the following solutions: In a first aspect, a method for monitoring abnormal state of a ball screw based on a current signal is provided, comprising: obtaining a first current signal of a motor in a driving system of the ball screw, wherein the first current signal is a historical current signal of the motor in the driving system of the ball screw; performing time domain feature extraction and frequency domain feature extraction on the first current signal to obtain a first feature set, wherein the first feature set includes time domain features and frequency domain features of the first current signal; based on the first feature set, performing fault diagnosis on the ball screw using a machine learning method to obtain a first fault diagnosis result, wherein the machine learning method is any one of a global K-neighbor GKNN algorithm and a support vector machine SVM algorithm; when the first fault diagnosis result indicates that the ball screw is faulty, outputting an alarm message, wherein the alarm message is used to prompt that the ball screw is faulty.
[0005] Optionally, in the case where the machine learning method is the GKNN algorithm, the ball screw is fault diagnosed using a machine learning method based on the first time domain data and the first frequency domain data to obtain a first fault diagnosis result, including: constructing a first training set based on the first feature set, the sample types in the first training set including spectral energy, current mean, and current peak factor, each sample in the first training set has a label, and the labels include normal and faulty; obtaining a second current signal, the second current signal being a current signal to be detected of a motor in a drive system of the ball screw; determining the top N first samples closest to the second current signal based on the Euclidean distance between the second current signal and each sample in the first training set, the first sample being a sample in the first training set; obtaining the first fault diagnosis result based on the labels of the top N first samples.
[0006] Optionally, in the case where the machine learning method is the SVM algorithm, the machine learning method is used to perform fault diagnosis on the ball screw based on the first time domain data and the first frequency domain data to obtain a first fault diagnosis result, including: constructing a second training set based on the first feature set, the second training set including multiple pairs of feature data, any pair of feature data including normal samples and faulty samples of the same type of feature data; training the SVM with the second training set; optimizing the parameters of the trained SVM with 5-fold cross validation; obtaining a second current signal, the second current signal being a current signal to be detected of the motor in the drive system of the ball screw; and inputting the second current signal into the SVM after optimizing the parameters to obtain the first fault diagnosis result.
[0007] Optionally, in the first feature set, the time domain characteristics of the first current signal include: mean, root mean square value, root amplitude, absolute mean, skewness, kurtosis, variance, maximum value, minimum value, peak-to-peak value, waveform index, peak index, pulse index, margin index, skewness index, and kurtosis index; in the first feature set, the frequency domain characteristics of the first current signal include: spectral density and harmonic components.
[0008] In the second aspect, a ball screw abnormal state monitoring device based on current signal is also provided, including: a first acquisition module, used to acquire a first current signal of a motor in a driving system of the ball screw, wherein the first current signal is a historical current signal of the motor in the driving system of the ball screw; a second acquisition module, used to perform time domain feature extraction and frequency domain feature extraction on the first current signal to obtain a first feature set, wherein the first feature set includes time domain features and frequency domain features of the first current signal; a fault diagnosis module, used to perform fault diagnosis on the ball screw based on the first feature set using a machine learning method to obtain a first fault diagnosis result, wherein the machine learning method is any one of a global K-neighbor GKNN algorithm and a support vector machine SVM algorithm; an output module, used to output an alarm message when the first fault diagnosis result indicates that the ball screw is faulty, wherein the alarm message is used to prompt that the ball screw is faulty.
[0009] Optionally, when the machine learning method is the GKNN algorithm, the fault diagnosis module is also used to: construct a first training set based on the first feature set, the sample types in the first training set include spectral energy, current mean, and current peak factor, and each sample in the first training set has a label, and the labels include normal and faulty; obtain a second current signal, and the second current signal is a current signal to be detected of the motor in the drive system of the ball screw; based on the Euclidean distance between the second current signal and each sample in the first training set, determine the first N first samples closest to the second current signal, and the first sample is a sample in the first training set; based on the labels of the first N first samples, obtain the first fault diagnosis result.
[0010] Optionally, when the machine learning method is the SVM algorithm, the fault diagnosis module is also used to: construct a second training set based on the first feature set, the second training set including multiple pairs of feature data, any pair of feature data including normal samples and fault samples of the same type of feature data; use the second training set to train the SVM; use 5-fold cross validation to optimize the parameters of the trained SVM; obtain a second current signal, the second current signal is a current signal to be detected of the motor in the drive system of the ball screw; input the second current signal to the SVM after optimizing the parameters to obtain the first fault diagnosis result.
[0011] In a third aspect, a computer device is also provided, comprising: a memory and a processor, wherein at least one computer program is stored in the memory, and the at least one computer program is loaded and executed by the processor, thereby executing the ball screw abnormal state monitoring method based on current signal described in the above embodiment.
[0012] In a fourth aspect, a computer-readable storage medium is also provided, in which at least one computer program is stored. The at least one computer program is loaded and executed by a processor, thereby executing the ball screw abnormal state monitoring method based on current signal described in the above embodiment.
[0013] In a fifth aspect, a computer program product is provided, comprising a computer program / instruction, wherein when the computer program / instruction is executed by a processor, the method described in the first aspect is implemented.
[0014] The beneficial effects brought by the technical solution provided by the embodiments of the present disclosure include at least: In the disclosed embodiment, the first current signal of the motor in the ball screw drive system is analyzed, wherein the first current signal is subjected to time domain feature extraction and frequency domain feature extraction to obtain a first feature set, and based on the first feature set, a machine learning method is used to diagnose the ball screw fault to obtain a first fault diagnosis result. The machine learning method is a GKNN algorithm or a SVM algorithm. Both of these machine learning methods can perform fault diagnosis more comprehensively and accurately, and thus can realize timely and accurate abnormal state monitoring of the ball screw to avoid the expansion of faults. BRIEF DESCRIPTION OF THE DRAWINGS
[0015] In order to more clearly illustrate the technical solutions in the embodiments of the present disclosure, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0016] Figure 1 A flow chart of a method for monitoring abnormal state of a ball screw based on a current signal provided by an exemplary embodiment of the present disclosure is shown; Figure 2 Schematic diagram of current signal before and after preprocessing Figure 3 A flow chart of a method for monitoring abnormal state of a ball screw based on a current signal provided by another exemplary embodiment of the present disclosure is shown; Figure 4 The prediction performance of the trained SVM on the training set; Figure 5 A schematic structural diagram of a ball screw abnormal state monitoring device based on a current signal provided by an exemplary embodiment of the present disclosure is shown; Figure 6 It is a schematic diagram of the structure of a computer device provided in an embodiment of the present disclosure. DETAILED DESCRIPTION
[0017] Unless otherwise defined, the technical terms or scientific terms used herein shall have the usual meanings understood by persons with ordinary skills in the field to which the present disclosure belongs. The words "first", "second", "third" and similar words used in the patent application specification and claims of the present disclosure do not indicate any order, quantity or importance, but are only used to distinguish different components. Similarly, words such as "one" or "one" do not indicate a quantity limitation, but indicate the existence of at least one. Words such as "include" or "comprise" and similar words mean that the elements or objects appearing before "include" or "comprise" include the elements or objects listed after "include" or "comprise" and their equivalents, and do not exclude other elements or objects. Words such as "connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right", etc. are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly. A and / or B means that there are three situations: A, B, and A and B.
[0018] In order to make the objectives, technical solutions and advantages of the present disclosure more clear, the embodiments of the present disclosure will be further described in detail below with reference to the accompanying drawings.
[0019] Figure 1 A flowchart of a method for monitoring abnormal state of a ball screw based on a current signal provided by an exemplary embodiment of the present disclosure is shown. The method can be executed by a computer device, see Figure 1 , the method comprising: In step 101, a first current signal of a motor in a driving system of a ball screw is acquired.
[0020] The first current signal is a historical current signal of a motor in the driving system of the ball screw.
[0021] Exemplarily, the driving system of the ball screw is a PLC system, in which the current signal is usually acquired in real time at a sampling frequency of 1 kHz to 10 kHz. The current signal in the PLC system includes all current changes in the operation of the motor, including but not limited to current fluctuations during load changes, starting, stopping, etc. Therefore, the first current signal can be obtained by reading the current signal in the PLC system over a period of time.
[0022] Optionally, before executing step 102, the method further includes: preprocessing the first current signal. The preprocessing includes denoising and smoothing.
[0023] Exemplarily, the Kalman filter technology is used to perform denoising on the first current signal to remove fluctuations caused by system noise (system noise includes noise from the power supply, external equipment or environment). The first current signal is smoothed by moving average method or Gaussian filtering to smooth the instantaneous fluctuations of the current signal and obtain a relatively stable signal sequence. It can also remove long-term current trends and remove long-term trend changes caused by factors such as performance degradation of the device itself.
[0024] There are many related technologies for implementing Kalman filtering technology, moving average method, and Gaussian filtering, so detailed description is omitted here.
[0025] Figure 2 Schematic diagram of the current signal before and after preprocessing. It can be seen that compared with the current signal before preprocessing, the current signal after preprocessing is smoother.
[0026] In step 102, time domain feature extraction and frequency domain feature extraction are performed on the first current signal to obtain a first feature set.
[0027] The first feature set includes time domain features and frequency domain features of the first current signal.
[0028] In step 103, based on the first feature set, a machine learning method is used to diagnose the fault of the ball screw to obtain a first fault diagnosis result.
[0029] The machine learning method is any one of a GKNN (Global K-Nearest Neighbors) algorithm and a SVM (Support Vector Machines) algorithm.
[0030] The first feature set can train a model in a machine learning method, and the trained model in the machine learning method can be used to diagnose faults of the ball screw.
[0031] In step 104, when the first fault diagnosis result indicates that the ball screw is faulty, an alarm message is output.
[0032] The alarm message is used to indicate that the ball screw is faulty.
[0033] In the disclosed embodiment, the first current signal of the motor in the ball screw drive system is analyzed, wherein the first current signal is subjected to time domain feature extraction and frequency domain feature extraction to obtain a first feature set, and based on the first feature set, a machine learning method is used to diagnose the ball screw fault to obtain a first fault diagnosis result. The machine learning method is a GKNN algorithm or a SVM algorithm. Both of these machine learning methods can perform fault diagnosis more comprehensively and accurately, and thus can realize timely and accurate abnormal state monitoring of the ball screw to avoid the expansion of faults.
[0034] Figure 3 A flowchart of a method for monitoring abnormal state of a ball screw based on a current signal provided by another exemplary embodiment of the present disclosure is shown. The method can be executed by a computer device, see Figure 3 , the method comprising: In step 301, a first current signal of a motor in a driving system of a ball screw is acquired.
[0035] The relevant contents of step 301 refer to the aforementioned step 101, and the detailed description is omitted here.
[0036] In step 302, time domain feature extraction and frequency domain feature extraction are performed on the first current signal to obtain a first feature set.
[0037] The first feature set includes time domain features and frequency domain features of the first current signal.
[0038] Optionally, in the first feature set, the time domain characteristics of the first current signal include: mean, root mean square value, root amplitude, absolute mean, skewness, kurtosis, variance, maximum value, minimum value, peak-to-peak value, waveform index, peak index, pulse index, margin index, skewness index, and kurtosis index.
[0039] Among them, the mean value can reflect the average level of current. The standard deviation can reflect the degree of current fluctuation. If the standard deviation increases frequently, it may indicate that the ball screw may have mechanical hysteresis or load fluctuation. The peak factor is the ratio of the maximum value of the signal to the mean value. If the peak factor of the signal is abnormal, it means that the ball screw may have severe load impact or unstable state.
[0040] Optionally, in the first feature set, the frequency domain features of the first current signal include: spectrum density and harmonic components.
[0041] Frequency domain features can be calculated by Fourier transforming the current signal. The spectrum density can reflect the energy distribution of the current signal at each frequency. Harmonic components can reveal abnormal vibration modes in the ball screw. In particular, when the ball screw is worn or stuck, the current signal may have significant harmonic components.
[0042] There are many methods for calculating time domain features and frequency domain features in the related technologies, so detailed description is omitted here.
[0043] Optionally, the first feature set may also include time-frequency features of the first current signal, which may be obtained by performing wavelet transform on the first current signal. The time-frequency features may simultaneously reveal the time domain and frequency domain features of the current signal, and are particularly suitable for fault detection of non-stationary signals.
[0044] In step 303, based on the first feature set, a machine learning method is used to diagnose the fault of the ball screw to obtain a first fault diagnosis result.
[0045] The machine learning method is any one of the GKNN algorithm and the support vector machine SVM algorithm.
[0046] When the machine learning method is the GKNN algorithm, step 303 includes the following steps ad.
[0047] Step a: construct a first training set based on the first feature set.
[0048] The sample types in the first training set include spectrum energy, current mean, and current peak factor. Each sample in the first training set has a label, and the labels include normal and fault.
[0049] Here, the sample type in the first training set may also be the same as the type of the features in the first feature set. For example, labels (such as normal labels or fault labels) may be added to the features in the first feature set one by one to obtain the first training set.
[0050] Step b: acquiring a second current signal.
[0051] The second current signal is a current signal to be detected of a motor in a driving system of the ball screw. Exemplarily, the second current signal is a real-time current signal of a motor in the driving system of the ball screw.
[0052] Step c: determining the first N first samples closest to the second current signal based on the Euclidean distance between the second current signal and each sample in the first training set.
[0053] The first sample is a sample in the first training set.
[0054] There are many methods for calculating the Euclidean distance in the related art, and a detailed description thereof is omitted here.
[0055] Wherein, N is a positive integer. Exemplarily, the value range of N is 10-20, for example, 10, 15, 20, etc.
[0056] Step d: obtaining a first fault diagnosis result based on the labels of the first N first samples.
[0057] Here, each first sample has a label, and the labels include normal and faulty. If the number of faulty samples in the first N first samples is greater than 0.5N, it means that the ball screw is faulty in the first fault diagnosis result; if the number of faulty samples in the first N first samples is less than or equal to 0.5N, it means that the ball screw is not faulty in the first fault diagnosis result.
[0058] Conventional K-neighbor algorithms can usually only compare features in a single dimension, that is, in conventional K-neighbor algorithms, only a single type of feature (such as only the current mean) can exist in the first training set, but only considering features in a single dimension is not comprehensive enough when performing fault diagnosis. The GKNN algorithm used in the embodiments of the present disclosure can compare features in multiple dimensions, so that multiple types of current features can be fully considered when performing fault diagnosis, effectively improving the accuracy of fault diagnosis.
[0059] When the machine learning method is the SVM algorithm, step 303 includes the following steps ei.
[0060] Step e: constructing a second training set based on the first feature set.
[0061] The second training set includes multiple pairs of feature data, and any pair of feature data includes a normal sample and a fault sample of the same type of feature data.
[0062] Here, each feature in the first feature set can be labeled so that each feature has a label (including normal and faulty), and the feature with a normal label is a normal sample, and the feature with a fault label is a faulty sample. Then, a normal sample and a faulty sample can be selected from the same type of features to construct a pair of feature data.
[0063] Step f: train the SVM using the second training set.
[0064] The basic idea of SVM is to separate different types of data through a hyperplane and maximize the margin to minimize the classification error. Before training SVM, you also need to select a suitable kernel function and overshoot parameters (such as penalty parameter C value, γ parameter, etc.), and then you can use the second training set to train SVM.
[0065] There are many methods for implementing SVM in the related art, so detailed description is omitted here.
[0066] Step g: Optimize the parameters of the trained SVM using 5-fold cross validation.
[0067] During the training process, the parameters of the SVM need to be continuously optimized. In the embodiment of the present disclosure, 5-fold cross validation is used to optimize the parameters of the SVM. The 5-fold cross validation can comprehensively and reliably evaluate the SVM and thus select appropriate parameters.
[0068] There are many methods for implementing 5-fold cross validation in the related art, so detailed description is omitted here.
[0069] By using the second training set to train the SVM and optimizing its parameters, a trained SVM can be obtained.
[0070] Figure 4 The prediction performance of the trained SVM on the training set. Figure 4 It can be seen that the training set prediction value of the trained SVM is basically the same as the true value of the training set, which has a good prediction effect.
[0071] Step h: obtaining a second current signal.
[0072] The second current signal is a current signal to be detected of a motor in a driving system of the ball screw.
[0073] For the relevant contents of the second current signal, please refer to the aforementioned step b, and the detailed description is omitted here.
[0074] Step i: input the second current signal into the SVM after optimizing the parameters to obtain the first fault diagnosis result.
[0075] After the SVM training is completed, the second current signal data can be input into the SVM for classification prediction to obtain the first fault diagnosis result. If the result is a "fault" category, step 304 is executed to trigger an alarm, indicating that the ball screw may have a fault. Different fault levels (for example, a minor fault, a serious fault, etc.) can also be given according to the prediction results, and some maintenance suggestions can be preset, and different fault suggestions can be given according to different fault levels.
[0076] In some embodiments, a threshold method may also be used to determine the first fault diagnosis result. The threshold method includes: processing the second current signal in the same manner as the first current signal to obtain a second feature set, wherein the types of features in the second feature set are the same as those in the first feature set; setting a threshold for each type of feature in the second feature set, and if at least one feature in the second feature set exceeds the corresponding threshold, the first fault diagnosis result is used to indicate that the ball screw is faulty.
[0077] In step 304, when the first fault diagnosis result indicates that the ball screw is faulty, an alarm message is output.
[0078] The alarm message is used to indicate that the ball screw is faulty.
[0079] The following are device embodiments of the present application. For details not described in detail in the device embodiments, reference may be made to the above method embodiments.
[0080] Figure 5 FIG. 1 shows a schematic diagram of a ball screw abnormal state monitoring device based on a current signal provided by an exemplary embodiment of the present disclosure, see FIG. Figure 5 The ball screw abnormal state monitoring device 500 based on current signal includes: a first acquisition module 501, a second acquisition module 502, a fault diagnosis module 503, and an output module 504.
[0081] The first acquisition module 501 is used to acquire a first current signal of a motor in a driving system of a ball screw, where the first current signal is a historical current signal of the motor in the driving system of the ball screw.
[0082] The second acquisition module 502 is used to extract time domain features and frequency domain features of the first current signal to obtain a first feature set, where the first feature set includes time domain features and frequency domain features of the first current signal.
[0083] The fault diagnosis module 503 is used to perform fault diagnosis on the ball screw based on the first feature set by using a machine learning method to obtain a first fault diagnosis result. The machine learning method is any one of a global K-neighborhood GKNN algorithm and a support vector machine SVM algorithm.
[0084] The output module 504 is used to output an alarm message when the first fault diagnosis result indicates that the ball screw has a fault, and the alarm message is used to prompt that the ball screw has a fault.
[0085] Optionally, when the machine learning method is the GKNN algorithm, the fault diagnosis module 503 is also used to: construct a first training set based on the first feature set, the sample types in the first training set include spectral energy, current mean, and current peak factor, and each sample in the first training set has a label, and the labels include normal and faulty; obtain a second current signal, and the second current signal is a current signal to be detected by the motor in the drive system of the ball screw; based on the Euclidean distance between the second current signal and each sample in the first training set, determine the first N first samples closest to the second current signal, and the first samples are samples in the first training set; based on the labels of the first N first samples, obtain a first fault diagnosis result.
[0086] Optionally, when the machine learning method is the SVM algorithm, the fault diagnosis module 503 is also used to: construct a second training set based on the first feature set, the second training set including multiple pairs of feature data, any pair of feature data including normal samples and fault samples of the same type of feature data; use the second training set to train the SVM; use 5-fold cross validation to optimize the parameters of the trained SVM; obtain a second current signal, the second current signal is a current signal to be detected of the motor in the drive system of the ball screw; input the second current signal to the SVM after optimizing the parameters to obtain a first fault diagnosis result.
[0087] It should be noted that: when the ball screw abnormal state monitoring device based on current signal provided in the above embodiment performs ball screw abnormal state monitoring, only the division of the above functional modules is used as an example. In actual application, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the ball screw abnormal state monitoring device based on current signal provided in the above embodiment and the ball screw abnormal state monitoring method embodiment based on current signal belong to the same concept. The specific implementation process is detailed in the method embodiment, which will not be repeated here.
[0088] The division of modules in the embodiments of the present disclosure is schematic and is only a logical function division. There may be other division methods in actual implementation. In addition, each functional module in each embodiment of the present disclosure may be integrated into a processor, or may exist physically separately, or two or more modules may be integrated into one module. The above-mentioned integrated modules may be implemented in the form of hardware or in the form of software functional modules.
[0089] If the integrated module is implemented in the form of a software function module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present disclosure is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a terminal device (which can be a personal computer, mobile phone, or communication device, etc.) or a processor (processor) to execute all or part of the steps of the method of each embodiment of the present disclosure. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc., various media that can store program codes.
[0090] Figure 6Schematic diagram of the structure of the computer device provided by the embodiment of the present disclosure. Figure 6 As shown, the computer device 600 includes: a processor 601 and a memory 602 .
[0091] The processor 601 may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor 601 may be implemented in at least one hardware form of DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array). The processor 601 may also include a main processor and a coprocessor. The main processor is a processor for processing data in the awake state, also known as a CPU (Central Processing Unit); the coprocessor is a low-power processor for processing data in the standby state. In some embodiments, the processor 601 may be integrated with a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the display screen. In some embodiments, the processor 601 may also include an AI (Artificial Intelligence) processor, which is used to process computing operations related to machine learning.
[0092] The memory 602 may include one or more computer-readable storage media, which may be non-transitory. The memory 602 may also include a high-speed random access memory, and a non-volatile memory, such as one or more disk storage devices, flash memory storage devices. In some embodiments, the non-transitory computer-readable storage medium in the memory 602 is used to store at least one instruction, which is used to be executed by the processor 601 to implement the ball screw abnormal state monitoring method based on the current signal provided in the embodiment of the present disclosure.
[0093] Those skilled in the art will understand that Figure 6 The structure shown in the figure does not constitute a limitation on the computer device 600, and the computer device 600 may include more or less components than those shown in the figure, or combine some components, or adopt a different arrangement of components.
[0094] The embodiment of the present disclosure also provides a non-temporary computer-readable storage medium. When the instructions in the storage medium are executed by the processor of a computer device, the computer device is able to execute the ball screw abnormal state monitoring method based on current signals provided in the embodiment of the present disclosure.
[0095] The embodiments of the present disclosure also provide a computer program product, including a computer program / instruction, which, when executed by a processor, implements the ball screw abnormal state monitoring method based on current signals provided in the embodiments of the present disclosure.
[0096] The above description is only an optional embodiment of the present disclosure and is not intended to limit the present disclosure. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present disclosure shall be included in the protection scope of the present disclosure.
Claims
1. A method for monitoring abnormal state of a ball screw based on current signal, characterized in that: The method comprises: Acquire a first current signal of a motor in a driving system of a ball screw, where the first current signal is a historical current signal of the motor in the driving system of the ball screw; Performing time domain feature extraction and frequency domain feature extraction on the first current signal to obtain a first feature set, where the first feature set includes time domain features and frequency domain features of the first current signal; Based on the first feature set, a machine learning method is used to perform fault diagnosis on the ball screw to obtain a first fault diagnosis result, wherein the machine learning method is any one of a global K-neighborhood GKNN algorithm and a support vector machine SVM algorithm; When the first fault diagnosis result indicates that the ball screw is faulty, an alarm message is output, where the alarm message is used to prompt that the ball screw is faulty.
2. The method according to claim 1, characterized in that In the case where the machine learning method is the GKNN algorithm, the method of using the machine learning method to perform fault diagnosis on the ball screw based on the first time domain data and the first frequency domain data to obtain a first fault diagnosis result includes: Based on the first feature set, a first training set is constructed, wherein the sample types in the first training set include spectrum energy, current mean, and current peak factor, and each sample in the first training set has a label, and the label includes normal and fault; Acquire a second current signal, where the second current signal is a current signal to be detected of a motor in a driving system of the ball screw; Based on the Euclidean distance between the second current signal and each sample in the first training set, determine the first N first samples closest to the second current signal, where the first samples are samples in the first training set; The first fault diagnosis result is obtained based on the labels of the first N first samples.
3. The method according to claim 1, characterized in that In the case where the machine learning method is the SVM algorithm, the method of using the machine learning method to perform fault diagnosis on the ball screw based on the first time domain data and the first frequency domain data to obtain a first fault diagnosis result includes: Based on the first feature set, construct a second training set, wherein the second training set includes multiple pairs of feature data, and any pair of feature data includes a normal sample and a fault sample of the same type of feature data; Using the second training set to train the SVM; Optimizing the parameters of the trained SVM using 5-fold cross validation; Acquire a second current signal, where the second current signal is a current signal to be detected of a motor in a driving system of the ball screw; The second current signal is input into the SVM after parameter optimization to obtain the first fault diagnosis result.
4. The method according to any one of claims 1 to 3, characterized in that: In the first feature set, the time domain features of the first current signal include: mean value, root mean square value, root square amplitude, absolute mean value, skewness, kurtosis, variance, maximum value, minimum value, peak-to-peak value, waveform index, peak index, pulse index, margin index, skewness index, and kurtosis index; In the first feature set, the frequency domain features of the first current signal include: spectrum density and harmonic components.
5. A ball screw abnormal state monitoring device based on current signal, characterized in that: The device comprises: A first acquisition module, used to acquire a first current signal of a motor in a driving system of a ball screw, wherein the first current signal is a historical current signal of the motor in the driving system of the ball screw; a second acquisition module, configured to perform time domain feature extraction and frequency domain feature extraction on the first current signal to obtain a first feature set, wherein the first feature set includes time domain features and frequency domain features of the first current signal; a fault diagnosis module, configured to perform fault diagnosis on the ball screw based on the first feature set by using a machine learning method to obtain a first fault diagnosis result, wherein the machine learning method is any one of a global K-neighborhood GKNN algorithm and a support vector machine SVM algorithm; An output module is used to output an alarm message when the first fault diagnosis result indicates that the ball screw has a fault, wherein the alarm message is used to prompt that the ball screw has a fault.
6. The device according to claim 5, characterized in that When the machine learning method is the GKNN algorithm, the fault diagnosis module is further used for: Based on the first feature set, a first training set is constructed, wherein the sample types in the first training set include spectrum energy, current mean, and current peak factor, and each sample in the first training set has a label, and the label includes normal and fault; Acquire a second current signal, where the second current signal is a current signal to be detected of a motor in a driving system of the ball screw; Based on the Euclidean distance between the second current signal and each sample in the first training set, determine the first N first samples closest to the second current signal, where the first samples are samples in the first training set; The first fault diagnosis result is obtained based on the labels of the first N first samples.
7. The device according to claim 5, characterized in that When the machine learning method is the SVM algorithm, the fault diagnosis module is further used for: Based on the first feature set, construct a second training set, wherein the second training set includes multiple pairs of feature data, and any pair of feature data includes a normal sample and a fault sample of the same type of feature data; Using the second training set to train the SVM; Optimizing the parameters of the trained SVM using 5-fold cross validation; Acquire a second current signal, where the second current signal is a current signal to be detected of a motor in a driving system of the ball screw; The second current signal is input into the SVM after parameter optimization to obtain the first fault diagnosis result.
8. A computer device, characterized in that: The computer device comprises: a memory and a processor, wherein at least one computer program is stored in the memory, and the at least one computer program is loaded and executed by the processor to implement the method according to any one of claims 1 to 4.
9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores at least one computer program, and the at least one computer program is loaded and executed by a processor to implement the method according to any one of claims 1 to 4.
10. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, the method according to any one of claims 1 to 4 is implemented.