A method and system for monitoring wear of a metal cutting tool

By combining multi-source sensor sequence filtering for noise reduction with machine learning models, the problem of low accuracy in tool wear monitoring was solved, enabling real-time and accurate monitoring and prediction of tool wear, thereby improving machining quality and production efficiency.

CN120038598BActive Publication Date: 2025-11-07DONGGUAN ZHENGHE CHUJI TECH CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202510445152.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2025-04-10
Publication Date
2025-11-07
Estimated Expiration
2045-04-10

AI Technical Summary

Technical Problem

Existing technologies for tool wear monitoring have low accuracy, making it difficult to detect wear in a timely manner. Furthermore, the sensor signals contain a large amount of noise and redundant information, which affects machining quality and production efficiency.

Method used

By real-time monitoring of multi-source sensor sequences of metal cutting tools, including cutting force, vibration, acoustic emission and temperature signals, a tool cutting monitoring matrix is ​​constructed after filtering and noise reduction. Wear detection, differential detection and image perception correction are combined with machine learning models to construct wear trend maps for health management.

Benefits of technology

It improves the accuracy of tool wear monitoring, enabling real-time and accurate monitoring and prediction of tool wear, thus ensuring the stability and safety of the machining process.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120038598B_ABST
    Figure CN120038598B_ABST
Patent Text Reader

Abstract

The application discloses a metal cutting tool wear monitoring method and system, and relates to the technical field of wear monitoring.The method comprises the following steps: real-time monitoring of a metal cutting tool is performed to obtain a multi-source sensing sequence; a tool cutting monitoring matrix is constructed; the tool cutting monitoring matrix is input into a tool wear detection channel to obtain a tool wear first coefficient; differential detection is performed on the tool wear first coefficient according to a predetermined cutting scheme to output a wear differential signal; image perception correction is performed on the tool wear first coefficient to obtain a tool wear second coefficient; a historical tool wear coefficient set of the metal cutting tool is obtained, wear trend prediction is performed in combination with the tool wear second coefficient, a tool wear trend graph is constructed, and health management is performed on the metal cutting tool according to the tool wear trend graph.The technical problem of low tool wear monitoring precision in the prior art is solved, and the technical effect of improving tool wear monitoring precision is achieved.
Need to check novelty before this filing date? Find Prior Art

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of wear monitoring, in particular to a metal cutting tool wear monitoring method and system. BACKGROUND

[0002] In the process of metal cutting, tool wear is one of the key factors affecting the processing quality, production efficiency and manufacturing cost.

[0003] In the process of metal cutting, the wear condition of the tool has an important influence on the processing efficiency, processing quality and production cost. The traditional tool wear monitoring method often relies on the experience of the operator and regular manual inspection, which not only has limited accuracy, but also is difficult to find the wear condition of the tool in time. In addition, the existing tool wear monitoring system often only uses single sensing information, such as cutting force or vibration signal, which is difficult to fully reflect the wear condition of the tool. At the same time, due to various interference factors in the cutting process, such as the unevenness of the workpiece material, the change of the cutting parameters, etc., a large amount of noise and redundant information is contained in the sensing signal, which brings challenges to the accurate monitoring of tool wear. SUMMARY

[0004] The present application provides a metal cutting tool wear monitoring method and system, which solves the technical problem of low precision of tool wear monitoring in the prior art.

[0005] The first aspect of the present application provides a metal cutting tool wear monitoring method, which comprises:

[0006] When the metal cutting tool executes a predetermined cutting scheme, the metal cutting tool is monitored in real time to obtain a multi-source sensing sequence, the multi-source sensing sequence comprising a cutting force sequence, a vibration signal sequence, an acoustic emission signal sequence and a temperature signal sequence; the multi-source sensing sequence is filtered and denoised to construct a tool cutting monitoring matrix; the tool cutting monitoring matrix is input into a tool wear detection channel to obtain a tool wear first coefficient; the tool wear first coefficient is differentially detected according to the predetermined cutting scheme to output a wear differential signal; based on the wear differential signal, the tool wear first coefficient is corrected by image perception to obtain a tool wear second coefficient; a historical tool wear coefficient set of the metal cutting tool is obtained, the tool wear second coefficient is combined for wear trend prediction, a tool wear trend graph is constructed, and the metal cutting tool is managed in health according to the tool wear trend graph.

[0007] The second aspect of the present application provides a metal cutting tool wear monitoring system, which comprises:

[0008] The monitoring module is used for monitoring the metal cutting tool in real time when the metal cutting tool performs a predetermined cutting scheme, and obtaining a multi-source sensing sequence, the multi-source sensing sequence including a cutting force sequence, a vibration signal sequence, an acoustic emission signal sequence and a temperature signal sequence; the denoising module is used for filtering and denoising according to the multi-source sensing sequence, and constructing a tool cutting monitoring matrix; the wear detection module is used for inputting the tool cutting monitoring matrix into a tool wear detection channel, and obtaining a tool wear first coefficient; the difference detection module is used for performing difference detection on the tool wear first coefficient according to the predetermined cutting scheme, and outputting a wear difference signal; the correction module is used for performing image perception correction on the tool wear first coefficient based on the wear difference signal, and obtaining a tool wear second coefficient; and the management module is used for obtaining a historical tool wear coefficient set of the metal cutting tool, combining the tool wear second coefficient to perform wear trend prediction, constructing a tool wear trend graph, and performing health management on the metal cutting tool according to the tool wear trend graph.

[0009] The one or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0010] When the metal cutting tool performs a predetermined cutting scheme, the metal cutting tool is monitored in real time, and a multi-source sensing sequence is obtained, the multi-source sensing sequence including a cutting force sequence, a vibration signal sequence, an acoustic emission signal sequence and a temperature signal sequence. Then, filtering and denoising are performed according to the multi-source sensing sequence, and a tool cutting monitoring matrix is constructed. Then, the tool cutting monitoring matrix is input into a tool wear detection channel, and a tool wear first coefficient is obtained. Next, difference detection is performed on the tool wear first coefficient according to the predetermined cutting scheme, and a wear difference signal is output. Further, image perception correction is performed on the tool wear first coefficient based on the wear difference signal, and a tool wear second coefficient is obtained. Finally, a historical tool wear coefficient set of the metal cutting tool is obtained, wear trend prediction is performed in combination with the tool wear second coefficient, a tool wear trend graph is constructed, and health management is performed on the metal cutting tool according to the tool wear trend graph. The technical problem of low tool wear monitoring precision in the prior art is solved, and the technical effect of improving tool wear monitoring precision is achieved. BRIEF DESCRIPTION OF DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative effort.

[0012] Figure 1 A metal cutting tool wear monitoring method flowchart is provided for the embodiments of the present application.

[0013] Figure 2 A metal cutting tool wear monitoring system structure schematic diagram is provided for the embodiment of the present application.

[0014] Reference signs: monitoring module 11, denoising module 12, wear detection module 13, differential detection module 14, correction module 15, management module 16. DETAILED DESCRIPTION

[0015] The present application provides a metal cutting tool wear monitoring method and system, which solves the technical problem of low tool wear monitoring precision in the prior art.

[0016] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of protection of the present application.

[0017] It should be noted that the terms "comprising" and "having" are intended to cover non-exclusive inclusion, for example, a process, method, system, product or server comprising a series of steps or units does not have to be limited to those clearly listed steps or units, but can include other steps or modules that are not clearly listed or inherent to these processes, methods, products or devices.

[0018] Embodiment one, as shown in the present application provides a metal cutting tool wear monitoring method, wherein the method comprises: Figure 1

[0019] When the metal cutting tool performs a predetermined cutting scheme, the metal cutting tool is monitored in real time to obtain a multi-source sensing sequence, and the multi-source sensing sequence includes a cutting force sequence, a vibration signal sequence, an acoustic emission signal sequence and a temperature signal sequence.

[0020] ​When the metal cutting tool performs a predetermined cutting scheme, the running state of the metal cutting tool is monitored in real time by various sensors arranged on the machining equipment to obtain a multi-source sensing sequence representing the tool wear state. Specifically, a three-axis dynamic cutting force sensor is installed on the metal cutting tool or the machine tool spindle to collect the X, Y and Z three-axis cutting force components in real time during the cutting process, and an analog-to-digital conversion is performed through a data acquisition module to generate a cutting force time sequence; an acceleration sensor is arranged near the tool clamping part or the machine tool spindle to obtain the vibration signal generated by the tool due to wear during the cutting process, the vibration signal contains X, Y and Z three-axis vibration data, and an amplification and filtering process is performed through a signal conditioning module to generate a vibration signal time sequence; an acoustic emission sensor is arranged near the tool to monitor the high-frequency acoustic signals generated during the interaction between the tool and the workpiece, the acoustic emission signal is enhanced through a preamplifier, and an analog-to-digital conversion is performed through a high-speed data acquisition module to generate an acoustic emission signal time sequence; an infrared temperature sensor or a thermocouple is installed near the tool-workpiece contact area to obtain the temperature change of the tool during the cutting process, and the collected temperature signal is converted into a standard electrical signal and input into the data acquisition system to generate a temperature signal time sequence.

[0021] According to the multi-source sensing sequence, filtering and denoising are performed to construct a tool cutting monitoring matrix.

[0022] The multi-source sensing sequence is filtered and denoised to eliminate environmental noise and system interference, thereby improving the data quality, and on this basis, a tool cutting monitoring matrix is constructed to provide high-quality input data for subsequent tool wear detection.

[0023] Further, according to the multi-source sensing sequence, filtering and denoising are performed to construct a tool cutting monitoring matrix, including:

[0024] The cutting force sequence is subjected to Kalman filtering to obtain a cutting force matrix; the vibration signal sequence is subjected to band-pass filtering to obtain a vibration signal matrix; the acoustic emission signal sequence is subjected to band-pass filtering to obtain an acoustic emission signal matrix; the temperature signal sequence is subjected to Kalman filtering to obtain a temperature signal matrix; and the cutting force matrix, the vibration signal matrix, the acoustic emission signal matrix and the temperature signal matrix are subjected to time window alignment splicing to generate the tool cutting monitoring matrix.

[0025] Specifically, the cutting force sequence is subjected to Kalman filtering processing. By establishing a state space model of the cutting force, the Kalman filtering method gradually estimates the true cutting force value according to the original data measured by the cutting force sensor, removes the errors introduced by noise and inaccurate measurement, and generates a denoised cutting force matrix. For the vibration signal sequence, a band-pass filter is used for processing. The band-pass filter can filter out low-frequency interference and high-frequency noise, and only retain the effective vibration signals related to tool wear in the cutting process, thereby obtaining a vibration signal matrix. For the acoustic emission signal sequence, a band-pass filter is also used to enhance the effective components in the signal and suppress irrelevant low-frequency and high-frequency noise, thereby generating an acoustic emission signal matrix. The temperature signal sequence is subjected to Kalman filtering processing to remove random fluctuations and environmental influences in the temperature sensor measurement, ensuring the stability of the temperature signal, and generating a temperature signal matrix.

[0026] After the above filtering and denoising processing, in order to ensure the time sequence consistency between different sensor data, a time window alignment method is used to perform time synchronization processing on the cutting force matrix, vibration signal matrix, acoustic emission signal matrix, and temperature signal matrix. Specifically, by setting a unified time reference, the data of each matrix is aligned according to the time window, ensuring that each sensor signal accurately corresponds in time. Finally, the aligned cutting force matrix, vibration signal matrix, acoustic emission signal matrix, and temperature signal matrix are spliced together according to a predetermined rule to generate a complete tool cutting monitoring matrix. This matrix contains the time sequence information of multiple sources of signals, providing accurate data support for subsequent tool wear analysis and trend prediction.

[0027] The tool cutting monitoring matrix is input into a tool wear detection channel to obtain a tool wear first coefficient.

[0028] The tool wear detection channel includes multiple tool wear detection models that are constructed based on machine learning and can identify the wear state of the tool according to different types of sensor data such as cutting force, vibration, acoustic emission, and temperature signals. The tool cutting monitoring matrix is input as input data, and the tool wear detection model processes the data to analyze the wear degree of the tool and generate a tool wear first coefficient, which represents the wear state or wear degree of the current tool.

[0029] Further, inputting the tool cutting monitoring matrix into the tool wear detection channel to obtain a tool wear first coefficient includes:

[0030] The tool wear detection channel includes Q tool wear detection models, Q being a positive integer greater than 1; the tool cutting monitoring matrix is input into the Q tool wear detection models to obtain Q tool wear detection coefficients; Q wear detection accuracy coefficients corresponding to the Q tool wear detection models are calculated to obtain Q detection output weights; the Q tool wear detection coefficients are weighted calculated according to the Q detection output weights to generate the tool wear first coefficient.

[0031] The tool wear detection channel includes Q tool wear detection models, Q being a positive integer greater than 1; the tool cutting monitoring matrix is input into the Q tool wear detection models to obtain Q tool wear detection coefficients. According to the performance of each tool wear detection model, the wear detection accuracy coefficient of each model is calculated. Specifically, a test data set with known wear states is used to evaluate the Q tool wear detection models. The test data set includes tool samples with different degrees of wear, and the wear state of each sample is labeled. By inputting the test data into each model, the prediction result is obtained, and the prediction result is compared with the actual wear state. Then, according to the standard evaluation index (such as accuracy, precision, recall, F1-score, etc.), the performance of each model is quantitatively evaluated, and the performance of each model on the test set is specifically calculated. Based on these evaluation results, the wear detection accuracy coefficient of each model is calculated, which usually considers the performance of each evaluation index by weighting and gives different weights according to the overall performance of the model. For example, if a model performs well on the test set with high accuracy and F1-score, the accuracy coefficient of the model is high, reflecting its high prediction reliability. Then, in order to ensure that the accuracy coefficients of different models are comparable, the accuracy coefficients of all models are usually normalized to between 0 and 1, ensuring that the sum of all accuracy coefficients is 1. According to the Q wear detection accuracy coefficients, the proportion calculation is performed to obtain Q detection output weights, which reflect the relative contribution of each tool wear detection model in the final wear evaluation. According to the Q detection output weights, the Q tool wear detection coefficients are weighted calculated to generate the final tool wear first coefficient. Through this weighted calculation method, the prediction results of different models are integrated to ensure that the obtained tool wear first coefficient can accurately reflect the wear state of the tool, providing a reliable basis for subsequent wear difference detection, wear trend prediction and tool health management.

[0032] Each tool wear detection model utilizes different algorithms and feature processing methods to extract characteristic information from the input tool cutting monitoring matrix, including cutting force, vibration, acoustic emission, temperature, and other signals, and assesses the wear degree of the tool based on these characteristic information, and then outputs the corresponding wear detection coefficient. Specifically, first, collect multi-source sensor signals such as cutting force, vibration, acoustic emission and temperature, and preprocess these signals, including filtering and denoising, outlier rejection and data normalization, etc.; after preprocessing, the signal data is used as an input matrix to enter the tool wear detection model; key feature information is extracted from each signal using signal processing techniques such as Fourier transform, wavelet transform, time-frequency analysis, etc., such as the oscillation frequency of the cutting force signal, the frequency characteristics of the vibration signal, the intensity and frequency distribution of the acoustic emission signal, and the thermal deformation of the temperature signal; use historical tool wear data to train multiple tool wear detection models, select appropriate machine learning algorithms (such as support vector machines, decision trees, neural networks, etc.) during training, and optimize the model through calibration data to ensure that it can accurately predict the wear state of the tool according to the signal characteristics; after training, the model is evaluated for performance through cross-validation and evaluation indicators (such as precision, recall rate, F1-score, etc.) to ensure that the model can effectively identify the tool state at different wear levels. The trained and evaluated model can be formally applied to the cutting process, and through the input tool cutting monitoring matrix, the model can output the corresponding wear detection coefficient in real time, providing accurate basis for subsequent wear analysis and health management.

[0033] According to the predetermined cutting scheme, the tool wear first coefficient is differentially detected, and a wear differential signal is output.

[0034] Based on the predetermined cutting scheme, the tool wear first coefficient is differentially detected, that is, the wear condition of the metal cutting tool is theoretically predicted, the tool wear prediction coefficient is compared with the tool wear first coefficient, and a wear detection differential coefficient is obtained to represent the deviation between the current tool wear condition and the theoretical prediction value; if the wear detection differential coefficient exceeds the wear detection differential threshold, it indicates that the actual tool wear condition has deviated from the normal prediction range, and there may be abnormal wear or working condition fluctuations, at which time a wear differential signal is generated to provide early warning information.

[0035] Further, according to the predetermined cutting scheme, the tool wear first coefficient is differentially detected, and a wear differential signal is output, including:

[0036] According to the predetermined cutting scheme, the metal cutting tool is predicted to obtain a tool wear prediction coefficient; according to the tool wear prediction coefficient and the tool wear first coefficient, a wear detection differential coefficient is calculated; it is judged whether the wear detection differential coefficient is greater than or equal to the wear detection differential threshold; if the wear detection differential coefficient is greater than or equal to the wear detection differential threshold, the wear differential signal is generated.

[0037] Firstly, based on the predetermined cutting scheme, the wear trend of the metal cutting tool is theoretically predicted, a tool wear prediction model is established, and the tool wear prediction coefficient is calculated by combining historical cutting data, tool material characteristics, cutting process parameters and the wear law of similar working condition tools. Subsequently, based on the first coefficient of tool wear obtained by real-time monitoring, the tool wear prediction coefficient is compared, the wear detection differential coefficient (that is, the absolute value of the difference between the first coefficient of tool wear and the tool wear prediction coefficient) is calculated to measure the deviation between the actual wear of the tool and the theoretical prediction value. Then, the wear detection differential threshold is set, and it is judged whether the wear detection differential coefficient is greater than or equal to the threshold. If the wear detection differential coefficient exceeds the set threshold, it indicates that the actual wear of the tool may be abnormal, such as accelerated wear, change of cutting conditions or process deviation. At this time, the wear differential signal is generated to provide a warning prompt, so that the system can adjust the cutting parameters, optimize the machining process or take maintenance measures, thereby improving the service life and machining quality of the tool and ensuring the stability and safety of the cutting process.

[0038] Further, according to the predetermined cutting scheme, the metal cutting tool is predicted to obtain a tool wear prediction coefficient, comprising:

[0039] The tool pre-wear coefficient of the metal cutting tool is read, wherein the tool pre-wear coefficient is the tool wear coefficient of the metal cutting tool before the predetermined cutting scheme is executed; the same type tool network association of the metal cutting tool is determined to obtain a same family tool set; the tool wear amplification sample retrieval of the same family tool set is performed according to the predetermined cutting scheme to obtain a tool wear amplification retrieval set; the central value calculation is performed according to the tool wear amplification retrieval set to obtain a predicted tool wear amplification; the tool pre-wear coefficient is stimulated and adjusted according to the predicted tool wear amplification to obtain the tool wear prediction coefficient.

[0040] Firstly, the tool pre-wear coefficient of the metal cutting tool is read, the tool pre-wear coefficient being a tool wear coefficient before the tool performs a predetermined cutting scheme, ensuring the accuracy of the prediction reference. Then, based on information such as the specifications, materials, manufacturing batches and use environments of the metal cutting tool, networking correlation of the same type of tool is performed, and a same-family tool set is constructed to fully utilize the historical wear data of similar tools and improve the prediction accuracy. On this basis, according to the predetermined cutting scheme, tool wear increment samples that meet the current cutting conditions are retrieved from the same-family tool set to form a tool wear increment retrieval set, for example, the historical tool wear coefficient increases by 20% after the predetermined cutting scheme is performed. Subsequently, data analysis is performed on the tool wear increment retrieval set, and a statistical method is used to calculate the central value, such as the mean, median or weighted average, to obtain the predicted tool wear increment. Finally, according to the predicted tool wear increment, the tool pre-wear coefficient is stimulated and adjusted to obtain the tool wear prediction coefficient, wherein the tool wear prediction coefficient = tool pre-wear coefficient + tool pre-wear coefficient x predicted tool wear increment.

[0041] Based on the wear difference signal, the tool wear first coefficient is corrected by image sensing to obtain a tool wear second coefficient.

[0042] Based on the wear differential signal, real-time image data of the metal cutting tool is called, and the real-time image is obtained by a visual detection system installed on the machining equipment. The visual detection system can include a high-resolution industrial camera, a light source system, and an image acquisition module, and continuously shoots the tool at a preset image acquisition frequency to ensure the real-time and integrity of the wear state. Then, the obtained real-time image of the tool is preprocessed, including but not limited to image enhancement, noise removal, contrast adjustment, and edge detection. The image enhancement can enhance the details of the tool surface using the Retinex algorithm, the noise removal can be smoothed using Gaussian filtering, and the edge detection can detect the boundary features of the tool edge wear area through the Canny or Roberts operator to improve the recognizability of the tool wear area. Based on the preprocessed tool image, a wear feature perception model is constructed. The model can be trained using deep learning methods such as convolutional neural network (CNN) or ResNet. During the training process, a large number of labeled tool wear image datasets are used, and the model parameters are optimized based on supervised learning to automatically extract the features of the tool wear area and output the image perception tool wear coefficient, which represents the visual level of tool wear. Then, the tool wear first coefficient and the image perception tool wear coefficient are fused, and the tool wear first correction coefficient and the tool wear second correction coefficient are obtained by proportion calculation to measure the weight of different detection methods and ensure the reliability of the final wear evaluation. Next, according to the tool wear first correction coefficient and the tool wear second correction coefficient, the tool wear first coefficient and the image perception tool wear coefficient are weighted calculated, and the weight can be dynamically adjusted based on historical detection data, model confidence and tool use environment to obtain the corrected tool wear second coefficient, thereby improving the accuracy of wear evaluation and enabling the tool wear monitoring system to more accurately predict tool life and provide more reliable decision basis for tool health management.

[0043] Further, based on the wear differential signal, the tool wear first coefficient is corrected by image perception to obtain a tool wear second coefficient, including:

[0044] Based on the wear differential signal, the tool real-time image of the metal cutting tool is called; the wear feature of the metal cutting tool is perceived according to the tool real-time image to obtain an image perception tool wear coefficient; the tool wear first correction coefficient and the tool wear second correction coefficient are obtained by adaptive correction parameter calculation according to the tool wear first coefficient and the image perception tool wear coefficient; the tool wear first coefficient and the image perception tool wear coefficient are weighted calculated according to the tool wear first correction coefficient and the tool wear second correction coefficient to generate the tool wear second coefficient.

[0045] Firstly, based on the wear differential signal, the tool real-time image of the metal cutting tool is called, which is collected by a high-precision industrial camera installed on the machine tool and combined with an adjustable light source system to ensure image clarity and stability. Secondly, the tool real-time image is preprocessed, including image denoising, contrast enhancement and edge detection. The denoising can use mean filtering or Gaussian filtering method, the contrast enhancement can be based on histogram equalization technology, and the edge detection can use Canny operator or Sobel operator to accurately extract the tool wear area. Subsequently, the preprocessed image is analyzed based on a deep learning model (such as CNN or Transformer architecture) to extract key feature information of the wear area, and matched with the existing tool wear database to generate the image-perceived tool wear coefficient. Then, according to the relative contribution of the tool wear first coefficient and the image-perceived tool wear coefficient, the adaptive weighting strategy is used to determine the tool wear first correction coefficient and the tool wear second correction coefficient. Specifically, α2=1-α1, where W1 is the tool wear first coefficient, W2 is the image-perceived tool wear coefficient, α1 is the tool wear first correction coefficient, α2 is the tool wear second correction coefficient, λ is the weight adjustment parameter for controlling the influence of the wear coefficient difference on the weight, and δ is the balance threshold for avoiding small errors leading to dramatic changes in weight. Finally, according to the tool wear first correction coefficient and the tool wear second correction coefficient, the tool wear first coefficient and the image-perceived tool wear coefficient are weighted to generate the tool wear second coefficient; specifically, corrected W2, where W corrected is the tool wear second coefficient, W1 is the tool wear first coefficient, W2 is the image-perceived tool wear coefficient, α1 is the tool wear first correction coefficient, and α2 is the tool wear second correction coefficient.

[0046] Further, the metal cutting tool is worn feature-aware based on the tool real-time image to obtain an image-perceived tool wear coefficient, including:

[0047] According to the tool real-time image, an adaptive wear feature convolution is performed to obtain a tool wear convolution image; a historical tool wear convolution image set and a historical image-perceived tool wear coefficient set are aligned to obtain an image-perceived tool wear evaluation construction set; a ResNet network is supervised trained based on the image-perceived tool wear evaluation construction set, and a tool wear evaluation loss coefficient is obtained every predetermined number of training times; if the tool wear evaluation loss coefficient is less than a tool wear evaluation loss threshold, an image-perceived tool wear evaluation model is generated; the tool wear convolution image is input into the image-perceived tool wear evaluation model, and the image-perceived tool wear coefficient is output.

[0048] Firstly, based on the tool real-time image, an adaptive wear feature convolution module is constructed by using a convolution neural network (CNN) to perform multi-scale feature extraction on the tool real-time image, and a tool wear convolution image is obtained, wherein the adaptive wear feature convolution module adopts a multi-layer convolution kernel combination structure to enhance the capture ability of different scale wear features; then, based on historical tool wear data, a historical tool wear convolution image set and a historical image-perceived tool wear coefficient set are aligned to form an image-perceived tool wear evaluation construction set, wherein the historical image-perceived tool wear coefficient set is an evaluation result obtained based on actual measurement data or calibration data; next, a deep residual neural network (ResNet) is used to supervise the training of the image-perceived tool wear evaluation construction set, and a tool wear evaluation loss coefficient is calculated every predetermined number of times of training, the tool wear evaluation loss coefficient is calculated based on a mean square error (MSE) loss function or a cross-entropy loss function to measure the deviation between the model prediction result and the true wear coefficient; if the tool wear evaluation loss coefficient is less than a tool wear evaluation loss threshold value, it is determined that the training has converged, and an image-perceived tool wear evaluation model is generated, wherein the tool wear evaluation loss threshold value is set according to experimental calibration or adaptive adjustment to ensure that the evaluation accuracy of the model meets the predetermined requirements; finally, the tool wear convolution image is input into the image-perceived tool wear evaluation model, and based on the feature mapping and the full connection layer output of the model, the image-perceived tool wear coefficient is generated and output to represent the wear degree of the metal cutting tool.

[0049] Further, the adaptive wear feature convolution is performed according to the tool real-time image to obtain a tool wear convolution image, comprising:

[0050] The tool real-time image is subjected to Retinex enhancement processing to obtain a first tool image; the first tool image is subjected to background separation according to a Fisher criterion function to obtain a second tool image; an adaptive wear feature convolution kernel is constructed according to the second tool image, and the second tool image is subjected to wear feature convolution according to the adaptive wear feature convolution kernel to obtain a tool wear initial convolution image; a Roberts operator is introduced to detect wear area edge features of the second tool image to obtain a wear edge detection operator; the tool wear initial convolution image is subjected to edge enhancement according to the wear edge detection operator to generate the tool wear convolution image.

[0051] Firstly, the real-time image of the tool is subjected to Retinex enhancement processing to obtain a first tool image, wherein the Retinex enhancement processing improves the local contrast and detail visibility of the image through multi-scale decomposition and reflection component estimation, thereby enhancing the recognizability of the tool wear area; then, the first tool image is subjected to background separation based on a Fisher criterion function to obtain a second tool image, wherein the Fisher criterion function is used to measure the inter-class variance between the tool area and the background area, and the image is binarized by maximizing the inter-class variance to accurately extract the tool wear area; next, an adaptive wear feature convolution kernel is constructed according to the second tool image, wherein the adaptive wear feature convolution kernel dynamically adjusts the size and weight of the convolution kernel by analyzing the texture features, gray gradient and spatial distribution pattern of the tool wear area, so as to enhance the extraction ability of the wear features, and the second tool image is subjected to wear feature convolution based on the adaptive wear feature convolution kernel to obtain a tool wear initial convolution image; further, a Roberts operator is introduced to detect the wear area edge features of the second tool image to obtain a wear edge detection operator, wherein the Roberts operator is a gradient-based edge detection method that accurately captures the boundary features of the tool wear area by calculating the gradient change rate of the image pixels; finally, the tool wear initial convolution image is subjected to edge enhancement according to the wear edge detection operator, wherein the edge enhancement technique is used to highlight the boundary information of the wear area to improve the contour clarity of the wear area, and a tool wear convolution image is generated for subsequent tool wear evaluation model to calculate and analyze the wear degree.

[0052] Further, the tool wear second coefficient is obtained, comprising:

[0053] It is judged whether the tool wear second coefficient is greater than or equal to a tool wear threshold value; if the tool wear second coefficient is greater than or equal to the tool wear threshold value, a tool wear warning signal is generated.

[0054] After obtaining the tool wear second coefficient, the tool wear second coefficient is subjected to threshold value judgment, specifically, a tool wear threshold value is obtained, which is a safety wear threshold value set based on historical tool wear data and tool failure mechanism modeling, and it is judged whether the tool wear second coefficient is greater than or equal to the tool wear threshold value; if the tool wear second coefficient is greater than or equal to the tool wear threshold value, a tool wear warning signal is generated, wherein the tool wear warning signal is used to indicate that the current tool has reached or exceeded the allowable wear limit, and triggers corresponding tool replacement, processing parameter adjustment or equipment shutdown protection measures to prevent the decline of machining precision or damage of equipment caused by severe tool wear.

[0055] Obtain the historical tool wear coefficient set of the metal cutting tool, combine it with the second tool wear coefficient to predict the wear trend, construct a tool wear trend map, and perform health management on the metal cutting tool based on the tool wear trend map.

[0056] First, a historical set of tool wear coefficients for metal cutting tools during past cutting processes is collected. This set contains data records of tool wear conditions in multiple different cutting cycles. Next, combined with a second tool wear coefficient—the current wear assessment result—wear trend prediction is performed using data analysis and modeling methods to predict the tool wear development trend over a certain period. This process can employ time series analysis, regression analysis, neural networks, and other methods to predict wear trends. For example, a Long Short-Term Memory (LSTM) neural network is used to predict tool wear. Historical tool wear data is collected and organized chronologically as time series data. To improve prediction accuracy, other relevant operating parameters (such as cutting force, temperature, vibration, etc.) can be combined as input features. Then, the dataset is divided into training, validation, and test sets. By training on the training set, the LSTM model learns the long-term and short-term dependencies in the time series data and predicts the tool wear development trend over a certain period. During training, the LSTM network optimizes model parameters by minimizing prediction errors (such as mean squared error). After training, the model can output predicted values ​​of future wear coefficients based on historical wear data and relevant operating parameters.

[0057] Next, based on the predicted wear trend, a tool wear trend map is constructed to show the wear change process of the tool in different cutting cycles and the predicted future wear change, helping to analyze the remaining tool life and the rate of wear development. Based on the tool wear trend map, health management of metal cutting tools is implemented. By monitoring changes in the tool wear trend map, the wear status of the tool can be assessed in real time, and the time of tool failure can be predicted in advance. This allows for the rational scheduling of tool replacement cycles, optimization of cutting parameters, ensuring stable tool performance during machining, and avoiding product quality degradation and reduced production efficiency due to excessive tool wear.

[0058] In summary, the embodiments of this application have at least the following technical effects:

[0059] When the metal cutting tool executes a predetermined cutting scheme, the metal cutting tool is monitored in real time to obtain a multi-source sensing sequence, the multi-source sensing sequence including a cutting force sequence, a vibration signal sequence, an acoustic emission signal sequence, and a temperature signal sequence. Next, the multi-source sensing sequence is filtered and denoised to construct a tool cutting monitoring matrix. Then, the tool cutting monitoring matrix is input into a tool wear detection channel to obtain a tool wear first coefficient. Next, the tool wear first coefficient is differentially detected according to the predetermined cutting scheme to output a wear differential signal. Further, based on the wear differential signal, the tool wear first coefficient is corrected through image perception to obtain a tool wear second coefficient. Finally, a historical tool wear coefficient set of the metal cutting tool is obtained, and the tool wear second coefficient is combined to predict a wear trend, construct a tool wear trend graph, and manage the health of the metal cutting tool according to the tool wear trend graph. The technical problem of low tool wear monitoring precision in the prior art is solved, and the technical effect of improving tool wear monitoring precision is achieved.

[0060] In the second embodiment, based on the same inventive concept as the tool wear monitoring method of the metal cutting tool in the foregoing embodiments, as shown in the following table, the present application provides a tool wear monitoring system of a metal cutting tool. Figure 2 As shown in the following table, the present application provides a tool wear monitoring system of a metal cutting tool.

[0061] The monitoring module 11 is configured to monitor the metal cutting tool in real time when the metal cutting tool executes a predetermined cutting scheme to obtain a multi-source sensing sequence, the multi-source sensing sequence including a cutting force sequence, a vibration signal sequence, an acoustic emission signal sequence, and a temperature signal sequence. The denoising module 12 is configured to filter and denoise the multi-source sensing sequence to construct a tool cutting monitoring matrix. The wear detection module 13 is configured to input the tool cutting monitoring matrix into a tool wear detection channel to obtain a tool wear first coefficient. The differential detection module 14 is configured to differentially detect the tool wear first coefficient according to the predetermined cutting scheme to output a wear differential signal. The correction module 15 is configured to correct the tool wear first coefficient through image perception based on the wear differential signal to obtain a tool wear second coefficient. The management module 16 is configured to obtain a historical tool wear coefficient set of the metal cutting tool, combine the tool wear second coefficient to predict a wear trend, construct a tool wear trend graph, and manage the health of the metal cutting tool according to the tool wear trend graph.

[0062] Further, the wear detection module 13 is configured to execute the following method:

[0063] The tool wear detection channel includes Q tool wear detection models, Q is a positive integer greater than 1; the tool cutting monitoring matrix is input into the Q tool wear detection models, and Q tool wear detection coefficients are obtained; the Q wear detection accuracy coefficients corresponding to the Q tool wear detection models are calculated by proportion, and Q detection output weights are obtained; the Q tool wear detection coefficients are calculated by weighting according to the Q detection output weights, and the tool wear first coefficient is generated.

[0064] Further, the difference detection module 14 is used to execute the following method:

[0065] According to the predetermined cutting scheme, the tool wear prediction coefficient is obtained by predicting the tool wear of the metal cutting tool; the wear detection difference coefficient is calculated according to the tool wear prediction coefficient and the tool wear first coefficient; it is judged whether the wear detection difference coefficient is greater than or equal to the wear detection difference threshold; if the wear detection difference coefficient is greater than or equal to the wear detection difference threshold, the wear difference signal is generated.

[0066] Further, the difference detection module 14 is used to execute the following method:

[0067] The tool pre-wear coefficient of the metal cutting tool is read, wherein the tool pre-wear coefficient is the tool wear coefficient of the metal cutting tool before executing the predetermined cutting scheme; the same type tool networking association is determined according to the metal cutting tool, and a same family tool set is determined; the tool wear amplification retrieval set is obtained by retrieving the tool wear amplification sample of the same family tool set according to the predetermined cutting scheme; the predicted tool wear amplification is obtained by calculating the central value according to the tool wear amplification retrieval set; the tool wear prediction coefficient is obtained by calculating the excitation adjustment of the tool pre-wear coefficient according to the predicted tool wear amplification.

[0068] Further, the correction module 15 is used to execute the following method:

[0069] Based on the wear difference signal, the tool real-time image of the metal cutting tool is called; the image perception tool wear coefficient is obtained by sensing the wear characteristics of the metal cutting tool according to the tool real-time image; the tool wear first correction coefficient and the tool wear second correction coefficient are obtained by calculating the adaptive correction parameters according to the tool wear first coefficient and the image perception tool wear coefficient; the tool wear first coefficient and the image perception tool wear coefficient are calculated by weighting according to the tool wear first correction coefficient and the tool wear second correction coefficient, and the tool wear second coefficient is generated.

[0070] Further, the correction module 15 is used to execute the following method:

[0071] According to the tool real-time image, adaptive wear feature convolution is performed to obtain a tool wear convolution image; a historical tool wear convolution image set and a historical image-perceived tool wear coefficient set are aligned to obtain a tool wear evaluation construction set; a ResNet network is supervised trained according to the tool wear evaluation construction set, and a tool wear evaluation loss coefficient is obtained when training a predetermined number of times; if the tool wear evaluation loss coefficient is less than a tool wear evaluation loss threshold, an image-perceived tool wear evaluation model is generated; and the tool wear convolution image is input into the image-perceived tool wear evaluation model, and the image-perceived tool wear coefficient is output.

[0072] Further, the correction module 15 is configured to perform the following method:

[0073] The Retinex enhancement processing is performed on the tool real-time image to obtain a first tool image; the Fisher criterion function is used to perform background separation on the first tool image to obtain a second tool image; an adaptive wear feature convolution kernel is constructed according to the second tool image, and wear feature convolution is performed on the second tool image according to the adaptive wear feature convolution kernel to obtain a tool wear initial convolution image; the Roberts operator is introduced to perform wear area edge feature detection on the second tool image to obtain a wear edge detection operator; and the edge of the tool wear initial convolution image is enhanced according to the wear edge detection operator to generate the tool wear convolution image.

[0074] Further, the denoising module 12 is configured to perform the following method:

[0075] The Kalman filtering is performed on the cutting force sequence to obtain a cutting force matrix; the band-pass filtering is performed on the vibration signal sequence to obtain a vibration signal matrix; the band-pass filtering is performed on the acoustic emission signal sequence to obtain an acoustic emission signal matrix; the Kalman filtering is performed on the temperature signal sequence to obtain a temperature signal matrix; and the time window alignment splicing is performed according to the cutting force matrix, the vibration signal matrix, the acoustic emission signal matrix and the temperature signal matrix to generate the tool cutting monitoring matrix.

[0076] Further, the management module 16 is configured to perform the following method:

[0077] It is judged whether the tool wear second coefficient is greater than or equal to a tool wear threshold; if the tool wear second coefficient is greater than or equal to the tool wear threshold, a tool wear early warning signal is generated.

[0078] It should be noted that the above-mentioned embodiment sequences of the present application are merely for description only, but not for representing the advantages and disadvantages of the embodiments. And the above-mentioned embodiment sequences of the present application are described in the specification. The processes depicted in the drawings do not necessarily require the particular sequence or continuous sequence shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or can be advantageous.

[0079] The above only describes the preferred embodiments of the present application, and does not limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0080] The specification and drawings are merely exemplary of the present application, and any and all modifications, variations, combinations or equivalents that are within the scope of the present application should be considered covered by the present application. Obviously, those skilled in the art can make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the present application and its equivalent technology, the present application is intended to include these modifications and variations.

Claims

1. A metal cutting tool wear monitoring method, c h a r a c t e r i s e d in that, The method comprises: real-time monitoring of the metal cutting tool when the metal cutting tool performs a predetermined cutting scheme, to obtain a multi-source sensing sequence, the multi-source sensing sequence comprising a cutting force sequence, a vibration signal sequence, an acoustic emission signal sequence and a temperature signal sequence; filtering and denoising according to the multi-source sensing sequence to construct a tool cutting monitoring matrix; inputting the tool cutting monitoring matrix into a tool wear detection channel to obtain a tool wear first coefficient; differential detection of the tool wear first coefficient according to the predetermined cutting scheme to output a wear differential signal; based on the wear differential signal, image perception correction is performed on the tool wear first coefficient to obtain a tool wear second coefficient; obtaining a historical tool wear coefficient set of the metal cutting tool, combining the tool wear second coefficient to perform wear trend prediction, constructing a tool wear trend graph, and performing health management on the metal cutting tool according to the tool wear trend graph; wherein inputting the tool cutting monitoring matrix into the tool wear detection channel to obtain the tool wear first coefficient comprises: the tool wear detection channel comprises Q tool wear detection models, Q being a positive integer greater than 1; inputting the tool cutting monitoring matrix into the Q tool wear detection models to obtain Q tool wear detection coefficients; occupancy calculation is performed according to Q wear detection accuracy coefficients corresponding to the Q tool wear detection models to obtain Q detection output weights; weighting calculation is performed on the Q tool wear detection coefficients according to the Q detection output weights to generate the tool wear first coefficient; wherein differential detection of the tool wear first coefficient according to the predetermined cutting scheme to output a wear differential signal comprises: wear prediction is performed on the metal cutting tool according to the predetermined cutting scheme to obtain a tool wear prediction coefficient; wear detection differential coefficient is calculated according to the tool wear prediction coefficient and the tool wear first coefficient; it is judged whether the wear detection differential coefficient is greater than or equal to a wear detection differential threshold value; if the wear detection differential coefficient is greater than or equal to the wear detection differential threshold value, the wear differential signal is generated; wherein, based on the wear differential signal, image perception correction is performed on the tool wear first coefficient to obtain a tool wear second coefficient, comprising: based on the wear differential signal, the tool real-time image of the metal cutting tool is called; wear feature perception is performed on the metal cutting tool according to the tool real-time image to obtain an image perception tool wear coefficient; adaptive correction parameter calculation is performed according to the tool wear first coefficient and the image perception tool wear coefficient to obtain a tool wear first correction coefficient and a tool wear second correction coefficient; weighting calculation is performed on the tool wear first coefficient and the image perception tool wear coefficient according to the tool wear first correction coefficient and the tool wear second correction coefficient to generate the tool wear second coefficient.

2. A metal cutting tool wear monitoring method according to claim 1, characterized in that, wear prediction is performed on the metal cutting tool according to the predetermined cutting scheme to obtain a tool wear prediction coefficient, comprising: reading a tool pre-wear coefficient of the metal cutting tool, wherein the tool pre-wear coefficient is a tool wear coefficient of the metal cutting tool before performing the predetermined cutting scheme; performing same-model tool network association according to the metal cutting tool to determine a same-family tool set; performing tool wear increment sample retrieval on the same-family tool set according to the predetermined cutting scheme to obtain a tool wear increment retrieval set; performing centralized value calculation according to the tool wear increment retrieval set to obtain a predicted tool wear increment; performing incentive adjustment calculation on the tool pre-wear coefficient according to the predicted tool wear increment to obtain the tool wear prediction coefficient.

3. A metal cutting tool wear monitoring method as claimed in claim 1, characterized by, performing wear feature perception on the metal cutting tool according to the tool real-time image to obtain an image-perceived tool wear coefficient, including: performing adaptive wear feature convolution according to the tool real-time image to obtain a tool wear convolution image; aligning a historical tool wear convolution image set and a historical image-perceived tool wear coefficient set to obtain an image-perceived tool wear evaluation construction set; performing supervised training on a ResNet network according to the image-perceived tool wear evaluation construction set, and obtaining a tool wear evaluation loss coefficient every predetermined number of training times; generating an image-perceived tool wear evaluation model if the tool wear evaluation loss coefficient is less than a tool wear evaluation loss threshold; inputting the tool wear convolution image into the image-perceived tool wear evaluation model to output the image-perceived tool wear coefficient.

4. A metal cutting tool wear monitoring method according to claim 3, characterized in that, performing adaptive wear feature convolution according to the tool real-time image to obtain a tool wear convolution image, including: performing Retinex enhancement processing on the tool real-time image to obtain a first tool image; performing background separation on the first tool image according to a Fisher criterion function to obtain a second tool image; constructing an adaptive wear feature convolution kernel according to the second tool image, and performing wear feature convolution on the second tool image according to the adaptive wear feature convolution kernel to obtain a tool wear initial convolution image; introducing a Roberts operator to perform wear area edge feature detection on the second tool image to obtain a wear edge detection operator; performing edge enhancement on the tool wear initial convolution image according to the wear edge detection operator to generate the tool wear convolution image.

5. A metal cutting tool wear monitoring method as claimed in claim 1, characterized by, performing filtering and denoising according to the multi-source sensing sequence to construct a tool cutting monitoring matrix, including: performing Kalman filtering on the cutting force sequence to obtain a cutting force matrix; performing band-pass filtering on the vibration signal sequence to obtain a vibration signal matrix; performing band-pass filtering on the acoustic emission signal sequence to obtain an acoustic emission signal matrix; performing Kalman filtering on the temperature signal sequence to obtain a temperature signal matrix; performing time window alignment splicing according to the cutting force matrix, the vibration signal matrix, the acoustic emission signal matrix, and the temperature signal matrix to generate the tool cutting monitoring matrix.

6. A metal cutting tool wear monitoring method as claimed in claim 1, characterized by, obtaining a tool wear second coefficient, including: determining whether the tool wear second coefficient is greater than or equal to a tool wear threshold value; If the second coefficient of tool wear is greater than or equal to the tool wear threshold, a tool wear early warning signal is generated.

7. A metal cutting tool wear monitoring system characterized by, A metal cutting tool wear monitoring method for implementing any one of claims 1-6, the system comprising: a monitoring module for monitoring the metal cutting tool in real time when the metal cutting tool performs a predetermined cutting scheme, obtaining a multi-source sensing sequence, the multi-source sensing sequence comprising a cutting force sequence, a vibration signal sequence, an acoustic emission signal sequence, and a temperature signal sequence; a denoising module for filtering and denoising according to the multi-source sensing sequence, constructing a tool cutting monitoring matrix; a wear detection module for inputting the tool cutting monitoring matrix into a tool wear detection channel, obtaining a first coefficient of tool wear; a difference detection module for differentiating the first coefficient of tool wear according to the predetermined cutting scheme, outputting a wear difference signal; a correction module for correcting the first coefficient of tool wear based on the wear difference signal, obtaining a second coefficient of tool wear; a management module for obtaining a historical tool wear coefficient set of the metal cutting tool, combining the second coefficient of tool wear to predict a wear trend, constructing a tool wear trend graph, and managing the health of the metal cutting tool according to the tool wear trend graph.

Citation Information

Patent Citations

  • CNC machine tool cutter wear monitoring method and system based on acoustic vibration sensing technology

    CN118404392A

  • Cutter residual life prediction method based on adaptive stage division

    CN119347540A