Method and system for detecting abnormal deflection of tool

Abstract

The invention relates to a method and system for detecting abnormal deflection of a tool. The method comprises the steps that first vibration signal data generated in the idling process of a tool spindle are collected through a vibration sensor, the first vibration signal data are processed, and training data are obtained; based on the training data, training a class of SVM models to obtain a trained classification model; second vibration signal data of the tool spindle in the real-time operation state are collected through a vibration sensor, the second vibration signal data are processed, and the processed second vibration signal data are obtained; on the basis of a classification model, predicting the processed second vibration signal data to obtain a classification result, and on the basis of the classification result, obtaining a proportion of a positive sample to a negative sample; calculating the deviation degree of the proportion of the positive sample and the negative sample in the classification result; the deviation degree is compared with a preset threshold value, and when the deviation degree is larger than the threshold value, the NC system sends out an alarm signal. And the abnormal deflection detection efficiency and the response speed of the cutter are improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent machining technology, and in particular to a method and system for detecting abnormal deflection of a tool. Background Art

[0002] As the manufacturing industry gradually develops towards high-end fields, the demand for CNC machine tools continues to increase, and the related demand for automated unmanned factories is also growing. In order to improve processing efficiency and reduce production costs, more and more factories and new production lines have begun to transform or build automated production lines, which has greatly liberated productivity. However, the transformation and use of automated production lines also bring many problems. For example, CNC machine tools use CNC programs to achieve automatic tool change during processing, but due to the large amount of fine metal chips generated during the processing, these metal chips may adhere to the tool handle of the cutter disc. When the machine tool performs automatic tool change operations, these tool handles with metal chips may be replaced on the spindle, resulting in an increase in the runout of the current tool, which in turn affects the processing accuracy, increases the tolerance of the product, and significantly increases the defective rate. In addition, when the spindle of the machine tool is worn, some rigid parts are used for a long time or poor lubrication causes excessive wear of the lead screw or guide rail, it will also lead to a decrease in the processing accuracy of the machine tool.

[0003] There are two existing solutions: The tool setting probe is used to measure the physical runout of the tool tip. This contact detection method requires the spindle to move to the specified position (i.e. the tool setting probe position) for detection, and then return to the origin after the detection is completed. Its advantage is high detection accuracy, but its disadvantage is that the detection time is long and the spindle needs to be moved repeatedly, which seriously affects the processing cycle.

[0004] Eddy current sensors are used to detect the runout displacement. Compared with the previous solution, this detection method is faster, but it requires higher sensor accuracy. Since the accuracy of the eddy current sensor is inversely proportional to the detection distance (that is, the higher the accuracy, the closer the distance), if the distance is too close, the sensor may be hit during the automatic tool change process. The measuring distance is usually required to be no less than 5mm, otherwise it will affect the automatic tool change operation. In order to consider the processing rhythm, the sensor is usually installed on the spindle, directly aligned with the measured position of the tool holder, and indirectly measures the runout of the tool holder to estimate the runout of the tool tip. However, the sensor installed on the spindle may be displaced due to the violent vibration of the spindle, or the oil and iron filings during the processing may cover the probe, resulting in loss of detection accuracy. In addition, the high cost of high-precision eddy current sensors limits their popularity and promotion. Summary of the invention

[0005] Therefore, the technical problem to be solved by the present invention is to overcome the problems of long detection time, inconvenient sensor installation, high cost, etc. in the prior art for the deflection detection of the tool handle.

[0006] In order to solve the above technical problems, the present invention provides a method for detecting abnormal deflection of a tool, the method comprising: Install vibration sensors on the tool spindles of CNC machine tools; The vibration sensor is used to collect first vibration signal data generated by the tool spindle during idling, and the first vibration signal data is processed to obtain training data; Based on the training data, training a class of SVM models to obtain a trained classification model; Collecting second vibration signal data of the tool spindle in a real-time running state by the vibration sensor, processing the second vibration signal data to obtain processed second vibration signal data; Based on the classification model, predicting the processed second vibration signal data to obtain a classification result; Based on the classification result, obtaining the ratio of positive samples to negative samples; Calculating the degree of deviation of the ratio of positive samples to negative samples in the classification result; The deviation degree is compared with a preset threshold value, and in response to the deviation degree being greater than the threshold value, the NC system issues an alarm signal.

[0007] In one embodiment of the present invention, processing the first vibration signal data to obtain training data includes: Performing denoising processing on the first vibration signal data; Performing normalization processing on the first vibration signal data after denoising; The first vibration signal data after normalization is subjected to feature extraction including time domain features, frequency domain features, and time-frequency domain features.

[0008] In one embodiment of the present invention, based on the training data, a type of SVM model is trained, including: Select radial basis function as the kernel function for training.

[0009] In one embodiment of the present invention, processing the second vibration signal data to obtain the processed second vibration signal data includes: Performing denoising processing on the second vibration signal data; performing normalization processing on the denoised second vibration signal data; The normalized second vibration signal data is subjected to feature extraction including time domain features, frequency domain features, and time-frequency domain features.

[0010] In one embodiment of the present invention, the ratio of initial positive samples to negative samples in the classification model is 1:1.

[0011] In one embodiment of the present invention, the operating frequency of the vibration sensor is set to 26666 pts / s.

[0012] In one embodiment of the present invention, the threshold value is between 0-50%.

[0013] In one embodiment of the present invention, the vibration sensor is adsorbed on the tool spindle via a neodymium magnet base.

[0014] The present invention also provides a system for detecting abnormal deflection of a tool, characterized in that it comprises: A training data acquisition unit, used for collecting first vibration signal data generated by the tool spindle during idling through the vibration sensor, and processing the first vibration signal data to obtain training data; A classification model acquisition module, used to train a class of SVM models based on the training data to obtain a trained classification model; A second vibration signal data acquisition module, used for collecting second vibration signal data of the tool spindle in a real-time running state through the vibration sensor, processing the second vibration signal data, and obtaining processed second vibration signal data; A classification result acquisition module, used to predict the processed second vibration signal data based on the classification model to obtain a classification result; A ratio acquisition module, used to acquire the ratio of positive samples to negative samples based on the classification result; A deviation degree acquisition module, used to calculate the deviation degree of the ratio of positive samples to negative samples in the classification result; The alarm module is used to compare the deviation degree with a preset threshold value, and in response to the deviation degree being greater than the threshold value, the NC system issues an alarm signal.

[0015] The above technical solution of the present invention has the following advantages compared with the prior art: The method and system for detecting abnormal tool deflection described in the present invention utilizes vibration signal data collected in real time to quickly analyze the tool status and detect abnormalities through a machine learning model (a type of SVM model). Compared with the traditional method with a longer detection time, the present invention can monitor the spindle vibration changes in real time during normal processing, quickly identify abnormal deflection, and significantly improve the detection efficiency and response speed. When the positive and negative sample ratios of the classification results deviate from the preset threshold, the NC system sends an alarm signal. The alarm threshold can be adjusted according to the actual usage scenario and equipment characteristics to ensure the sensitivity and accuracy of the detection.

[0016] The acceleration sensor in the present invention is firmly adsorbed on the spindle through the neodymium magnet base, eliminating the step of drilling the spindle, making installation and removal simpler and faster. This installation method not only avoids additional vibration or noise interference caused by improper installation, but also ensures that the sensor will not be displaced or loosened due to the high-speed rotation of the spindle. Compared with the traditional tool setting method, the installation process of the present invention does not require the spindle to be moved to a specific position, avoiding the interruption of the processing beat and improving the overall production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to make the contents of the present invention more clearly understood, the present invention is further described in detail below based on specific embodiments of the present invention in conjunction with the accompanying drawings.

[0018] Figure 1 It is a flow chart of the method for detecting abnormal deflection of a tool according to the present invention.

[0019] Figure 2 It is a schematic diagram of the vibration sensor of the present invention installed on the main shaft. DETAILED DESCRIPTION

[0020] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments so that those skilled in the art can better understand the present invention and implement it, but the embodiments are not intended to limit the present invention.

[0021] In the present invention, "several" means one or more, "multiple" means more than two, "greater than", "less than", "exceed" and the like are understood to exclude the number itself; "above", "below", "within" and the like are understood to include the number itself. In the description of the present invention, if there is a description of "first" or "second", it is only used for the purpose of distinguishing technical features, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features or implicitly indicating the order of the indicated technical features.

[0022] Example 1 Reference Figure 1 As shown, a method for detecting abnormal deflection of a tool in an embodiment of the present invention comprises: S1. Install a vibration sensor on the tool spindle of the CNC machine tool; the vibration sensor has a high sampling frequency, specifically 26666 pts / s and sufficient dynamic range to capture the vibration characteristics of the spindle when idling and running, and has high sensitivity to ensure that small vibration changes can be detected; In addition, refer to Figure 2As shown in the figure, the vibration sensor is firmly attached to the tool spindle through the neodymium magnet base. The high magnetic force of the neodymium magnet ensures that the sensor will not move or loosen at high speeds, thereby avoiding additional vibration or noise interference caused by improper installation, and avoiding measurement errors caused by oil and coolant splashing.

[0023] This installation method is simple and easy, no need to drill holes in the main shaft, and can be quickly installed and disassembled, greatly facilitating the maintenance and replacement of the sensor.

[0024] The data acquisition of the vibration sensor is connected to the edge device (such as IO / Ethernet module) through the machine tool NC program. This not only enables real-time data transmission, but also can be integrated with the machine tool control system to further improve the real-time response of the system.

[0025] S2. Collecting first vibration signal data generated by the tool spindle during idling (no load) through the vibration sensor, and processing the first vibration signal data to obtain training data.

[0026] Vibration signals are often mixed with a variety of noises, such as environmental noise, equipment noise, and signal transmission noise. The first vibration signal collected is denoised. The method used can use a low-pass filter to remove high-frequency noise and a band-pass filter to filter out signals outside the natural frequency range of the non-spindle. Empirical mode decomposition (EMD) can also be used to separate the main vibration modes and remove high-frequency noise and low-frequency drift. The EMD method can adaptively process nonlinear and non-stationary signals, so it is particularly suitable for vibration signal processing of machine tool spindles.

[0027] The denoised data is normalized to scale the signal value to a standard range (e.g., 0 to 1) to facilitate subsequent machine learning model processing. Normalization can reduce the numerical differences between different data groups and ensure that the subsequent machine learning model training and detection are more stable and efficient.

[0028] Extract the time domain features of the vibration signal, such as mean, root mean square (RMS), peak, skewness, kurtosis, etc. These features can reflect the amplitude and waveform characteristics of the vibration signal, and reflect the vibration intensity and volatility of the tool spindle.

[0029] Perform fast Fourier transform (FFT) on the signal to extract frequency domain features, such as main frequency, spectrum energy, bandwidth, frequency center, etc. These features can reflect the frequency characteristics and energy distribution of the main shaft vibration.

[0030] Use methods such as wavelet transform or short-time Fourier transform (STFT) to extract time-frequency domain features and capture the changing characteristics of the signal in time and frequency.

[0031] S3. Based on the training data, train a class of SVM models to obtain a trained classification model.

[0032] One-Class SVM is suitable for scenarios with only normal state data (single-class samples) and can be used for anomaly detection. It is used for anomaly detection in the present invention. Its basic principle is to identify abnormal samples outside the distribution boundary of normal state samples by learning the distribution boundary of normal state samples. The radial basis function (RBF) kernel function is used to process nonlinear data. The RBF kernel function can map low-dimensional input data to high-dimensional space, so that linearly inseparable data becomes linearly separable in high-dimensional space.

[0033] Adjust key hyperparameters of SVM, such as nu (controls the number of support vectors, indicating tolerance to outliers) and gamma (width of the RBF kernel function). Select the optimal hyperparameter combination through cross-validation or grid search to improve the detection accuracy of the model.

[0034] The training data is used to train the model and obtain the classification model. The goal of the training process is to find a hyperplane (or hypersphere) to maximize the boundary of normal samples and minimize the number of samples outside the boundary.

[0035] S4. Collect the second vibration signal data of the tool spindle in the real-time operation state through the vibration sensor, process the second vibration signal data, and obtain the processed second vibration signal data. The second vibration signal data is the vibration characteristics of the spindle under actual processing conditions, including any possible abnormal vibration (such as tool wear, imbalance, resonance, etc.).

[0036] The acquisition time and conditions of the second vibration signal data are consistent with those in the idling state to facilitate comparative analysis.

[0037] The second vibration data in real-time operation is processed in the same way as the training data, including denoising, normalization, feature extraction, etc., to ensure the consistency of data format and features.

[0038] S5. Based on the classification model, predict the processed second vibration signal data to obtain a classification result.

[0039] S6. Based on the classification result, obtain the ratio of positive samples to negative samples.

[0040] The ratio calculation formula is as follows: Ppositive = number of positive samples / total number of samples; Pnegative = number of negative samples / total number of samples; the initial ratio of positive samples to negative samples in the classification model is set to 1:1.

[0041] It is understood that positive samples generally refer to data that are recognized by the model as being in accordance with the expected state. In this embodiment, the positive samples represent the vibration signal data of the CNC machine tool spindle under normal working conditions. The negative samples represent the vibration signal data of the CNC machine tool spindle under abnormal working conditions (such as tool wear, imbalance, resonance, or the runout of the tool holder exceeds the normal range).

[0042] S7. Calculate the degree of deviation of the ratio of positive samples to negative samples in the classification result; the degree of deviation is the absolute value of the difference between Ppositive and Pnegative.

[0043] S8. Compare the degree of deviation with a preset threshold value. In response to the degree of deviation being greater than the threshold value, the NC system issues an alarm signal. The alarm threshold value (0-50%, adjusted according to the actual use scenario and equipment characteristics) indicates that if the degree of deviation exceeds this threshold value, the equipment status may be abnormal. According to the alarm signal, the NC system can execute preset response measures, such as shutdown protection, speed reduction, maintenance prompt, etc. The alarm information can also be recorded in the log for subsequent analysis and improvement.

[0044] In addition, it should be noted that, depending on the installation position of the sensor, the location for detection is not limited to the spindle, but can also be installed at the bearing or the screw guide position.

[0045] Example 2 Based on the same inventive concept, this embodiment provides a system for detecting abnormal tool deflection. The principle of solving the problem is similar to the method for detecting abnormal tool deflection, and the repeated parts will not be repeated.

[0046] This embodiment provides a system for detecting abnormal deflection of a tool, comprising: The present invention also provides a system for detecting abnormal deflection of a tool, characterized in that it comprises: A training data acquisition unit, used for collecting first vibration signal data generated by the tool spindle during idling through the vibration sensor, and processing the first vibration signal data to obtain training data; A classification model acquisition module, used to train a class of SVM models based on the training data to obtain a trained classification model; A second vibration signal data acquisition module, used for collecting second vibration signal data of the tool spindle in a real-time running state through the vibration sensor, processing the second vibration signal data, and obtaining processed second vibration signal data; A classification result acquisition module, used to predict the processed second vibration signal data based on the classification model to obtain a classification result; A ratio acquisition module, used to acquire the ratio of positive samples to negative samples based on the classification result; A deviation degree acquisition module, used to calculate the deviation degree of the ratio of positive samples to negative samples in the classification result; The alarm module is used to compare the deviation degree with a preset threshold value, and in response to the deviation degree being greater than the threshold value, the NC system issues an alarm signal.

[0047] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0048] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0049] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0050] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0051] Finally, it should be noted that the above specific implementation methods are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to examples, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention, which should be included in the scope of the claims of the present invention.

Claims

1. A method for detecting abnormal deflection of a tool, characterized in that: The method comprises: Install vibration sensors on the tool spindles of CNC machine tools; The vibration sensor is used to collect first vibration signal data generated by the tool spindle during idling, and the first vibration signal data is processed to obtain training data; Based on the training data, training a class of SVM models to obtain a trained classification model; Collecting second vibration signal data of the tool spindle in a real-time running state by the vibration sensor, processing the second vibration signal data to obtain processed second vibration signal data; Based on the classification model, predicting the processed second vibration signal data to obtain a classification result; Based on the classification result, obtaining the ratio of positive samples to negative samples; Calculating the degree of deviation of the ratio of positive samples to negative samples in the classification result; The deviation degree is compared with a preset threshold value, and in response to the deviation degree being greater than the threshold value, the NC system issues an alarm signal.

2. A method for detecting abnormal tool deflection according to claim 1, characterized in that: Processing the first vibration signal data to obtain training data includes: Performing denoising processing on the first vibration signal data; Performing normalization processing on the first vibration signal data after denoising; The first vibration signal data after normalization is subjected to feature extraction including time domain features, frequency domain features, and time-frequency domain features.

3. A method for detecting abnormal tool deflection according to claim 1, characterized in that: Based on the training data, a SVM model is trained, including: Select radial basis function as the kernel function for training.

4. A method for detecting abnormal tool deflection according to claim 1, characterized in that: Processing the second vibration signal data to obtain processed second vibration signal data includes: Performing denoising processing on the second vibration signal data; performing normalization processing on the denoised second vibration signal data; The normalized second vibration signal data is subjected to feature extraction including time domain features, frequency domain features, and time-frequency domain features.

5. The method for detecting abnormal tool deflection according to claim 1, characterized in that: The initial ratio of positive samples to negative samples in the classification model is 1:

1.

6. A method for detecting abnormal tool deflection according to claim 1, characterized in that: The vibration sensor's adopted frequency is set to 26666 pts / s.

7. A method for detecting abnormal tool deflection according to claim 1, characterized in that: The threshold is between 0-50%.

8. The method for detecting abnormal tool deflection according to claim 1, characterized in that: The vibration sensor is adsorbed on the tool spindle via a neodymium magnet base.

9. A system for detecting abnormal deflection of a tool, characterized in that: include: A training data acquisition unit, used for collecting first vibration signal data generated by the tool spindle during idling through the vibration sensor, and processing the first vibration signal data to obtain training data; A classification model acquisition module, used to train a class of SVM models based on the training data to obtain a trained classification model; A second vibration signal data acquisition module, used for collecting second vibration signal data of the tool spindle in a real-time running state through the vibration sensor, processing the second vibration signal data, and obtaining processed second vibration signal data; A classification result acquisition module, used to predict the processed second vibration signal data based on the classification model to obtain a classification result; A ratio acquisition module, used to acquire the ratio of positive samples to negative samples based on the classification result; A deviation degree acquisition module, used to calculate the deviation degree of the ratio of positive samples to negative samples in the classification result; The alarm module is used to compare the deviation degree with a preset threshold value, and in response to the deviation degree being greater than the threshold value, the NC system issues an alarm signal.

Images