Milling process vibration state monitoring method and system for machine tool milling cutter

Abstract

The invention discloses a milling process vibration state monitoring method and system for a machine tool milling cutter, and belongs to the technical field of intelligent monitoring. The method comprises the following steps: acquiring a vibration signal of a milling cutter under a reference processing condition, and establishing a reference template containing normal vibration characteristics; dynamically adjusting an alarm threshold value, and starting fault diagnosis analysis; performing envelope spectrum analysis on the vibration signal triggering diagnosis; preliminarily judging the affiliation fault type of the abnormal vibration through a preset fault classification model; and carrying out association verification by combining the attribution fault type and the spatial position information when the abnormality occurs, confirming the root cause of the fault, and outputting a targeted classification alarm instruction. According to the invention, the cutter handle single sensor is linked with phase slicing and dynamic threshold and coordinate verification, so that the faults of the cutter, the cutter rest and the workpiece are accurately distinguished and directionally alarmed, and the problems that a fault source cannot be distinguished by a single channel and the false alarm rate is high in the prior art are solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent monitoring, and in particular to a milling process vibration state monitoring method and system for a machine tool milling cutter. BACKGROUND

[0002] With the transformation of manufacturing industry to high precision, high efficiency and intelligentization, real-time state monitoring of machine tool processing process has become a key link to ensure product quality, prolong equipment life and improve production safety. As a widely used cutting method in machining, the vibration state of the milling cutter directly affects the machining precision, surface quality and tool wear rate during the milling process. Therefore, accurate monitoring and abnormal diagnosis of the vibration state of the milling process are of great practical significance.

[0003] Existing machine tool milling cutter vibration monitoring systems mostly collect vibration signals by installing acceleration sensors on the milling cutter handle, spindle or workbench, and judge whether the vibration is abnormal based on a pre-set fixed threshold, or extract limited feature parameters through simple frequency spectrum analysis for preliminary fault judgment.

[0004] In the prior art, due to the dynamic changes of processing parameters (such as rotation speed and cutting depth), the normal vibration level will be significantly different, and the fixed threshold is difficult to adapt to different working conditions, often resulting in false positives or missed detections. At the same time, the feature extraction of the vibration signal is not fine enough, and it is difficult to distinguish the specific vibration characteristics of different fault types such as tool chipping, tool holder loosening and workpiece vibration, resulting in ambiguous fault discrimination, making it difficult to accurately trace the root cause of the fault and affecting the efficiency of subsequent targeted treatment. SUMMARY

[0005] The embodiments of the present application provide a milling process vibration state monitoring method and system for a machine tool milling cutter, which solves the problem that a single channel cannot distinguish between tool, tool holder and workpiece faults in the prior art, and realizes accurate positioning of the fault source and gives directional maintenance instructions.

[0006] The embodiments of the present application provide a milling process vibration state monitoring method for a machine tool milling cutter, comprising: acquiring vibration signals of the milling cutter under a reference processing condition, and establishing a reference template containing normal vibration characteristics based on the vibration signals; Real-time acquisition of vibration signals of the milling cutter during the milling process, and dynamic adjustment of the alarm threshold, when the energy feature of the real-time vibration signal is not less than the dynamic alarm threshold, the fault diagnosis analysis is started; Envelope spectrum analysis is performed on the vibration signal triggering the diagnosis, and frequency domain feature parameters representing the tool, tool holder and workpiece state are extracted; Quantitative comparison of the extracted frequency domain feature parameters with the features in the reference template, and preliminary identification of the attribution fault type of the vibration abnormality through a pre-set fault classification model; The real-time coordinate data of the machine tool numerical control system is acquired, and the spatial position information at the time of the abnormality is associated and verified in combination with the fault type, the root cause of the fault is confirmed, and a targeted classified alarm instruction is output.

[0007] Further, the step of establishing a reference template containing normal vibration characteristics based on the vibration signal comprises: Under standard machining conditions, the machine tool is started and a preset reference cutting program is executed; The vibration signal of the milling cutter during the execution of the reference cutting program is continuously collected to obtain reference time series data; The reference time series data is processed to obtain the average value of the overall vibration energy, and the average value is stored as the reference energy; The reference time series data is envelope demodulated to generate a reference envelope spectrum; In the reference envelope spectrum, the tooth passing frequency feature is obtained; The reference envelope spectrum is divided into a low frequency band below the tooth passing frequency, a neighboring frequency band centered on the tooth passing frequency, and a high frequency band above the tooth passing frequency, to obtain the energy distribution characteristics of each frequency band; The recorded reference energy, tooth passing frequency feature, and energy distribution characteristics of each frequency band are combined to form a structured data set, and a reference template is obtained.

[0008] Further, the step of collecting the vibration signal of the milling cutter during the milling process and dynamically adjusting the alarm threshold value comprises: When performing an actual machining task, real-time vibration time series signals are acquired, and real-time machining parameters are simultaneously acquired from the machine tool numerical control system; A built-in basic threshold value is set based on the reference energy level in the reference template; Based on the acquired real-time machining parameters, a rotational speed adjustment multiplier and a cutting depth adjustment multiplier are obtained; The basic threshold value is multiplied by the rotational speed adjustment multiplier and the cutting depth adjustment multiplier to obtain a dynamic alarm threshold value.

[0009] Further, the step of performing envelope spectrum analysis on the vibration signal triggering diagnosis to extract frequency domain feature parameters representing the state of the tool, tool holder, and workpiece comprises: The real-time vibration time series signal triggering fault diagnosis analysis is intercepted; An analytical time series signal corresponding to the time series signal segment is constructed, which contains the amplitude and phase information of the original time series signal; The modulus of the analytical time series signal at each time point is obtained, and a vibration envelope signal without carrier frequency is extracted; The vibration envelope signal is subjected to fast Fourier transform to generate an envelope spectrum. Based on the generated envelope spectrum, predefined frequency domain feature parameters are automatically identified and extracted.

[0010] Furthermore, the step of quantizing and comparing the extracted frequency domain feature parameters with the features in the benchmark template includes: Extracted frequency domain feature parameters are retrieved from the real-time vibration time series signal that triggers fault diagnosis analysis; Retrieve reference feature parameters corresponding to the frequency domain feature parameters from the reference template. The reference feature parameters include reference cutter passing frequency energy and reference low-frequency band energy. The extracted frequency domain feature parameters are quantitatively compared with the benchmark feature parameters to obtain the energy ratio, low-frequency energy drift, sideband energy ratio, and high-frequency concentration index.

[0011] Furthermore, the preset fault classification model is as follows: The input vector of the fault classification model is the feature mutation vector: ; In the formula, The relative deviation of the frequency amplitude of the cutting teeth. This represents the absolute difference in the proportion of low-frequency energy. This represents the peak growth rate of high-frequency narrowband bandwidth. The decision function of the fault classification model is: ; In the formula, The threshold for the amplitude of tool chipping. The low-frequency energy exceeding the threshold for tool holder loosening is a concern. To eliminate the upper limit of high-frequency growth of flutter, The high-frequency spike threshold for workpiece chatter. This is the negative bias threshold for the amplitude of the tool without damage.

[0012] Further steps to preliminarily determine the fault type to which the vibration anomaly belongs include: The input vector generated after quantization and comparison is transmitted to the preset fault classification model; Following a preset order, the input vectors are matched and verified one by one with each group of logical rules in the fault classification model. If the input vector completely matches all the conditions of one set of rules, the subsequent rule verification is stopped, and the fault type uniquely corresponding to that set of rules is determined as the preliminary judgment result of this diagnosis.

[0013] Furthermore, the step of acquiring real-time coordinate data from the machine tool CNC system and verifying its correlation with the attributed fault type and the spatial location information at the time of the anomaly includes: Based on the preliminary judgment results, the three-dimensional Cartesian coordinates of the milling cutter cutting point at the moment when the vibration anomaly occurred were obtained; By pre-storing a digital geometric model of the currently processed workpiece, an association judgment is performed; If the initial fault type is determined to be a tool holder or spindle malfunction, then determine whether the acquired real-time coordinates fall within the high rigidity area marked in the digital geometric model. If the determination is yes, then the fault is confirmed. If the initial fault type is workpiece or fixture chatter, the system determines whether the acquired real-time coordinates fall within the marked low-rigidity area. If so, the fault is confirmed.

[0014] Furthermore, the steps for outputting targeted, categorized alarm commands include: After verifying and confirming the root cause of the fault through spatial location information, a corresponding text alarm command is selected and generated from the preset command library based on the finally confirmed fault type. If the root cause of the fault is tool damage, the generated alarm instruction will prompt you to check and replace the milling cutter insert. If the root cause of the fault is an abnormality in the tool holder or spindle, the generated alarm command will prompt you to check the clamping status of the tool holder or the clearance of the spindle bearing. If the root cause of the fault is workpiece or fixture chatter, the generated alarm command will prompt you to check whether the workpiece is securely clamped or suggest adjusting the cutting parameters at that specific position.

[0015] This application provides a vibration monitoring system for the milling process of a machine tool milling cutter, which is used to implement a method for monitoring the vibration state of the milling process of a machine tool milling cutter, including: a reference template establishment module, a diagnostic triggering module, a frequency domain feature extraction module, a preliminary discrimination module, and a fault root cause verification module; The reference template establishment module is used to acquire the vibration signal of the milling cutter under reference machining conditions, and establish a reference template containing normal vibration characteristics based on the vibration signal. The diagnostic trigger module is used to collect the vibration signal of the milling cutter in real time during the milling process and dynamically adjust the alarm threshold. When the energy characteristics of the real-time vibration signal are not less than the dynamic alarm threshold, the fault diagnosis analysis is started. The frequency domain feature extraction module is used to perform envelope spectrum analysis on the vibration signal that triggers diagnosis, and extract frequency domain feature parameters that characterize the state of the tool, tool holder and workpiece. The preliminary discrimination module quantizes and compares the extracted frequency domain feature parameters with the features in the reference template, and preliminarily determines the fault type to which the vibration abnormality belongs through the preset fault classification model. The fault root cause verification module is used to acquire real-time coordinate data of the machine tool CNC system, and perform correlation verification with the fault type and spatial location information when the abnormality occurs to confirm the root cause of the fault and output targeted classification alarm instructions.

[0016] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages: 1. By acquiring the vibration signal of the milling cutter under reference machining conditions and establishing a reference template containing normal vibration characteristics based on the vibration signal, a comparison basis is provided for real-time monitoring, thereby realizing the accurate differentiation of the causes of vibration abnormalities. This effectively solves the problem in the prior art that it is difficult to determine the specific fault source of tool damage, tool holder problems or workpiece chatter in the monitoring of vibration status during the milling process.

[0017] 2. By establishing a benchmark template containing normal vibration characteristics and dynamically adjusting the alarm threshold based on real-time machining parameters, vibration monitoring can adapt to the differences in normal vibration under different machining conditions such as rotational speed and cutting depth, thereby achieving accurate triggering diagnosis of abnormal vibration. This effectively solves the problem of high false alarm rate or missed detection caused by the use of fixed thresholds in the existing technology when machining parameters change.

[0018] 3. By obtaining real-time coordinate data of the machine tool CNC system and verifying the spatial location association with the fault type, the root cause of the fault can be confirmed by combining the rigid area information of the workpiece geometric model. This enables secondary verification and root cause analysis of the fault diagnosis results, significantly reducing the misjudgment rate and providing more accurate fault location and targeted maintenance guidance. Attached Figure Description

[0019] Figure 1 This is a flowchart of a method for monitoring vibration status during milling using a machine tool milling cutter, provided in an embodiment of this application. Figure 2 This is a schematic diagram of the structure of a vibration monitoring system for a machine tool milling cutter provided in an embodiment of this application. Detailed Implementation

[0020] This application provides a method and system for monitoring the vibration status of a milling process using a machine tool milling cutter. This solves the problem in the prior art that a single channel cannot distinguish between tool, tool holder, and workpiece faults. By using sensor dynamic threshold triggering, envelope spectrum coordinate coupling verification, and directional alarm, it achieves accurate location of the fault source and provides direct maintenance instructions.

[0021] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0022] like Figure 1The diagram shows a flowchart of a method for monitoring the vibration state of a milling process using a machine tool milling cutter, provided in an embodiment of this application. This method is applied to a vibration state monitoring system for a milling process using a machine tool milling cutter. The method includes the following steps: fixing a miniature vibration sensor to the milling cutter shank, acquiring the vibration signal of the milling cutter under reference machining conditions, including normal tooth impact rhythm, overall vibration energy level, energy distribution characteristics of low and high frequency bands, and establishing a reference template containing normal vibration characteristics based on the vibration signal. The vibration signal of the milling cutter during the milling process is collected in real time, and the alarm threshold is dynamically adjusted according to real-time machining parameters such as rotation speed and depth of cut. When the energy characteristics of the real-time vibration signal are not less than the dynamic alarm threshold, the fault diagnosis analysis is initiated. Envelope spectrum analysis is performed on the vibration signal that triggers the diagnosis to extract frequency domain characteristic parameters that characterize the state of the tool, tool holder and workpiece, such as the amplitude change of the tooth passing frequency, sideband noise, low frequency energy and high frequency spikes. The extracted frequency domain feature parameters are quantitatively compared with the features in the benchmark template. Through the preset fault classification model, the fault type to which the vibration anomaly belongs is initially determined. The system acquires real-time coordinate data from the CNC system of the machine tool, and performs correlation verification by combining the fault type and spatial location information at the time of the abnormality to confirm the root cause of the fault. It then outputs targeted alarm commands, such as prompting to replace the cutting teeth for tool damage, prompting to tighten and maintain the tool holder for tool holder problems, and suggesting to adjust the local feed parameters for workpiece chatter.

[0023] Since environmental interference suppression is implemented throughout the entire process, noise characteristics of fixed equipment in the workshop are collected in advance and a fingerprint database is established. Interference energy in these frequency bands is filtered in real time, and instantaneous sudden vibrations are automatically ignored. The reference database is automatically updated after each maintenance, and a single calibration is performed when a new tool or workpiece is switched. Through these steps, from reference acquisition to threshold triggering, feature extraction, comparison and judgment, and coordinate verification, the causes of abnormal vibrations can be accurately distinguished and targeted warnings can be issued, solving the problem of difficulty in identifying specific fault sources in vibration monitoring during the milling process.

[0024] Furthermore, the step of establishing a reference template containing normal vibration characteristics based on the vibration signal includes: Under standard machining conditions consisting of a brand-new milling cutter, a secure tool holder, and a stably clamped workpiece, start the machine tool and execute the preset reference cutting program; Vibration signals of the milling cutter during the execution of the reference cutting program are continuously collected to obtain reference time series data; The baseline time series data is processed to obtain the average value of the overall vibration energy, and this average value is stored as the baseline energy. The average value of the overall vibration energy is calculated using the following formula: ; In the formula, Reference energy (unit: m² / s²). Vibration velocity signal (unit: m / s) in the reference time series data. Given the total length of the time series (in seconds), the squared average of the signal was calculated by integration to characterize the energy level.

[0025] The reference time series data is envelope demodulated and transformed to the frequency domain to generate the reference envelope spectrum; In the reference envelope spectrum, the number of theoretical impacts per revolution of the main shaft is used as an index to locate the tooth passing frequency spectrum peak corresponding to the impact, and the center frequency value and the energy amplitude of the spectrum peak are recorded to obtain the tooth passing frequency characteristics. The reference envelope spectrum is divided into a low-frequency band below the blade passage frequency, a neighboring frequency band centered on the blade passage frequency, and a high-frequency band above the blade passage frequency. The total energy in each frequency band is calculated to obtain the energy distribution values ​​of the low-frequency band, the neighboring frequency band, and the high-frequency band, thus obtaining the energy distribution characteristics of each frequency band. The recorded baseline energy and blades are combined with frequency features and energy distribution features of each frequency band to form a structured dataset, which is used to obtain a baseline template for subsequent state comparison.

[0026] Furthermore, the steps of collecting vibration signals from the milling cutter during the milling process and dynamically adjusting the alarm threshold include: During actual machining tasks, real-time vibration time-series signals are acquired from vibration sensors, and real-time machining parameters are simultaneously acquired from the machine tool CNC system. These machining parameters include the current spindle speed and the current depth of cut. Set the built-in baseline threshold by using the baseline energy level in the baseline template; Based on the acquisition of real-time machining parameters, a speed adjustment multiplier is derived according to the ratio of the real-time spindle speed to the reference speed. This speed adjustment multiplier increases non-linearly with the increase of speed. The speed regulation multiplier is calculated using the following formula: ; In the formula, This is the speed regulation multiplier (dimensionless). Real-time spindle speed (unit: rpm). The reference speed (unit: rpm) is represented by an exponent of 1.5, which reflects the nonlinear growth relationship of vibration energy with speed.

[0027] Based on the ratio of the real-time cutting depth to the reference cutting depth, a cutting depth adjustment multiplier is derived, which increases exponentially with the increase of cutting depth. The depth-of-cut adjustment multiplier is calculated using the following formula: ; In the formula, This is a dimensionless multiplier for adjusting the depth of cut. Real-time cutting depth (unit: mm). Reference cutting depth (unit: mm).

[0028] The dynamic alarm threshold is obtained by multiplying the base threshold by the speed adjustment multiplier and the depth of cut adjustment multiplier. The dynamic alarm threshold is obtained by multiplying the basic threshold by the speed adjustment multiplier and the cutting depth adjustment multiplier. The above calculation process continues to ensure that the dynamic alarm threshold is updated in real time with each change in the working conditions, so as to accurately define the upper limit of normal vibration under the current working conditions.

[0029] Furthermore, the steps of performing envelope spectrum analysis on the vibration signal that triggered the diagnosis to extract frequency domain characteristic parameters characterizing the state of the tool, tool holder, and workpiece include: Extract the real-time vibration time series signal that triggers fault diagnosis analysis; By applying the Hilbert transform to the time series signal segment, a corresponding analytical time series signal is constructed, which contains the amplitude and phase information of the original time series signal. Obtain the magnitude of the analytical time series signal at each time point, and extract the vibration envelope signal without carrier frequency; Perform a fast Fourier transform on the vibration envelope signal to convert it from the time domain to the frequency domain and generate an envelope spectrum. Based on the generated envelope spectrum, predefined frequency domain feature parameters are automatically identified and extracted; The frequency domain feature parameters include the peak energy amplitude extracted based on the generated envelope spectrum, with the passing frequency of the blade as the center; Based on the peak energy amplitude, a fixed bandwidth is extended to the left and right sides, and the energy of all spectral lines within the bandwidth is accumulated as the sideband energy. Sideband energy is calculated using the following formula: ; In the formula, Sideband energy (unit: m² / s). Frequency of the cutting teeth (unit: Hz). Fixed bandwidth (unit: Hz) The amplitude spectrum of the envelope spectrum is obtained by integrating and accumulating the energy density within the bandwidth.

[0030] In the low-frequency range below the passing frequency of the cutting teeth, the total low-frequency energy is obtained by integration; In the high-frequency range above the passing frequency of the cutting teeth, search for narrowband high-energy spectral peaks with amplitudes exceeding a set threshold, and record their amplitudes and frequencies. The blade teeth are combined with the peak frequency energy amplitude, sideband energy, low-frequency total energy, and high-frequency narrow peak information to form a frequency domain description vector of the current vibration state, which is used for subsequent fault classification.

[0031] These extracted parameters together constitute a complete frequency domain description of the current vibration state.

[0032] Furthermore, the step of quantizing and comparing the extracted frequency domain feature parameters with the features in the benchmark template includes: Extracted frequency domain feature parameters are retrieved from the real-time vibration time series signal that triggers fault diagnosis analysis. The frequency domain feature parameters include real-time blade passing frequency energy, sideband energy, low-frequency energy, and high-frequency abnormal spectral peak information. Retrieve reference feature parameters corresponding to the frequency domain feature parameters from the reference template. The reference feature parameters include reference cutter passing frequency energy and reference low-frequency band energy. The extracted frequency domain feature parameters are quantized and compared with the reference feature parameters. The energy ratio is obtained by dividing the real-time cutting tooth passing frequency energy by the reference cutting tooth passing frequency energy. The difference between the real-time low-frequency energy and the reference low-frequency energy is calculated to obtain the low-frequency energy drift. The ratio of real-time sideband energy to real-time cutter tooth passing frequency energy is calculated to obtain the sideband energy proportion; And calculate the ratio of the energy of the most significant spectral peak in the high-frequency band to the average energy of its frequency band to obtain the high-frequency concentration index; The calculated energy ratio, energy drift, energy percentage, and concentration index together form a set of quantitative input vectors for fault classification.

[0033] Furthermore, the preset fault classification model is as follows: The input vector of the fault classification model is the feature mutation vector: ; In the formula, The relative deviation of the frequency amplitude of the cutting teeth. This represents the absolute difference in the proportion of low-frequency energy. This represents the peak growth rate of high-frequency narrowband bandwidth. The decision function of the fault classification model is: ; In the formula, The threshold for the amplitude of tool chipping. The low-frequency energy exceeding the threshold for tool holder loosening is a concern. To eliminate the upper limit of high-frequency growth of flutter, The high-frequency spike threshold for workpiece chatter. This is the negative bias threshold for the amplitude of the tool without damage.

[0034] Further steps to preliminarily determine the fault type to which the vibration anomaly belongs include: The input vector generated after quantization and comparison is transmitted to the preset fault classification model. The input vector includes energy ratio, energy drift, energy proportion and high frequency concentration index. Following a preset order, the input vectors are matched and verified one by one with each group of logical rules in the fault classification model. The verification process is as follows: First, check whether the input vector meets all the conditions of the tool damage rule; if not, continue to check whether it meets all the conditions of the tool holder or spindle system abnormality rule. If it still does not meet the requirements, then finally check whether it meets all the conditions of the workpiece or fixture chatter rule. If the input vector completely matches all the conditions of one set of rules, the subsequent rule verification is stopped, and the fault type uniquely corresponding to that set of rules, such as "tool damage" or "tool holder abnormality", is determined as the preliminary judgment result of this diagnosis, and this result is output to the next correlation verification stage.

[0035] The logical verification of the matching between the input vector and the rule can be quantified by calculating the Euclidean distance: ; In the formula, For the matching distance (dimensionless), if ( If the preset tolerance threshold is dimensionless, then it is considered a match; parameter , , With threshold , , All are dimensionless ratios.

[0036] Furthermore, the step of acquiring real-time coordinate data from the machine tool CNC system and verifying its correlation with the attributed fault type and the spatial location information at the time of the anomaly includes: Based on the preliminary judgment results, the three-dimensional Cartesian coordinates of the milling cutter cutting point at the moment of vibration abnormality were obtained by sending a request to the machine tool CNC system. By pre-storing a digital geometric model of the currently processed workpiece, which has been divided and marked with regions of different structural rigidity, clearly defining high-rigidity regions and low-rigidity regions, an association judgment is performed. The steps for defining high-rigidity and low-rigidity regions are as follows: After importing the 3D CAD model file of the workpiece into the computer, the minimum radius of curvature of each surface patch, the thickness characteristic value of the solid region, and the aspect ratio parameter of the cantilever structure are automatically analyzed based on the geometric topology. Regions with a thickness characteristic value greater than a set thickness threshold, a minimum radius of curvature greater than a set curvature threshold, and no cantilever features are marked as high-rigidity regions. Regions with a thickness characteristic value less than a set thickness threshold, a minimum radius of curvature less than a set curvature threshold, or with cantilever structures having an aspect ratio greater than a set ratio threshold are marked as low-rigidity regions. Finally, the marking results are embedded into the 3D model data file. When the system obtains the real-time coordinates of the tool, the position association verification is achieved by calculating the distance to the nearest triangular facet and retrieving its rigidity attribute.

[0037] If the initial fault type is determined to be a tool holder or spindle malfunction, then determine whether the acquired real-time coordinates fall within the high rigidity area marked in the digital geometric model. If the determination is yes, then the fault is confirmed. If the initial fault type is workpiece or fixture chatter, the system determines whether the acquired real-time coordinates fall within the marked low-rigidity area. If it does, the fault is confirmed. This step ultimately identifies the root cause of the fault by establishing logical consistency between the spatial location and the physical cause of the fault.

[0038] Furthermore, the steps for outputting targeted, categorized alarm commands include: After verifying and confirming the root cause of the fault through spatial location information, a corresponding text alarm command is selected and generated from the preset command library based on the finally confirmed fault type. If the root cause of the fault is tool damage, the generated alarm instruction will prompt you to check and replace the milling cutter insert. If the root cause of the fault is an abnormality in the tool holder or spindle system, the generated alarm command will prompt you to check the clamping status of the tool holder or the clearance of the spindle bearing. If the root cause of the fault is workpiece or fixture chatter, the generated alarm instruction will prompt you to check whether the workpiece is securely clamped or suggest adjusting the cutting parameters at that specific position. The generated text alarm command is sent to the human-machine interface for display, ensuring that the operator receives specific and directly executable maintenance action suggestions, rather than a general vibration over-limit warning.

[0039] like Figure 2The diagram shown is a structural schematic of a milling process vibration monitoring system for machine tool milling cutters provided in this application embodiment. This system is used to implement the milling process vibration monitoring method for machine tool milling cutters. The milling process vibration monitoring system for machine tool milling cutters provided in this application embodiment includes: a reference template establishment module, a diagnostic triggering module, a frequency domain feature extraction module, a preliminary discrimination module, and a fault root cause verification module. The reference template establishment module is used to acquire the vibration signal of the milling cutter under reference machining conditions, and establish a reference template containing normal vibration characteristics based on the vibration signal. The diagnostic trigger module is used to collect the vibration signal of the milling cutter in real time during the milling process and dynamically adjust the alarm threshold. When the energy characteristics of the real-time vibration signal are not less than the dynamic alarm threshold, the fault diagnosis analysis is started. The frequency domain feature extraction module is used to perform envelope spectrum analysis on the vibration signal that triggers diagnosis, and extract frequency domain feature parameters that characterize the state of the tool, tool holder and workpiece. The preliminary discrimination module quantizes and compares the extracted frequency domain feature parameters with the features in the reference template, and preliminarily determines the fault type to which the vibration abnormality belongs through the preset fault classification model. The fault root cause verification module is used to acquire real-time coordinate data of the machine tool CNC system, and perform correlation verification with the fault type and spatial location information when the abnormality occurs to confirm the root cause of the fault and output targeted classification alarm instructions.

[0040] Those skilled in the art should understand that the above formulas are all dimensionless numerical calculations. The formulas are derived from software simulations using a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0041] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0042] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0043] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0044] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0045] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for monitoring vibration during the milling process of a machine tool milling cutter, characterized in that, Includes the following steps: The vibration signal of the milling cutter under reference machining conditions is acquired, and a reference template containing normal vibration characteristics is established based on the vibration signal. The vibration signal of the milling cutter during the milling process is collected in real time, and the alarm threshold is dynamically adjusted. When the energy characteristics of the real-time vibration signal are not less than the dynamic alarm threshold, the fault diagnosis analysis is initiated. Envelope spectrum analysis was performed on the vibration signal that triggered the diagnosis to extract frequency domain characteristic parameters that characterize the state of the tool, tool holder and workpiece. The extracted frequency domain feature parameters are quantitatively compared with the features in the benchmark template. Through the preset fault classification model, the fault type to which the vibration anomaly belongs is initially determined. The system acquires real-time coordinate data from the CNC system of the machine tool, and performs correlation verification by combining the fault type and the spatial location information at the time of the abnormality to confirm the root cause of the fault and output targeted classification alarm commands.

2. The method for monitoring vibration status during milling process of a machine tool milling cutter as described in claim 1, characterized in that, The steps for establishing a reference template containing normal vibration characteristics based on the vibration signal include: Under standard machining conditions, start the machine tool and execute the preset reference cutting program; Vibration signals of the milling cutter during the execution of the reference cutting program are continuously collected to obtain reference time series data; The baseline time series data is processed to obtain the average value of the overall vibration energy, and this average value is stored as the baseline energy. Envelope demodulation is performed on the baseline time series data to generate the baseline envelope spectrum; In the baseline envelope spectrum, obtain the passing frequency characteristics of the cutting teeth; The reference envelope spectrum is divided into a low-frequency band below the blade passage frequency, a neighboring frequency band centered on the blade passage frequency, and a high-frequency band above the blade passage frequency, thus obtaining the energy distribution characteristics of each frequency band. The recorded baseline energy and cutting teeth are combined with frequency features and energy distribution features of each frequency band to form a structured dataset, thus obtaining the baseline template.

3. The method for monitoring vibration status during milling using a machine tool milling cutter as described in claim 1, characterized in that, The steps for collecting vibration signals from the milling cutter during milling and dynamically adjusting alarm thresholds include: During actual machining tasks, real-time vibration time sequence signals are acquired, and real-time machining parameters are simultaneously acquired from the machine tool CNC system. Set the built-in baseline threshold by using the baseline energy level in the baseline template; Based on the acquisition of real-time machining parameters, the speed adjustment multiplier and the depth of cut adjustment multiplier are derived. The dynamic alarm threshold is obtained by multiplying the basic threshold by the speed adjustment multiplier and the cutting depth adjustment multiplier.

4. The method for monitoring vibration status during milling process of a machine tool milling cutter as described in claim 1, characterized in that, The steps for performing envelope spectrum analysis on the vibration signal that triggers the diagnosis, and extracting frequency domain characteristic parameters representing the state of the tool, tool holder, and workpiece, include: Extract the real-time vibration time series signal that triggers fault diagnosis analysis; By constructing a corresponding analytical time series signal from the time series signal segments, the analytical time series signal contains the amplitude and phase information of the original time series signal; Obtain the magnitude of the analytical time series signal at each time point, and extract the vibration envelope signal without carrier frequency; Perform a Fast Fourier Transform on the vibration envelope signal to generate an envelope spectrum. Based on the generated envelope spectrum, predefined frequency domain feature parameters are automatically identified and extracted.

5. The method for monitoring vibration during milling using a machine tool milling cutter as described in claim 1, characterized in that, The steps for quantizing and comparing the extracted frequency domain feature parameters with the features in the benchmark template include: Extracted frequency domain feature parameters are retrieved from the real-time vibration time series signal that triggers fault diagnosis analysis; Retrieve reference feature parameters corresponding to the frequency domain feature parameters from the reference template. The reference feature parameters include reference cutter passing frequency energy and reference low-frequency band energy. The extracted frequency domain feature parameters are quantitatively compared with the benchmark feature parameters to obtain the energy ratio, low-frequency energy drift, sideband energy ratio, and high-frequency concentration index.

6. The method for monitoring vibration status during milling process of a machine tool milling cutter as described in claim 1, characterized in that, The preset fault classification model is: The input vector of the fault classification model is the feature mutation vector: ； In the formula, The relative deviation of the frequency amplitude of the cutting teeth. This represents the absolute difference in the proportion of low-frequency energy. This represents the peak growth rate of high-frequency narrowband bandwidth. The decision function of the fault classification model is: ； In the formula, The threshold for the amplitude of tool chipping. The low-frequency energy exceeding the threshold for tool holder loosening is a concern. To eliminate the upper limit of high-frequency growth of flutter, The high-frequency spike threshold for workpiece chatter. This is the negative bias threshold for the amplitude of the tool without damage.

7. The method for monitoring vibration status during milling process of a machine tool milling cutter as described in claim 1, characterized in that, The steps for initially determining the fault type of the vibration anomaly include: The input vector generated after quantization and comparison is transmitted to the preset fault classification model; Following a preset order, the input vectors are matched and verified one by one with each group of logical rules in the fault classification model. If the input vector completely matches all the conditions of one set of rules, the subsequent rule verification is stopped, and the fault type uniquely corresponding to that set of rules is determined as the preliminary judgment result of this diagnosis.

8. The method for monitoring vibration status during milling process of a machine tool milling cutter as described in claim 1, characterized in that, The steps of acquiring real-time coordinate data from the machine tool CNC system and verifying its correlation with the attributed fault type and the spatial location information at the time of the anomaly include: Based on the preliminary judgment results, the three-dimensional Cartesian coordinates of the milling cutter cutting point at the moment when the vibration anomaly occurred were obtained; By pre-storing a digital geometric model of the currently processed workpiece, an association judgment is performed; If the initial fault type is determined to be a tool holder or spindle malfunction, then determine whether the acquired real-time coordinates fall within the high rigidity area marked in the digital geometric model. If the determination is yes, then the fault is confirmed. If the initial fault type is workpiece or fixture chatter, the system determines whether the acquired real-time coordinates fall within the marked low-rigidity area. If so, the fault is confirmed.

9. The method for monitoring vibration status during milling process of a machine tool milling cutter as described in claim 1, characterized in that, The steps for outputting targeted, categorized alarm commands include: After verifying and confirming the root cause of the fault through spatial location information, a corresponding text alarm command is selected and generated from the preset command library based on the finally confirmed fault type. If the root cause of the fault is tool damage, the generated alarm instruction will prompt you to check and replace the milling cutter insert. If the root cause of the fault is an abnormality in the tool holder or spindle, the generated alarm command will prompt you to check the clamping status of the tool holder or the clearance of the spindle bearing. If the root cause of the fault is workpiece or fixture chatter, the generated alarm command will prompt you to check whether the workpiece is securely clamped or suggest adjusting the cutting parameters at that specific position.

10. A vibration monitoring system for a milling process using a machine tool milling cutter, used to implement the vibration monitoring method for a milling process using a machine tool milling cutter as described in any one of claims 1-9, characterized in that, include: The module includes a benchmark template establishment module, a diagnostic triggering module, a frequency domain feature extraction module, a preliminary discrimination module, and a fault root cause verification module. The reference template establishment module is used to acquire the vibration signal of the milling cutter under reference machining conditions, and establish a reference template containing normal vibration characteristics based on the vibration signal. The diagnostic trigger module is used to collect the vibration signal of the milling cutter in real time during the milling process and dynamically adjust the alarm threshold. When the energy characteristics of the real-time vibration signal are not less than the dynamic alarm threshold, the fault diagnosis analysis is started. The frequency domain feature extraction module is used to perform envelope spectrum analysis on the vibration signal that triggers diagnosis, and extract frequency domain feature parameters that characterize the state of the tool, tool holder and workpiece. The preliminary discrimination module quantizes and compares the extracted frequency domain feature parameters with the features in the reference template, and preliminarily determines the fault type to which the vibration abnormality belongs through the preset fault classification model. The fault root cause verification module is used to acquire real-time coordinate data of the machine tool CNC system, and perform correlation verification with the fault type and spatial location information when the abnormality occurs to confirm the root cause of the fault and output targeted classification alarm instructions.

Images