A method and system for adaptive parameters of turning machining of thin-walled parts

Abstract

The present application provides a method and system for adaptively adjusting machining parameters for turning thin-walled parts. The method generates a background spectrum baseline by collecting background vibration signals during the non-cutting period, and uses the baseline to purify the cutting force signal during the cutting period. Cutting state identification and parameter adjustment are performed based on the purified signal, thereby effectively solving the problem of interference of machine tool background vibration on cutting force signal. The method has the advantages of being able to effectively remove interference of machine tool background vibration on cutting force signal, improve the accuracy of cutting state identification, thereby achieving more accurate adaptive adjustment of machining parameters and improving machining quality and stability of thin-walled parts.

Description

Technical Field

[0001] The present application relates to the fields of mechanical processing, numerical control technology and adaptive control, and in particular to a method and system for adaptively adjusting machining parameters for turning of thin-walled parts. Background Art

[0002] In the field of precision manufacturing, especially for the production of ultra-thin-walled aluminum alloy pipes for high-precision applications such as optical instruments, the wall thickness tolerance is usually required to reach the micron level, and chatter marks are not allowed. In order to meet these stringent processing requirements, high-end CNC lathes on automated production lines generally integrate adaptive control systems for processing parameters based on the frequency domain characteristics of the cutting force signal. This type of system uses high-sensitivity sensors to monitor the dynamic cutting force during the turning process in real time, performs time-frequency analysis on the signal, extracts frequency domain characteristics such as the dominant frequency and harmonic components, and combines expert databases with optimization algorithms to adjust the spindle speed, feed rate, and cutting depth online to actively suppress processing chatter and ensure part quality. In the automated mass production process, the robotic arm automatically completes the clamping of tubular blanks and the unloading of finished products.

[0003] However, even with tools in good condition and workpiece material batches consistent, even after long periods of continuous high-load operation, individual lathes on the production line may occasionally exhibit quality defects such as subtle, irregular surface ripples or a loss of finish in specific areas. These defects appear with no discernible regularity, posing a challenge to quality control. This is because CNC machine tool components, such as the spindle system and feed system, inevitably experience thermal deformation and accumulated micro-wear over extended periods of operation. These factors can cause slow, imperceptible drift in the dynamic characteristics of the entire machine tool (such as natural frequency and damping ratio). This drift often appears as unexpected background vibration noise or specific interfering frequency components, superimposing and contaminating the raw cutting force signals collected in real time by the force sensor.

[0004] Existing parameter adaptive control systems are typically designed based on the assumption that the machine tool's state is relatively stable, lacking effective mechanisms for sensing and compensating for dynamic changes in this background state. Consequently, when the cutting force signal is contaminated by this type of contamination, the system struggles to accurately distinguish whether changes in the frequency domain characteristics are due to actual tool-workpiece cutting interactions (such as early chatter signals) or to gradual changes in the machine tool's own background state. This confusion can lead the system to misinterpret the contaminated signal, resulting in inappropriate parameter adjustments. For example, overly conservative adjustments can reduce efficiency, or erroneous interventions can lead to new machining issues, ultimately compromising the machining quality and production stability of precision thin-walled parts. Summary of the Invention

[0005] The present application provides a method and system for adaptively adjusting machining parameters for thin-walled parts turning, which has the advantages of being able to effectively remove the interference of machine tool background vibration on cutting force signals, improve the accuracy of cutting state identification, thereby achieving more accurate adaptive adjustment of machining parameters and improving the machining quality and stability of thin-walled parts.

[0006] On the one hand, the present application provides a method for adaptively adjusting machining parameters of thin-walled parts turning, comprising:

[0007] collecting background vibration signals of a CNC machine tool during a non-cutting period of automated turning, wherein the CNC machine tool is used for automated turning of thin-walled parts;

[0008] generating a target background spectrum baseline according to the background vibration signal, wherein the target background spectrum baseline represents the current background vibration characteristics of the CNC machine tool;

[0009] During the cutting process of automated turning, a cutting force signal is collected in real time, a spectrum analysis is performed on the cutting force signal to obtain an original cutting force spectrum, and a spectrum difference operation is performed on the original cutting force spectrum using the target background spectrum baseline to obtain a purified cutting force spectrum;

[0010] performing cutting state identification based on the spectrum characteristics of the purified cutting force spectrum to obtain an identification result;

[0011] Based on the recognition result and according to a preset process rule library, the processing parameters of the CNC machine tool are adjusted, wherein the processing parameters include spindle speed, feed rate and cutting depth.

[0012] Optionally, the step of generating a target background spectrum baseline based on the background vibration signal, wherein the target background spectrum baseline characterizes the current background vibration characteristics of the CNC machine tool includes:

[0013] Performing a fast Fourier transform on the background vibration signal to generate background vibration spectrum data;

[0014] According to the preset machine tool background spectrum baseline and the background vibration spectrum data, an exponentially weighted moving average algorithm is used to update and generate a target background spectrum baseline.

[0015] Optionally, the step of updating the preset machine tool background spectrum baseline and the background vibration spectrum data using an exponentially weighted moving average algorithm to generate a target background spectrum baseline comprises:

[0016] Among them, B is the target background spectrum baseline, w is the update weight factor, A1 is the machine tool background spectrum baseline, and A2 is the background vibration spectrum data.

[0017] Optionally, the step of generating a target background spectrum baseline based on the background vibration signal, wherein the target background spectrum baseline characterizes the current background vibration characteristics of the CNC machine tool includes:

[0018] determining whether the amplitude of the background vibration signal exceeds a preset threshold, and if the amplitude of the background vibration signal exceeds the preset threshold, determining that the background vibration signal is a disturbed signal;

[0019] removing the background vibration signal of the disturbed signal from the background vibration signal to obtain a residual background vibration signal;

[0020] A target background spectrum baseline is generated using the remaining background vibration signal.

[0021] Optionally, the step of generating a target background spectrum baseline by using the remaining background vibration signal includes:

[0022] Performing spectrum analysis on the remaining background vibration signal to obtain a current background vibration spectrum;

[0023] Comparing the current background vibration spectrum with the machine tool background spectrum baseline to determine the degree of difference between the current background vibration spectrum and the machine tool background spectrum baseline;

[0024] adjusting update parameters of the machine tool background spectrum baseline according to the degree of the difference;

[0025] The update parameters and the current background vibration spectrum are used to update the machine tool background spectrum baseline to generate a target background spectrum baseline.

[0026] Optionally, the step of performing a spectrum difference operation on the original cutting force spectrum using the target background spectrum baseline to obtain a purified cutting force spectrum includes:

[0027] According to the corresponding relationship between the original cutting force spectrum and the target background spectrum baseline at each frequency point, identifying the frequency interval in which the spectrum components of the original cutting force spectrum overlap with the target background spectrum baseline;

[0028] Determining a spectrum processing strategy for processing the original cutting force spectrum component within the frequency interval based on a preset comparison condition between the amplitude of the original cutting force spectrum component within the frequency interval and the amplitude of the corresponding component of the target background spectrum baseline;

[0029] Based on the spectrum processing strategy, the original cutting force spectrum components in the frequency interval are processed to obtain the overlapped region spectrum components; and

[0030] The original cutting force spectrum components that are not in the frequency range are processed using a preset reference spectrum processing method to obtain non-overlapping area spectrum components;

[0031] The overlapping region spectrum component and the non-overlapping region spectrum component are combined to generate the cleanup cutting force spectrum.

[0032] Optionally, the step of adjusting the update parameters of the machine tool background spectrum baseline according to the degree of difference includes:

[0033] Performing a validity judgment on the degree of difference, the validity judgment comprising judging whether the instantaneous change characteristics of the degree of difference conform to a preset drift model;

[0034] Processing the difference degree according to the validity judgment result to obtain a target difference degree, wherein the processing includes correcting the difference degree when the difference degree does not conform to a preset drift model;

[0035] The target difference degree is used to adjust update parameters of the machine tool background spectrum baseline.

[0036] Optionally, the step of correcting the degree of difference includes:

[0037] analyzing the instantaneous change characteristics of the difference degree;

[0038] Dynamically adjusting a correction parameter for correcting the degree of difference based on the instantaneous change characteristic;

[0039] The correction parameter is used to correct the degree of difference.

[0040] On the other hand, the present application provides a thin-walled part turning parameter adaptive system, characterized in that the system includes:

[0041] A background vibration signal acquisition module is used to collect background vibration signals of a CNC machine tool during a non-cutting period of automated turning, wherein the CNC machine tool is used for automated turning of thin-walled parts;

[0042] A background spectrum baseline generating module is used to generate a target background spectrum baseline according to the background vibration signal, wherein the target background spectrum baseline represents the current background vibration characteristics of the CNC machine tool;

[0043] a cutting force signal processing module, configured to collect a cutting force signal in real time during the cutting process of the automated turning operation, perform spectrum analysis on the cutting force signal to obtain an original cutting force spectrum, and perform spectrum difference calculation on the original cutting force spectrum using the target background spectrum baseline to obtain a purified cutting force spectrum;

[0044] an identification module, configured to identify the cutting state based on the spectrum characteristics of the purified cutting force spectrum to obtain an identification result;

[0045] A parameter adjustment module is used to adjust the processing parameters of the CNC machine tool based on the recognition result and according to a preset process rule library, wherein the processing parameters include spindle speed, feed rate and cutting depth.

[0046] Optionally, the background spectrum baseline generation module includes a background spectrum processing module and a machine tool background spectrum baseline storage module, wherein:

[0047] A machine tool background spectrum baseline storage module is used to store a preset machine tool background spectrum baseline;

[0048] The background spectrum processing module is used to perform fast Fourier transform on the background vibration signal to generate background vibration spectrum data; and

[0049] After obtaining the preset machine tool background spectrum baseline from the machine tool background spectrum baseline storage module, the preset machine tool background spectrum baseline and the background vibration spectrum data are updated using an exponentially weighted moving average algorithm to generate a target background spectrum baseline.

[0050] The present application provides a method and system for adaptively adjusting machining parameters for turning thin-walled parts. The method generates a background spectrum baseline by collecting background vibration signals during the non-cutting period, and uses the baseline to purify the cutting force signal during the cutting period. Cutting state identification and parameter adjustment are performed based on the purified signal, thereby effectively solving the problem of interference of machine tool background vibration on cutting force signal. The method has the advantages of being able to effectively remove interference of machine tool background vibration on cutting force signal, improve the accuracy of cutting state identification, thereby achieving more accurate adaptive adjustment of machining parameters and improving machining quality and stability of thin-walled parts. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the technical solution of the present application, the following is a brief introduction to the drawings required for use in the embodiments. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0052] Figure 1 : A schematic diagram of a thin-walled part turning parameter adaptive method according to an embodiment is exemplarily shown;

[0053] Figure 2 Schematically shows a module configuration block diagram of a thin-walled part turning processing parameter adaptive system 100 according to an embodiment. DETAILED DESCRIPTION

[0054] The technical solutions in this application will be clearly and completely described below in conjunction with the drawings in this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all of the embodiments. The components of the present application generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application for which protection is claimed, but merely represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of this application.

[0055] It should be noted that similar reference numerals and letters represent similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings. At the same time, in the description of this application, the terms "first", "second", etc. are only used to distinguish the description and should not be understood as indicating or implying relative importance.

[0056] Conventional adaptive machining parameter control systems based on the frequency domain characteristics of cutting force signals, when used for automated turning of thin-walled parts, suffer from a problem where the dynamic characteristics of the CNC machine tool's structural components slowly drift after prolonged operation. This causes background vibration noise to contaminate the real-time cutting force signal, making it difficult for the system to accurately distinguish between the actual cutting state and background interference. This contamination affects the system's accurate judgment of the cutting state, which can lead to inappropriate parameter adjustments.

[0057] For example, imagine a high-end CNC lathe on an automated production line continuously processing large quantities of ultra-thin-walled aluminum alloy pipes, requiring micron-level wall thickness tolerances and a surface free of chatter marks. The lathe is equipped with a real-time cutting force monitoring and parameter adaptation system. After long periods of continuous operation, the machine's spindle bearings or feed screw may experience minor wear or thermal deformation, causing the machine to generate weak but persistent background vibrations at specific frequencies. This background vibration is superimposed on the cutting force signal collected by the force sensor. Especially in the early chatter stage, when the cutting force signal itself is weak, the background noise may approach or overlap with the frequency components of the chatter signal, resulting in false peaks in the original cutting force spectrum caused by the background noise or masking the true cutting state characteristics.

[0058] If these issues are not addressed, the parameter adaptive system may make erroneous judgments based on cutting force signals contaminated by background noise. For example, the system may misinterpret background noise as chatter and over-adjust machining parameters, resulting in reduced machining efficiency. Alternatively, the system may fail to identify true early chatter signals, leading to quality defects such as irregular ripples on the part surface or reduced finish. This uncertainty impacts the stability of automated production and the consistency of product quality.

[0059] like Figure 1 As shown, the present application exemplarily shows a flow chart of a method for self-adapting machining parameters for thin-walled parts turning. The present application proposes a method for self-adapting machining parameters for thin-walled parts turning, comprising:

[0060] S10 , collecting background vibration signals of a CNC machine tool during a non-cutting period of automated turning, wherein the CNC machine tool is used for automated turning of thin-walled parts.

[0061] Among them, the background vibration signal refers to the vibration signal of the CNC machine tool itself collected during the non-cutting period of automated turning. It can be collected in the non-cutting state using a vibration sensor or force sensor installed on the structural components of the machine tool to obtain background noise information reflecting the dynamic characteristics of the machine tool itself.

[0062] S20 , generating a target background spectrum baseline according to the background vibration signal, wherein the target background spectrum baseline represents current background vibration characteristics of the CNC machine tool.

[0063] The target background spectrum baseline refers to the spectrum data generated based on the background vibration signal, which represents the current background vibration characteristics of the CNC machine tool. It can be generated by performing spectrum analysis on the background vibration signal and updating it in combination with historical data. It is used as a reference for subsequent cutting force signal spectrum purification.

[0064] S30, during the cutting period of automated turning, collecting cutting force signals in real time, performing spectrum analysis on the cutting force signals to obtain an original cutting force spectrum, and performing spectrum difference operation on the original cutting force spectrum using the target background spectrum baseline to obtain a purified cutting force spectrum.

[0065] The original cutting force spectrum refers to the spectrum data obtained after spectrum analysis of the cutting force signal collected in real time during the cutting process of automated turning. It can be obtained by using a force sensor installed on the tool holder or workbench to collect the cutting force signal and perform fast Fourier transform. It is mainly used to include cutting process information and superimposed background noise information.

[0066] This application uses a target background spectrum baseline to perform a spectral differential operation on the original cutting force spectrum to eliminate the influence of background noise. Specifically, this is achieved by subtracting the amplitude of the target background spectrum baseline at the corresponding frequency point from the amplitude of the original cutting force spectrum at each frequency point, thereby stripping the background noise component from the original cutting force spectrum and ultimately obtaining a purified cutting force spectrum. This purified cutting force spectrum is a cutting force spectrum obtained after the spectral differential operation, in which the background noise component is effectively suppressed, accurately reflecting the actual cutting interaction state between the tool and the workpiece.

[0067] S40: Perform cutting state identification based on the spectrum characteristics of the purified cutting force spectrum to obtain an identification result.

[0068] In some embodiments, cutting state identification is achieved by judging the state of the current cutting process based on the spectral characteristics of the purified cutting force spectrum. The amplitude, frequency or change trend of specific frequency components in the purified cutting force spectrum are analyzed and compared with preset thresholds or models to accurately determine whether there are abnormal or specific states such as chatter and tool wear.

[0069] S50, based on the recognition result and according to a preset process rule library, adjusting the processing parameters of the CNC machine tool, wherein the processing parameters include spindle speed, feed rate and cutting depth.

[0070] Machining parameters refer to the process parameters that affect the turning process, including spindle speed, feed rate and cutting depth. Machining parameters are controlled and adjusted by the CNC system to optimize the cutting process, suppress undesirable conditions and ensure machining quality.

[0071] In some embodiments, a CNC lathe machining ultra-thin-walled aluminum alloy pipe fittings on an automated production line is used as an example. During the non-cutting gap between the robotic arm unloading a workpiece and loading the next blank, the system automatically triggers background vibration signal acquisition. During this time, the spindle stops, and a force sensor mounted on the turret acquires a segment of the machine tool's background vibration signal. The acquired time-domain signal is fed into the industrial control computer, where it is transformed through a fast Fourier transform (FFT) to generate the current background spectrum data. The system maintains a machine tool background spectrum baseline and updates it using an exponentially weighted moving average (EWMA) algorithm. For example, a weighted average of the currently acquired background spectrum data and the previous baseline is calculated using a weighting factor to generate a new target background spectrum baseline. A new workpiece is clamped and cutting begins. The force sensor acquires the cutting force signal in real time and performs a FFT to obtain the raw cutting force spectrum. The purification module performs a spectral difference operation, subtracting the amplitude of the current target background spectrum baseline at the corresponding frequency point from the amplitude of the raw cutting force spectrum at each frequency point to obtain a purified cutting force spectrum. The parameter adaptive system analyzes the purified spectrum, for example, by detecting amplitude peaks or rates of change within a specific frequency range to identify the presence of early-stage chatter. If chatter is detected, the system adjusts the CNC machine's spindle speed and feed rate based on pre-set process rules (for example, reducing the spindle speed or feed rate when the chatter frequency is within a certain range and the amplitude exceeds a threshold). During the next non-cutting interval of the machining cycle, the system repeats the above background signal acquisition and baseline update process to ensure that the baseline consistently tracks changes in the machine's status.

[0072] This application collects background vibration signals from a CNC machine tool during the non-cutting period of automated turning to determine the machine tool's inherent vibration characteristics when not affected by cutting forces. Based on the collected background vibration signals, a target background spectrum baseline is generated. This baseline dynamically represents the machine tool's current background vibration characteristics and can reflect changes in its dynamic characteristics due to factors such as prolonged operation. During the subsequent cutting period, a cutting force signal containing cutting information and background noise is collected in real time and spectrally analyzed to obtain the original cutting force spectrum. The key is to perform a spectral difference operation on the original cutting force spectrum using the previously generated dynamic target background spectrum baseline, effectively removing or compensating for the machine tool's background noise components from the original spectrum, resulting in a purified cutting force spectrum. This purified spectrum more accurately reflects the cutting interaction between the tool and the workpiece. Based on the spectral characteristics of the purified cutting force spectrum, accurate cutting state identification is performed to obtain an identification result. Finally, based on the identification results and combined with a pre-set process rule library, the CNC machine tool's machining parameters (such as spindle speed, feed rate, and depth of cut) are adaptively adjusted. The entire process forms a closed loop, periodically updating the background baseline and purifying the cutting signal to ensure that parameter adjustments are always based on pure signals, thereby coping with slow drifts in the machine tool state and ensuring the processing quality of thin-walled parts.

[0073] In some embodiments, step S20 specifically includes:

[0074] S201 , performing fast Fourier transform on the background vibration signal to generate background vibration spectrum data.

[0075] Among them, fast Fourier transform refers to an algorithm for efficiently calculating discrete Fourier transform, which can convert time domain signals into frequency domain representation and can be implemented using the Cooley-Tukey algorithm to obtain the frequency components of the background vibration signal.

[0076] S202 : Based on the preset machine tool background spectrum baseline and the background vibration spectrum data, an exponentially weighted moving average algorithm is used to update and generate a target background spectrum baseline.

[0077] The preset machine tool background spectrum baseline refers to a background vibration spectrum reference collected and established when the machine tool is in a normal or stable state, and can be stored in a storage unit of the system.

[0078] This application performs a fast Fourier transform on the collected background vibration signal to obtain background vibration spectrum data, thereby converting the time domain signal into frequency domain information. Then, using an exponentially weighted moving average algorithm, combined with a preset machine tool background spectrum baseline and real-time background vibration spectrum data, the target background spectrum baseline is dynamically updated. This update method assigns higher weight to the latest background vibration spectrum data, allowing the target background spectrum baseline to quickly respond to instantaneous changes in the machine tool state. At the same time, by weighted averaging historical data, it maintains a certain degree of smoothness and stability, effectively suppressing the interference of occasional noise. The generated dynamically updated target background spectrum baseline can more accurately characterize the current background vibration characteristics of the CNC machine tool. Applying this target background spectrum baseline, which accurately reflects the current machine tool background noise characteristics, to the subsequent spectral differential purification process of the cutting force signal can more effectively remove the background noise component from the original cutting force spectrum, resulting in a purer cutting force spectrum. Cutting state identification and parameter adjustment based on the purified cutting force spectrum can avoid misjudgments caused by background noise contamination, thereby improving the accuracy and reliability of adaptive adjustment of processing parameters and ultimately ensuring the processing quality of thin-walled parts.

[0079] In some embodiments, the system collects background vibration signals of the CNC machine tool during non-cutting periods. For example, a three-axis force sensor installed on the turret can be used to collect signals. The collected time-domain background vibration signal is sent to the signal processing unit. The signal processing unit performs a fast Fourier transform on the signal to generate background vibration spectrum data. It is assumed that the system maintains a preset machine tool background spectrum baseline, which can be established when the machine tool leaves the factory or in the initial stage of stable operation. The system uses an exponentially weighted moving average algorithm to update the target background spectrum baseline. In this way, the target background spectrum baseline can be dynamically adjusted according to the real-time background vibration situation, while retaining the influence of historical information to achieve smooth updates.

[0080] The above technical solution uses a fast Fourier transform to acquire background vibration spectrum data, combined with an exponentially weighted moving average algorithm for dynamic updating. This allows for the rapid and accurate generation and update of a target background spectrum baseline representing the current background vibration characteristics of a CNC machine tool. This method effectively smooths transient noise while rapidly responding to slow changes in the machine tool's state. This allows the generated baseline to accurately reflect the dynamic characteristics of the machine tool's background vibration, providing an accurate reference for subsequent cutting force signal purification, thereby improving the effectiveness and stability of the cutting force signal purification.

[0081] In some embodiments, the step of updating the target background spectrum baseline using an exponentially weighted moving average algorithm based on the preset machine tool background spectrum baseline and the background vibration spectrum data includes:

[0082] Among them, B is the target background spectrum baseline, w is the update weight factor, A1 is the machine tool background spectrum baseline, and A2 is the background vibration spectrum data.

[0083] The machine tool background spectrum baseline A1 refers to a preset or previously updated machine tool background spectrum baseline, which characterizes the background vibration spectrum characteristics of the machine tool in a relatively stable state and serves as a starting point or reference benchmark for this update.

[0084] Background vibration spectrum data A2 refers to the background vibration spectrum data obtained by collecting background vibration signals and performing spectrum analysis during the non-cutting period. It represents the current background vibration spectrum characteristics of the machine tool and aims to provide the latest machine tool status information.

[0085] The update weight factor w is a weight factor ranging from 0 to 1. Its purpose is to control the proportion of the background vibration spectrum data A2 collected this time during the update process, thereby adjusting the response speed of the target background spectrum baseline B to the current machine tool state change.

[0086] The target background spectrum baseline B refers to the target background spectrum baseline generated after this update. Its purpose is to fuse the information of the old baseline A1 and the new data A2 to more accurately characterize the current background vibration characteristics of the CNC machine tool and serve as the benchmark for subsequent cutting force signal purification.

[0087] It refers to a specific calculation method that uses an exponentially weighted moving average algorithm for updating. Its purpose is to smoothly integrate new background vibration information into the original baseline through weighted averaging, thereby achieving dynamic tracking and updating of the baseline.

[0088] This application uses the exponentially weighted moving average algorithm to perform weighted fusion of the preset machine tool background spectrum baseline A1 and the newly collected background vibration spectrum data A2 to generate the target background spectrum baseline B. Specifically, using the formula The calculation is performed, where the updated weight factor w controls the degree of influence of the new data A2 on the updated result B. When the w value is large, the target background spectrum baseline B adapts more quickly to the current background vibration characteristics; when the w value is small, the target background spectrum baseline B changes more smoothly, effectively filtering out transient interference. As a result, the target background spectrum baseline B can dynamically and smoothly track the slow drift of the background vibration characteristics of the CNC machine tool. Applying this dynamically updated target background spectrum baseline B to the differential operation of the original cutting force spectrum collected during cutting effectively removes background noise components, resulting in a purer cutting force spectrum. This combination of dynamic baseline update and spectrum cleanup makes subsequent cutting state identification based on the cleaned spectrum more accurate, enabling adjustment of machining parameters based on accurate identification results. This effectively addresses the problem of cutting force signal contamination caused by machine background noise drift, which in turn affects the accuracy of parameter adaptive control.

[0089] In some embodiments, specifically, at a certain frequency point, it is assumed that the amplitude of the machine tool background spectrum baseline A1 obtained in the previous update at this frequency point is 0.1 units. During the current non-cutting period, after collecting the background vibration signal and performing spectrum analysis, the amplitude of the background vibration spectrum data A2 obtained at this frequency point is 0.5 units. Select the update weight factor w = 0.2. According to the formula B = (1-w)* A1 + w *A2 of the exponentially weighted moving average algorithm, the amplitude of the target background spectrum baseline B after this update at this frequency point is calculated: B = (1 - 0.2) * 0.1 + 0.2 *0.5 = 0.8 * 0.1 + 0.2 * 0.5 = 0.08 + 0.1 = 0.18 units. In the next update cycle, this amplitude of 0.18 units will be used as the new machine tool background spectrum baseline A1, and will be calculated together with the new background vibration spectrum data A2 to continuously and dynamically update the baseline.

[0090] Through the above technical solution, the exponentially weighted moving average algorithm is adopted The machine tool background spectrum baseline A1 is updated to generate the target background spectrum baseline B. This effectively integrates the original machine tool background spectrum baseline A1 with the newly acquired background vibration spectrum data A2, and adjusts and updates the weight factor w based on actual conditions to generate the target background spectrum baseline B that accurately reflects the current machine tool background vibration characteristics, providing a reliable benchmark for subsequent cutting force signal purification.

[0091] In some embodiments, the step of generating a target background spectrum baseline based on the background vibration signal, wherein the target background spectrum baseline characterizes the current background vibration characteristics of the CNC machine tool, comprises:

[0092] determining whether the amplitude of the background vibration signal exceeds a preset threshold, and if the amplitude of the background vibration signal exceeds the preset threshold, determining that the background vibration signal is a disturbed signal;

[0093] removing the background vibration signal of the disturbed signal from the background vibration signal to obtain a residual background vibration signal;

[0094] A target background spectrum baseline is generated using the remaining background vibration signal.

[0095] Among them, the preset threshold refers to a reference value used to determine whether the amplitude of the background vibration signal is abnormal. It can be set according to the normal operating status of the machine tool, the environmental noise level and experience, and is used to distinguish normal background vibration from abnormal signals caused by external disturbances.

[0096] If the amplitude of the background vibration signal exceeds a preset threshold, it is determined to be a disturbed signal, which is used to identify signals affected by abnormal factors.

[0097] The disturbed signal is removed from the collected background vibration signal set to achieve elimination, which can be achieved by data filtering, data cleaning or data screening to eliminate the impact of abnormal data on subsequent processing.

[0098] The residual background vibration signal refers to the background vibration signal retained after the disturbed signal is removed, thereby obtaining a signal that is more representative of the normal background vibration of the machine tool.

[0099] In some embodiments, a reference baseline reflecting the spectral characteristics of the machine tool's current background noise is constructed based on the processed background vibration signal, also known as a target background spectrum baseline. This baseline can be achieved by performing spectral analysis on the remaining background vibration signal, calculating an average spectrum, or employing other statistical methods. This baseline provides a background noise reference for subsequent purification of the cutting force signal.

[0100] This application adds a preprocessing step to the background vibration signal before generating a target background spectrum baseline from it. Specifically, it determines whether the amplitude of the background vibration signal exceeds a preset threshold, thereby identifying disturbed signals. By identifying and eliminating these disturbed signals, the background vibration signal used to generate the target background spectrum baseline becomes purer and more accurately reflects the background vibration characteristics of the machine tool. This application addresses the problem of external disturbances when directly generating a baseline using all background vibration signals because, by setting a preset threshold and performing amplitude determination, it effectively distinguishes abnormal vibration signals not generated by the machine tool itself. By eliminating these disturbed signals from the background vibration signal, the remaining background vibration signal more accurately represents the background noise level of the machine tool under normal conditions. The target background spectrum baseline generated based on this more accurate residual background vibration signal can more precisely characterize the current background vibration characteristics of the CNC machine tool. During the subsequent cutting process, using this more accurate target background spectrum baseline to perform a spectral difference calculation on the real-time collected cutting force signal can more effectively remove background noise interference, resulting in a cutting force spectrum that more accurately reflects the cutting process. It is precisely because a cutting force spectrum that more truly reflects the cutting process is obtained that the cutting state identification based on its spectrum characteristics is more reliable, and the adaptive adjustment of processing parameters based on the identification results is more effective, which ultimately helps to ensure the processing quality and production stability of thin-walled parts.

[0101] In some embodiments, after collecting a background vibration signal, the system first processes it. For example, the signal's root mean square (RMS) amplitude or peak amplitude can be calculated. The calculated amplitude is then compared with a preset threshold. This threshold can be an empirical value determined through long-term monitoring and statistical analysis during normal machine operation, for example, a multiple of the normal background vibration amplitude. If the amplitude (e.g., the RMS value) of the current background vibration signal exceeds this threshold, the system determines that this signal segment has been disturbed and marks it as a disturbed signal. Next, the system removes these disturbed signal segments from all collected background vibration signal data. For example, if a 10-second background vibration signal is collected and segmented into several smaller segments for analysis, and a 2-second segment is found to have an abnormal amplitude, this 2-second segment is removed. The remaining 8-second segment is considered the remaining background vibration signal. Finally, the system uses this remaining background vibration signal to generate a target background spectrum baseline. For example, a fast Fourier transform (FFT) can be performed on the residual background vibration signal to obtain a spectrum. The spectrum of the residual background vibration signal over multiple cycles can then be averaged, or methods such as exponentially weighted moving average can be used to generate or update the target background spectrum baseline. This ensures that the data used to generate the baseline is relatively pure and unaffected by sudden external interference.

[0102] Through the above technical solution, when generating the target background spectrum baseline, background vibration signals affected by external disturbances can be effectively identified and eliminated. This avoids the influence of abnormal signals on baseline generation and improves the accuracy of the target background spectrum baseline. Because the target background spectrum baseline can more realistically reflect the background vibration characteristics of the machine tool, the subsequent use of it to perform spectral differential purification of the cutting force signal can more effectively remove background noise and obtain a cutting force spectrum that more realistically reflects the cutting process. This makes the identification of cutting status based on the purified spectrum more reliable, thereby enabling more effective adaptive adjustment of processing parameters, ultimately helping to ensure the processing quality and production stability of thin-walled parts.

[0103] In some embodiments, the step of generating a target background spectrum baseline using the remaining background vibration signal includes:

[0104] Performing spectrum analysis on the remaining background vibration signal to obtain a current background vibration spectrum;

[0105] Comparing the current background vibration spectrum with the machine tool background spectrum baseline to determine the degree of difference between the current background vibration spectrum and the machine tool background spectrum baseline;

[0106] adjusting update parameters of the machine tool background spectrum baseline according to the degree of the difference;

[0107] The update parameters and the current background vibration spectrum are used to update the machine tool background spectrum baseline to generate a target background spectrum baseline.

[0108] Among them, the time domain signal is converted into a frequency domain signal to realize spectrum analysis, which can be implemented by using algorithms such as fast Fourier transform (FFT) to reveal the energy distribution of the signal at different frequencies.

[0109] The machine tool background spectrum baseline refers to the reference spectrum that characterizes the inherent vibration characteristics of the machine tool in the non-cutting state, which can be obtained by pre-measurement or by accumulating and updating historical data.

[0110] The degree of difference refers to the degree of deviation between the current background vibration spectrum and the machine tool background spectrum baseline in spectrum amplitude or energy distribution. It can be specifically quantified as the difference in amplitude of corresponding frequency points, the similarity index of the entire spectrum, or the energy ratio of a specific frequency interval.

[0111] The update parameter refers to the coefficient or factor used to control the update process of the machine tool background spectrum baseline. It can be a weight factor, a learning rate, or a coefficient in an adjustment function, which is used to determine the degree of influence of the current background vibration spectrum on the update of the machine tool background spectrum baseline.

[0112] This application performs spectral analysis on the residual background vibration signal to obtain the current background vibration spectrum. This spectrum is then compared with the existing machine tool background spectrum baseline to quantify the degree of difference between the two. Because this degree of difference reflects the deviation of the current machine tool background vibration state from the baseline, the application dynamically adjusts the parameters used to update the machine tool background spectrum baseline based on this degree of difference. For example, when the difference is large, more aggressive update parameters can be used to enable the baseline to adapt more quickly to the current state; when the difference is small, more conservative update parameters are used to maintain baseline stability and suppress the impact of transient fluctuations. Subsequently, the machine tool background spectrum baseline is updated using the adjusted update parameters and the current background vibration spectrum to generate a new target background spectrum baseline. This adaptive adjustment of update parameters based on the degree of difference enables the generated target background spectrum baseline to more accurately and timely reflect the actual background vibration characteristics of the machine tool, thereby providing a more reliable reference for subsequent purification of the cutting force signal. This solution further optimizes the baseline generation process by eliminating disturbed signals, improving the baseline's accuracy and adaptability.

[0113] In some embodiments, specifically, after performing spectral analysis on the remaining background vibration signal to obtain the current background vibration spectrum, the system compares the current background vibration spectrum with the currently stored machine tool background spectrum baseline point by point or segment by segment to calculate the degree of difference between the two. For example, the mean absolute difference or mean square difference between the current spectrum amplitude and the baseline amplitude within key frequency points or frequency intervals can be calculated as a quantitative indicator of the degree of difference. Assume that the calculated degree of difference is a numerical value D. The system presets a function or lookup table that outputs a corresponding update parameter w(D) based on the degree of difference D. For example, when D is small, w(D) takes a small value (e.g., 0.05), indicating that the current spectrum has little impact on the baseline and the update process is smooth. When D is large, w(D) takes a large value (e.g., 0.2), indicating that the current spectrum differs significantly from the baseline and a faster baseline update is required to adapt to the current state. The system then uses this update parameter w(D), dynamically determined based on the difference D, combined with the current background vibration spectrum and the original machine tool background spectrum baseline, to calculate a new target background spectrum baseline using an update algorithm (such as an exponentially weighted moving average or other weighted averaging method). For example, the new baseline = (1 - w(D)) * original machine tool background spectrum baseline + w(D) * current background vibration spectrum. This approach eliminates the need for a fixed update speed and amplitude for the machine tool background spectrum baseline, allowing it to adaptively adjust based on the actual deviation of the machine tool background vibration state from the baseline.

[0114] Through this technical solution, when generating a target background spectrum baseline from the residual background vibration signal, the difference between the current background vibration spectrum and the machine tool background spectrum baseline is taken into account, and the update parameters of the machine tool background spectrum baseline are dynamically adjusted based on this difference. This enables the generated target background spectrum baseline to more accurately capture the slowly changing characteristics of the machine tool background vibration, improving the real-time nature and accuracy of the baseline. This provides a more precise reference for subsequent spectral purification of the cutting force signal, helping to improve the accuracy of cutting state identification and ultimately ensuring the machining quality of thin-walled parts.

[0115] In some embodiments, the step of performing a spectrum difference operation on the original cutting force spectrum using the target background spectrum baseline to obtain a purified cutting force spectrum includes:

[0116] According to the corresponding relationship between the original cutting force spectrum and the target background spectrum baseline at each frequency point, identifying the frequency interval in which the spectrum components of the original cutting force spectrum overlap with the target background spectrum baseline;

[0117] Determining a spectrum processing strategy for processing the original cutting force spectrum component within the frequency interval based on a preset comparison condition between the amplitude of the original cutting force spectrum component within the frequency interval and the amplitude of the corresponding component of the target background spectrum baseline;

[0118] Based on the spectrum processing strategy, the original cutting force spectrum components in the frequency interval are processed to obtain the overlapped region spectrum components; and

[0119] The original cutting force spectrum components that are not in the frequency range are processed using a preset reference spectrum processing method to obtain non-overlapping area spectrum components;

[0120] The overlapping region spectrum component and the non-overlapping region spectrum component are combined to generate the cleanup cutting force spectrum.

[0121] The frequency interval refers to the frequency range in which the original cutting force spectrum and the target background spectrum baseline have similar amplitudes or influence each other in the frequency domain, which can be identified by setting an amplitude difference threshold or an amplitude ratio threshold.

[0122] The preset comparison condition refers to the rule used to judge the relative strength relationship between the original cutting force spectrum components and the corresponding components of the target background spectrum baseline, which can be defined by amplitude difference, amplitude ratio or a combination of the two.

[0123] The spectrum processing strategy refers to the specific processing method adopted for the spectrum components in the overlapping frequency range, which can be implemented by weighted difference, adaptive filtering or signal separation.

[0124] The preset reference spectrum processing method refers to a standard processing method used for spectrum components in non-overlapping frequency intervals, which can be implemented by simple spectrum difference or directly retaining the original spectrum components.

[0125] In some embodiments, by identifying the frequency intervals in the original cutting force spectrum where the spectrum components overlap with the target background spectrum baseline based on the correspondence between the original cutting force spectrum and the target background spectrum baseline at each frequency point, it is possible to specifically process the areas where the background noise and the effective signal overlap, avoiding unnecessary complex processing of the non-overlapping areas. Based on the preset comparison conditions between the amplitudes of the original cutting force spectrum components in the frequency interval and the amplitudes of the corresponding components of the target background spectrum baseline, the spectrum processing strategy for processing the original cutting force spectrum components in the frequency interval is determined. It is precisely because of the introduction of the comparison conditions and strategy selection that the processing of overlapping areas is no longer a simple subtraction, but a more appropriate method is adopted based on the relative strength of the signal and the noise. For example, when the original signal is much stronger than the background, only a small amount of compensation may be required; when the two are close, more sophisticated separation technology may be required, which improves the intelligence and adaptability of noise elimination. Based on a spectrum processing strategy, the original cutting force spectrum components within the frequency interval are processed to obtain the spectral components in the overlapping region. Simultaneously, the original cutting force spectrum components outside the frequency interval are processed using a preset baseline spectrum processing method to obtain the spectral components in the non-overlapping region. The different processing methods for the overlapping and non-overlapping regions effectively suppress the noise in the overlapping region while maximally preserving the effective signal in the non-overlapping region, thus avoiding the signal loss or distortion that may be caused by simple differentiation. Finally, the spectral components in the overlapping and non-overlapping regions are combined to generate a purified cutting force spectrum, resulting in a purified spectrum that suppresses background noise while retaining effective cutting force information. This region-specific, adaptive processing method, combined with the basic background baseline differentiation framework, improves the accuracy of cutting force signal purification and provides a more reliable data foundation for subsequent cutting state identification.

[0126] In some embodiments, it is assumed that the amplitude of the raw cutting force spectrum at a certain frequency point f1 is A_raw(f1), and the amplitude of the target background spectrum baseline at the frequency point f1 is A_base(f1).

[0127] First, the system identifies frequency intervals based on pre-set rules. For example, a threshold T is set, and if |A_raw(f) - A_base(f)| < T, the frequency point f is considered to be in the overlapping frequency interval.

[0128] For a frequency point f within the overlapping frequency range, a spectrum processing strategy is determined based on a preset comparison condition. For example, the preset comparison condition may be defined as an amplitude ratio R = A_raw(f) / A_base(f).

[0129] If R is less than a threshold R1 (for example, 1.5), it means that the amplitude of the original signal is close to that of the background noise. A weighted difference strategy may be used. For example, the purified amplitude A_purified(f) = A_raw(f) - k * A_base(f), where the weight factor k is dynamically adjusted according to R.

[0130] If R is between R1 and R2 (e.g., 2.5), it means that the raw signal is slightly stronger than the background, and a simple difference strategy may be used, such as A_purified(f) = max(0, A_raw(f) - A_base(f)).

[0131] If R is greater than R2, it means that the original signal is much stronger than the background, and a threshold filtering strategy may be used. For example, if A_raw(f) is greater than a certain threshold T_signal, then A_purified(f) = A_raw(f), otherwise A_purified(f) = 0.

[0132] Based on the established spectrum processing strategy, the raw cutting force spectrum components within the overlapping region are processed to obtain the overlapping region spectrum components. For frequency points f' that are not within the overlapping frequency interval, a preset baseline spectrum processing method, such as simple difference, is used to obtain the non-overlapping region spectrum components A_purified(f') = max(0, A_raw(f') - A_base(f')). Finally, the overlapping region spectrum components and the non-overlapping region spectrum components are combined to form a complete purified cutting force spectrum.

[0133] Through the above technical solution, the present application can more accurately separate the effective cutting force signal from the original cutting force spectrum, reduce the interference of background noise on the effective signal, avoid the loss or distortion of the effective signal that may be caused by simple differentiation, improve the accuracy of the purified cutting force spectrum, and provide a more reliable data basis for subsequent cutting state identification and parameter adjustment.

[0134] In some embodiments, the step of adjusting the update parameters of the machine tool background spectrum baseline according to the degree of difference includes:

[0135] Performing a validity judgment on the degree of difference, the validity judgment comprising judging whether the instantaneous change characteristics of the degree of difference conform to a preset drift model;

[0136] Processing the difference degree according to the validity judgment result to obtain a target difference degree, wherein the processing includes correcting the difference degree when the difference degree does not conform to a preset drift model;

[0137] The target difference degree is used to adjust update parameters of the machine tool background spectrum baseline.

[0138] Among them, validity judgment is the process of evaluating the reliability or validity of the degree of difference, which is achieved by setting thresholds, analyzing change trends, or comparing with historical data.

[0139] The instantaneous change characteristic refers to the short-term change pattern or fluctuation feature of the difference degree during the continuous acquisition cycle, which is characterized by calculating the difference between adjacent difference degrees, the rate of change or analyzing the fluctuation frequency. The preset drift model refers to a pre-established mathematical model or behavior pattern that describes the slow and gradual change of the machine tool background noise over time. It can be a linear or nonlinear trend model, a probability model based on statistical distribution or a set of rules to limit instantaneous fluctuations.

[0140] The target difference degree refers to a difference degree value that is more valuable for reference in subsequent adjustment and update of parameters after validity judgment and processing. It can be obtained from the original difference degree by filtering, smoothing, correction or selective retention.

[0141] In some embodiments, the degree of difference judged to be invalid or abnormal is corrected and adjusted to make it more consistent with the actual drift law of the machine tool background noise. This can be achieved by smoothing based on historical data, predictive adjustment based on a preset model, or weighted averaging based on instantaneous change characteristics.

[0142] When adjusting the update parameters of the machine tool background spectrum baseline based on the degree of difference, the present application does not directly use the original degree of difference value, but first performs a validity judgment on the degree of difference. This judgment process analyzes the instantaneous change characteristics of the degree of difference and compares it with the preset machine tool background noise drift model, thereby identifying abnormal fluctuations caused by instantaneous interference or random noise that do not conform to the slow drift law of the machine tool background. Subsequently, based on the results of the validity judgment, the degree of difference is processed to obtain a target degree of difference. If the degree of difference is judged to be inconsistent with the preset drift model, it is corrected to make it closer to the actual drift trend, avoiding the problems that may arise from directly using abnormal data or simply discarding data. Finally, the target degree of difference after validity judgment and processing is used to adjust the update parameters of the machine tool background spectrum baseline. By using a more reliable target degree of difference, it can be ensured that the baseline update parameters can accurately reflect the actual, slow drift of the machine tool background noise, avoiding frequent or erroneous baseline updates caused by instantaneous fluctuations. This preprocessing mechanism for the degree of difference, a key step in the entire adaptive control method, ensures the stability and accuracy of the machine tool background spectrum baseline, thereby improving the precision of subsequent spectrum differentiation operations. This makes it more effective to remove background noise from the original cutting force spectrum, ultimately enhancing the accuracy and robustness of cutting state identification and machining parameter adjustment based on the purified cutting force spectrum. This intelligent judgment and correction mechanism, introduced before adjusting and updating parameters, makes the entire adaptive system more adaptable to dynamic changes in machine tool background noise, improving the reliability of the entire method.

[0143] In one specific embodiment, the update parameters of the machine tool background spectrum baseline can be adjusted based on the degree of difference. First, the degree of difference between the currently acquired background vibration spectrum and the current machine tool background spectrum baseline is calculated. For example, the spectral energy difference between the two within a critical frequency range can be calculated. Then, the validity of this degree of difference is determined. A change rate threshold can be set. If the rate of change of the current degree of difference relative to the previously calculated degree of difference exceeds this threshold, its instantaneous change characteristics are considered to not conform to the preset slow drift model and the degree of difference is determined to be invalid. Next, the degree of difference is processed based on the validity determination result. If the degree of difference is determined to be invalid, the previously determined valid degree of difference value can be used as the target degree of difference for the current cycle, or the average of several previously determined valid degree of difference can be used as the target degree of difference, thereby correcting the invalid degree of difference. If the degree of difference is determined to be valid, the target degree of difference is directly used as the target degree of difference. Finally, the target degree of difference is used to adjust the update parameters of the machine tool background spectrum baseline. For example, a predefined functional relationship can be set so that the update parameters increase as the target degree of difference increases, thereby achieving adaptive adjustment of the baseline update rate.

[0144] Through the above technical solution, the effectiveness of the difference degree used to adjust the machine tool background spectrum baseline update parameters is judged and processed. This effectively eliminates the influence of interference factors such as instantaneous impact and random noise on the difference degree value, avoiding frequent or erroneous baseline updates caused by irregular fluctuations in the difference degree. The processed target difference degree can more accurately reflect the true slow drift trend of the machine tool background noise, making the baseline update more stable and accurate. This improves the tracking accuracy and robustness of the machine tool background spectrum baseline, providing a reliable benchmark for subsequent spectrum differentiation and cutting state identification, thereby improving the performance and reliability of the entire thin-walled part turning parameter adaptation method.

[0145] In some embodiments, the step of correcting the degree of difference includes:

[0146] analyzing the instantaneous change characteristics of the difference degree;

[0147] Dynamically adjusting a correction parameter for correcting the degree of difference based on the instantaneous change characteristic;

[0148] The correction parameter is used to correct the degree of difference.

[0149] The correction parameter refers to a parameter used to adjust the value of the difference degree, which can be a weight factor, a smoothing coefficient or a cutoff frequency of a filter.

[0150] Dynamic adjustment refers to the real-time or periodic change of the correction parameters according to the instantaneous change characteristics of the difference degree, which can be achieved by rule-based judgment, lookup table mapping or adaptive algorithm.

[0151] Correction refers to processing the original degree of difference to reduce the impact of its instantaneous fluctuation or mutation, which can be achieved by using methods such as sliding average, exponential smoothing or median filtering.

[0152] Through the above explanation of the instantaneous variation characteristics of the degree of difference, correction parameters, dynamic adjustment, and correction, the present application specifically analyzes the instantaneous variation characteristics of the degree of difference to identify possible abnormal fluctuations or sudden changes in the degree of difference. Based on the analysis results of the instantaneous variation characteristics, the system can determine the reliability level and fluctuation type of the current degree of difference. Based on the instantaneous variation characteristics, the system dynamically adjusts the correction parameters used to correct the degree of difference. This means that the correction parameters are not fixed but are adjusted according to the actual fluctuations in the degree of difference. For example, when a sharp fluctuation or sudden change in the degree of difference is detected, a stronger correction strategy can be adopted, such as setting a larger smoothing coefficient or using a stronger filtering algorithm, to quickly suppress the impact of outliers. When the degree of difference changes more slowly, a weaker correction strategy can be adopted, such as setting a smaller smoothing coefficient or not performing correction, to preserve the actual change information of the degree of difference. The adjusted correction parameters are used to correct the degree of difference to obtain a stable and reliable target degree of difference. This corrected target degree of difference is used to adjust the update parameters of the machine tool background spectrum baseline. By providing an accurate and stable difference degree as input, we can avoid incorrect parameter adjustments caused by transient fluctuations in the original difference degree, thereby ensuring that the background spectrum baseline accurately tracks the slow drift of the machine tool state without being affected by transient interference. This correction mechanism, combined with a scheme that adjusts the update parameters based on the target difference degree, can improve the accuracy and stability of the background spectrum baseline update, thereby enhancing the accuracy and robustness of subsequent cutting state identification and machining parameter adjustment.

[0153] On the other hand, Figure 2 As shown, the present application further proposes a thin-walled part turning parameter adaptive system 100, which includes:

[0154] A background vibration signal acquisition module 10 is used to collect background vibration signals of a CNC machine tool during a non-cutting period of automated turning, wherein the CNC machine tool is used for automated turning of thin-walled parts;

[0155] A background spectrum baseline generating module 20 is configured to generate a target background spectrum baseline according to the background vibration signal, wherein the target background spectrum baseline represents the current background vibration characteristics of the CNC machine tool;

[0156] a cutting force signal processing module 30 for collecting cutting force signals in real time during the cutting process of the automated turning operation, performing spectrum analysis on the cutting force signals to obtain an original cutting force spectrum, and performing spectrum difference calculation on the original cutting force spectrum using the target background spectrum baseline to obtain a purified cutting force spectrum;

[0157] an identification module 40 for identifying a cutting state based on the spectrum characteristics of the purified cutting force spectrum to obtain an identification result;

[0158] The parameter adjustment module 50 is used to adjust the processing parameters of the CNC machine tool based on the recognition result and according to a preset process rule library, wherein the processing parameters include spindle speed, feed rate and cutting depth.

[0159] Here, a module refers to a unit with specific functions, which can be implemented by hardware circuits, software programs, or a combination of hardware circuits and software programs.

[0160] The present application collects background vibration signals during non-cutting periods through a background vibration signal acquisition module, and a background spectrum baseline generation module generates a target background spectrum baseline that characterizes the current background vibration characteristics of the machine tool based on the collected background vibration signals. During cutting, the cutting force signal processing module collects cutting force signals in real time, performs spectrum analysis to obtain the original cutting force spectrum, and uses the target background spectrum baseline to perform spectrum difference operation on the original cutting force spectrum to obtain a purified cutting force spectrum. The recognition module identifies the cutting state based on the spectrum characteristics of the purified cutting force spectrum. The parameter adjustment module adjusts the processing parameters based on the recognition results and the preset process rule library. Through the orderly collaboration of these modules, the entire system realizes real-time monitoring, state identification and parameter adaptive adjustment of the cutting process, and solves the problem that each functional step requires the implementation of corresponding modules and the collaborative work between modules.

[0161] Through the above technical solution, this application provides a thin-walled part turning parameter adaptive system. This system achieves effective integration and collaborative work of functions such as background vibration signal acquisition, background spectrum baseline generation, cutting force signal processing, cutting state identification, and parameter adjustment by clearly dividing functional modules and specifying the information flow and collaborative relationship between modules. This solves the technical problem of how to effectively cooperate with each module to realize the overall function of the adaptive system in practical applications, ensuring that the system can operate stably and reliably and realize adaptive adjustment of thin-walled part turning parameters.

[0162] In some embodiments, the background spectrum baseline generation module includes a background spectrum processing module and a machine tool background spectrum baseline storage module, wherein:

[0163] A machine tool background spectrum baseline storage module is used to store a preset machine tool background spectrum baseline;

[0164] The background spectrum processing module is used to perform fast Fourier transform on the background vibration signal to generate background vibration spectrum data; and

[0165] After obtaining the preset machine tool background spectrum baseline from the machine tool background spectrum baseline storage module, the preset machine tool background spectrum baseline and the background vibration spectrum data are updated using an exponentially weighted moving average algorithm to generate a target background spectrum baseline.

[0166] The above description is merely an embodiment of the present application and is not intended to limit the scope of protection of the present application. For those skilled in the art, various modifications and variations of the present application are possible. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the scope of protection of the present application.

Claims

1. A method for adaptive parameters in turning of thin-walled parts, characterized in that: include: collecting background vibration signals of a CNC machine tool during a non-cutting period of automated turning, wherein the CNC machine tool is used for automated turning of thin-walled parts; generating a target background spectrum baseline according to the background vibration signal, wherein the target background spectrum baseline represents the current background vibration characteristics of the CNC machine tool; Determining whether the amplitude of the background vibration signal exceeds a preset threshold, and if the amplitude of the background vibration signal exceeds the preset threshold, determining that the background vibration signal is a disturbed signal; removing the background vibration signal of the disturbed signal from the background vibration signal to obtain a residual background vibration signal; and generating a target background spectrum baseline using the residual background vibration signal; Performing spectrum analysis on the remaining background vibration signal to obtain a current background vibration spectrum; comparing the current background vibration spectrum with a machine tool background spectrum baseline to determine a degree of difference between the current background vibration spectrum and the machine tool background spectrum baseline; adjusting an update parameter of the machine tool background spectrum baseline based on the degree of difference; and updating the machine tool background spectrum baseline using the updated parameter and the current background vibration spectrum to generate a target background spectrum baseline. During the cutting process of automated turning, a cutting force signal is collected in real time, a spectrum analysis is performed on the cutting force signal to obtain an original cutting force spectrum, and a spectrum difference operation is performed on the original cutting force spectrum using the target background spectrum baseline to obtain a purified cutting force spectrum; According to the correspondence between the original cutting force spectrum and the target background spectrum baseline at each frequency point, the frequency interval in the original cutting force spectrum where the spectrum components overlap with the target background spectrum baseline is identified; based on a preset comparison condition between the amplitude of the original cutting force spectrum components in the frequency interval and the amplitude of the corresponding components of the target background spectrum baseline, a spectrum processing strategy for processing the original cutting force spectrum components in the frequency interval is determined; based on the spectrum processing strategy, the original cutting force spectrum components in the frequency interval are processed to obtain overlapping area spectrum components; and, a preset reference spectrum processing method is used to process the original cutting force spectrum components that are not in the frequency interval to obtain non-overlapping area spectrum components; the overlapping area spectrum components and the non-overlapping area spectrum components are combined to generate the purified cutting force spectrum; performing cutting state identification based on the spectrum characteristics of the purified cutting force spectrum to obtain an identification result; Based on the recognition result and according to a preset process rule library, the processing parameters of the CNC machine tool are adjusted, wherein the processing parameters include spindle speed, feed rate and cutting depth.

2. The thin-walled part turning parameter adaptive method according to claim 1, characterized in that: The step of generating a target background spectrum baseline according to the background vibration signal, wherein the target background spectrum baseline characterizes the current background vibration characteristics of the CNC machine tool comprises: Performing a fast Fourier transform on the background vibration signal to generate background vibration spectrum data; According to the machine tool background spectrum baseline and the background vibration spectrum data, an exponentially weighted moving average algorithm is adopted to update and generate a target background spectrum baseline.

3. The thin-walled part turning parameter adaptive method according to claim 2, characterized in that: The step of updating the target background spectrum baseline by using an exponentially weighted moving average algorithm based on the preset machine tool background spectrum baseline and the background vibration spectrum data includes: B=(1-w)*A1+w*A2,(0<w<1) Among them, B is the target background spectrum baseline, w is the update weight factor, A1 is the machine tool background spectrum baseline, and A2 is the background vibration spectrum data.

4. The thin-walled part turning parameter adaptive method according to claim 1, characterized in that: The step of adjusting the update parameters of the machine tool background spectrum baseline according to the degree of difference includes: Performing a validity judgment on the degree of difference, the validity judgment comprising judging whether the instantaneous change characteristics of the degree of difference conform to a preset drift model; Processing the difference degree according to the validity judgment result to obtain a target difference degree, wherein the processing includes correcting the difference degree when the difference degree does not conform to a preset drift model; The target difference degree is used to adjust update parameters of the machine tool background spectrum baseline.

5. The thin-walled part turning parameter adaptive method according to claim 4, characterized in that: The step of correcting the degree of difference comprises: analyzing the instantaneous change characteristics of the difference degree; Dynamically adjusting a correction parameter for correcting the degree of difference based on the instantaneous change characteristic; The correction parameter is used to correct the degree of difference.

6. A thin-walled part turning parameter adaptive system, used to execute the thin-walled part turning parameter adaptive method according to any one of claims 1 to 5, characterized in that: The system includes: A background vibration signal acquisition module is used to collect background vibration signals of a CNC machine tool during a non-cutting period of automated turning, wherein the CNC machine tool is used for automated turning of thin-walled parts; A background spectrum baseline generating module is used to generate a target background spectrum baseline according to the background vibration signal, wherein the target background spectrum baseline represents the current background vibration characteristics of the CNC machine tool; a cutting force signal processing module, configured to collect a cutting force signal in real time during the cutting process of the automated turning operation, perform spectrum analysis on the cutting force signal to obtain an original cutting force spectrum, and perform spectrum difference calculation on the original cutting force spectrum using the target background spectrum baseline to obtain a purified cutting force spectrum; an identification module, configured to identify the cutting state based on the spectrum characteristics of the purified cutting force spectrum to obtain an identification result; A parameter adjustment module is used to adjust the processing parameters of the CNC machine tool based on the recognition result and according to a preset process rule library, wherein the processing parameters include spindle speed, feed rate and cutting depth.

7. The thin-walled parts turning parameter adaptive system according to claim 6, characterized in that: The background spectrum baseline generation module includes a background spectrum processing module and a machine tool background spectrum baseline storage module, wherein: A machine tool background spectrum baseline storage module is used to store a preset machine tool background spectrum baseline; The background spectrum processing module is used to perform fast Fourier transform on the background vibration signal to generate background vibration spectrum data; and After obtaining the preset machine tool background spectrum baseline from the machine tool background spectrum baseline storage module, the preset machine tool background spectrum baseline and the background vibration spectrum data are updated using an exponentially weighted moving average algorithm to generate a target background spectrum baseline.

Images