Fault monitoring method and system for numerical control machine tool

Abstract

The invention relates to the technical field of numerical control system fault pre-diagnosis, and particularly discloses a fault monitoring method and system for a numerical control machine tool. The method comprises the steps that a spindle vibration signal, a cutting force signal and a spindle current signal are collected; outputting an alignment signal set; respectively executing adaptive filtering processing according to the alignment signal set; constructing a multi-dimensional feature vector; constructing a health index under the processing behavior period; then accumulated trend deviation analysis is executed through a CUSUM algorithm; and judging whether the tool wear trend is abnormal. The technical problem that in the prior art, static threshold judgment is carried out based on a single signal source, and especially in a numerical control machining scene with signal noise interference, accurate early warning of the early wear trend of a tool cannot be achieved is solved. Due to the fact that a period calibration mechanism fusing spindle vibration, cutting force and current signals, feature vector entropy weight optimization and a CUSUM trend detection mechanism are adopted, the pre-diagnosis level of the numerical control machine tool on cutter state changes is improved.

Description

Technical Field

[0001] This invention relates to the field of fault pre-diagnosis technology for CNC systems, and in particular to a fault monitoring method and system for CNC machine tools. Background Technology

[0002] Currently, cutting tools, as core actuators in CNC machine tool machining, have a significant impact on machining accuracy, workpiece surface quality, and equipment operational safety due to their wear condition. With the increasing complexity and continuity of machining tasks, cutting tools are prone to wear forms such as dulling, chipping, and chipping during prolonged operation. If the wear trend cannot be identified and warned in advance, it can easily lead to increased dimensional errors, deterioration of machining quality, and even machine tool failure or workpiece scrapping, resulting in production interruptions and increased maintenance costs.

[0003] Existing technologies mainly include monitoring methods based on single sensor signals, judgment methods based on multi-feature statistical analysis, and prediction methods based on artificial intelligence models. Among them, monitoring methods based on single signals, such as spindle current, vibration signals, or cutting force signals, extract mean, volatility, or spectral features to determine the degree of wear. Although simple to implement, these methods are easily affected by machine tool load fluctuations, process parameter changes, and environmental noise, resulting in poor feature extraction stability and a high false alarm rate. Some studies have introduced multi-dimensional signal fusion technology to improve recognition accuracy, but they lack effective processing mechanisms for the temporal synchronization and dynamic alignment of multi-source signals. Especially when there are slight disturbances in the machining cycle, time delay offsets exist between signals, affecting the reliability of fused features. Although some studies have attempted to use machine learning methods such as neural networks and support vector machines for wear prediction, these methods usually require a large amount of labeled data for offline training. The models are sensitive to changes in working conditions, are difficult to deploy, and their stability and interpretability are hard to guarantee. They are still difficult to meet the combined requirements of real-time performance and lightweight design in actual production sites. Especially in the early stages of tool wear, due to the weak amplitude and slow change trend of the degradation signal, existing methods often cannot accurately capture the subtle wear evolution trend, resulting in a significant lag in wear trend identification.

[0004] Therefore, there is an urgent need for a fault monitoring method for CNC machine tools to achieve stable perception and accurate identification of early tool wear, thereby improving the intelligence level of equipment operation and maintenance efficiency. Summary of the Invention

[0005] To address the aforementioned technical shortcomings, the purpose of this invention is to propose a fault monitoring method for CNC machine tools. This method aims to solve the technical problem that existing technologies, which rely on static threshold discrimination based on a single signal source, are unable to accurately predict early tool wear trends, especially in CNC machining scenarios with signal noise interference.

[0006] To solve the above-mentioned technical problems, the present invention adopts the following technical solution: The present invention provides a fault monitoring method for CNC machine tools. The fault monitoring method for the CNC machine tool includes: Step S10: Synchronously acquire the spindle vibration signal at time t during CNC machine tool operation. Cutting force signal and spindle current signal Based on the spindle vibration signal Cutting force signal and spindle current signal A cross-correlation synchronization mechanism based on joint verification of processing behavior cycles is used to perform time delay calibration, and a set of timing-consistent aligned signals is output. Step S20: Perform adaptive filtering on the output aligned signal set to output a normalized signal set; Step S30: Obtain the processing behavior cycle The normalized signal set is arranged according to the processing behavior period. Sliding segmentation to construct the k-th processing cycle eigenvectors within ; Step S40: Use the information entropy weight optimization mechanism to optimize the feature vector. Perform weighted fusion and output a health index under the processing behavior cycle. Based on health index Calculate the cumulative trend offset using the CUSUM algorithm. ; Step S50: Preset cumulative trend offset threshold When the cumulative trend shifts Greater than or equal to the cumulative trend offset threshold If an abnormal tool wear trend is detected, an early tool wear alarm signal will be output. When the cumulative trend shifts Less than the cumulative trend offset threshold At that time, output an early tool wear alarm signal. .

[0007] Preferably, in step S10, based on the spindle vibration signal... Cutting force signal and spindle current signal The steps of performing time delay calibration processing using a cross-correlation synchronization mechanism based on joint verification of processing behavior cycles, and outputting a time-consistent set of aligned signals, specifically include: Obtain the machine tool spindle speed n, and calculate the machining cycle at the machine tool spindle speed n. , and the collected spindle vibration signal Cutting force signal and spindle current signal According to the processing cycle Segmentation is performed to obtain signal segments. ;in, This represents the spindle vibration signal during the kth machining cycle. The cutting force signal for the kth machining cycle; This is the spindle current signal for the kth machining cycle; The spindle vibration signal during the kth machining cycle Using the reference anchor signal, an improved Hilbert autocorrelation endpoint enhancement detection algorithm is employed to extract the main cutting pulse position, thus obtaining the main cutting pulse position. Output the set of periodic anchor indexes. ; Based on periodic anchor point index set For cutting force signals and spindle current signal A local phase-matching calibration method is used to perform time delay correction, and the final output is a set of aligned signals with consistent timing.

[0008] Preferably, in step S10, an improved Hilbert autocorrelation endpoint enhancement detection algorithm is used to extract the main cutting pulse position from the reference anchor point signal, thereby obtaining the main cutting pulse position. The steps specifically include: First, the reference anchor signal is subjected to envelope enhancement processing using the Hilbert transform method, and the Hilbert transform signal is output. Selecting the processing behavior period based on the Hilbert transform signal The first signal range and processing cycle of the head The second signal interval at the tail is used to calculate the autocorrelation enhancement between the first and second signal intervals. The position of the main cutting pulse is extracted from the reference anchor signal based on the autocorrelation enhancement. .

[0009] Preferably, step S20, which involves performing adaptive filtering on the output aligned signal set to output a normalized signal set, specifically includes: Step S201: Obtain the alignment signal set, which includes the spindle alignment signal, cutting force alignment signal, and spindle current alignment signal; perform adaptive filtering on the spindle alignment signal using wavelet packet decomposition; output the normalized spindle signal. Step S202: Obtain spindle speed The cutting force alignment signal is based on the spindle speed. Dynamically set the center frequency of the bandpass filter , Based on the center frequency of the bandpass filter Adaptive filtering is performed using a variable step size empirical mode bandpass filter; the output is a normalized cutting force signal. Step S203: Adaptive filtering is performed on the spindle current alignment signal using baseline drift elimination combined with sliding window mean filtering; output the normalized spindle current signal. Step S204: Merge the normalized spindle signal, the normalized cutting force signal, and the normalized spindle current signal, and normalize the normalized spindle signal, the normalized cutting force signal, and the normalized spindle current signal to a zero-mean, unit-variance form, and finally output the normalized signal set.

[0010] Preferably, in step S30, the feature vector It includes vibration signal kurtosis characteristics, vibration high-frequency band energy characteristics, cutting force main frequency band energy characteristics, and current signal fluctuation coefficient characteristics. Among them, vibration signal kurtosis characteristics are used to characterize the instantaneousness and abnormality of impact signals; vibration high-frequency band energy characteristics are used to measure the intensity of high-frequency disturbances during high-speed cutting; cutting force main frequency band energy characteristics are used to reflect the cutting load intensity and changes of the tool within the main frequency range; and current signal fluctuation coefficient characteristics are used to evaluate the stability of the spindle drive system within this cycle.

[0011] Preferably, in step S40, an information entropy weight optimization mechanism is used to optimize the feature vector. Perform weighted fusion and output a health index under the processing behavior cycle. The steps specifically include: first calculating the feature vectors respectively. The relative proportion of each dimension feature in all processing behavior cycles is used to set an information entropy value; then, an information redundancy index is constructed based on the set information entropy value; finally, a weighted calculation is performed based on the information redundancy index and the features of the corresponding dimension to output the health index under the processing behavior cycle. Health Index It is used to comprehensively reflect abnormal changes in multiple dimensions within the current period.

[0012] Preferably, in step S40, based on the health index Calculate the cumulative trend offset using the CUSUM algorithm. The steps specifically include: pre-setting a health index benchmark value and a micro-offset tolerance parameter. The health index benchmark value is used to characterize the health level reference standard under normal processing conditions, and the micro-offset tolerance parameter is used to define the allowable fluctuation range of the health index; and within the processing cycle, based on the health index... The health index benchmark and micro-offset tolerance parameters are used to calculate the cumulative trend offset using the CUSUM algorithm. .

[0013] The present invention also provides a fault monitoring system for CNC machine tools, comprising: The multi-source signal synchronization and alignment module is used to synchronously acquire the spindle vibration signal at time t during CNC machine tool operation. Cutting force signal and spindle current signal Based on the spindle vibration signal Cutting force signal and spindle current signal A cross-correlation synchronization mechanism based on joint verification of processing behavior cycles is used to perform time delay calibration, and a set of timing-consistent aligned signals is output. The normalization filtering module is used to perform adaptive filtering on the output aligned signal set and output a normalized signal set. The processing cycle feature extraction module is used to obtain the processing behavior cycle. The normalized signal set is arranged according to the processing behavior period. Sliding segmentation to construct the k-th processing cycle eigenvectors within ; The health index construction and offset analysis module is used to optimize feature vectors using an information entropy weighting mechanism. Perform weighted fusion and output a health index under the processing behavior cycle. Based on health index Calculate the cumulative trend offset using the CUSUM algorithm. ; The trend anomaly detection and alarm module is used to preset the cumulative trend deviation threshold. When the cumulative trend shifts Greater than or equal to the cumulative trend offset threshold If an abnormal tool wear trend is detected, an early tool wear alarm signal will be output. When the cumulative trend shifts Less than the cumulative trend offset threshold At that time, output an early tool wear alarm signal. .

[0014] The present invention also provides a fault monitoring device for CNC machine tools, comprising: a memory, a processor, and a fault monitoring program for CNC machine tools stored in the memory and executable on the processor. When the fault monitoring program for CNC machine tools is executed by the processor, a fault monitoring method for CNC machine tools is implemented.

[0015] The present invention also provides a computer program product, including a fault monitoring program for CNC machine tools, wherein the fault monitoring program for CNC machine tools, when executed by a processor, implements the fault monitoring method for CNC machine tools.

[0016] The beneficial effects of this invention are as follows: This invention uses a cross-correlation synchronization mechanism based on joint verification of machining behavior cycles to perform time delay calibration on spindle vibration signals, cutting force signals, and spindle current signals, solving the problem of time offset in multi-source sensor acquisition and resulting distortion of state judgment in the prior art; furthermore, through wavelet packet decomposition, adaptive filtering, and normalization processing, the consistency of the signals and the sensitivity to fault characteristics are significantly improved, providing a higher quality raw data foundation for subsequent fault feature extraction and trend judgment.

[0017] This invention employs an information entropy weight optimization mechanism to construct a health index that integrates multi-dimensional features, and combines it with the CUSUM algorithm for trend deviation detection. This can effectively identify minute health changes during the processing and issue early warnings of tool wear. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating the first embodiment of a fault monitoring method for CNC machine tools according to the present invention.

[0020] Figure 2 This is a schematic diagram of the equipment for a fault monitoring method for CNC machine tools according to the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Example 1: As Figure 1 The diagram shown is a flowchart of the first embodiment of the fault monitoring method for CNC machine tools of the present invention, which presents the first embodiment of the fault monitoring method for CNC machine tools of the present invention.

[0023] In the first embodiment, the fault monitoring method for the CNC machine tool includes: Step S10: Synchronously acquire the spindle vibration signal at time t during CNC machine tool operation. Cutting force signal and spindle current signal Based on the spindle vibration signal Cutting force signal and spindle current signal A cross-correlation synchronization mechanism based on joint verification of processing behavior cycles is used to perform time delay calibration, and a set of timing-consistent aligned signals is output. It should be noted that the "cross-correlation synchronization mechanism based on joint verification of machining behavior cycles" in this step refers to the precise calibration of the time delay of spindle vibration signals, cutting force signals, and spindle current signals by calculating the cross-correlation between the signals during the operation of the CNC machine tool, based on the relative timing of these signals within each machining cycle. Specifically, this mechanism divides the signals into multiple machining cycle segments and precisely aligns these segments to ensure that the relative position of each signal is consistent within the same machining cycle. This method is not merely simple signal synchronization; it also utilizes the dynamic characteristics within the machining cycle, considering factors such as signal amplitude and frequency, to ensure optimal signal alignment accuracy under different machining conditions. This technical solution can significantly improve the fusion quality of multi-source signals, providing high-quality data support for subsequent feature extraction and trend analysis.

[0024] It should be understood that traditional methods often rely on fixed time delay settings or simple signal synchronization methods. These methods cannot cope with complex factors such as signal delay changes and operating condition fluctuations during machining, leading to inaccurate wear trend identification. In contrast, this invention solves the problem of timing inconsistencies between different signal sources by dynamically calibrating the time delay deviations between signals in each machining cycle. This ensures that the health index after signal fusion is more accurate and stable, enabling early identification of subtle changes in tool wear trends, thereby providing more reliable fault warnings for CNC machine tools.

[0025] For example, taking a practical application on a CNC milling machine as an example, when the spindle speed is 600 rpm, the traditional method, by cross-correlating and aligning the spindle vibration signal and cutting force signal, yields a maximum time delay deviation of 6 milliseconds. However, the method of this invention, through periodic calibration and joint signal verification, can reduce the time delay deviation to less than 1 millisecond. Through this refined alignment process, even extremely small signal fluctuations (such as minute changes in cutting force and vibration signals) can be accurately reflected in the health index in the early stages of tool wear. Experimental results show that this invention can identify tool wear trends earlier than traditional methods, and under different machining conditions, the sensitivity and accuracy of fault warning are significantly improved, successfully reducing the false alarm rate and missed alarm rate.

[0026] Step S20: Perform adaptive filtering on the output aligned signal set to output a normalized signal set; It should be noted that "adaptive filtering" refers to dynamically selecting the optimal filtering parameters and methods to achieve noise reduction and smoothing of signals from multiple sources, addressing non-stationary noise, abrupt interference, and multi-scale disturbances introduced during processing. The adaptive filter automatically adjusts the filter type (e.g., wavelet threshold filtering, exponentially weighted moving average, variable-parameter Kalman filtering) and its parameter settings based on the statistical characteristics of the signal within the current processing cycle (e.g., mean, variance, rate of change of slope), ensuring optimal noise suppression and feature preservation for different signal channels (vibration, cutting force, current). Simultaneously, normalizing the signal set refers to converting the filtered multi-channel signals into a dimensionless form with a unified dimension and scale range. This typically employs methods such as range normalization and Z-score normalization to normalize all signal values ​​to a preset interval (e.g., [0,1] or a mean of 0 and a standard deviation of 1) to improve the consistency of subsequent feature extraction and fusion analysis.

[0027] Understandably, by introducing an adaptive filtering mechanism after signal alignment, this invention effectively suppresses the effects of sensor noise, environmental interference, and abnormal peak values, enabling the machining signal to more accurately reflect the physical contact characteristics between the tool and the workpiece. Compared to traditional fixed-parameter filtering methods, this invention can automatically adjust the filtering intensity according to different machining states. While preserving key features of the machining signal (such as amplitude changes and frequency drift caused by minor wear), it significantly improves signal quality, providing a more stable and reliable input data foundation for subsequent feature construction and health index calculation.

[0028] It should be understood that traditional machining signal processing methods often employ fixed windows and static filter coefficients, making it difficult to handle the abrupt and gradual signal characteristics in complex nonlinear machining processes. This can easily lead to the filtering out of early tool wear characteristics or the retention of a large amount of noise that masks subtle changes. The adaptive filtering mechanism proposed in this invention can dynamically adjust the processing method based on the local statistical properties of the signal. Especially in the early tool wear stage, it can accurately identify weak and gradually changing trends in the signal, effectively improving the detectability and trend judgment ability of micro-damage states. The normalization operation further enhances the versatility and comparability of multi-channel signals during fusion processing, reducing the interference of differences in amplitude scales between different channels on subsequent analysis.

[0029] For example, taking milling test data as an example, frequent machine tool reversing interference and environmental vibration noise were superimposed during the machining process. The unfiltered spindle current signal contained a large number of abrupt pulses, directly affecting the stability of subsequent health indicators. Although traditional low-pass filtering methods can suppress some high-frequency interference, they lose effective signal energy, making it difficult to capture early wear changes. The adaptive filtering mechanism used in this invention automatically uses a wavelet soft thresholding algorithm to retain high-frequency local changes in the initial detection stage, and switches to exponential smoothing to enhance trend stability in the later stable machining stage. The standard deviation fluctuation of the processed signal is controlled at 38.6% of the original signal, and the normalized vibration signal and current signal have a highly aligned dynamic trend, ultimately improving the sensitivity of the health index to minor wear conditions by about 23.4%, significantly improving detection accuracy and early warning capability.

[0030] Step S30: Obtain the processing behavior cycle The normalized signal set is arranged according to the processing behavior period. Sliding segmentation to construct the k-th processing cycle eigenvectors within ; It should be noted that the "machining behavior cycle" refers to the time window corresponding to a stable and repetitive unit (such as one milling feed, one turning trajectory, or one drilling rotation) in a complete machining process of a CNC machine tool. It can usually be automatically identified by spindle speed, tool path commands, or characteristic signal patterns. The normalized signal set includes the standardized forms of spindle vibration signals, cutting force signals, and spindle current signals obtained after time alignment and filtering. Sliding the normalized signal into segments according to the machining behavior cycle means extracting local signal segments on the time series according to the length of the machining behavior cycle with a set step size or overlap ratio, ensuring that each segment corresponds to a complete machining behavior. Subsequently, based on the statistical and frequency domain characteristics (such as mean, variance, peak factor, kurtosis, spectral centroid, etc.) of each channel signal within the segment, a feature vector of the k-th machining behavior cycle is constructed as a multi-dimensional expression of the machining state in that cycle.

[0031] Understandably, by constructing sliding segments and feature vectors based on the machining behavior cycle, this invention can transform the original continuous signal into a discrete feature sequence with a temporal structure and multi-channel fusion, laying the foundation for subsequent health assessment and trend analysis. This method significantly enhances the timeliness and accuracy of machining status identification, enabling gradual faults such as tool wear to be dynamically tracked and quantified on a periodic basis. The dimension and type of the feature vector can be flexibly configured according to different tools and process scenarios, achieving a balance between versatility and adaptability.

[0032] Step S40: Use the information entropy weight optimization mechanism to optimize the feature vector. Perform weighted fusion and output a health index under the processing behavior cycle. Based on health index Calculate the cumulative trend offset using the CUSUM algorithm. ; It should be noted that the "information entropy weight optimization mechanism" refers to calculating the information entropy value of each dimension element of the feature vector within the k-th processing cycle to measure its uncertainty and information content in the sample sequence, and then assigning corresponding weights to each feature dimension. A larger weight indicates richer effective information and stronger ability to discern health status. This mechanism avoids the uncertainty of manually setting weights, making the weighted fusion process more objective and reasonable. The resulting health index is a normalized indicator representing the overall health level of the current cycle's processing status. The "CUSUM algorithm," or cumulative sum control chart method, is used to progressively accumulate the offset between the health index sequence and the baseline value, and calculates the cumulative trend offset value based on a set micro-offset tolerance parameter to capture long-term, weak performance degradation trends. The cumulative trend offset reflects the cumulative degradation trend over multiple cycles and is a key quantitative basis for judging whether early abnormalities exist in tool wear.

[0033] Understandably, this step achieves optimal fusion of multi-dimensional features through the information entropy weighting mechanism. This not only preserves the effectiveness of key features but also suppresses the interference of redundant features, thereby improving the sensitivity of the health index to machining anomalies. The CUSUM algorithm further introduces the time accumulation effect, enabling even small single-cycle fluctuations to gradually amplify their detection signals over multiple cycles, thus improving the response capability to the early evolution of tool wear.

[0034] It should be understood that, unlike traditional health monitoring methods that often employ linear dimensionality reduction techniques such as simple averaging or principal component analysis (PCA), this invention enhances the responsiveness of the health index to feature representation through a non-parametric weighting method guided by information entropy, avoiding the risk of dilution or loss of key features. Furthermore, compared to traditional thresholding or static control chart methods, the CUSUM algorithm exhibits greater sensitivity to gradual faults, effectively capturing even when signals are weak and changes are slow in the early stages of wear evolution.

[0035] For example, in actual CNC machining scenarios, information entropy analysis of multidimensional features extracted from spindle vibration, cutting force, and spindle current signals reveals significant differences in the amount of information carried by each feature at different machining stages. By employing an information entropy weighting mechanism, features with stronger volatility or more pronounced degradation trends receive higher influence weights during the health index fusion process, making the fusion results more closely reflect actual working conditions. Furthermore, as the machining process continues, the health index exhibits a slow and continuous downward trend over multiple cycles. At this point, the cumulative trend offset calculated using the CUSUM algorithm shows a clear upward trend even before reaching the traditional discrimination threshold, revealing potential tool performance degradation signals in advance. This method demonstrates strong sensitivity to early tool wear or potential abnormal evolution processes in practical deployments, providing more sufficient response time and higher reliability for maintenance strategies.

[0036] Step S50: Preset cumulative trend offset threshold When the cumulative trend shifts Greater than or equal to the cumulative trend offset threshold If an abnormal tool wear trend is detected, an early tool wear alarm signal will be output. When the cumulative trend shifts Less than the cumulative trend offset threshold At that time, output an early tool wear alarm signal. .

[0037] It should be noted that the "cumulative trend offset threshold" in this step refers to the deviation limit between the trend offset obtained by accumulating the trend of health index changes during the machining process using the CUSUM algorithm and the statistical baseline under normal machining conditions. This threshold can be pre-set by offline modeling based on different types of tools, machining materials, and process parameters. Once the cumulative trend offset reaches or exceeds this threshold, it is considered that the tool wear has entered an abnormal evolution stage, triggering an early warning mechanism to avoid sudden tool failure or irreversible impact on workpiece quality.

[0038] Example 2: Furthermore, the fault monitoring system for CNC machine tools provided by the present invention employs a fault monitoring method for CNC machine tools as described in the above embodiments, and can solve the technical problem of fault monitoring for CNC machine tools. Compared with the prior art, the beneficial effects of the fault monitoring system for CNC machine tools provided by the present invention are the same as the beneficial effects of the fault monitoring method for CNC machine tools provided in the above embodiments, and other technical features of the fault monitoring system for CNC machine tools are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0039] Example 3: This invention provides a fault monitoring device for CNC machine tools. Please refer to... Figure 2A fault monitoring device for a CNC machine tool includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, which are executed by the at least one processor to enable the at least one processor to perform a fault monitoring method for a CNC machine tool as described in Embodiment 1 above. The fault monitoring device for a CNC machine tool in this embodiment may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital radio receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle terminals (e.g., vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. This fault monitoring device for a CNC machine tool is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this invention. A fault monitoring device for a CNC machine tool may include a processing device 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory 1002 or a program loaded from a storage device 1003 into a random access memory 1004. Random access memory 1004 also stores various programs and data required for the operation of a fault monitoring device for a CNC machine tool. Processing unit 1001, read-only memory 1002, and random access memory 1004 are interconnected via bus 1005. I / O interface 1006 is also connected to the bus. Typically, the following systems can be connected to I / O interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. Communication device 1009 allows a fault monitoring device for a CNC machine tool to communicate wirelessly or wiredly with other devices to exchange data. Although a fault monitoring device for a CNC machine tool with various systems is shown in the figure, it should be understood that it is not required to implement or possess all the systems shown. More or fewer systems may be implemented or possessed alternatively.

[0040] Example 4: This invention also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the fault monitoring method for CNC machine tools described above. The computer program product provided by this invention can solve the technical problem of fault monitoring in CNC machine tools. Compared with the prior art, the beneficial effects of the computer program product provided by this invention are the same as those of the fault monitoring method for CNC machine tools provided in the above embodiments, and will not be repeated here.

[0041] In particular, according to the embodiments disclosed in this invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from read-only memory 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this invention.

[0042] It should be understood that the various parts disclosed in this invention can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics may be combined in any suitable manner in one or more embodiments or examples.

[0043] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A fault monitoring method for CNC machine tools, characterized in that, The methods include: Step S10: Synchronously acquire the spindle vibration signal at time t during CNC machine tool operation. Cutting force signal and spindle current signal Based on the spindle vibration signal Cutting force signal and spindle current signal A cross-correlation synchronization mechanism based on joint verification of processing behavior cycles is used to perform time delay calibration, and a set of timing-consistent aligned signals is output. Step S20: Perform adaptive filtering on the output aligned signal set to output a normalized signal set; Step S30: Obtain the processing behavior cycle The normalized signal set is arranged according to the processing behavior period. Sliding segmentation to construct the k-th processing cycle eigenvectors within ; Step S40: Use the information entropy weight optimization mechanism to optimize the feature vector. Perform weighted fusion and output a health index under the processing behavior cycle. Based on health index Calculate the cumulative trend offset using the CUSUM algorithm. ; Step S50: Preset cumulative trend offset threshold When the cumulative trend shifts Greater than or equal to the cumulative trend offset threshold If an abnormal tool wear trend is detected, an early tool wear alarm signal will be output. When the cumulative trend shifts Less than the cumulative trend offset threshold At that time, output an early tool wear alarm signal. .

2. The fault monitoring method for CNC machine tools as described in claim 1, characterized in that, In step S10, based on the spindle vibration signal Cutting force signal and spindle current signal The steps of performing time delay calibration processing using a cross-correlation synchronization mechanism based on joint verification of processing behavior cycles, and outputting a time-consistent set of aligned signals, specifically include: Obtain the machine tool spindle speed n, and calculate the machining cycle at the machine tool spindle speed n. , and the collected spindle vibration signal Cutting force signal and spindle current signal According to the processing cycle Segmentation is performed to obtain signal segments. ;in, This represents the spindle vibration signal during the kth machining cycle. The cutting force signal for the kth machining cycle; This is the spindle current signal for the kth machining cycle; The spindle vibration signal during the kth machining cycle Using the reference anchor signal, an improved Hilbert autocorrelation endpoint enhancement detection algorithm is employed to extract the main cutting pulse position, thus obtaining the main cutting pulse position. Output the set of periodic anchor indexes. ; Based on periodic anchor point index set For cutting force signals and spindle current signal A local phase-matching calibration method is used to perform time delay correction, and the final output is a set of aligned signals with consistent timing.

3. The fault monitoring method for CNC machine tools as described in claim 2, characterized in that, In step S10, an improved Hilbert autocorrelation endpoint enhancement detection algorithm is used to extract the main cutting pulse position from the reference anchor point signal, thereby obtaining the main cutting pulse position. The steps specifically include: First, the reference anchor signal is subjected to envelope enhancement processing using the Hilbert transform method, and the Hilbert transform signal is output. Selecting the processing behavior period based on the Hilbert transform signal The first signal range and processing cycle of the head The second signal interval at the tail is used to calculate the autocorrelation enhancement between the first and second signal intervals. The position of the main cutting pulse is extracted from the reference anchor signal based on the autocorrelation enhancement. .

4. The fault monitoring method for CNC machine tools as described in claim 1, characterized in that, Step S20, which involves performing adaptive filtering on the output aligned signal set to output a normalized signal set, specifically includes: Step S201: Obtain the alignment signal set, which includes the spindle alignment signal, cutting force alignment signal, and spindle current alignment signal; perform adaptive filtering on the spindle alignment signal using wavelet packet decomposition; output the normalized spindle signal. Step S202: Obtain spindle speed The cutting force alignment signal is based on the spindle speed. Dynamically set the center frequency of the bandpass filter , Based on the center frequency of the bandpass filter Adaptive filtering is performed using a variable step size empirical mode bandpass filter; the output is a normalized cutting force signal. Step S203: Adaptive filtering is performed on the spindle current alignment signal using baseline drift elimination combined with sliding window mean filtering; output the normalized spindle current signal. Step S204: Merge the normalized spindle signal, the normalized cutting force signal, and the normalized spindle current signal, and normalize the normalized spindle signal, the normalized cutting force signal, and the normalized spindle current signal to a zero-mean, unit-variance form, and finally output the normalized signal set.

5. The fault monitoring method for CNC machine tools as described in claim 1, characterized in that, In step S30, the feature vector It includes vibration signal kurtosis characteristics, vibration high-frequency band energy characteristics, cutting force main frequency band energy characteristics, and current signal fluctuation coefficient characteristics. Among them, vibration signal kurtosis characteristics are used to characterize the instantaneousness and abnormality of impact signals; vibration high-frequency band energy characteristics are used to measure the intensity of high-frequency disturbances during high-speed cutting; cutting force main frequency band energy characteristics are used to reflect the cutting load intensity and changes of the tool within the main frequency range; and current signal fluctuation coefficient characteristics are used to evaluate the stability of the spindle drive system within this cycle.

6. The fault monitoring method for CNC machine tools as described in claim 1, characterized in that, In step S40, an information entropy weight optimization mechanism is used to optimize the feature vector. Perform weighted fusion and output a health index under the processing behavior cycle. The steps specifically include: first calculating the feature vectors respectively. The relative proportion of each dimension feature in all processing behavior cycles is used to set an information entropy value; then, an information redundancy index is constructed based on the set information entropy value; finally, a weighted calculation is performed based on the information redundancy index and the features of the corresponding dimension to output the health index under the processing behavior cycle. Health Index It is used to comprehensively reflect abnormal changes in multiple dimensions within the current period.

7. The fault monitoring method for CNC machine tools as described in claim 1, characterized in that, In step S40, based on the health index Calculate the cumulative trend offset using the CUSUM algorithm. The steps specifically include: pre-setting a health index benchmark value and a micro-offset tolerance parameter. The health index benchmark value is used to characterize the health level reference standard under normal processing conditions, and the micro-offset tolerance parameter is used to define the allowable fluctuation range of the health index; and within the processing cycle, based on the health index... The health index benchmark and micro-offset tolerance parameters are used to calculate the cumulative trend offset using the CUSUM algorithm. .

8. A fault monitoring system for a CNC machine tool, applied to the fault monitoring method for a CNC machine tool according to any one of claims 1 to 7, characterized in that, The fault monitoring system of the CNC machine tool includes: The multi-source signal synchronization and alignment module is used to synchronously acquire the spindle vibration signal at time t during CNC machine tool operation. Cutting force signal and spindle current signal Based on the spindle vibration signal Cutting force signal and spindle current signal A cross-correlation synchronization mechanism based on joint verification of processing behavior cycles is used to perform time delay calibration, and a set of timing-consistent aligned signals is output. The normalization filtering module is used to perform adaptive filtering on the output aligned signal set and output a normalized signal set. The processing cycle feature extraction module is used to obtain the processing behavior cycle. The normalized signal set is arranged according to the processing behavior period. Sliding segmentation to construct the k-th processing cycle eigenvectors within ; The health index construction and offset analysis module is used to optimize feature vectors using an information entropy weighting mechanism. Perform weighted fusion and output a health index under the processing behavior cycle. Based on health index Calculate the cumulative trend offset using the CUSUM algorithm. ; The trend anomaly detection and alarm module is used to preset the cumulative trend deviation threshold. When the cumulative trend shifts Greater than or equal to the cumulative trend offset threshold If an abnormal tool wear trend is detected, an early tool wear alarm signal will be output. When the cumulative trend shifts Less than the cumulative trend offset threshold At that time, output an early tool wear alarm signal. .

9. A fault monitoring device for CNC machine tools, characterized in that, The fault monitoring device for the CNC machine tool includes: a memory, a processor, and a fault monitoring program for the CNC machine tool stored in the memory and executable on the processor. When the fault monitoring program for the CNC machine tool is executed by the processor, it implements a fault monitoring method for a CNC machine tool according to any one of claims 1 to 7.

10. A computer program product, characterized in that, The computer program product includes a fault monitoring program for CNC machine tools, which, when executed by a processor, implements a fault monitoring method for CNC machine tools according to any one of claims 1 to 7.

Images