Wear monitoring and service life prediction method for TiAISiN coated cutting tool

Abstract

The invention discloses a wear monitoring and service life prediction method for a TiAlSiN coated cutter, and belongs to the technical field of cutter state monitoring. Quantitative evaluation of the coating state is realized through frequency domain characteristic analysis of the acoustic emission signal. The method comprises the following specific steps: collecting acoustic emission signals in a cutting process, carrying out band-pass filtering and background noise removal, carrying out FFT (Fast Fourier Transform) on the preprocessed signals to obtain a power spectrum, respectively calculating coating characteristic frequency band energy, matrix characteristic frequency band energy and total energy, and calculating a coating integrity index CII according to a weighting formula, and judging that the coating is in a complete, local peeling or failure state according to a CII threshold value, and further predicting the residual life of the tool. According to the method, the frequency domain separation model of the coating and the matrix is established for the first time, accurate recognition of the failure state of the coating is achieved, the correlation coefficient of CII and the coating stripping area reaches 0.93, early warning can be conducted 2-5 hours in advance, the method has the advantages of being online and real-time, clear in physical significance, wide in applicability and the like, the production cost can be effectively reduced, and the machining efficiency can be effectively improved.

Description

Technical Field

[0001] This invention relates to the field of tool condition monitoring technology, and in particular to a method for wear monitoring and life prediction of TiAlSiN coated tools. Background Technology

[0002] TiAlSiN coating is an advanced physical vapor deposition (PVD) hard coating characterized by high hardness (above 38 GPa), excellent high-temperature oxidation resistance, and a low coefficient of friction. It is widely used on the surfaces of cemented carbide cutting tools for high-speed machining of difficult-to-machine materials such as titanium alloys and high-temperature alloys. In aerospace, automotive manufacturing, and mold making, the service life and performance stability of TiAlSiN coated tools directly affect machining efficiency and product quality. However, during cutting, the coating gradually peels off due to high temperature, high stress, and friction. When a large area of ​​the coating peels off, the cemented carbide substrate is directly exposed to the cutting environment, leading to a sharp increase in tool wear rate. If the tool is not replaced in time, it will cause a decline in workpiece surface quality or even scrap. Therefore, real-time monitoring of the coating status and accurate prediction of remaining tool life are of great significance for optimizing tool management and reducing production costs.

[0003] Existing tool wear monitoring methods mainly include cutting force monitoring, vibration monitoring, and machine vision monitoring. However, these methods have significant shortcomings: First, traditional methods treat coated tools as a homogeneous material as a whole, failing to distinguish between coating peeling and substrate wear, two different failure mechanisms. This leads to an inability to accurately identify the critical point of coating failure, resulting in premature tool replacement leading to tool waste or delayed tool replacement leading to workpiece scrap. Second, single sensor signals (such as monitoring only cutting force or only vibration) are not sensitive to microscopic damage to the coating, making it difficult to provide effective early warning before complete coating failure. Third, existing methods lack physical models that correlate coating material properties and failure mechanisms, relying mainly on empirical thresholds or purely data-driven black-box models, resulting in poor adaptability and predictive accuracy under different machining conditions. To address these issues, there is an urgent need to develop a monitoring method that can accurately distinguish coating states, quantitatively assess coating integrity in real time, and reliably predict the remaining tool life. Summary of the Invention

[0004] The purpose of this invention is to provide a method for wear monitoring and life prediction of TiAlSiN coated cutting tools, so as to solve the problems existing in the prior art.

[0005] This invention provides a method for wear monitoring and life prediction of TiAlSiN coated cutting tools, characterized by the following steps: (1) During the cutting process, a high-frequency acoustic emission sensor is used to collect the acoustic emission signal when the tool is cutting, and the sampling frequency is not less than 2MHz; (2) Perform band-pass filtering on the collected acoustic emission signals, with the filtering frequency range being 20 - 500 kHz; (3) Perform fast Fourier transform (FFT) on the filtered signals to obtain the frequency-domain power spectrum P[k]; (4) Calculate the energy of the coating characteristic frequency band, the energy of the substrate characteristic frequency band, and the total energy respectively according to the power spectrum P[k], where: The coating characteristic frequency band is 150 - 400 kHz, and the calculation formula for the energy of the coating characteristic frequency band is: The substrate characteristic frequency band is 50 - 150 kHz, and the calculation formula for the energy of the substrate characteristic frequency band is: The total energy frequency band is 20 - 500 kHz, and the calculation formula for the total energy is: where k is the frequency index and Δf is the frequency resolution; (5) Calculate the coating integrity index CII according to the following formula: where α is the coating contribution weight coefficient, with a value range of 1.0 - 1.5; β is the substrate wear penalty weight coefficient, with a value range of 0.6 - 1.0; (6) Judge the coating state according to the CII value: When CII > 0.75, it is determined that the coating is in a complete state; When 0.4 < CII ≤ 0.75, it is determined that the coating is in a partial spalling state; When CII ≤ 0.4, it is determined that the coating has failed.

[0006] Through the frequency-domain feature analysis of the acoustic emission signals of the above technical solutions, an energy separation model between the coating characteristic frequency band (150 - 400 kHz) and the substrate characteristic frequency band (50 - 150 kHz) is established, and a quantitative evaluation index, the coating integrity index CII, is proposed, realizing real-time, accurate, and non-contact monitoring of the coating spalling state. This method can clearly distinguish three different stages: complete coating, partial coating spalling, and coating failure, providing clear discrimination criteria through the CII thresholds (0.75 and 0.4), overcoming the defect that traditional methods cannot distinguish the failure states of the coating and the substrate. Experimental verification shows that there is a good linear correlation between CII and the actual spalling area of the coating (correlation coefficient R With a resolution of 0.93, the accuracy rate exceeds 95%, providing early warning 2-5 hours before complete coating failure, thus offering a reliable basis for timely tool replacement decisions. This method boasts advantages such as online real-time monitoring, clear physical meaning, and wide applicability. It effectively avoids workpiece scrap losses due to excessive tool wear, while also preventing tool waste caused by premature tool replacement, optimizing tool life utilization, significantly improving machining efficiency, and reducing production costs.

[0007] Furthermore, the high-frequency acoustic emission sensor has a resonant frequency of 150kHz, an operating frequency range of 100-400kHz, and a sampling frequency of 2-5MHz.

[0008] Furthermore, the bandpass filter in step (2) uses a fourth-order Butterworth bandpass filter with a lower cutoff frequency of 20kHz and an upper cutoff frequency of 500kHz.

[0009] 4. The wear monitoring and life prediction method for TiAlSiN coated cutting tools according to claim 1, characterized in that, before step (3), a background noise removal step is further included: before the tool starts cutting, an acoustic emission signal of 10-20 seconds is collected as a background noise template in the machine tool idling state, and the noise signal is subjected to Fast Fourier Transform (FFT) to obtain the noise power spectrum. The power spectrum is obtained by performing an FFT transform on the acquired signal during the cutting process. Then follow the formula Perform spectral subtraction, where α is the over-subtraction factor, with a value ranging from 1.0 to 2.0.

[0010] Furthermore, the preferred value for the over-reduction factor α is 1.5.

[0011] Furthermore, the preferred value of the coating contribution weighting coefficient α is 1.2, and the preferred value of the substrate wear penalty weighting coefficient β is 0.8.

[0012] Furthermore, after calculating the original CII value in step (5), the process also includes normalization and smoothing steps: first, according to the formula... Normalization was performed, where The original CII value; then according to the formula Exponential moving average smoothing is applied, where n is the sampling number and λ is the smoothing coefficient, ranging from 0.6 to 0.9.

[0013] Furthermore, the preferred value for the smoothing coefficient λ is 0.7.

[0014] Furthermore, the method also includes a tool remaining life prediction step: recording CII values ​​at fixed time intervals to obtain the CII evolution curve CII(t) over time, and establishing a model of CII changing over time by fitting historical data, wherein the model includes a linear model. or exponential model ,in denoted as the initial coating integrity index, k as the linear decay coefficient, λ as the exponential decay coefficient, and t as time. The remaining tool life is defined as the time required for the CII to decrease to the failure threshold of 0.4, as predicted by the fitted model.

[0015] Furthermore, the acoustic emission sensor is mounted on the tool holder or workpiece fixture at a distance of 50-150mm from the cutting area. The sensor is connected to the data acquisition system via a preamplifier, with the preamplifier gain set to 40-60dB. The beneficial effects of this invention are: 1. This invention, through spectral analysis of acoustic emission signals, discovered and verified that the acoustic emission signals generated by TiAlSiN coating peeling are mainly concentrated in the 150-400kHz frequency band, while the acoustic emission signals generated by cemented carbide substrate wear are mainly concentrated in the 50-150kHz frequency band. Based on this physical mechanism, a Coating Integrity Index (CII) was established, achieving quantitative characterization of the coating state. Compared to traditional methods that treat the coated tool as a whole, this invention can accurately identify the occurrence and development process of coating peeling and accurately determine the critical point of coating failure. Experimental results show that when the CII drops to 0.4, the coating peeling area reaches 60-80%, at which point the tool should be replaced immediately; while traditional methods often wait until the flank wear VB reaches 0.4-0.5mm before identifying tool failure, by which time the workpiece surface quality has already been severely deteriorated. This invention can provide early warning of large-area coating peeling 2-5 hours in advance, gaining sufficient time for tool replacement decisions and avoiding workpiece scrapping caused by sudden tool failure, resulting in significant economic benefits.

[0016] 2. This invention utilizes the high sensitivity of acoustic emission technology to microscopic damage in materials, enabling the detection of early failure signs such as microcrack initiation and interface separation in coatings. These microscopic damages occur before the macroscopic wear (e.g., flank wear VB) increases significantly, making them difficult to detect using traditional cutting force or vibration monitoring methods. Furthermore, the CII index established in this invention has a clear physical meaning: a higher CII value indicates a larger proportion of energy in the coating's characteristic frequency bands and a more intact coating; a lower CII value indicates a larger proportion of energy in the matrix's characteristic frequency bands and more severe coating peeling. This physics-based index is more reliable than purely empirical thresholds or black-box models. The results obtained by different operators using this method show good consistency and repeatability, with a coefficient of variation of less than 8%. In addition, the weighting coefficients α and β (preferred values ​​of 1.2 and 0.8, respectively) introduced into the CII calculation formula are optimized based on a large amount of experimental data. These two coefficients reflect the relative importance of coating contribution and matrix penalty, making the CII index highly sensitive to changes in coating state.

[0017] 3. This invention belongs to a non-contact, online, real-time monitoring method. The acoustic emission sensor is installed on the tool holder or workpiece fixture, which does not affect the normal machining process and does not require machine downtime for detection. The monitoring frequency can be flexibly adjusted as needed (from once per second to once per minute). The entire monitoring and calculation process can be completed within 2 seconds, meeting real-time requirements. The system hardware configuration is simple; the total cost of the acoustic emission sensor, preamplifier, and data acquisition card is typically between 50,000 and 100,000 yuan, which is more economical than high-precision force gauges or machine vision systems. This method is applicable to various types of TiAlSiN coated cutting tools (including milling cutters, turning tools, drills, etc.) and can also be applied to other PVD coatings (such as TiN, TiAlN, CrN, etc.) by adjusting the characteristic frequency band range and weighting coefficients. Experimental verification shows that the method can work reliably under different cutting speeds (60-120m / min), different feed rates (0.05-0.15mm / z), and different workpiece materials (titanium alloy, high-temperature alloy, hardened steel). The overall law of CII evolution remains consistent, and it has good robustness and adaptability to working conditions. It can be promoted and applied in aerospace, automotive, mold and other industries. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in this 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 for this 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 schematic diagram of the method flow of the present invention; Figure 2This is a schematic diagram of the system structure of the present invention; Figure 3 This is a schematic diagram of the spectrum analysis principle of the present invention; Figure 4 The graph shows the experimental results of this invention. Detailed Implementation

[0020] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. It should also be noted that, to make the embodiments more comprehensive, the following embodiments are the best and preferred embodiments, and those skilled in the art can use other alternative methods to implement some well-known technologies; moreover, the accompanying drawings are only for more specific description of the embodiments and are not intended to specifically limit the present invention.

[0021] It should be noted that the use of terms such as "an embodiment," "an embodiment," "an exemplary embodiment," and "some embodiments" in the specification indicates that the described embodiment may include a specific feature, structure, or characteristic, but not every embodiment necessarily includes that specific feature, structure, or characteristic. Furthermore, when a specific feature, structure, or characteristic is described in connection with an embodiment, implementing such a feature, structure, or characteristic in conjunction with other embodiments (whether explicitly described or not) should be within the knowledge of those skilled in the art.

[0022] Generally, terms can be understood at least partly from their use in context. For example, depending at least partly on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in a singular sense, or a combination of features, structures, or characteristics in a plural sense. Additionally, the term "based on" can be understood not necessarily to convey an exclusive set of factors, but rather, alternatively, depending at least partly on the context, to allow for the presence of other factors that are not necessarily explicitly described.

[0023] See Figures 1 to 4 As shown This invention provides a method for wear monitoring and life prediction of TiAlSiN coated cutting tools. This method achieves real-time quantitative assessment of coating integrity through frequency domain characteristic analysis of acoustic emission signals, thereby predicting the remaining tool life. The technical solution of this invention is based on the following findings: the acoustic emission signals generated when the TiAlSiN coating peels off during cutting are mainly concentrated in the 150-400 kHz frequency band, while the acoustic emission signals generated when the cemented carbide substrate undergoes plastic deformation and wear are mainly concentrated in the 50-150 kHz frequency band. By analyzing the energy distribution ratio of these two frequency bands, the integrity status of the coating can be accurately determined. The invention will be described in detail below with reference to specific embodiments.

[0024] The method of this invention includes the following steps: First, during the cutting process, a high-frequency acoustic emission sensor is used to collect the acoustic emission signals generated during cutting in real time. The acoustic emission sensor is preferably a piezoelectric ceramic sensor with a resonant frequency of 150 kHz and an operating frequency range of 100-400 kHz. This sensor is mounted on the tool holder or workpiece fixture, and the distance between it and the cutting area is controlled within the range of 50-150 mm to ensure signal strength. The sensor is connected to the data acquisition system via a preamplifier, with the preamplifier gain set to 40-60 dB. The sampling frequency of the data acquisition system is set to no less than 2 MHz, preferably 2-5 MHz, to meet the requirements of the Nyquist sampling theorem and ensure accurate capture of all frequency components within 500 kHz. The acquisition time window is set to 1 second, i.e., 2 × 10⁻⁶ signals are acquired per acquisition. 6 Each data point can be collected at intervals of 1-10 seconds, depending on actual needs. It is recommended to shorten the collection interval during periods of increased tool wear to improve monitoring sensitivity.

[0025] The acquired acoustic emission signal undergoes preprocessing. The first step in preprocessing is bandpass filtering, using a fourth-order Butterworth bandpass filter with a lower cutoff frequency of 20 kHz and an upper cutoff frequency of 500 kHz. The purpose of filtering is to remove low-frequency mechanical vibration interference (typically below 20 kHz) and high-frequency electronic noise (typically above 500 kHz). In practice, a digital filter can be used, with the filter's transfer function being... ,in rad / s is the lower limit angular frequency. rad / s is the upper limit angular frequency, and s is the Laplace variable. After bandpass filtering, the signal retains the frequency components directly related to the tool wear mechanism.

[0026] The second step in preprocessing is background noise removal. In actual machining environments, machine tool spindle rotation, coolant flow, and workpiece vibration all generate background noise, which can affect the identification of coating peeling signals. This invention uses spectral subtraction for noise removal. Specifically, before the tool begins cutting, a 10-20 second acoustic emission signal is collected while the machine tool is idling as a background noise template. This noise signal is then subjected to a Fast Fourier Transform (FFT) to obtain the noise power spectrum. In subsequent cutting processes, the power spectrum is obtained by performing FFT transformation on each acquired signal. Then perform spectrum subtraction: ,in This is an over-reduction factor, with a value ranging from 1.0 to 2.0, preferably 1.5. Its function is to moderately over-reduce noise to ensure sufficient removal. The function ensures that the power spectrum after subtraction will not have negative values. After spectral subtraction, the obtained... This is the power spectrum of the denoised signal.

[0027] The preprocessed acoustic emission signal is subjected to a Fast Fourier Transform (FFT) to obtain its frequency domain characteristics. The FFT transform converts the time domain signal... Convert to frequency domain signal The conversion formula is:

[0028] in Let be the frequency (Hz), t be the time (s), j be the imaginary unit, and e be the base of the natural logarithm. In practical calculations, the Discrete Fourier Transform (DFT) is used for discrete signals containing N sampling points. ( Its DFT expression is: ,in This is the frequency index, and the corresponding actual frequency is... , This refers to the sampling frequency. For example, when... MHz At that time, the frequency resolution is 1Hz. To improve computational efficiency, the Fast Fourier Transform (FFT) algorithm is used in practical applications, reducing the computational complexity from... Reduce to The complex spectrum obtained by FFT calculation It contains amplitude and phase information, and its amplitude spectrum is as follows: ,in and These represent the real and imaginary parts, respectively. The power spectral density (PSD) is the square of the amplitude spectrum, i.e. The power spectrum reflects the distribution of signal energy at various frequencies.

[0029] Based on the power spectrum obtained from FFT, the energy of the coating's characteristic frequency band, the substrate's characteristic frequency band, and the total energy were calculated. The coating's characteristic frequency band was defined as 150-400 kHz. This band corresponds to the frequency range of elastic waves released when the TiAlSiN coating material undergoes failure behaviors such as spalling and microcrack propagation during machining. During coating spalling, high-frequency transient stress waves are generated due to the interface separation between the coating and the substrate, and the frequencies of these stress waves are mainly concentrated in this frequency band. The coating's characteristic frequency band energy... The calculation formula is: , in The frequency index corresponding to 150kHz is calculated as follows: , The frequency index corresponding to 400 kHz is calculated as follows: , Indicates rounding down. For frequency resolution, This represents the power value at the k-th frequency point. The matrix characteristic frequency band is defined as 50-150kHz. This band corresponds to the frequency range of acoustic emission signals generated when the cemented carbide matrix undergoes plastic deformation, abrasive wear, etc., during cutting. The matrix material has a relatively large grain size, typically 1-5μm, and the acoustic emission signals generated by its deformation and fracture are at lower frequencies. The energy of the matrix characteristic frequency band... The calculation formula is: , in Other symbols have the same meaning as above. Total energy Defined as the energy across the entire frequency band of 20-500 kHz, the calculation formula is:

[0030] ,in , It is important to note that when calculating the energy of each frequency band, if the power spectrum has undergone spectral subtraction noise removal, the denoised power spectrum should be used. Perform the calculation.

[0031] Based on the extracted characteristic frequency band energy, the coating integrity index (CII) is calculated using the following formula: ,in The coating contribution weighting coefficient represents the degree of contribution of the coating's characteristic frequency band energy proportion to CII. The physical meaning of this coefficient is: when the coating is intact, the energy proportion of the coating's characteristic frequency band should be relatively high, and therefore a larger positive weight is given. The value range is 1.0-1.5, with a preferred value of 1.2. The substrate wear penalty weighting coefficient represents the degree of penalty imposed on the proportion of energy in the substrate's characteristic frequency band by the CII. The physical meaning of this coefficient is: when the substrate is exposed and wear occurs, the proportion of energy in the substrate's characteristic frequency band will increase. At this time, the CII value should be reduced to reflect the coating failure state, so a negative weight is given. The value range is 0.6-1.0, with a preferred value of 0.8. This indicates the proportion of energy in the coating's characteristic frequency band to the total energy. This proportion is higher when the coating is intact (usually greater than 0.6) and decreases when the coating peels off. The CII (Compatibility Index) represents the proportion of energy in the characteristic frequency band of the substrate to the total energy. This proportion is lower when the coating is intact (typically less than 0.2%), and increases when the coating peels off and the substrate is exposed. Through this weighted subtraction method, the CII can comprehensively reflect the integrity status of the coating.

[0032] To improve the stability and anti-interference capability of CII, normalization and smoothing are required after calculating the original CII value. Normalization restricts the CII value to the range of 0-1, and the specific method is as follows: ,in The original CII value is obtained by directly calculating according to the above formula. The function ensures that CII does not exceed 1.0. The function ensures that the CII is not lower than 0. Normalization aims to give the CII a clear physical meaning: CII=1 indicates the coating is completely intact, and CII=0 indicates the coating is completely failed. Smoothing is performed using the exponential moving average (EMA) method, calculated as follows: ,in For sampling sequence number, The smoothed CII value of the nth sample. The smoothed CII value from the previous sample. The normalized CII value of the nth sample. This is a smoothing coefficient, ranging from 0.6 to 0.9, with a preferred value of 0.7. When... When the size is large, the smoothness is high but the response speed is slow; when When the value is small, the response speed is fast but it is easily affected by instantaneous fluctuations. The advantages of the exponential moving average method are its simple calculation and small memory footprint, making it suitable for real-time online applications. The equivalent length of the smoothing time window is approximately... ,in For example, when the sampling interval is... =1 second When CII = 0.7, the equivalent time window is approximately 3.3 seconds, which means that the CII value mainly reflects the coating state within the most recent 3-4 seconds.

[0033] The integrity status of the coating is determined based on the smoothed CII value. The criterion is: when... When the coating is deemed intact, there is no obvious peeling on the coating surface, the coating provides good protection for the substrate, and the tool can continue to be used normally; when When the coating is in a state of localized peeling, localized peeling areas begin to appear on the coating surface, with the peeling area accounting for approximately 20%-60% of the total coating area. The substrate is exposed, and the tool is in a stage of accelerated wear. Close monitoring and preparation for tool replacement are necessary. When the coating has peeled off significantly or completely (over 60%), the substrate is exposed and directly participates in the cutting process, causing a sharp increase in tool wear. The machine should be stopped immediately and the tool replaced to avoid workpiece quality degradation or tool breakage. These three thresholds (0.75 and 0.4) are empirical values ​​determined through extensive experimental statistics. In practical applications, they can be fine-tuned based on specific tool materials, coating thickness, and machining conditions, typically within ±0.05.

[0034] Based on the assessment of the coating condition, the remaining tool life can be further predicted. The specific method is as follows: during the cutting process, the CII value is recorded at fixed time intervals, such as every 1-5 minutes, to obtain the CII evolution curve over time. The curve typically shows a downward trend. The rate of coating peeling can be determined based on the rate of change of CII, which is defined as: ,in For time intervals. When The absolute value is small (e.g. When the rate is ( / min), it indicates that the coating peeling rate is slow and the tool is in a stable wear stage; when The absolute value is large (e.g.) When the rate reaches ( / min), it indicates that the coating peeling rate is accelerating, and the tool is entering the rapid failure stage. By fitting historical CII data, an empirical model of CII changing over time can be established. Commonly used models include linear models. Applicable to steady wear stage and exponential model Applicable to the accelerated failure stage, where The initial coating integrity index is typically close to 1.0, and k is the linear decay coefficient. These parameters, representing the exponential decay coefficients, can be obtained by fitting historical data using the least squares method. Based on the fitted model, the time required for the CII to decrease to the failure threshold of 0.4 is predicted; this time is the tool's remaining useful life (RUL). For example, if a linear model is used, then... ,in This is the current CII value; if an exponential model is used, then... To improve prediction accuracy, other monitoring signals, such as cutting force and temperature, can be combined for multivariate joint prediction. However, CII, as a direct indicator reflecting the coating condition, should play a dominant role in the prediction model.

[0035] In practical applications, the method of this invention can be integrated into the monitoring system of CNC machine tools to achieve automatic monitoring and alarm of tool status. Specifically, the acoustic emission sensor is connected to an industrial computer or embedded controller via a data acquisition card. The data acquisition card has high-speed AD conversion capabilities, a resolution of at least 12 bits, and a sampling rate of at least 2 MHz. Monitoring software runs on the industrial computer or controller, which calculates the CII value in real time according to the method of this invention and displays the current value and historical curves of the CII on the human-machine interface. When the CII drops to a preset warning threshold, such as 0.6, the system issues a warning signal to remind the operator to pay attention to the tool status; when the CII drops to the failure threshold of 0.4, the system issues a serious alarm signal, recommending immediate machine shutdown and tool replacement. The monitoring software can also record CII data, raw acoustic emission signals, cutting parameters, and other information into a database for subsequent data analysis and model optimization. Through long-term data accumulation, a database of CII evolution patterns under different workpiece materials and cutting parameters can be established, providing decision support for tool life management.

[0036] The technical advantages of this invention are as follows: First, by analyzing the spectrum of acoustic emission signals, the coating state and the substrate state can be effectively distinguished, overcoming the shortcomings of traditional methods that treat the coated tool as a whole and cannot identify the coating failure stage. Second, the CII index has clear physical meaning and quantitative characterization capability. Compared with qualitative methods that rely on experience, this invention provides an objective and repeatable quantitative evaluation method. Third, this method is a non-contact, online, real-time monitoring method that does not affect the normal processing, does not require machine downtime for testing, and the monitoring frequency can be flexibly adjusted as needed. Fourth, acoustic emission signals are very sensitive to microscopic damage to the coating and can provide early warning before the coating completely fails, giving time for timely tool replacement. Fifth, the hardware cost of this method is relatively low, with a low total cost for the acoustic emission sensor and data acquisition equipment, making it suitable for promotion and application in small and medium-sized manufacturing enterprises.

[0037] It should be noted that although this invention focuses on TiAlSiN coatings, the basic principles of this method are also applicable to other types of PVD-coated cutting tools, such as TiN, TiAlN, and CrN coatings. The only difference is that the range and weighting coefficients of the characteristic frequency bands need to be adjusted according to the characteristics of different coating materials. For example, for TiN coatings, the coating characteristic frequency band can be adjusted to 120-300 kHz, and the substrate characteristic frequency band to 40-120 kHz; for TiAlN coatings, the coating characteristic frequency band can be adjusted to 140-350 kHz. These parameters can be adjusted through prior experimental calibration. The calibration method involves conducting cutting experiments using brand-new coated cutting tools, simultaneously acquiring acoustic emission signals, and periodically stopping the machine to observe the coating state (using a microscope or SEM). The acoustic emission spectrum characteristics under different coating states are analyzed to determine the frequency range that best distinguishes between coating peeling and substrate wear, as well as the optimal weighting coefficients α and β.

[0038] Furthermore, to further improve the reliability of monitoring, it is recommended to adopt a multi-sensor arrangement when implementing the method of this invention, that is, to install an acoustic emission sensor on both the tool holder and the workpiece fixture, forming a dual-channel monitoring system. The signals collected by the two sensors can be mutually verified. When the CII values ​​calculated by the two channels are consistent, it indicates that the monitoring results are reliable; when there is a large deviation between the CII values ​​of the two channels, such as a difference exceeding 0.1, it may be that one of the sensors is faulty or the signal transmission path is interfered with. In this case, the installation status and signal quality of the sensors should be checked. For the dual-channel system, the final CII can be taken as a weighted average of the CII values ​​of the two channels, with the weights determined according to the signal strength and signal-to-noise ratio of each channel.

[0039] Example 1 To illustrate the specific implementation process and effects of the present invention, a detailed embodiment is given below. This embodiment takes milling titanium alloy Ti-6Al-4V with a TiAlSiN coated cemented carbide end mill as an example to demonstrate the complete process from acoustic emission signal acquisition to coating condition judgment.

[0040] The experimental equipment and conditions are as follows: The machine tool is a three-axis CNC vertical milling machine with a maximum spindle speed of 12000 rpm and a spindle power of 7.5 kW; the cutting tool is a φ10 mm four-flute carbide end mill with an ultra-fine grain carbide substrate. The tool surface is coated with a TiAlSiN coating by physical vapor deposition (PVD) with a coating thickness of 3.5 μm and a coating hardness of 38 GPa; the workpiece material is a Ti-6Al-4V titanium alloy forging with a hardness of HRC 35; the cutting parameters are set as follows: spindle speed 3000 rpm, corresponding to a cutting speed of... m / min, feed per tooth mm / z, corresponding to feed per revolution mm / r, axial cutting depth mm, radial cutting width mm, using reverse milling, dry cutting without coolant.

[0041] The monitoring system is configured as follows: The acoustic emission sensor is a PAC R15α, with a resonant frequency of 150 kHz, an operating frequency range of 50-400 kHz, a sensitivity of -65 dB, and a reference value of 1 V / μbar. The sensor is mounted on the side of the workpiece fixture via a magnetic base, approximately 80 mm away from the cutting area. Silicon grease is applied between the sensor and the workpiece to improve acoustic coupling efficiency. The preamplifier is a PAC 2 / 4 / 6, with a gain of 40 dB and a bandwidth of 20 kHz-1.2 MHz. The data acquisition card is an NI USB-6366, with 8 analog input channels, a 16-bit resolution, and a maximum sampling rate of 2 MS / s (2 MHz). The acquisition software is developed using NI LabVIEW, and the data storage format is TDMS. The data processing and analysis software is written in MATLAB R2021b and runs on an industrial computer.

[0042] The experimental procedure is as follows: First, the acoustic emission monitoring system was calibrated and tested. With the machine tool spindle idling at 3000 rpm and no cutting being performed, background noise signals were collected for 30 seconds. Analysis of its spectral characteristics revealed that the main noise components were concentrated in the 0-20 kHz and above 600 kHz frequency bands. This verified the rationality of the bandpass filter parameters (20-500 kHz) set in this invention. Then, a new cutting tool was installed. At this point, the tool coating was intact. The first cut was performed using the brand-new tool, with a cutting length of 100 mm and a cutting time of approximately 19 seconds. Acoustic emission signals were continuously collected during the cutting process. The sampling frequency was set to 2 MHz, with one sample collected every 1 second, each lasting 1 second. Therefore, each sample yielded 2 × 10⁻⁶ signals. 6 Data points. After the first cut, the machine was stopped to check the tool condition. Using a tool microscope at 50x magnification, the flank face was observed, and the wear amount VB on the flank face was found to be 0.02 mm. The coating surface was smooth and intact, with no obvious peeling marks. The second and third cuts were then performed, each with a cut length of 100 mm. The machine was stopped between each cut to check and record the tool wear condition.

[0043] During the 10th cut, the cumulative cutting length was 1000 mm and the cumulative cutting time was approximately 190 seconds. Inspection after stopping the machine revealed that the flank wear (VB) had increased to 0.12 mm, and the coating began to peel off locally at the tool tip and flank edge, presenting as irregular patches, covering approximately 15% of the total flank area. During the 20th cut, the cumulative cutting length was 2000 mm and the cumulative cutting time was approximately 380 seconds. The flank wear (VB) was 0.25 mm, and the coating peeling area significantly expanded, covering approximately 45% of the total flank area. The peeled area revealed a dark gray substrate color, contrasting sharply with the surrounding golden coating. During the 28th cut, the cumulative cutting length was 2800 mm and the cumulative cutting time was approximately 532 seconds. The flank wear (VB) was 0.38 mm, and the coating had almost completely peeled off, with a peeling area exceeding 80%, exposing a large area of ​​the substrate. The tool wear rate accelerated significantly. When cutting continued for the 32nd time, the wear on the back face VB reached 0.52 mm. At the same time, the surface roughness Ra of the workpiece deteriorated from the initial 0.8 μm to 2.5 μm, which exceeded the machining requirement Ra<1.6 μm. Therefore, the tool was determined to have reached a failure state and the experiment was stopped.

[0044] The collected acoustic emission signals were processed and analyzed. Taking the 15th cut, with a cumulative cutting length of 1500 mm, as an example, the data processing procedure was illustrated. This cut lasted 19 seconds, and a total of 19 sets of acoustic emission data were collected. Each set of data contained 2×10⁻⁶ data points. 6 10 sampling points. Data from the 10th second was analyzed in detail. First, the original signal was bandpass filtered with the following parameters: lower cutoff frequency 20 kHz, upper cutoff frequency 500 kHz, filter order 4, and Butterworth bandpass filter. Before filtering, the signal amplitude ranged from -2.5 V to +2.5 V, containing a large amount of low-frequency vibration components, mainly the spindle rotation frequency of 50 Hz and its harmonics, as well as high-frequency electronic noise. After bandpass filtering, the low-frequency and high-frequency components were effectively suppressed, and the filtered signal amplitude range was reduced to -1.8 V to +1.8 V, resulting in a clearer signal waveform.

[0045] Then, background noise removal is performed. The noise signal collected during the idling state is used as a noise template. FFT transformation is performed on both the filtered signal and the noise template. The FFT is implemented using the fft function in MATLAB, and the FFT length is 2×10. 6 The obtained spectral resolution is Hz. The amplitude spectrum is calculated from the FFT result of the filtered signal using the following formula:

[0046] Then calculate the power spectrum. The power spectrum of the noise template was calculated using the same method. Perform spectral subtraction, over-subtracting factor. Taking 1.5, we obtain the power spectrum after denoising. Observing the power spectrum after denoising, it was found that there are obvious energy peaks in the 150-400 kHz frequency band, which correspond to the acoustic emission signals generated by coating peeling; there is also an energy distribution in the 50-150 kHz frequency band, but the intensity is lower than that of the coating characteristic frequency band.

[0047] Next, the energy of each frequency band is calculated. The frequency index range corresponding to the coating's characteristic frequency band of 150-400 kHz is as follows: , Within this range, the power spectrum is summed. In the specific calculation, (Frequency resolution) Hz), the calculation result is V 2 The frequency index range corresponding to the matrix characteristic frequency band of 50-150 kHz is as follows: , Calculations yielded V 2 The total energy across the entire frequency band of 20-500 kHz is: , Calculations yielded V 2 ·s.

[0048] Based on the energy calculation results, CII is calculated according to the formula of this invention, with the weighting coefficients taken as follows: , ,but:

[0049]

[0050] The original CII values ​​were normalized because... It is already in the range of 0-1, therefore Then, exponential moving average smoothing is performed. Assuming the previous sample had a smoothed CII value of 9 seconds... Smoothing coefficient Then the smoothed CII value at the 10th second is: According to the discrimination criteria, Therefore, it was determined that the coating was in a state of partial peeling at this time. Actual shutdown inspection also confirmed this judgment. During the 15th cut, the peeling area of ​​the coating was about 30%, which is consistent with the state reflected by the CII value.

[0051] The evolution of CII throughout the experiment was analyzed, and curves showing the change of CII with cutting time or cutting length were plotted. Experimental data show that in the first 5 cuts (0-95 seconds, 0-500 mm), the CII remained between 0.85 and 0.92, with an average of 0.88 and a standard deviation of 0.03, indicating that the coating was intact. In the 6th-15th cuts (95-285 seconds, 500-1500 mm), the CII gradually decreased from 0.85 to 0.52, with an average decrease rate of approximately 0.0017 / second, corresponding to the beginning of localized peeling and gradual expansion of the coating. In the 16th-25th cuts (285-475 seconds, 1500-2500 mm), the CII gradually decreased. In the mm stage, CII continued to decrease from 0.52 to 0.35, with the average rate of decrease accelerating to 0.0009 / second. Note that this is an absolute value; the actual rate of decrease is negative. CII first dropped below 0.4 during the 22nd cut, reaching 0.38. At this point, a stop inspection revealed that the coating had peeled off 65%, consistent with the failure criteria. In the 26th-32nd cuts, 475-608 seconds, 2500-3200 mm stage, CII rapidly decreased to between 0.15 and 0.25, indicating that the coating had almost completely failed, and a large area of ​​the substrate was exposed.

[0052] Correlation analysis was performed between CII data and actual coating peeling area. A scatter plot was plotted with the percentage of coating peeling area on the x-axis and the CII value on the y-axis, and linear fitting was performed. The fitting results show that there is a good linear correlation between CII and coating peeling area, with a correlation coefficient of [value missing]. The fitted straight line equation is ,in The value represents the percentage of coating peeling area, ranging from 0% to 100%. The fitted equation indicates that when the coating is completely intact, When the coating peels off, the CII is approximately 0.92; when 50% of the coating has peeled off, the CII is approximately 0.48; when the coating is completely peeled off, At that time, the CII was approximately 0.05. This linear relationship verifies the effectiveness of the method of the present invention, proving that the CII can quantitatively reflect the integrity of the coating.

[0053] To verify the adaptability of the method of the present invention to different cutting conditions, comparative experiments were also conducted. In the comparative experiments, other conditions were kept constant, only the cutting speed was reduced from 94.2 m / min to 60 m / min, the spindle speed was increased to 1910 rpm, or the feed per tooth was increased from 0.08 mm / z to 0.12 mm / z. The experimental results show that although the tool wear rate and coating peeling rate differ under different cutting conditions, the overall evolution law of CII remains consistent, that is, the CII remains above 0.75 in the coating integrity stage, between 0.4 and 0.75 in the coating partial peeling stage, and below 0.4 in the coating failure stage. This indicates that the method of the present invention has good robustness and is suitable for different machining conditions.

[0054] In addition, repeatability experiments were conducted using five new cutting tools of the same specifications under the same cutting conditions to compare the CII evolution curves of different tools. The experimental results showed that the CII evolution curves of the five tools were basically consistent, and the coefficient of variation of the CII values ​​was less than 8% under the same cutting length, indicating that the method of this invention has good repeatability and consistency. The minor differences between individual tools mainly stemmed from small variations in coating thickness (tolerance approximately ±0.5 μm) and manufacturing errors in the tool geometry.

[0055] As demonstrated in this embodiment, the coating integrity assessment method based on acoustic emission spectrum analysis provided by this invention can effectively distinguish different wear states of the coating. A good correlation exists between the CII index and the coating peeling area, and the discrimination criteria are accurate and reliable. This method enables online real-time monitoring of tool condition, providing a scientific basis for tool life management and timely tool replacement decisions, and has significant practical value and promising application prospects. Applying this method in actual production can avoid workpiece scrapping due to excessive tool wear, and also avoid tool waste caused by premature tool replacement, thereby optimizing tool life utilization, reducing production costs, and improving processing efficiency and product quality.

[0056] This invention encompasses any substitutions, modifications, equivalent methods, and solutions made within the spirit and scope of this invention. To provide the public with a thorough understanding of this invention, specific details are described in detail in the following preferred embodiments; however, those skilled in the art will fully understand the invention even without these details. Furthermore, to avoid unnecessary misunderstanding of the essence of this invention, well-known methods, processes, procedures, components, and circuits are not described in detail.

[0057] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for wear monitoring and life prediction of TiAlSiN coated cutting tools, characterized in that, It includes the following steps: (1) During the cutting process, a high-frequency acoustic emission sensor is used to collect the acoustic emission signals during tool cutting, and the sampling frequency is not less than 2 MHz; (2) The collected acoustic emission signals are subjected to band-pass filtering, and the filtering frequency range is 20 - 500 kHz; (3) The filtered signals are subjected to fast Fourier transform FFT to obtain the frequency-domain power spectrum P[k]; (4) Calculate the energy of the coating characteristic frequency bands based on the power spectrum P[k]. Matrix characteristic frequency band energy and total energy , Where: The coating characteristic frequency band is 150 - 400 kHz, and the calculation formula for the energy of the coating characteristic frequency band is: The matrix's characteristic frequency band is 50-150kHz, and the formula for calculating the energy of the matrix's characteristic frequency band is: The total energy frequency band is 20-500kHz, and the formula for calculating the total energy is: Where k is the frequency index and Δf is the frequency resolution; (5) Calculate the coating integrity index CII according to the following formula: Where α is the coating contribution weighting coefficient, with a value ranging from 1.0 to 1.5; β is the substrate wear penalty weighting coefficient, with a value ranging from 0.6 to 1.0; (6) Judge the coating state according to the CII value: When CII > 0.75, it is determined that the coating is in a complete state; When 0.4 < CII ≤ 0.75, it is determined that the coating is in a state of partial spalling; When CII ≤ 0.4, it is determined that the coating has failed.

2. The method for wear monitoring and life prediction of TiAlSiN coated cutting tools according to claim 1, characterized in that, The resonance frequency of the high-frequency acoustic emission sensor is 150 kHz, the working frequency range is 100 - 400 kHz, and the sampling frequency is 2 - 5 MHz.

3. The method for wear monitoring and life prediction of TiAlSiN coated cutting tools according to claim 1, characterized in that, The band-pass filtering in step (2) uses a 4th-order Butterworth band-pass filter, the lower cut-off frequency is 20 kHz, and the upper cut-off frequency is 500 kHz.

4. The method for wear monitoring and life prediction of TiAlSiN coated cutting tools according to claim 1, characterized in that, Before step (3), a background noise removal step is also included: before the tool starts cutting, the acoustic emission signal for 10-20 seconds is collected as a background noise template while the machine tool is idling, and the noise signal is subjected to Fast Fourier Transform (FFT) to obtain the noise power spectrum. The power spectrum is obtained by performing an FFT transform on the acquired signal during the cutting process. Then follow the formula Perform spectral subtraction, where α is the over-subtraction factor, with a value ranging from 1.0 to 2.

0.

5. The method for wear monitoring and life prediction of TiAlSiN coated cutting tools according to claim 4, characterized in that, The preferred value of the over-subtraction factor α is 1.

5.

6. The method for wear monitoring and life prediction of TiAlSiN coated cutting tools according to claim 1, characterized in that, The preferred value of the coating contribution weight coefficient α is 1.2, and the preferred value of the matrix wear penalty weight coefficient β is 0.

8.

7. The method for wear monitoring and life prediction of TiAlSiN coated cutting tools according to claim 1, characterized in that, After the original CII value is calculated in step (5), the normalization and smoothing steps are also included: first, according to the formula Normalization was performed, where The original CII value; then according to the formula Exponential moving average smoothing is applied, where n is the sampling number and λ is the smoothing coefficient, ranging from 0.6 to 0.

9.

8. The method for wear monitoring and life prediction of TiAlSiN coated cutting tools according to claim 7, characterized in that, The preferred value of the smoothing coefficient λ is 0.

7.

9. The method for wear monitoring and life prediction of TiAlSiN coated cutting tools according to claim 1, characterized in that, It also includes a tool remaining life prediction step: recording CII values ​​at fixed time intervals to obtain the CII evolution curve CII(t) over time, and establishing a model of CII changing over time by fitting historical data, wherein the model includes a linear model. or exponential model ,in denoted as the initial coating integrity index, k as the linear decay coefficient, λ as the exponential decay coefficient, and t as time. The remaining tool life is defined as the time required for the CII to decrease to the failure threshold of 0.4, as predicted by the fitted model.

10. The method for wear monitoring and life prediction of TiAlSiN coated cutting tools according to claim 1, characterized in that, The acoustic emission sensor is installed on the tool holder or workpiece fixture, the distance from the cutting area is 50 - 150 mm, and the sensor is connected to the data acquisition system through a preamplifier, and the preamplifier gain is set to 40 - 60 dB.

Images