A method for identifying the chipping state of an ultrafine grain cemented carbide circuit board drill bit

Abstract

The application provides a microcrystalline hard alloy circuit board drill bit chipping state recognition method, comprising: extracting the amplitude at the main shaft rotation fundamental frequency and the amplitude at each harmonic from the current spectrum record, representing the cutting load strength of the corresponding blade edge according to the position of each harmonic to obtain the harmonic amplitude record of each blade cutting load; evaluating the uniformity of each blade cutting resistance by a preset resistance threshold, identifying the existence of multi-blade micro-chipping of the multi-blade drill bit when the uniformity exceeds the resistance threshold to obtain the first chipping recognition conclusion and the confidence; comprehensively evaluating the confidence of the first chipping recognition conclusion and the confidence of the second chipping recognition conclusion to assess the final chipping recognition conclusion of the multi-blade drill bit as the discrimination result output of the chipping state of the multi-blade drill bit.

Description

Technical Field

[0001] This invention relates to the field of information technology, and in particular to a method for identifying the chipping condition of drill bits for ultrafine-grained cemented carbide circuit boards. Background Technology

[0002] The application of ultra-fine grain cemented carbide micro drills has become standard in precision machining in high-end electronic manufacturing, especially in drilling multilayer circuit boards with a thickness of only a few tenths of a millimeter. The drill diameter is often less than 0.3 millimeters, and its machining state directly determines the roughness, roundness, and reliability of interlayer conductivity of the hole wall. In high-speed drilling environments with speeds of tens of thousands or even hundreds of thousands of revolutions per minute, the drill cutting edge is subjected to extremely high cutting temperatures and alternating stresses. Accurately identifying the chipping state of the drill bit has become a core technical challenge to ensure machining accuracy and avoid mass scrapping. Current methods for monitoring drill bit damage largely rely on the conventional assumption that cutting load and damage degree are positively correlated, assuming that the more severe the tool damage, the more severe the resulting electrical signal abnormalities. This judgment method based on a linear increasing law ignores the load transfer effect between the cutting edges of multi-edged tools during rotary cutting, which easily leads to misjudgment. After chipping occurs, the unevenness of the cutting load causes fluctuations in the spindle current. The conventional understanding is that the larger the chipped area, the greater the current fluctuation amplitude should be. However, when a single cutting edge suffers extensive chipping, the number of intact cutting edges remaining in the drill bit becomes a crucial factor affecting the stress state. Intact cutting edges spontaneously share the cutting load of the damaged edge, and this redistribution weakens the asymmetry of stress during spindle rotation. For example, when a single cutting edge experiences extensive chipping, the strong support from other intact cutting edges results in a relatively smooth change in resistance felt by the spindle motor within a single rotation cycle, leading to smaller current fluctuations. Conversely, if multiple cutting edges simultaneously experience small-scale chipping, the uneven cutting resistance of each edge fails to provide effective compensation, resulting in severe current fluctuations. This anomalous inversion between chipping area and current fluctuation intensity renders traditional judgment mechanisms completely ineffective. Accurately capturing the current amplitude and spectral characteristics within the spindle motor's rotation cycle, and combining this with the remaining intact cutting edges and chipping distribution location, to dynamically clarify the contradiction between chipping area and current fluctuations, becomes a key issue in achieving precise identification of chipping conditions in ultra-fine-grained cemented carbide circuit board drill bits. Summary of the Invention

[0003] This invention provides a method for identifying the chipping condition of ultrafine-grained cemented carbide circuit board drill bits, the method comprising: The spindle current signal of the spindle motor in a single rotation cycle is acquired, the spindle current signal is frequency domain converted, and the amplitude and sideband components of each frequency component in the current spectrum are extracted to obtain a current spectrum record reflecting the cutting process of the multi-blade drill bit. The amplitude at the fundamental frequency of the spindle rotation and the amplitude at each harmonic are extracted from the current spectrum record. The position of each harmonic characterizes the strength of the cutting load on the corresponding cutting edge, and the harmonic amplitude record of the cutting load of each cutting edge is obtained. Based on the sideband components in the current spectrum record, the asymmetry of the cutting edge load torque within one rotation is evaluated, the chipping distribution location is extracted, and the asymmetry and the harmonic amplitude record are converted into the uniformity of cutting resistance of each blade and the load sharing pattern among multiple blades. The uniformity of cutting resistance of each cutting edge is evaluated by a preset resistance threshold. When the uniformity exceeds the resistance threshold, the presence of micro-chipping of the multi-edge drill bit is identified, and the first chipping identification conclusion and its confidence level are obtained. Based on the location of the chipping edge distribution, the number of intact edges and the chipping area of ​​a single edge are extracted from the load sharing pattern between the harmonic amplitude record and the multi-edge. When the chipping area of ​​a single edge exceeds a preset area threshold and the number of intact edges exceeds a preset number of edges threshold, the multi-edge drill bit is identified as being in a state of large-area chipping of a single edge, and a second chipping edge identification conclusion and its confidence level are obtained. The confidence levels of the first and second chipping identification conclusions are combined to determine the final chipping identification conclusion of the multi-bladed drill bit, which is then output as the discrimination result of the chipping state of the multi-bladed drill bit.

[0004] Furthermore, acquiring the spindle current signal of the spindle motor within a single rotation cycle includes: A Hall current sensor is connected to the power supply circuit of the spindle motor. The spindle current signal is continuously sampled at a sampling frequency higher than the spindle rotation frequency. The sampled data is truncated according to the single-turn trigger pulse output by the spindle encoder to obtain the current timing segment aligned with the single-turn rotation cycle of the spindle. Low-pass filtering is used to suppress high-frequency noise and power frequency interference in the current timing segment to obtain the denoised current timing segment.

[0005] Furthermore, the step of performing frequency domain conversion on the spindle current signal, extracting the amplitude and sideband components of each frequency component in the current spectrum, and obtaining a current spectrum record reflecting the multi-blade drill cutting process includes: The spindle current signal is weighted at both ends of the sampling window using a Hanning window, and converted from the time domain to the frequency domain using a fast Fourier transform. The position of the spindle rotation fundamental frequency and the position of the harmonics located at integer multiples of the fundamental frequency are read. The modulation sideband components are identified within the frequency search interval on both sides of the fundamental frequency and the harmonics. The frequency offset and amplitude values ​​of the sideband components relative to the fundamental frequency are extracted. The fundamental frequency amplitude, harmonic amplitude, sideband offset and sideband amplitude are collected in order of frequency from low to high to obtain the current spectrum record.

[0006] Furthermore, the extraction of the amplitude at the fundamental frequency of spindle rotation and the amplitude at each harmonic from the current spectrum record, and the characterization of the cutting load intensity of the corresponding cutting edge based on the position of each harmonic, to obtain the harmonic amplitude record of the cutting load of each cutting edge, includes: Locate the frequency position of the spindle rotation fundamental spectrum peak from the current spectrum record, sequentially search for each harmonic spectrum peak located at an integer multiple of the fundamental frequency along the frequency axis, read the amplitude values ​​of the fundamental spectrum peak and each harmonic spectrum peak, establish a mapping relationship between the number of each harmonic and the corresponding cutting edge number based on the total number of cutting edges of the multi-blade drill bit, and use the amplitude value of each harmonic to characterize the load strength of the mapped cutting edge during a single-turn cutting process to obtain the harmonic amplitude record.

[0007] Furthermore, the step of evaluating the asymmetry of the cutting edge load torque within one rotation based on the sideband components in the current spectrum record, and extracting the chipping distribution location, includes: The sideband components located on both sides of the fundamental frequency and harmonics are read from the current spectrum record. The frequency offset of each sideband component is divided by the main shaft rotation frequency to obtain the modulation order. The modulation order reflects the number of times the cutting edge torque fluctuation occurs in a single revolution. The amplitudes of each sideband component are weighted and summed with the modulation order as the weight to obtain a torque fluctuation measure that describes the torque deviation from the uniform distribution amplitude. The torque fluctuation measure is the degree of asymmetry. To address the phase difference between the sideband components and the fundamental frequency peak, a phase positioning method is used to map each sideband component to an angle range of a single rotation of the spindle. The angle range corresponds one-to-one with the installation orientation of each cutting edge of the multi-blade drill bit on the spindle end face. Based on the phase mapping result, the cutting edge number in the angle range where the torque fluctuation measurement exceeds a preset threshold is found. The cutting edge number is the location of the chipping distribution.

[0008] Furthermore, the conversion of the asymmetry and harmonic amplitude records into the uniformity of cutting resistance of each cutting edge and the load sharing pattern among multiple cutting edges includes: Based on the location of the chipping edge distribution, the harmonic amplitude values ​​of the corresponding cutting edges are retrieved from the harmonic amplitude record. The harmonic amplitude values ​​of each cutting edge are arranged one by one with the degree of asymmetry according to the cutting edge position number. The range of each cutting edge value in the arrangement result represents the uniformity of the cutting resistance of each cutting edge. The proportion of each cutting edge value to the total value of all cutting edges represents the load sharing pattern among the multiple cutting edges.

[0009] Furthermore, the evaluation of the uniformity of cutting resistance of each cutting edge using a preset resistance threshold, and the identification of multi-edge micro-chipping in a multi-edge drill bit when the uniformity exceeds the resistance threshold, yields a first chipping identification conclusion and its confidence level, including: The range of each cutting edge value is used as the uniformity measure, and the variance of each cutting edge value around its mean is used as the fluctuation measure. The uniformity measure and the fluctuation measure are weighted and superimposed according to a pre-calibrated weight coefficient to obtain a comprehensive uniformity index. When the comprehensive uniformity index exceeds the resistance threshold and the number of cutting edge numbers involved exceeds a preset cutting edge number threshold, multi-edge micro-breakage is identified. The extent to which the comprehensive uniformity index exceeds the resistance threshold and the proportion of the cutting edge number to the total number of cutting edges of the multi-edge drill bit are substituted into a pre-established confidence metric mapping table to retrieve the confidence level of the first breakage identification conclusion.

[0010] Furthermore, the step of extracting the number of intact blades and the single-blade damage area from the load-sharing pattern between the harmonic amplitude record and the multi-blade, based on the location of the chipping distribution, includes: Read the blade position number included in the blade position number set from the chipped blade distribution location as the chipped blade position. Subtract the number of chipped blade positions from the total number of multi-blade drill bit cutting edges to obtain the set of intact blade position numbers. The total number of blade position numbers in the set of intact blade position numbers is the number of intact blades. Based on the load sharing pattern among the multiple blades, the deviation of the sum of the load proportions of intact blade positions from the uniform sharing benchmark and the attenuation of the damaged blade position relative to the normal cutting amplitude in the harmonic amplitude record are converted according to the pre-calibrated inversion coefficient to obtain the single blade damage area.

[0011] Furthermore, the step of identifying a multi-flute drill bit as being in a state of large-area single-flute damage when the single-flute chipping area exceeds a preset area threshold and the number of intact flutes exceeds a preset number of flutes threshold, and obtaining a second chipping identification conclusion and its confidence level, includes: The confidence level of the second broken blade identification conclusion is obtained by substituting the difference between the extent by which the single-blade chipped area exceeds the preset area threshold and the number of intact blades and the preset blade number threshold into a pre-established confidence metric mapping table.

[0012] Furthermore, the final assessment of the chipping identification conclusion of the multi-bladed drill bit, based on the combined confidence levels of the first and second chipping identification conclusions, includes: The confidence levels of the first and second blade breakage identification conclusions are compared and aggregated according to their corresponding blade breakage state categories to obtain a confidence level comparison record. The evaluation is based on the difference between the two confidence values ​​in the confidence comparison record and a preset difference threshold. When the difference exceeds the preset difference threshold, the blade breakage state corresponding to the one with higher confidence is taken as the superior conclusion. When the difference does not exceed the preset difference threshold, the two chipping states are included in the arbitration conclusion. The arbitration conclusion is based on the ratio of the number of blade position numbers in the chipping distribution position to the number of intact blades. When the ratio of the number of blade position numbers to the total number of multi-blade drill bit edges exceeds the preset ratio threshold, they are merged into a multi-blade micro-chipping state; otherwise, they are merged into a single-blade large-area chipping state, thus obtaining the final chipping identification conclusion.

[0013] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects: This invention discloses a method for identifying the chipping condition of ultrafine-grained cemented carbide circuit board drill bits. This method addresses the problem of uneven cutting load distribution and decreased machining quality caused by micro-chipping or large-area chipping of a single cutting edge during high-speed cutting of multi-flute drill bits. It collects the current signal within a single rotation cycle of the spindle motor and performs frequency domain conversion, extracting the amplitude of each frequency component and sideband components in the current spectrum to reflect the multi-flute cutting process. The core solution of this invention is to assess the strength of the cutting load on each cutting edge from the fundamental frequency and harmonic amplitude of the spindle rotation, and analyze the asymmetry of the cutting edge load torque by combining the sideband components, thereby calculating the uniformity of cutting resistance and load distribution pattern of each cutting edge, and then identifying the chipping location, the number of intact cutting edges, and the chipping area of ​​a single cutting edge. This invention distinguishes between two typical failure modes—multi-flute micro-chipping and large-area chipping of a single cutting edge—by setting resistance thresholds, area thresholds, and cutting edge number thresholds, respectively. The final judgment result is formed by combining the confidence levels of the two identification conclusions, achieving accurate identification and quantitative evaluation of drill bit chipping conditions, effectively improving tool monitoring accuracy and machining process reliability. Attached Figure Description

[0014] Figure 1 This is a flowchart of a method for identifying the chipping state of an ultrafine-grained cemented carbide circuit board drill bit according to the present invention.

[0015] Figure 2 This is a schematic diagram of a method for identifying the chipping state of an ultrafine-grained cemented carbide circuit board drill bit according to the present invention.

[0016] Figure 3 This is another schematic diagram of a method for identifying the chipping state of an ultrafine-grained cemented carbide circuit board drill bit according to the present invention. Detailed Implementation

[0017] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this specification.

[0018] like Figures 1-3 This embodiment of a method for identifying the chipping state of an ultrafine-grained cemented carbide circuit board drill bit may specifically include: S101. Acquire the spindle current signal of the spindle motor in a single rotation cycle, perform frequency domain conversion on the spindle current signal, extract the amplitude and sideband components of each frequency component in the current spectrum, and obtain the current spectrum record reflecting the cutting process of the multi-blade drill bit.

[0019] A Hall current sensor is connected to the power supply circuit of the spindle motor. The Hall current sensor continuously samples the spindle current at a sampling frequency tens of times higher than the spindle rotation frequency. The sampled data is truncated according to the single-turn trigger pulse output by the spindle encoder to obtain a current timing segment aligned with the single-turn rotation cycle of the spindle. A low-pass filter is used to suppress high-frequency noise and power frequency interference in the current timing segment to obtain a denoised current timing segment. For the denoised current timing segment, a Hanning window is used to weight the two ends of the sampling window to suppress spectral leakage. Then, a fast Fourier transform is used to convert the current timing segment from the time domain to the frequency domain. The position of the spindle rotation fundamental frequency and the harmonic positions located at integer multiples of the fundamental frequency are read from the frequency domain transformation result. The amplitude values ​​at the fundamental frequency and each harmonic position are extracted to obtain a set of spectral amplitudes reflecting the cutting pulses of each cutting edge of the multi-blade drill bit. A symmetrical frequency search interval is set on both sides of the fundamental frequency and harmonics of the frequency spectrum amplitude set. The modulation sideband component caused by uneven cutting edge load is identified from the frequency search interval. The frequency offset and amplitude value of the sideband component relative to the fundamental frequency are extracted. The fundamental frequency amplitude, harmonic amplitude, sideband offset and sideband amplitude are collected in order of frequency from low to high to obtain the current spectrum record reflecting the cutting process of the multi-blade drill bit.

[0020] In high-speed drilling operations using ultra-fine-grained cemented carbide multi-bladed micro drills to machine multilayer circuit boards, the spindle motor rotates at tens of thousands of revolutions per minute. Each cutting edge that penetrates the material triggers a cutting resistance pulse, which is transmitted to the motor rotor through the spindle system. Ultimately, this pulse leaves a periodic imprint in the stator winding current that is closely related to the number and condition of the cutting edges. In one embodiment, one phase cable from the three-phase power supply circuit of the spindle motor is randomly selected and threaded into the core window of an open-loop Hall current sensor. The output voltage signal of the Hall current sensor is then conditioned by an instrumentation amplifier and sent to a data acquisition card.

[0021] For example, if the spindle speed is 120,000 revolutions per minute, the spindle frequency is 2 kHz. The sampling frequency is set to be several tens of times higher than this frequency. The spindle current is continuously sampled at a sampling frequency of 100 kHz so that each single rotation cycle contains a sufficient number of sampling points.

[0022] Specifically, the spindle encoder outputs a zero-position trigger pulse after each rotation. The zero-position trigger pulse serves as a timing reference to segment the continuously sampled current data. The sampling points between one zero-position pulse and the next zero-position pulse constitute a current timing segment aligned with the single rotation cycle of the spindle.

[0023] Preferably, the current timing segment is filtered using a low-pass filter, wherein the highest-priority harmonic is set to 5 to 10 times the spindle frequency, and the cutoff frequency f of the low-pass filter is... lp Set the cutoff frequency to 1.2 to 1.5 times the highest harmonic of interest... For example, when the spindle frequency is 2000 Hz and the highest harmonic of interest is the 5th harmonic, the cutoff frequency is set to 12000 to 15000 Hz. This is the core step for the frequency domain conversion of the current timing segment after noise reduction.

[0024] It should be noted that the two ends of a single-cycle sampling segment are often discontinuous when spliced ​​into a periodic signal. Directly performing a Fourier transform will produce significant frequency leakage in the spectrum, causing the amplitude at the fundamental frequency and harmonics to be contaminated by side lobes.

[0025] In one embodiment, the denoised current time series segment is multiplied by a Hanning window function of the same length for windowing. The Hanning window function is calculated as: w = 0.5 - 0.5 × cosine function, where the independent variable of the cosine function is 2π × n ÷ (N-1), where w is the Hanning window function value, n is the current sampling point number ranging from 0 to N-1, N is the total number of sampling points in a single-turn current time series segment, and π is pi. The function value of the Hanning window decays to zero at both ends of the time series segment, enabling a smooth transition at the periodic splicing points after windowing. The windowed current time series segment is then converted from the time domain to the frequency domain using a Fast Fourier Transform to obtain the frequency component distribution reflecting the single-turn rotation process. The spindle rotation frequency is represented by the fundamental frequency peak in the frequency domain, denoted as f0. For multi-flute drill bits with K cutting edges, K ranges from 2 to 6, with common drill bits having 2 or 3 cutting edges. During drilling, K cutting edges sequentially cut into the sheet metal within one rotation of the spindle. Each cutting edge generates periodic pulse forces during cutting, and these pulse forces, when superimposed, excite harmonic components located at integer multiples of the fundamental frequency in the frequency domain. Since the K cutting edges are uniformly distributed and participate in cutting synchronously, the frequency characteristics of the cutting force present a harmonic sequence of f0, 2f0, 3f0 up to Kf0. The Kth harmonic is located at frequency Kf0, which corresponds to the repetition frequency of the combined cutting pulses from each cutting edge, thus carrying the main energy of the combined cutting force of each edge. In the frequency domain results, starting from the fundamental frequency f0, the positions of the 2nd harmonic 2f0, the 3rd harmonic 3f0, up to the 6th harmonic 6f0 are sequentially located. The amplitude values ​​A1, A2, A3 to A6 at each harmonic position are read one by one, and the read amplitude values ​​are compiled to form a spectral amplitude set {A1, A2, A3, A4, A5, A6} for subsequent drilling state analysis.

[0026] It is understandable that if a multi-bladed drill bit has a chipped edge, its cutting resistance will deviate relative to the other intact edges, and this deviation will cause periodic changes in impact load.

[0027] Specifically, if the drill bit rotation frequency is denoted as fr and the number of cutting edges is z, then under normal circumstances, the cutting force signal is mainly concentrated at the cutting edge frequency f. z and its harmonics, where f z =fr×z. When a single cutting edge breaks, an abnormal impact is generated once per revolution. This impact signal modulates the cutting edge frequency harmonics with the revolution frequency fr as the period, which manifests in the frequency domain as at the cutting edge frequency f. z and its multiplier 2f z 3f z Symmetrical modulation sidebands with an interval of fr appear on both sides of the harmonic peak.

[0028] For example, at the blade frequency fc, sideband components with fc minus fr and fc plus fr appear on both sides, and the amplitude of the sidebands is positively correlated with the degree of breakage. By extracting the amplitude of these modulated sidebands and comparing them with the normal state, single-blade breakage faults can be identified.

[0029] In one embodiment, a symmetrical frequency search interval is extended to both sides of the identified fundamental frequency and each harmonic peak as the center, with a width three times the main shaft rotation frequency. Within the frequency search interval, a local maximum detection algorithm is used to find local peak points deviating from the central spectral peak. During detection, an amplitude threshold is set to 0.1 times the amplitude of the central spectral peak. Local peak points higher than this threshold are identified as modulation sideband components. The frequency position fs of each modulation sideband component is extracted, and the frequency position fc of the corresponding central spectral peak is subtracted to obtain the frequency offset Δf, where Δf = fs - fc. The amplitude value at this position is then read. The ratio n = Δf / fr of the frequency offset to the main shaft rotation frequency fr is calculated, and the integer obtained by rounding this ratio is the modulation order of the sideband. For example, when the spindle frequency is 2000 Hz, the detected sideband component frequency is 16850 Hz, and the center spectral peak frequency is 12000 Hz, the frequency offset is 4850 Hz, and the ratio is 2.425. Taking the nearest integer gives the modulation order as 2. The judgment rule is: if the deviation of the ratio from the nearest integer is less than 0.3, it is considered a valid modulation sideband and rounded to the modulation order; if the deviation is greater than or equal to 0.3, the component is considered noise interference and is discarded.

[0030] For example, if the sideband appears at exactly one frequency offset from the fundamental frequency, it indicates a single rotational load asymmetry between the cutting edges; if the sideband appears at a higher order offset, it indicates that the load asymmetry exhibits multiple harmonic coupling, meaning that the cutting force fluctuations of multiple cutting edges act on the spindle with different frequency components superimposed. The fundamental frequency amplitude, harmonic amplitudes, frequency offsets of each modulation sideband, and sideband amplitude are arranged into an ordered record in ascending order of frequency to obtain a characteristic spectrum sequence reflecting the multi-flute drill bit cutting process, which serves as the frequency domain representation of the multi-flute drill bit cutting state within that single rotational cycle.

[0031] S102. Extract the amplitude at the fundamental frequency of the spindle rotation and the amplitude at each harmonic from the current spectrum record. Based on the position of each harmonic, characterize the strength of the cutting load on the corresponding cutting edge to obtain the harmonic amplitude record of the cutting load of each cutting edge.

[0032] The frequency position of the spindle rotation fundamental spectrum peak is located from the current spectrum record. Harmonic spectrum peaks located at integer multiples of the fundamental frequency are sequentially found along the frequency axis. The amplitude values ​​of the fundamental spectrum peak and each harmonic spectrum peak are read to obtain a harmonic amplitude sequence arranged in harmonic order. Based on the total number of cutting edges of the multi-blade drill bit, a mapping relationship is established between the order of each harmonic in the harmonic amplitude sequence and the corresponding cutting edge number. The amplitude value of each harmonic represents the load strength of the mapped cutting edge during a single cutting cycle, thus obtaining a harmonic amplitude record of the cutting load for each cutting edge.

[0033] In one implementation, starting from the frequency position of the spindle rotation fundamental spectrum peak identified in step S101, the frequency is recursively pushed along the frequency axis towards higher frequencies, sequentially searching for local maxima near frequencies that are 2 times, 3 times, and up to multiples of the total number of cutting edges. These local maxima are the harmonic spectrum peaks. The amplitude values ​​at the fundamental spectrum peak and each harmonic spectrum peak are read and arranged in ascending order of harmonics to form a harmonic amplitude sequence. For multi-blade drills, each cutting edge sequentially cuts into the material during a single spindle rotation. The energy concentration at the harmonic corresponding to the number of cutting edges in the spindle current spectrum represents the multi-blade combined cutting pulse, while the amplitude at the fundamental frequency reflects the load difference between each cutting edge. Taking a two-flute carbide micro drill bit commonly used in high-speed drilling of multilayer circuit boards as an example, the cutting pulses of the two cutting edges are mainly reflected in the second harmonic. The fundamental frequency amplitude reflects the degree of load imbalance between the two cutting edges. When the fundamental frequency amplitude reaches or exceeds 30% of the second harmonic amplitude, it indicates that the wear difference between the two cutting edges has reached the critical state requiring tool replacement. This 30% value is determined by statistically analyzing the distribution of the "fundamental frequency / second harmonic amplitude ratio" of normal cutting and wear failure samples of the same type of tool. The lower quantile of this ratio distribution of the failure samples (or the mean of the normal samples plus k times the standard deviation) is taken as the basis. A one-to-one mapping relationship is established between the number k of the kth harmonic in the harmonic amplitude sequence and the kth cutting edge, where the value of k ranges from 1 to the number of drill cutting edges N. This mapping relationship uses the harmonic number to indicate the load strength of the corresponding cutting edge in a single cutting process. The larger the harmonic amplitude value, the stronger the cutting load borne by the corresponding cutting edge. Based on the mapping relationship, the amplitude values ​​of each harmonic in the harmonic amplitude sequence are reorganized according to the mapped cutting edge number to obtain the harmonic amplitude record of each cutting edge load. Each item in the harmonic amplitude record corresponds to a specific cutting edge on a multi-flute drill bit. For example, for a 3-flute drill bit, the first harmonic amplitude corresponds to the load of cutting edge number 1, the second harmonic amplitude corresponds to the load of cutting edge number 2, and the third harmonic amplitude corresponds to the load of cutting edge number 3, forming a harmonic amplitude record sequence containing 3 values.

[0034] S103. Evaluate the asymmetry of the cutting edge load torque within one revolution based on the sideband components in the current spectrum record, extract the chipping distribution location, and convert the asymmetry and harmonic amplitude records together into the uniformity of cutting resistance of each blade and the load sharing pattern among multiple blades.

[0035] Sideband components located on both sides of the fundamental frequency and harmonics are read from the current spectrum record. The frequency offset of each sideband component is divided by the spindle rotation frequency to obtain the modulation order. The modulation order reflects the number of times the cutting edge torque fluctuation occurs in a single revolution. The amplitudes of each sideband component are weighted and summed using the modulation order as the weight to obtain the torque asymmetry of the cutting edge load torque in one revolution. The higher the torque asymmetry value, the greater the deviation of the torque from the uniform distribution in a single revolution, thus obtaining a torque fluctuation metric reflecting the torque asymmetry. Based on the phase difference between the sideband components and the fundamental frequency peak, a phase positioning method is used to map each sideband component to an angle range of a single revolution of spindle rotation. This angle range corresponds one-to-one with the installation orientation of each cutting edge of the multi-flute drill bit on the spindle end face. Based on the phase mapping result, the cutting edge number in the angle range where the torque fluctuation metric exceeds a preset threshold is found. This cutting edge number is the location of the chipping distribution. Based on the location of the chipping edge distribution, the harmonic amplitude value of the corresponding cutting edge is retrieved from the harmonic amplitude record. The harmonic amplitude values ​​of each cutting edge and the torque fluctuation measurement are arranged one by one according to the cutting edge position number. The smaller the range of the values ​​of each cutting edge in the arrangement result, the higher the uniformity of the cutting resistance of each cutting edge. The proportion of each cutting edge value to the total value of all cutting edges is used to characterize the load sharing pattern among multiple cutting edges, thus obtaining the uniformity of inter-cutting resistance and the load sharing pattern.

[0036] In the high-speed drilling process of ultrafine-grained cemented carbide multi-flute micro drills, the asymmetry of the cutting edge load torque during a single spindle rotation essentially reflects the deviation direction and magnitude of the cutting resistance of each cutting edge from a uniform distribution. This asymmetry is defined as an inter-edge load imbalance index based on the spindle current spectrum, obtained by weighting the normalized values ​​of the amplitudes of each modulated sideband component relative to the amplitude of the fundamental spectrum peak according to the modulation order. When a cutting edge breaks, the cutting resistance corresponding to that edge position drops sharply, resulting in a significant sideband peak at integer multiples of the spindle rotation frequency. When the sum of the normalized sideband amplitudes reaches or exceeds 0.15, it indicates significant uneven cutting edge load, a key clue revealing the location of the broken edge. In addition to the fundamental frequency and harmonics, which reflect the periodic cutting process, the spindle current spectrum also contains sideband components with lower energy adjacent to the fundamental frequency and each harmonic. These sideband components originate from the amplitude modulation of the periodic cutting current by the cutting edge torque fluctuations. When the cutting resistance of each cutting edge is completely uniform, the main current contains only the fundamental frequency and harmonics. When a load deviation occurs in one or more cutting edges, the deviation will modulate the main spectrum peak with the rotational frequency or its harmonics, forming sideband components symmetrically distributed on both sides of the main spectrum peak. Dividing the frequency offset of each sideband component by the spindle rotational frequency yields the modulation order of that sideband component. A higher modulation order means a denser distribution of torque fluctuation within a single revolution, and a more significant contribution to the cutting edge asymmetry. Therefore, the amplitudes of each sideband component are weighted and summed using the modulation order as the weight. If the calculated modulation order is not an integer, it is rounded to the nearest integer value. The torque fluctuation measure D is calculated using the following formula: D=Σ(k×(A) s / A0)), where k is the modulation order of the sideband component, A s A0 represents the amplitude of the corresponding sideband component, and A0 represents the amplitude of the fundamental frequency peak. s / A0 represents the normalized sideband amplitude, Σ represents the summation of all sideband components, and D is the metric, which is comparable under different spindle speeds and load conditions. The higher the torque fluctuation metric value, the greater the deviation of the torque from the uniform distribution within a single revolution. The torque fluctuation metric alone is insufficient to determine which specific cutting edge has broken; spatial positioning is achieved using the phase difference θ between the sideband component and the fundamental frequency peak, where the phase difference θ reflects the angular position of the broken cutting edge on the tool circumference.

[0037] Specifically, the current spectrum record is subjected to a Fourier transform to obtain complex spectral data. Complex spectra are then read from each sideband component and the fundamental frequency peak to obtain the complex spectrum of the sideband and the complex spectrum of the fundamental frequency. The phase difference between these two spectra is the phase difference of the sideband component relative to the fundamental frequency peak. The system acquires the spindle rotation angle information in real time through an encoder mounted on the spindle. Using the angle corresponding to the zero-position pulse of the spindle encoder as the zero-degree reference, a single rotation of 360 degrees is divided into several angle intervals equal to the total number of cutting edges of the multi-flute drill bit. For example, for a four-flute drill bit, this is divided into four angle intervals: 0° to 90°, 90° to 180°, 180° to 270°, and 270° to 360°. These angle intervals correspond one-to-one with the installation orientation of each cutting edge of the multi-flute drill bit on the spindle end face. The phase difference is converted into an angle value θ using the formula θ = φ × 360 ÷ 2π, where φ is the phase difference in radians. The converted angle value falling within a certain angle interval corresponds to the spatial orientation of the cutting edge. If the cutting edges are not evenly distributed, the angle intervals are dynamically divided according to the measured interval angles of each cutting edge, and the boundary of each interval is the midpoint of the center angle of two adjacent cutting edges.

[0038] In one embodiment, the sideband complex spectrum is merged according to its modulation order. The merging method is to sum the squares of the amplitudes of all sideband components of the same modulation order to obtain the torque fluctuation energy under that modulation order. For all cutting edge positions within each angle interval, their normalized sideband amplitudes are weighted and accumulated according to the modulation order to obtain the cumulative torque fluctuation metric S for that interval, which maintains the same measurement caliber as the torque fluctuation metric D defined in S103 above. The cumulative metric S is compared with a preset threshold T, which is calibrated by the 95th percentile of the cumulative metric distribution of each angle interval in the normal cutting sample, with a typical value range of 0.05 to 0.20, preferably 0.10. If the cumulative value S exceeds the preset threshold T, the cutting edge number corresponding to that angle interval is included in the chipping distribution location set, and the cutting edge number is the chipping distribution location. The chipping distribution location gives the cutting edge number where load deviation occurs. To further characterize the relative load relationship between each cutting edge, it is necessary to combine the cutting load strength of each cutting edge in the harmonic amplitude record. Based on the chipping distribution location, the harmonic amplitude values ​​of the corresponding cutting edges are extracted from the harmonic amplitude record. The harmonic amplitude values ​​of each cutting edge are then matched one-to-one with the torque fluctuation metric according to the cutting edge position number, forming an arrangement result with the cutting edge position number as the row index and the harmonic amplitude and torque fluctuation metric as the column items. The range is obtained by subtracting the minimum value from the maximum value among the cutting edge values ​​in the arrangement result. The smaller the range, the closer the cutting resistance of each cutting edge is to a uniform state; the range is the measure of the uniformity of the cutting resistance of each cutting edge. Simultaneously, the proportion of each cutting edge value to the total value of all cutting edges is calculated. This proportion reflects the load share borne by that cutting edge during the overall multi-edge cutting process. The set of proportions for all cutting edges represents the load sharing pattern among the multiple cutting edges. The uniformity of the inter-cutting resistance and the load sharing pattern together characterize the load distribution state of the multi-edge drill bit within a single rotation.

[0039] S104. Evaluate the uniformity of cutting resistance of each cutting edge using a preset resistance threshold. When the uniformity exceeds the resistance threshold, identify the presence of micro-chipping of the multi-bladed drill bit and obtain the first chipping identification conclusion and its confidence level.

[0040] The range of each cutting edge value in the uniformity of inter-cutting resistance is obtained as a uniformity measure, and the variance of each cutting edge value around its mean is used as a fluctuation measure. The uniformity measure and the fluctuation measure are weighted and superimposed according to a pre-calibrated weighting coefficient based on the total number of cutting edges of the multi-flute drill bit, to obtain a comprehensive uniformity index characterizing the uniformity of cutting resistance of each cutting edge. The comprehensive uniformity index is evaluated against a preset resistance threshold. If the comprehensive uniformity index exceeds the resistance threshold, the number of cutting edge numbers in the chipping distribution location is verified. When the number of involved cutting edge numbers exceeds a preset number of cutting edge numbers threshold, multi-edge micro-chipping is identified in the multi-flute drill bit, and a first chipping identification conclusion is obtained. Based on the extent to which the comprehensive uniformity index exceeds the resistance threshold and the proportion of the number of involved cutting edge numbers in the chipping distribution location to the total number of cutting edges of the multi-flute drill bit, the extent of the exceedance and the proportion are substituted into a pre-established confidence metric mapping table to retrieve the corresponding values, thus obtaining the confidence level of the first chipping identification conclusion.

[0041] In high-speed drilling with multi-flute drills, multi-flute micro-chipping refers to the simultaneous occurrence of small-scale chipping on multiple cutting edges within the same machining cycle. Its manifestation in the spindle current is not a drastic fluctuation in a single cutting edge, but rather a deterioration in overall uniformity caused by uneven cutting resistance among the cutting edges. Therefore, precise characterization of the uniformity of inter-cutting resistance becomes the core aspect of identifying multi-flute micro-chipping. This phenomenon is particularly pronounced in high-precision drilling scenarios such as ultra-fine-grained cemented carbide micro-circuit board drills, where the cutting edge diameter is typically in the range of 0.1 mm to 0.5 mm, and the rotational speed can reach over 100,000 revolutions per minute. Even minute differences in inter-cutting resistance can lead to a decrease in machining quality. During the cutting process, the harmonic amplitudes of each cutting edge are recorded using the spindle current spectrum, and the normalized harmonic amplitude of the i-th cutting edge is taken as H. i The torque fluctuation metric of the cutting edge obtained from the current sideband of S103 is taken as σ. i H i With σ i Arranged according to the cutting edge position number i, two dimensionless sequences are formed. The harmonic amplitude value H and the torque fluctuation measure σ are arranged according to the cutting edge position number i to obtain the basic data sequence for the uniformity of inter-cutting resistance, where H... i and σ i Let R represent the harmonic amplitude and torque fluctuation value of the i-th cutting edge position, respectively. Each cutting edge position number corresponds to a value representing the current cutting resistance of that cutting edge. The range R is calculated by subtracting the minimum value from the maximum value within the normalized harmonic amplitude sequence. hAnd take the standard deviation S of the sequence. h To eliminate dimensions, both are divided by the range R of the sequence mean. h With dispersion S h The comprehensive uniformity index is calculated as U=α×R h +β×S h Weighting is applied, where α and β are determined based on the number of cutting edges to avoid directly adding the range and variance due to their different dimensions. The weighting coefficients are pre-calibrated based on the total number of cutting edges of the multi-flute drill bit. For two-flute drill bits, due to the fewer cutting edges, the focus is more on the range, so α is higher than β; for three- or four-flute drill bits, due to the more cutting edges, the focus is more on overall dispersion, so β ​​is higher than α. In one embodiment, α is 0.7 and β is 0.3 for two-flute drill bits, α is 0.5 and β is 0.5 for three-flute drill bits, and α is 0.4 and β is 0.6 for four-flute drill bits. Further, a preset resistance threshold T is introduced for the comprehensive uniformity index U. This resistance threshold is obtained before the multi-flute drill bit leaves the factory by statistically analyzing current spectrum samples under normal cutting conditions. At least 100 sets of normal cutting samples are collected to calculate the comprehensive uniformity index, and the 95th percentile of the index distribution is taken as the threshold value T. If the comprehensive uniformity index U exceeds the resistance threshold T, it is determined that there is significant non-uniformity in the inter-cutting resistance.

[0042] Understandably, unevenness between blades alone is insufficient to distinguish between multi-blade micro-chipping and single-blade chipping, necessitating the introduction of a verification process for the number of blade positions.

[0043] Preferably, the number N of the included cutting edge numbers is read from the set of chipping locations obtained in the aforementioned resistance deviation determination process. This set includes all cutting edge numbers whose resistance deviation exceeds a deviation threshold. The number N of cutting edge numbers is compared with a preset cutting edge number threshold M. If N exceeds M, it indicates that a large number of cutting edges exhibit load deviation, thus identifying multi-flute micro-chipping in a multi-flute drill bit. Specifically, the preset cutting edge number threshold M is set to 1 for two-flute drill bits, 2 for three-flute drill bits, and 2 for four-flute drill bits. When both the comprehensive uniformity index U exceeds the resistance threshold T and the number N of cutting edge numbers exceeds the preset cutting edge number threshold M, a first chipping identification conclusion is obtained, with the multi-flute micro-chipping state as its content.

[0044] It should be noted that the credibility of the first blade breakage identification conclusion is not a binary result, but a quantitative confidence level needs to be given based on the strength of the conditions met.

[0045] Specifically, let K be the total number of cutting edges of the multi-flute drill bit. Calculate the extent to which the comprehensive uniformity index U exceeds the resistance threshold T, δ=(UT) / T. Then calculate the proportion ρ=N / K of the number of cutting edge numbers N in the total number of cutting edges K. A larger δ indicates a more severe degree of non-uniformity between cutting edges, and a larger ρ indicates a wider range of cutting edges involved in the breakage. A pre-established confidence metric mapping table uses δ and ρ as two-dimensional input indices and confidence values ​​between 0 and 1 as output values. The mapping table is obtained through the following process: collect no less than 500 sets of historical drilling samples, and record the corresponding δ value, ρ value and label of whether actual chipping occurred for each set of samples; discretize the δ axis into several intervals at 0.1 intervals and the ρ axis into several intervals at 0.2 intervals to form a two-dimensional grid; calculate the ratio of the number of samples with chipping in each grid cell to the total number of samples in that cell, and this ratio is the confidence output value of the corresponding grid; for sparse grid cells, a weighted average of neighboring grids is used to supplement them, and the weight coefficient is inversely proportional to the distance between grids.

[0046] For example, the confidence metric mapping table outputs the confidence level C using δ and ρ as two-dimensional indices. The δ and ρ axes are discretized into complete grids with increments of 0.1 and 0.2, respectively, covering all value combinations, including intersecting units where δ and ρ are at different levels. The value is taken according to the calibration ratio of the grids they fall on, with boundaries always left-open and right-closed. When δ≤0.3 and ρ≤0.25, C is approximately 0.15 to 0.30; when 0.3<δ≤0.6 and 0.25<ρ≤0.55, C is approximately 0.35 to 0.65; when δ>0.6 and ρ>0.55, C is approximately 0.85 to 0.98. Other intersecting units are directly given by the calibration grid. Substituting the calculated δ and ρ values ​​into the confidence metric mapping table to retrieve the corresponding values ​​yields the confidence level of the first chipping blade identification conclusion. This confidence level, together with the first chipping blade identification conclusion, constitutes the quantitative judgment result of the multi-blade micro-chipping state.

[0047] S105. Combining the location of chipping, extract the number of intact cutting edges and the chipping area of ​​a single cutting edge from the load sharing pattern between the harmonic amplitude record and the multi-cutting edge. Evaluate using a preset area threshold and a preset cutting edge number threshold. When the chipping area of ​​a single cutting edge exceeds the area threshold and the number of intact cutting edges exceeds the cutting edge number threshold, identify the multi-cutting drill bit as being in a state of large-area chipping of a single cutting edge, and obtain the second chipping edge identification conclusion and its confidence level.

[0048] The chipped edge numbers are read from the chipped edge distribution locations and included in the chipped edge number set as the chipped edge positions. The total number of chipped edge positions is subtracted from the total number of multi-flute drill bit cutting edges to obtain the set of intact edge number positions. The total number of edge number positions in the intact edge number set is the number of intact edges, which represents the number of cutting edges that still perform normal cutting functions. Based on the load sharing pattern among the multi-flute drill bits, the deviation of the sum of the load proportions of the intact edge positions from the uniform sharing benchmark is converted into the attenuation of the chipped edge position relative to the normal cutting amplitude in the harmonic amplitude record using a pre-calibrated inversion coefficient. The inversion coefficient is pre-calibrated based on the correspondence between the intact cutting edge area and the current amplitude attenuation, thus obtaining the single-edge chipping area of ​​the chipped edge position. The single-edge chipping area is evaluated using a preset area threshold, and the number of intact edges is evaluated using a preset edge number threshold. If the single-edge chipping area exceeds the area threshold and the number of intact edges exceeds the edge number threshold, the multi-edge drill bit is identified as being in a state of large-area single-edge chipping, and a second chipping identification conclusion is obtained. The extent by which the single-edge chipping area exceeds the area threshold and the difference between the number of intact edges and the edge number threshold are substituted into a pre-established confidence metric mapping table to retrieve the corresponding values, thereby obtaining the confidence level of the second chipping identification conclusion.

[0049] In high-speed drilling, multi-flute micro drills exhibit a load-sharing pattern among their individual blades due to cutting force fluctuations within the rotation cycle. However, in typical applications such as ultrafine-grained carbide circuit board drills, the current characteristics exhibited by large-area chipping of a single blade differ fundamentally from those of multi-flute micro-chipping. When a blade experiences large-area chipping, the remaining intact blades spontaneously take over the cutting area of ​​the damaged blade in the next spindle rotation cycle. This load redistribution weakens the asymmetry of force during spindle rotation, preventing the spindle current fluctuation from increasing linearly with the chipped area. The load-sharing pattern shifts from an initial uniform distribution to a non-uniform state where intact blades bear an additional load. Therefore, identifying large-area chipping of a single blade requires simultaneously examining the excluded chipped blade locations, the number of remaining intact blades, and the share of additional load borne by the intact blades.

[0050] It should be noted that the chipping distribution locations are recorded in the preceding steps as a set of cutting edge numbers indicating the occurrence of load deviations. In one embodiment, the chipping cutting edge number set is read from the chipping distribution locations to obtain all cutting edge numbers in the set; the total number of pre-calibrated cutting edges of the multi-blade drill bit is denoted as N, and the number of cutting edge numbers in the chipping cutting edge number set is denoted as M. Then, the number of intact cutting edges is N minus M, where N is the total number of cutting edges and M is the number of chipped edges.

[0051] Understandably, simply obtaining the number of intact blades is insufficient to infer the damage area of ​​the broken blade edge; it is also necessary to use the load sharing pattern and harmonic amplitude records to jointly invert the broken area.

[0052] Specifically, under normal cutting conditions of a multi-flute drill bit, each cutting edge should equally share the total cutting load under ideal and uniform conditions. For a multi-flute drill bit with a total of K cutting edges, the uniform sharing benchmark for each cutting edge is 1 / K. Using the harmonic amplitude records of each cutting edge obtained from the spindle current spectrum and the relationship between the load and the cutting edge, the ratio of the normalized harmonic amplitude of each cutting edge to the sum of the normalized harmonic amplitudes of all cutting edges is taken as the load proportion of that cutting edge. The load proportions of all cutting edges constitute the load sharing pattern. The load proportion corresponding to each intact cutting edge is read from this load sharing pattern. The load proportions of the intact cutting edges are added together to obtain the sum S of the intact cutting edge load proportions. Then, the product of the number of intact cutting edges n and the uniform sharing benchmark 1 / K is subtracted from the sum S of the intact cutting edge load proportions to obtain the deviation dS = Sn / K. The larger the value of the deviation dS, the more of the load share that the intact cutting edges bear to compensate for the broken cutting edges. The cutting force signal is converted to the frequency domain using a Fast Fourier Transform (FFT). The amplitude values ​​corresponding to the drill bit rotation frequency and its harmonics are extracted to form a harmonic amplitude record. The amplitude of each harmonic component in this record corresponds to the cutting state of each cutting edge, with the i-th harmonic amplitude primarily reflecting the cutting characteristics of the i-th cutting edge. Further, the amplitude values ​​corresponding to the chipped cutting edge in the harmonic amplitude record are processed.

[0053] Preferably, the harmonic amplitude value corresponding to the chipped cutting edge is compared with the reference amplitude value of the same cutting edge under normal cutting conditions of the same type of drill bit, and the difference between the two is the attenuation of the chipped cutting edge relative to the normal cutting amplitude. The reference amplitude value refers to the average value of the harmonic amplitude of each cutting edge collected under the same cutting parameters when the same type of drill bit is in its factory condition or with an intact cutting edge. The reference data collection process is as follows: select 3 to 5 unused drill bits of the same type, and perform 10 cutting tests under high-speed conditions consistent with actual drilling, i.e., spindle speed of 80,000 to 120,000 rpm and corresponding feed parameters. Extract the peak value of the harmonic amplitude of each cutting edge in each test, and calculate the arithmetic mean of the same cutting edge in all tests as the reference amplitude value A0 of that cutting edge. The attenuation dA is calculated as the difference between the current measured amplitude value A1 and the reference amplitude value A0, i.e., dA = A0 - A1. The reference data is stored in the monitoring system database with the cutting edge number as the index, and each cutting edge corresponds to a reference amplitude value. The attenuation dA directly reflects the decrease in pulse energy caused by the reduction in cutting edge length and the decrease in the intact area of ​​the chipped cutting edge. Neither the deviation dS nor the attenuation dA alone can adequately characterize the chipped area; both need to be calculated together.

[0054] In one embodiment, the inversion coefficients are pre-calibrated based on the correspondence between the intact cutting edge area and the harmonic amplitude attenuation. During the factory inspection process, more than 30 sets of standard chipping samples of varying degrees are manufactured on the cutting edge of the same model of ultra-fine grain cemented carbide drill bit using precision grinding equipment. The actual chipping area of ​​each sample is measured using an industrial microscope as the calibration value. For each sample, the deviation dS and the harmonic amplitude attenuation dA0 are measured. Dividing dA0 by the reference harmonic amplitude I0 of the intact cutting edge yields the normalized attenuation dA, which is equal to dA0 divided by I0, thus converting both dS and dA into ratios. Using the known chipping area as the target value and the deviation dS and normalized attenuation dA as inputs, linear least squares fitting is used to obtain the inversion coefficients α and β, with a correlation coefficient greater than 0.85. The formula for calculating the chipping area is A equal to α multiplied by dS plus β multiplied by dA, where A is the chipping area of ​​a single blade in square millimeters, and α and β are the deviation coefficient and attenuation coefficient, respectively, which are fixed in the controller memory during the factory inspection stage. Substituting the dS and dA measured in this test into the calculation formula, the chipping area of ​​the single blade at the chipped blade position is obtained. Further, a preset area threshold S is set. th Based on the complete cutting edge area S of a single blade e The value is taken as 0.1 to 0.2 times (typically 0.15 × S). e ), S e This refers to the nominal area of ​​a single, complete cutting edge of the same type of drill bit; typically, for micro drill bits, this is approximately 0.1 to 0.5 square millimeters. The preset number of cutting edges threshold is taken as 60% of the total number of cutting edges, rounded up. For example, for a four-flute drill bit, S... e When the area is about 0.3 square millimeters, the area threshold is about 0.045 square millimeters. The number of cutting edges threshold is 3. The load sharing and compensation capacity is considered sufficient only when more than 3 of the 4 cutting edges are intact.

[0055] Specifically, the excess range of the single-edge chipping area A relative to the area threshold AT is calculated and denoted as the excess ratio R = (A - AT) / AT; simultaneously, the difference between the number of intact edges n and the edge number threshold nT is calculated and denoted as the intact edge difference D = n - nT. A pre-established confidence mapping table is a two-dimensional discrete grid table, using the excess ratio R and the intact edge difference D as row and column indices. Each grid point in the table stores the corresponding confidence value, ranging from 0 to 1. The construction process of this mapping table is as follows: At least 500 confirmed single-edge large-area chipping instances are collected from historical drilling samples. The proportion of correctly judged samples within each grid interval is calculated to the total number of samples in that interval; this proportion is the confidence value of that grid point. When dividing the grid, the excess ratio R is divided at 0.1 intervals, and the intact edge difference D is divided at intervals of one edge. When looking up the table, if the input value falls exactly on a grid point, the value is directly taken; if it falls between grid points, bilinear interpolation is used to calculate the confidence level. For example, for a six-flute drill bit, when the chipped area A of a single flute is 0.054 square millimeters, the area threshold AT is 0.045 square millimeters, the number of intact flutes n is 5, and the number of flutes threshold nT is 3, the ratio R is exceeded. out =(A AT) / AT≈0.2, Perfect edge difference D n =n With nT=2, after looking up the table and interpolating, the confidence level is approximately 0.87.

[0056] S106. Based on the confidence levels of the first and second chipping identification conclusions, evaluate the final chipping identification conclusion of the multi-bladed drill bit and output it as the judgment result of the chipping state of the multi-bladed drill bit.

[0057] The confidence levels of the first and second chipping identification conclusions are obtained. These two confidence levels are then grouped together according to their corresponding chipping state categories to obtain a confidence comparison record. This record retains the confidence values ​​for both multi-edge micro-chipping and single-edge large-area chipping states. The difference between the two confidence values ​​in the comparison record is compared to a preset difference threshold. If the difference exceeds the threshold, the chipping state corresponding to the higher confidence level is considered the dominant conclusion. If the difference does not exceed the threshold, both chipping states are included in the arbitration conclusion. The arbitration is based on the ratio of the number of chipping edge numbers to the number of intact edges. When the ratio of the number of edge numbers to the total number of multi-edge drill bit edges exceeds a preset threshold, the two states are merged into a multi-edge micro-chipping state; otherwise, they are merged into a single-edge large-area chipping state, resulting in a state merging result. The state merging result is used as the final chipping identification conclusion for the multi-bladed drill bit, and the final chipping identification conclusion is output as the discrimination result of the chipping state of the multi-bladed drill bit.

[0058] In the high-speed drilling process of ultra-fine grain cemented carbide multi-blade micro circuit board drill bit, the two abnormal states of multi-blade micro-chipping and single-blade large-area chipping often exhibit partially overlapping characteristics in the spindle current. The first chipping identification conclusion and the second chipping identification conclusion obtained in the previous step may both have non-zero confidence. Therefore, it is necessary to comprehensively evaluate the two identification conclusions in order to give a clear final chipping identification conclusion.

[0059] It should be noted that the first chipping identification conclusion uses the multi-blade micro-chipping state as its content, while the second chipping identification conclusion uses the single-blade large-area chipping state as its content. Specifically, the first chipping identification conclusion identifies a multi-blade micro-chipping state when the comprehensive uniformity index U of the inter-blade resistance obtained from the spindle current spectrum exceeds the resistance threshold T, and the number of chipped blade positions exceeds a preset blade number threshold. The second chipping identification conclusion identifies a single-blade large-area chipping state when the single-blade chipping area obtained from the spindle current harmonic amplitude and sidefrequency inversion exceeds an area threshold, and the number of intact blades exceeds a blade number threshold. Both conclusions use the spindle current spectrum as the sole signal source, obtained from two channels: the uniformity of inter-blade resistance and the single-blade chipping area. In one implementation, the confidence level of the first chipping blade identification conclusion is denoted as C1, and the confidence level of the second chipping blade identification conclusion is denoted as C2. C1 and C2 are grouped together into a set of binary data according to their corresponding chipping blade state categories. The binary data is the confidence level comparison record. The confidence level comparison record retains the confidence level values ​​of the multi-blade micro-chipping state and the single-blade large-area chipping state, respectively.

[0060] Specifically, C1 and C2 are read from the confidence comparison record, and their absolute difference D = |C1 - C2| is calculated, where D represents the degree of difference in confidence between the two collapse states. A preset difference threshold D is then set. T The preset difference threshold D T The threshold value is calculated from the 10th to 25th percentiles of the statistical distribution of confidence differences in historical drilling samples that clearly indicate a single chipping condition. The 15th percentile is preferred, with a typical range of 0.15 to 0.25. When the difference D exceeds the preset difference threshold D... T When the two identification conclusions show a significant difference in confidence, the one with higher confidence can be directly used as the output basis; when the difference D does not exceed the preset difference threshold D T When the two identification conclusions are close in confidence, additional information needs to be introduced for arbitration.

[0061] Preferably, if the difference D exceeds the preset difference threshold D T The magnitudes of C1 and C2 are compared from the confidence comparison records, and the broken blade state corresponding to the one with higher confidence is taken as the dominant conclusion.

[0062] For example, when C1 is greater than C2 and the difference exceeds the threshold, the multi-blade micro-collapse state is taken as the dominant conclusion; when C2 is greater than C1 and the difference exceeds the threshold, the single-blade large-area collapse state is taken as the dominant conclusion.

[0063] In one embodiment, if the difference D does not exceed the preset difference threshold D T The multi-edge micro-chipping state and the single-edge large-area chipping state are both included in the arbitration conclusion. The arbitration is based on the chipping distribution location and the number of intact edges determined in the previous stage. The proportion ρ of the number of edge position numbers in the chipping distribution location to the total number of multi-edge drill bit edges is calculated, and the proportion ρ is compared with a preset proportion threshold ρ. T The preset proportion threshold ρ is compared. T Take half of the total number of cutting edges of a multi-bladed drill bit. If the percentage ρ exceeds the preset percentage threshold ρ... T This indicates that there are many blade positions with load deviations, and the arbitration conclusion is classified into a multi-blade micro-collapse state; if the proportion ρ does not exceed the preset proportion threshold ρ T This indicates that only a small number of blade positions have load deviations, and the arbitration conclusions are merged into a single blade large-area breakage state, thus obtaining the state merging result.

[0064] It is understandable that the state merging result obtained through either the dominant conclusion path or the arbitration path represents a unique determination of the current chipping state of the multi-flute drill bit. Using this state merging result as the final chipping identification conclusion for the multi-flute drill bit, the final chipping identification conclusion, along with its corresponding confidence level, is output. This final chipping identification conclusion is the judgment result of the chipping state of the multi-flute drill bit.

[0065] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to any specific implementation. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A method for identifying the chipping condition of ultrafine-grained cemented carbide circuit board drill bits, characterized in that, The method includes: The spindle current signal of the spindle motor in a single rotation cycle is acquired, the spindle current signal is frequency domain converted, and the amplitude and sideband components of each frequency component in the current spectrum are extracted to obtain a current spectrum record reflecting the cutting process of the multi-blade drill bit. The amplitude at the fundamental frequency of the spindle rotation and the amplitude at each harmonic are extracted from the current spectrum record. The position of each harmonic characterizes the strength of the cutting load on the corresponding cutting edge, and the harmonic amplitude record of the cutting load of each cutting edge is obtained. Based on the sideband components in the current spectrum record, the asymmetry of the cutting edge load torque within one rotation is evaluated, the chipping distribution location is extracted, and the asymmetry and the harmonic amplitude record are converted into the uniformity of cutting resistance of each blade and the load sharing pattern among multiple blades. The uniformity of cutting resistance of each cutting edge is evaluated by a preset resistance threshold. When the uniformity exceeds the resistance threshold, the presence of micro-chipping of the multi-edge drill bit is identified, and the first chipping identification conclusion and its confidence level are obtained. Based on the location of the chipping edge distribution, the number of intact edges and the chipping area of ​​a single edge are extracted from the load sharing pattern between the harmonic amplitude record and the multi-edge. When the chipping area of ​​a single edge exceeds a preset area threshold and the number of intact edges exceeds a preset number of edges threshold, the multi-edge drill bit is identified as being in a state of large-area chipping of a single edge, and a second chipping edge identification conclusion and its confidence level are obtained. The confidence levels of the first and second chipping identification conclusions are combined to determine the final chipping identification conclusion of the multi-bladed drill bit, which is then output as the discrimination result of the chipping state of the multi-bladed drill bit.

2. The method according to claim 1, characterized in that, The acquisition of the spindle current signal within a single rotation cycle of the spindle motor includes: A Hall current sensor is connected to the power supply circuit of the spindle motor. The spindle current signal is continuously sampled at a sampling frequency higher than the spindle rotation frequency. The sampled data is truncated according to the single-turn trigger pulse output by the spindle encoder to obtain the current timing segment aligned with the single-turn rotation cycle of the spindle. Low-pass filtering is used to suppress high-frequency noise and power frequency interference in the current timing segment to obtain the denoised current timing segment.

3. The method according to claim 1, characterized in that, The step of performing frequency domain conversion on the spindle current signal, extracting the amplitude and sideband components of each frequency component in the current spectrum, and obtaining a current spectrum record reflecting the cutting process of the multi-blade drill bit includes: The spindle current signal is weighted at both ends of the sampling window using a Hanning window, and converted from the time domain to the frequency domain using a fast Fourier transform. The position of the spindle rotation fundamental frequency and the position of the harmonics located at integer multiples of the fundamental frequency are read. The modulation sideband components are identified within the frequency search interval on both sides of the fundamental frequency and the harmonics. The frequency offset and amplitude values ​​of the sideband components relative to the fundamental frequency are extracted. The fundamental frequency amplitude, harmonic amplitude, sideband offset and sideband amplitude are collected in order of frequency from low to high to obtain the current spectrum record.

4. The method according to claim 1, characterized in that, The amplitude at the fundamental frequency of spindle rotation and the amplitude at each harmonic are extracted from the current spectrum record. Based on the position of each harmonic, the strength of the cutting load on the corresponding cutting edge is characterized, resulting in a harmonic amplitude record for each cutting edge, including: Locate the frequency position of the spindle rotation fundamental spectrum peak from the current spectrum record, sequentially search for each harmonic spectrum peak located at an integer multiple of the fundamental frequency along the frequency axis, read the amplitude values ​​of the fundamental spectrum peak and each harmonic spectrum peak, establish a mapping relationship between the number of each harmonic and the corresponding cutting edge number based on the total number of cutting edges of the multi-blade drill bit, and use the amplitude value of each harmonic to characterize the load strength of the mapped cutting edge during a single-turn cutting process to obtain the harmonic amplitude record.

5. The method according to claim 1, characterized in that, The step of evaluating the asymmetry of the cutting edge load torque during one rotation based on the sideband components in the current spectrum record and extracting the chipping distribution location includes: The sideband components located on both sides of the fundamental frequency and harmonics are read from the current spectrum record. The frequency offset of each sideband component is divided by the main shaft rotation frequency to obtain the modulation order. The modulation order reflects the number of times the cutting edge torque fluctuation occurs in a single revolution. The amplitudes of each sideband component are weighted and summed with the modulation order as the weight to obtain a torque fluctuation measure that describes the torque deviation from the uniform distribution amplitude. The torque fluctuation measure is the degree of asymmetry. To address the phase difference between the sideband components and the fundamental frequency peak, a phase positioning method is used to map each sideband component to an angle range of a single rotation of the spindle. The angle range corresponds one-to-one with the installation orientation of each cutting edge of the multi-blade drill bit on the spindle end face. Based on the phase mapping result, the cutting edge number in the angle range where the torque fluctuation measurement exceeds a preset threshold is found. The cutting edge number is the location of the chipping distribution.

6. The method according to claim 1, characterized in that, The process of converting the asymmetry and harmonic amplitude records into the uniformity of cutting resistance of each cutting edge and the load sharing pattern among multiple cutting edges includes: Based on the location of the chipping edge distribution, the harmonic amplitude values ​​of the corresponding cutting edges are retrieved from the harmonic amplitude record. The harmonic amplitude values ​​of each cutting edge are arranged one by one with the degree of asymmetry according to the cutting edge position number. The range of each cutting edge value in the arrangement result represents the uniformity of the cutting resistance of each cutting edge. The proportion of each cutting edge value to the total value of all cutting edges represents the load sharing pattern among the multiple cutting edges.

7. The method according to claim 1, characterized in that, The process involves evaluating the uniformity of cutting resistance of each cutting edge using a preset resistance threshold. When the uniformity exceeds the resistance threshold, it identifies micro-chipping in the multi-edge drill bit, resulting in a first chipping identification conclusion and its confidence level, including: The range of each cutting edge value is used as the uniformity measure, and the variance of each cutting edge value around its mean is used as the fluctuation measure. The uniformity measure and the fluctuation measure are weighted and superimposed according to a pre-calibrated weight coefficient to obtain a comprehensive uniformity index. When the comprehensive uniformity index exceeds the resistance threshold and the number of cutting edge numbers involved exceeds a preset cutting edge number threshold, multi-edge micro-breakage is identified. The extent to which the comprehensive uniformity index exceeds the resistance threshold and the proportion of the cutting edge number to the total number of cutting edges of the multi-edge drill bit are substituted into a pre-established confidence metric mapping table to retrieve the confidence level of the first breakage identification conclusion.

8. The method according to claim 1, characterized in that... The step of extracting the number of intact blades and the single-blade damage area from the load-sharing pattern between the harmonic amplitude record and the multi-blade, based on the location of the chipping distribution, includes: Read the blade position number included in the blade position number set from the chipped blade distribution location as the chipped blade position. Subtract the number of chipped blade positions from the total number of multi-blade drill bit cutting edges to obtain the set of intact blade position numbers. The total number of blade position numbers in the set of intact blade position numbers is the number of intact blades. Based on the load sharing pattern among the multiple blades, the deviation of the sum of the load proportions of intact blade positions from the uniform sharing benchmark and the attenuation of the damaged blade position relative to the normal cutting amplitude in the harmonic amplitude record are converted according to the pre-calibrated inversion coefficient to obtain the single blade damage area.

9. The method according to claim 1, characterized in that, When the single-edge chipping area exceeds a preset area threshold and the number of intact edges exceeds a preset edge number threshold, the multi-edge drill bit is identified as being in a state of large-area single-edge chipping, and a second chipping identification conclusion and its confidence level are obtained, including: The confidence level of the second broken blade identification conclusion is obtained by substituting the difference between the extent by which the single-blade chipped area exceeds the preset area threshold and the number of intact blades and the preset blade number threshold into a pre-established confidence metric mapping table.

10. The method according to claim 1, characterized in that, The final assessment of the chipping identification conclusion for a multi-flute drill bit, based on the combined confidence levels of the first and second chipping identification conclusions, includes: The confidence levels of the first and second blade breakage identification conclusions are compared and aggregated according to their corresponding blade breakage state categories to obtain a confidence level comparison record. The evaluation is based on the difference between the two confidence values ​​in the confidence comparison record and a preset difference threshold. When the difference exceeds the preset difference threshold, the blade breakage state corresponding to the one with higher confidence is taken as the superior conclusion. When the difference does not exceed the preset difference threshold, the two chipping states are included in the arbitration conclusion. The arbitration conclusion is based on the ratio of the number of blade position numbers in the chipping distribution position to the number of intact blades. When the ratio of the number of blade position numbers to the total number of multi-blade drill bit edges exceeds the preset ratio threshold, they are merged into a multi-blade micro-chipping state; otherwise, they are merged into a single-blade large-area chipping state, thus obtaining the final chipping identification conclusion.

Images