Deep hole drilling numerical control correction method and system based on force sense feedback

Abstract

The invention provides a deep hole drilling numerical control correction method and system based on force sense feedback, and relates to the technical field of machining numerical control, and the method comprises the steps: monitoring the vibration amplitude and frequency spectrum features of the surface of a drill rod, comparing the vibration amplitude and frequency spectrum features with a cutting vibration feature library, and extracting a chip attachment signal strength index; fusing the index with a resultant force time domain signal to generate a data pair, and separating an additional load time domain sequence from a pure cutting force time domain sequence by using an interference stripping algorithm; if the additional load time domain sequence exceeds the threshold value, a cleaning signal is issued, and cuttings are discharged; comparing the pure cutting force time domain sequence with a corresponding reference, calculating a force deviation, and generating a compensation instruction according to a feed compensation strategy; the instruction is executed, machining continues, the hole diameter and guiding deviation are fed back, the correction effect is verified, and parameters are adjusted till the machining precision is met, so that chip accumulation can be accurately recognized, cleaning is triggered, the real cutting force and the additional interference load are effectively separated, the feeding parameters are dynamically compensated, and the deep hole machining precision and stability are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of CNC machining technology, and in particular to a CNC correction method and system for deep hole drilling based on force feedback. Background Technology

[0002] Deep hole drilling, a key process in mechanical manufacturing, is susceptible to various interferences due to its enclosed machining environment and difficulties in chip removal. During drilling, chips tend to accumulate between the drill rod and the hole wall, causing additional friction and vibration, leading to abnormal fluctuations in cutting load. Traditional machining methods lack the ability to monitor chip accumulation in real time, making it difficult to distinguish between the actual cutting force and the additional interference load, resulting in deviations in cutting force measurement.

[0003] Because the actual cutting state cannot be accurately obtained, feed parameter adjustments lack reliable basis, easily leading to problems such as hole diameter deviation and guide misalignment. Furthermore, existing technologies rely heavily on manual intervention or fixed-cycle operations for chip removal, making it difficult to dynamically trigger based on actual chip accumulation. This results in untimely or excessive chip removal, affecting both machining accuracy and production efficiency. In addition, post-machining accuracy verification lacks a closed-loop linkage with the front-end machining process, making real-time error correction difficult. Ultimately, this leads to unstable deep hole machining quality, failing to meet the demands of high-precision manufacturing.

[0004] Therefore, it is necessary to provide a CNC correction method and system for deep hole drilling based on force feedback to solve the above-mentioned technical problems. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a CNC correction method and system for deep hole drilling based on force feedback, which solves the problem of insufficient machining accuracy and poor stability caused by the lack of dynamic monitoring and real-time correction mechanisms in existing deep hole drilling processes.

[0006] The present invention provides a CNC correction method for deep hole drilling based on force feedback, the method comprising: The vibration amplitude and spectral characteristics of the drill pipe surface are monitored by a micro vibration sensor array and compared with a normal cutting vibration feature library to extract the chip adhesion signal intensity index. The chip adhesion signal intensity index is fused with the collected resultant force time domain signal to generate a vibration-force coupling data pair, and a preset interference stripping algorithm is applied to the vibration-force coupling data pair to separate the additional load time domain sequence and the pure cutting force time domain sequence. If the additional load time-domain sequence exceeds a predefined threshold, a self-cleaning trigger signal is sent to drive the multi-blade scraper at the tail of the drill pipe and the directional jet cleaning nozzle to work together to remove the chips accumulated in the deep hole. The pure cutting force time-domain sequence is compared with the ideal cutting force reference and the force deviation is calculated. Based on the force deviation, a local feed compensation command is generated according to the preset feed compensation strategy. The local feed compensation command is executed and processing continues. The hole diameter deviation and guide deviation after processing are fed back to the force separation module through the online measurement device to verify the correction effect and adjust the peeling and compensation parameter set until the processing accuracy target is achieved.

[0007] Preferably, the step of using a micro vibration sensor array to monitor the vibration amplitude and spectral characteristics of the drill pipe surface, and comparing them with a normal cutting vibration characteristic database to extract the chip adhesion signal intensity index, specifically includes: N miniature piezoelectric vibration sensors are arranged at equal intervals along the axial and circumferential directions on the drill pipe shell, and the N miniature piezoelectric vibration sensors are in close contact with the deep hole wall. According to the sampling rate U s With a sampling duration T, the output of each miniature piezoelectric vibration sensor is synchronously sampled using a high-speed analog-to-digital converter to generate a vibration time-domain signal matrix O=[O1(t), ..., O i (t), ..., O I [t], where I represents the total number of vibration frequency domain signals in the vibration time domain signal matrix, i represents the index of the vibration frequency domain signal in the vibration time domain signal matrix, t represents the current time, and the collected vibration time domain signal matrix O is stored in a circular buffer to form a C1×C2 dimensional matrix, C2=U s ×T, U s ≥10kHz; A bandpass filter is used to remove the time-domain signal O of each vibration. i The low-frequency workpiece vibration and high-frequency noise in (t) are used to calculate the filtered vibration time-domain signal O. i The envelope amplitude A of (t) i =max|O i (t)|, for the filtered vibration time-domain signal O i (t) Perform a fast Fourier transform to obtain the amplitude spectrum h i [k], k=1,…,M, where k represents the spectral line index and M represents the total number of spectral lines in the amplitude spectrum. The current spectrum H is generated by aggregating the spectral lines according to their average values. k ; Obtain the normal spectrum from the normal cutting vibration feature library obtained through offline calibration. And calculate the current spectrum H k With the normal spectrum The correlation, i.e., the chip adhesion signal intensity index S; If the decrease in the chip adhesion signal intensity index S exceeds the amplitude threshold, it is determined that there are accumulated chips in the deep hole, and a chip removal request signal is issued.

[0008] Preferably, the current spectrum H k The calculation formula is as follows:

[0009] In the formula, FFT() represents the Fast Fourier Transform operation; The formula for calculating the chip adhesion signal intensity index S is as follows: .

[0010] Preferably, the step of fusing the chip adhesion signal intensity index with the acquired resultant force time-domain signal to generate a vibration-force coupled data pair, and applying a preset interference stripping algorithm to the vibration-force coupled data pair to separate the additional load time-domain sequence and the pure cutting force time-domain sequence, specifically includes: The resultant force time-domain signal F is acquired by a force sensor at the drill pipe spindle end. total (t), the bandpass filter is used to remove the resultant force time-domain signal F. total The low-frequency workpiece vibration and high-frequency noise in (t) are determined according to the sampling rate U. s The time-domain signal F of the resultant force after filtering will be obtained from the sampling duration T. total (t) is cut into equal-length sequences; The chip adhesion signal intensity index S is updated to the chip adhesion vibration index S(t) according to the actual vibration sampling period; The chip adhesion vibration index S(t) and the resultant force time domain signal F total (t) Perform time synchronization alignment to generate the vibration-force coupling data pair; The preset interference stripping algorithm employs a vibration-force coupling model. The vibration-force coupling data is input into the vibration-force coupling model to calculate the time-domain sequence F of the additional load. int (t), and then combined with the resultant force time domain signal F total (t), the time-domain sequence F of the pure cutting force is calculated. cut (t); If F cut (t) is negative or F cut (t) If it exceeds the normal cutting force range, then F cut (t) is marked as an outlier and subjected to amplitude limiting.

[0011] Preferably, the additional load time-domain sequence F int The formula for calculating (t) is as follows:

[0012] In the formula, α represents the additional load scaling factor obtained from offline calibration, β represents the nonlinear correction coefficient obtained from offline fitting, and both α and β are parameters of the vibration-force coupling model; The pure cutting force time-domain sequence F cut The formula for calculating (t) is as follows: .

[0013] Preferably, the step of comparing the pure cutting force time-domain sequence with the ideal cutting force reference and calculating the force deviation, and generating a local feed compensation command based on the force deviation according to a preset feed compensation strategy, specifically includes: At the current machining position, the cutting time T is recorded. w The pure cutting force time-domain sequence F within cut (t), calculate the cutting time T w The pure cutting force time-domain sequence F within cut The time average of (t), i.e., the average cutting force F avg :

[0014] Where, Represents the integral variable; The ideal cutting force reference F, calibrated offline, is retrieved from the process database for the current machining condition. ref Combined with the average cutting force F avg Calculate the force deviation ΔF:

[0015] According to the preset feed compensation strategy, the force deviation ΔF is mapped to the feed rate adjustment amount ΔV. f :

[0016] In the formula, K f This indicates the feed rate compensation gain calibrated offline. If the feed rate adjustment amount ΔV f If the maximum adjustment range is exceeded, the feed rate adjustment amount ΔV is adjusted. f Amplitude limiting is applied; The force deviation ΔF is converted into the cutting depth adjustment amount ΔL according to the proportional relationship: ΔL=-K d ΔF In the formula, K d This indicates the depth of cut compensation gain calibrated offline; If the depth of cut adjustment ΔL exceeds the machining tolerance range, then the depth of cut adjustment ΔL is truncated to the tolerance limit; Insert the original feed segment at the current CNC trajectory point position, and adjust the original feed speed V0 to V0 + ΔV. f The local feed rate command is generated, and the original depth of cut L0 is adjusted to L0+ΔL. The depth of cut increment command is generated, and the local feed rate command and the depth of cut increment command are combined to generate the local feed compensation command.

[0017] Preferably, the step of executing the local feed compensation command and continuing processing involves feeding back the machined hole diameter deviation and guide deviation to the force separation module via an online measuring device to verify the correction effect and adjust the peeling and compensation parameter set until the machining accuracy target is achieved. Specifically, this includes: The local feed compensation command is received and executed, and the diameter D of the machined deep hole is measured by an electromagnetic diameter sensor mounted at the deep hole exit. meas (t), and the offset G of the drill pipe centerline is measured by a laser triangular guide sensor installed at the front end of the drill pipe. meas (t), combined with aperture reference D ref and guiding benchmark G ref Calculate the aperture deviation ΔD(t) = D meas (t)- D ref With guidance deviation ΔG(t) = G meas (t)- G ref : The maximum absolute deviation E is calculated based on the aperture deviation ΔD(t) and the guide deviation ΔG(t). max Obtain the accuracy tolerance R tol If E max ≤R tol If the correction is successful, the correction effect is acceptable; otherwise, the correction effect is unacceptable, and the peeling and compensation parameter set needs to be readjusted to continue the next round of peeling and compensation operations until the processing accuracy target is achieved.

[0018] Preferably, the maximum absolute deviation E max The calculation formula is as follows: .

[0019] A force feedback-based CNC correction system for deep hole drilling, the system comprising: The index extraction module is used to monitor the vibration amplitude and spectral characteristics of the drill pipe surface using a micro vibration sensor array, and compare them with a normal cutting vibration feature library to extract the chip adhesion signal intensity index. The sequence separation module is used to fuse the chip adhesion signal intensity index with the collected resultant force time domain signal to generate a vibration-force coupling data pair, and apply a preset interference stripping algorithm to the vibration-force coupling data pair to separate the additional load time domain sequence and the pure cutting force time domain sequence. The chip cleaning module is used to send a self-cleaning trigger signal if the additional load time domain sequence exceeds a predefined threshold, thereby driving the multi-blade scraper at the tail of the drill pipe and the directional jet cleaning nozzle to work together to remove the chips accumulated in the deep hole. The instruction generation module is used to compare the pure cutting force time-domain sequence with the ideal cutting force reference and calculate the force deviation, and generate a local feed compensation instruction based on the force deviation according to a preset feed compensation strategy. The effect verification module is used to execute the local feed compensation command and continue processing. It feeds back the hole diameter deviation and guide deviation after processing to the force separation module through an online measurement device to verify the correction effect and adjust the peeling and compensation parameter set until the processing accuracy target is achieved.

[0020] Compared with related technologies, the CNC correction method and system for deep hole drilling based on force feedback provided by this invention has the following beneficial effects: This invention utilizes a miniature vibration sensor array to monitor the vibration amplitude and spectral characteristics of the drill pipe surface, compares them with a normal cutting vibration feature library, and extracts the chip adhesion signal intensity index. The chip adhesion signal intensity index is then fused with the acquired resultant force time-domain signal to generate a vibration-force coupled data pair. A preset interference stripping algorithm is applied to the vibration-force coupled data pair to separate the additional load time-domain sequence and the pure cutting force time-domain sequence. If the additional load time-domain sequence exceeds a predefined threshold, a self-cleaning trigger signal is issued, driving the multi-blade scraper at the drill pipe tail and the directional jet cleaning nozzle to work together to peel off and discharge the chip. Chips accumulate inside deep holes; the time-domain sequence of pure cutting force is compared with the ideal cutting force reference and the force deviation is calculated. According to the preset feed compensation strategy, a local feed compensation command is generated based on the force deviation; the local feed compensation command is executed and machining continues. The machined hole diameter deviation and guide deviation are fed back to the force separation module through an online measurement device to verify the correction effect and adjust the peeling and compensation parameter set until the machining accuracy target is achieved. This allows for accurate identification of chip accumulation and triggering of cleaning, effectively separating the real cutting force from the additional interference load, and dynamically compensating the feed parameters, significantly improving the machining accuracy and stability of deep holes.

[0021] This invention effectively solves the accuracy fluctuation problem in deep hole drilling by constructing a complete dynamic correction system. Through a vibration sensor array, it can accurately capture abnormal vibration characteristics caused by chip accumulation, achieving real-time identification of the chip accumulation state and providing a reliable basis for subsequent intervention. This invention uses a vibration-force coupling model to separate the load signal, eliminating additional interference from chip accumulation and accurately extracting the true cutting force, laying the foundation for feed parameter adjustment. When chip accumulation exceeds the limit, the self-cleaning device of this invention can be triggered in time, clearing the accumulated chips through the coordinated action of scrapers and jets, preventing the continuous accumulation of interference. Based on the feed compensation command generated from the actual cutting force deviation, this invention can dynamically adjust the feed rate and cutting depth, correct the machining trajectory in real time, and reduce hole diameter and guide errors. The closed-loop verification mechanism of this invention continuously optimizes the separation parameters and compensation strategy through online measurement feedback results, ensuring stable and compliant machining accuracy, and significantly improving the accuracy consistency and machining efficiency of deep hole drilling. Attached Figure Description

[0022] Figure 1 This is a flowchart of the CNC correction method for deep hole drilling based on force feedback according to the present invention; Figure 2 This is a system block diagram of the CNC correction system for deep hole drilling based on force feedback according to the present invention. Detailed Implementation

[0023] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0024] Example 1

[0025] like Figure 1 As shown, a CNC correction method for deep hole drilling based on force feedback is described, the method comprising: S1. The vibration amplitude and spectral characteristics of the drill pipe surface are monitored by a micro vibration sensor array and compared with the normal cutting vibration characteristic library to extract the chip adhesion signal intensity index. S2, the chip adhesion signal intensity index is fused with the collected resultant force time domain signal to generate a vibration-force coupling data pair, and a preset interference stripping algorithm is applied to the vibration-force coupling data pair to separate the additional load time domain sequence and the pure cutting force time domain sequence; S3, if the additional load time domain sequence exceeds the predefined threshold, a self-cleaning trigger signal is sent to drive the multi-blade scraper at the tail of the drill pipe and the directional jet cleaning nozzle to work together to remove the chips accumulated in the deep hole. S4, compare the pure cutting force time-domain sequence with the ideal cutting force reference and calculate the force deviation, and generate a local feed compensation command based on the force deviation according to the preset feed compensation strategy; S5, execute the local feed compensation command and continue processing. The machined hole diameter deviation and guide deviation are fed back to the force separation module through the online measurement device to verify the correction effect and adjust the peeling and compensation parameter set until the processing accuracy target is achieved.

[0026] In practical applications, miniature vibration sensor arrays can be used to capture the vibration state of the drill pipe surface in real time. These sensors are closely attached to the contact surface between the drill pipe and the borehole wall, allowing for comprehensive acquisition of vibration amplitude and spectral characteristics. By comparing and analyzing the acquired vibration information with a pre-built database of normal cutting vibration characteristics, the chip adhesion signal intensity index can be accurately extracted. This index quantifies the degree of chip accumulation between the drill pipe and the borehole wall, providing a reliable basis for subsequent intervention measures and enabling early identification of chip accumulation status.

[0027] Then, the chip adhesion signal intensity index and the resultant force time-domain signal collected by the spindle end force sensor can be fused to form a vibration-force coupled data pair. Based on a preset interference removal algorithm, this coupled data pair is deeply analyzed to separate two types of key sequences: the time-domain sequence of additional loads caused by chip accumulation and the time-domain sequence of pure cutting force that truly reflects the cutting state. This effectively eliminates the interference load caused by chip accumulation, ensuring that subsequent parameter adjustments are based solely on the real cutting force, laying the foundation for accurate compensation.

[0028] When the time-domain sequence of the additional load exceeds a predefined threshold, the self-cleaning mechanism is immediately activated. A self-cleaning trigger signal is sent to drive the multi-bladed scraper at the drill pipe tail and the directional jet cleaning nozzle to work in tandem. The scraper mechanically removes the attached chips, while the cleaning nozzle further removes residual chips through directional jetting and discharges them out of the hole, preventing further processing interference caused by continuous chip accumulation and ensuring the stability of the processing environment.

[0029] For example, through 50 sets of normal machining experiments, the maximum natural fluctuation value of the additional load was measured to be 90N, with an upper limit of 100N for the 95% confidence interval. To eliminate interference from material impurities, slight tool vibrations, etc., and to reserve a safety margin of 20N, the predefined threshold was set to 120N to accurately identify load anomalies caused by chip accumulation, ensuring that the self-cleaning mechanism is activated promptly and without error. During normal cutting, the time-domain sequence of the additional load fluctuates slightly, such as {42N, 38N, 45N}; when chips accumulate, the sequence value increases significantly, such as {135N, 148N, 155N}, all exceeding the predefined threshold of 120N, requiring immediate activation of the self-cleaning mechanism.

[0030] Furthermore, the separated pure cutting force time-domain sequence can be compared with the offline calibrated ideal cutting force reference to accurately calculate the force deviation between the two. Based on the preset feed compensation strategy, the force deviation is converted into specific local feed compensation commands, covering the dynamic adjustment of feed rate and depth of cut. By correcting the machining trajectory parameters in real time, the hole diameter deviation and guide offset caused by cutting force fluctuations can be effectively offset, improving machining accuracy.

[0031] Finally, a feed compensation command can be executed to continue machining, while the hole diameter deviation and guide deviation data after machining are collected in real time by an online measuring device and fed back to the force separation module. The correction effect is verified based on the feedback results. If the deviation does not meet the accuracy requirements, the interference stripping parameters and compensation parameter set are dynamically adjusted, and the aforementioned correction process is repeated. Through continuous iterative optimization, until the machining accuracy reaches the preset target, a complete dynamic correction closed loop is formed, ensuring high precision and stability in deep hole machining.

[0032] In the specific implementation process, the method of using a micro vibration sensor array to monitor the vibration amplitude and spectral characteristics of the drill pipe surface, and comparing them with a normal cutting vibration characteristic library to extract the chip adhesion signal intensity index, specifically includes: N miniature piezoelectric vibration sensors are arranged at equal intervals along the axial and circumferential directions on the drill pipe shell, and the N miniature piezoelectric vibration sensors are in close contact with the deep hole wall. According to the sampling rate U s With a sampling duration T, the output of each miniature piezoelectric vibration sensor is synchronously sampled using a high-speed analog-to-digital converter to generate a vibration time-domain signal matrix O=[O1(t), ..., O i (t), ..., O I [(t)], where I represents the total number of vibration frequency domain signals in the vibration time domain signal matrix, i represents the index of the vibration frequency domain signal in the vibration time domain signal matrix, t represents the current time, F s The kHz value is ≥10, and the acquired vibration time-domain signal matrix O is stored in a circular buffer to form a C1×C2 dimensional matrix, where C2=F. s ×T; A bandpass filter is used to remove the time-domain signal O of each vibration. i The low-frequency workpiece vibration and high-frequency noise in (t) are used to calculate the filtered vibration time-domain signal O. i The envelope amplitude A of (t) i =max|O i (t)|, for the filtered vibration time-domain signal O i (t) Perform a fast Fourier transform to obtain the amplitude spectrum h i [k], k=1,…,M, where k represents the spectral line index and M represents the total number of spectral lines in the amplitude spectrum. The current spectrum H is generated by aggregating the spectral lines according to their average values.k ; Obtain the normal spectrum from the normal cutting vibration feature library obtained through offline calibration. And calculate the current spectrum H k With the normal spectrum The correlation, i.e., the chip adhesion signal intensity index S; If the decrease in the chip adhesion signal intensity index S exceeds the amplitude threshold, it is determined that there are accumulated chips in the deep hole, and a chip removal request signal is issued.

[0033] The current spectrum H k The calculation formula is as follows:

[0034] In the formula, FFT() represents the Fast Fourier Transform operation; The formula for calculating the chip adhesion signal intensity index S is as follows: .

[0035] Understandably, miniature piezoelectric vibration sensors can be evenly spaced along the axial and circumferential directions of the drill pipe casing, with all sensors tightly fitted to the contact surface of the deep hole wall. This arrangement ensures that the sensors can comprehensively capture the local vibration responses caused by friction between the drill pipe and the hole wall, chip compression, and other behaviors, providing multi-dimensional raw vibration information for subsequent signal analysis and guaranteeing the spatial coverage and signal authenticity of the monitoring.

[0036] Then, using preset sampling parameters as a reference, the output signals of each miniature piezoelectric vibration sensor are synchronously acquired through a high-speed analog-to-digital converter. During the acquisition process, a stable sampling frequency and duration must be maintained. The acquired multi-channel continuous vibration time-domain signals are integrated into a matrix form and stored in a ring-shaped buffer area. This buffer area has dynamic update capabilities, which can retain the latest vibration signal data in real time, providing continuous and complete raw information support for subsequent processing, and ensuring the timeliness and continuity of signal analysis.

[0037] Furthermore, bandpass filtering technology can be used to purify the vibration time-domain signal, filtering out low-frequency overall workpiece vibration interference and high-frequency electromagnetic noise, thereby improving the signal-to-noise ratio. The envelope amplitude of the filtered signal is then calculated, which quantifies the intensity of the vibration impact. Simultaneously, a fast Fourier transform is performed on the filtered signal to convert the time-domain signal into a frequency-domain amplitude spectrum. Then, spectral features under the current processing state are generated through spectral line averaging and aggregation. This achieves the goal of extracting multi-dimensional features in both the time and frequency domains from the vibration signal, laying a quantitative foundation for identifying chip accumulation states.

[0038] In addition, a normal cutting vibration feature library pre-built through offline calibration can be accessed to extract corresponding normal spectral features. This library requires offline calibration, meaning that benchmark tests are conducted under stable machining conditions without chip buildup, targeting the specific materials and tools used in the machining process, as different materials and tools exhibit different cutting vibration characteristics. Furthermore, the calibration environment must be consistent with the real-time machining environment, including using the same machine tool, cutting parameters, and cooling conditions, to eliminate environmental interference with the vibration signal. Normal spectral features refer to the vibration spectrum characteristics during stable cutting, obtained through repeated experiments and statistical analysis. They reflect the vibration state without chip buildup, and their spectral shape and main frequency components are significantly different from the abnormal vibrations under chip buildup conditions, serving as a benchmark for determining whether the cutting state is normal.

[0039] By calculating the correlation between the spectrum under the current machining state and the normal spectrum, this correlation is defined as the chip adhesion signal intensity index. This index essentially reflects the degree of agreement between the current vibration characteristics and the normal cutting state. Its value change can intuitively reflect the degree of interference of chip accumulation on vibration characteristics. That is, the more severe the chip accumulation, the more significant the deviation between the current spectrum and the normal spectrum, and the more obvious the decrease in the index value.

[0040] If the decrease in this indicator exceeds the preset amplitude threshold, it indicates that the current vibration characteristics have significantly deviated from the normal cutting state. Based on this, it can be determined that there is chip accumulation inside the deep hole, and a chip removal request signal is issued. This provides a decision basis for the subsequent activation of the self-cleaning device, ensuring that the chip accumulation problem can be detected and dealt with in a timely manner, thereby maintaining the stability of the deep hole drilling process.

[0041] For example, through offline calibration experiments, under stable machining conditions without chip accumulation, the fluctuation range of the chip adhesion signal intensity index of 50 sets of normal data was statistically analyzed, and the maximum natural fluctuation amplitude did not exceed 15%. To avoid false triggering by normal fluctuations, a 5% safety margin was reserved, so the amplitude threshold was set to 20%. This threshold can cover normal interference such as slight differences in material hardness and minor tool wear, and chip removal is only triggered when chip accumulation causes a significant deviation in vibration characteristics. During normal cutting, the chip adhesion signal intensity index is stable between 0.85 and 0.95. If the current chip adhesion signal intensity index drops sharply from 0.9 to 0.65, the decrease is approximately 27.8%, which is significantly greater than the amplitude threshold of 20%, indicating that the current vibration characteristics have significantly deviated from the normal cutting state.

[0042] The above methods enable real-time and accurate identification of chip accumulation, providing crucial pre-monitoring support for dynamic correction in deep hole drilling.

[0043] The process of fusing the chip adhesion signal intensity index with the acquired resultant force time-domain signal to generate a vibration-force coupled data pair, and applying a preset interference stripping algorithm to the vibration-force coupled data pair to separate the additional load time-domain sequence and the pure cutting force time-domain sequence, specifically includes: The resultant force time-domain signal F is acquired by a force sensor at the drill pipe spindle end. total (t), the bandpass filter is used to remove the resultant force time-domain signal F. total The low-frequency workpiece vibration and high-frequency noise in (t) are determined according to the sampling rate U. s The time-domain signal F of the resultant force after filtering will be obtained from the sampling duration T. total (t) is cut into equal-length sequences; The chip adhesion signal intensity index S is updated to the chip adhesion vibration index S(t) according to the actual vibration sampling period; The chip adhesion vibration index S(t) and the resultant force time domain signal F total (t) Perform time synchronization alignment to generate the vibration-force coupling data pair; The preset interference stripping algorithm employs a vibration-force coupling model. The vibration-force coupling data is input into the vibration-force coupling model to calculate the time-domain sequence F of the additional load. int (t), and then combined with the resultant force time domain signal F total (t), the time-domain sequence F of the pure cutting force is calculated. cut (t); If F cut (t) is negative or F cut (t) If it exceeds the normal cutting force range, then F cut (t) is marked as an outlier and subjected to amplitude limiting.

[0044] The additional load time-domain sequence F int The formula for calculating (t) is as follows:

[0045] In the formula, α represents the additional load scaling factor obtained from offline calibration, β represents the nonlinear correction coefficient obtained from offline fitting, and both α and β are parameters of the vibration-force coupling model; The pure cutting force time-domain sequence F cut The formula for calculating (t) is as follows: .

[0046] First, the resultant force time-domain signal can be acquired by a force sensor deployed at the drill pipe spindle end. This signal contains all force information during the cutting process and needs to be bandpass filtered to remove low-frequency workpiece vibration interference and high-frequency electromagnetic noise, ensuring signal purity. Then, according to the sampling frequency and duration consistent with the vibration signal, the filtered resultant force time-domain signal is divided into equal-duration sequences, ensuring that the force signal matches the vibration signal on the time scale, laying the foundation for subsequent fusion analysis.

[0047] Then, the extracted chip adhesion signal intensity index can be updated in real time according to the sampling period of the actual vibration signal, transforming it into a chip adhesion vibration index that changes continuously over time. This transformation endows the originally static intensity index with dynamic characteristics, enabling it to reflect the chip accumulation state at different times in real time, and providing the possibility for correlation analysis with dynamically changing force signals.

[0048] Next, timestamp calibration technology can be used to precisely match the dynamic chip adhesion vibration index with the preprocessed resultant force time-domain signal in the time dimension, ensuring that the vibration state index and force signal at the same moment form a one-to-one vibration-force coupled data pair. This strict synchronization in time can effectively avoid correlation deviations caused by signal delay or misalignment, ensuring the reliability of subsequent model analysis.

[0049] It should be noted that the core of the preset interference removal algorithm is the application of the vibration-force coupling model. After inputting the vibration-force coupling data into the vibration-force coupling model, the model is constructed based on the physical correlation laws calibrated offline. It can separate the time-domain sequence of additional loads caused by chip accumulation through quantitative calculation. These additional loads are interference components and are not the forces required for actual cutting. By combining the original resultant force time-domain signal, the pure cutting force time-domain sequence reflecting the actual cutting state can be obtained through force component decomposition, thus achieving accurate separation of interference from effective signals.

[0050] The offline calibration of the vibration-force coupling model must be performed in a stable machining environment to ensure consistency with actual machining conditions, including identical machine tools, materials, cutting tools, and cutting parameters, avoiding environmental interference. Vibration signals from the vibration sensor array and the resultant force signal from the spindle end are simultaneously acquired under conditions of no chip buildup and varying degrees of chip buildup. The acquired vibration-force data is preprocessed to remove noise, and the correlation between vibration characteristics and the additional load is analyzed. Through repeated experiments, vibration-force coupling model parameters reflecting the relationship between the two are fitted, and the accuracy of the parameters is verified to ensure that the vibration-force coupling model can accurately quantify the mapping relationship between vibration and additional load. Finally, stable model parameters are determined, providing a basis for online separation of additional load and pure cutting force.

[0051] If the separated pure cutting force time-domain sequence contains negative values, or if its value exceeds the pre-calibrated normal cutting force range, it is considered an outlier. In this case, the outlier needs to be limited to a reasonable range to prevent abnormal data caused by sensor errors, model calculation deviations, etc., from affecting the generation of subsequent feed compensation commands, thus ensuring the robustness of the entire correction system.

[0052] Therefore, the above process, through deep fusion and modeling analysis of multiple signals, successfully eliminated the interference load caused by chip accumulation, providing accurate data support for subsequent parameter compensation based on real cutting forces.

[0053] The step of comparing the pure cutting force time-domain sequence with the ideal cutting force reference and calculating the force deviation, and generating a local feed compensation command based on the force deviation according to a preset feed compensation strategy, specifically includes: At the current machining position, the cutting time T is recorded. w The pure cutting force time-domain sequence F within cut (t), calculate the cutting time T w The pure cutting force time-domain sequence F within cut The time average of (t), i.e., the average cutting force F avg :

[0054] Where, Represents the integral variable; The ideal cutting force reference F, calibrated offline, is retrieved from the process database for the current machining condition. ref Combined with the average cutting force F avg Calculate the force deviation ΔF:

[0055] According to the preset feed compensation strategy, the force deviation ΔF is mapped to the feed rate adjustment amount ΔV. f :

[0056] In the formula, K f This indicates the feed rate compensation gain calibrated offline. If the feed rate adjustment amount ΔV f If the maximum adjustment range is exceeded, the feed rate adjustment amount ΔV is adjusted. f Amplitude limiting is applied; The force deviation ΔF is converted into the cutting depth adjustment amount ΔL according to the proportional relationship: ΔL=-K d ΔF In the formula, K dThis indicates the depth of cut compensation gain calibrated offline; If the depth of cut adjustment ΔL exceeds the machining tolerance range, then the depth of cut adjustment ΔL is truncated to the tolerance limit; Insert the original feed segment at the current CNC trajectory point position, and adjust the original feed speed V0 to V0 + ΔV. f The local feed rate command is generated, and the original depth of cut L0 is adjusted to L0+ΔL. The depth of cut increment command is generated, and the local feed rate command and the depth of cut increment command are combined to generate the local feed compensation command.

[0057] In practical applications, for the current machining position, a time-domain sequence of pure cutting force within a set cutting duration is extracted. By averaging the pure cutting force signal over time within this period, the average cutting force reflecting the overall cutting state of the machining segment is obtained. This average value can smooth out instantaneous force fluctuations, objectively characterize the force level of the current cutting process, provide a stable quantitative basis for subsequent deviation analysis, and avoid misjudgments caused by sudden changes in instantaneous signals.

[0058] Then, the ideal cutting force benchmark matching the current machining condition can be retrieved from the process database. The current machining condition includes material properties, hole diameter specifications, etc. The ideal cutting force benchmark is obtained through offline calibration, which must be performed in a stable machining environment without chip accumulation. Specifically, first, the specific machining condition is determined, including parameters such as the material to be machined, the type and specifications of the cutting tool used, and the preset hole diameter. Second, under the same machine tool and cooling conditions, cutting operations are performed according to standard machining parameters to ensure no additional interference. Finally, cutting force data is continuously collected using a force sensor. After eliminating abnormal fluctuations, the cutting force characteristics in the stable phase are statistically analyzed. Through repeated experiments, the cutting force range that enables high-precision machining is extracted as the benchmark and stored in the process database, forming the ideal cutting force benchmark matching the current working condition. The extracted average cutting force is compared with the ideal cutting force benchmark; the difference between the two is the force deviation. Its magnitude directly reflects the degree of deviation between the current cutting state and the ideal state and is the core indicator for triggering compensation adjustments.

[0059] Based on a preset feed compensation strategy, force deviation is first mapped to a feed rate adjustment. This mapping process relies on an offline calibrated feed rate compensation gain, which quantifies the correlation between force deviation and feed rate adjustment, ensuring that the adjustment amount matches the degree of deviation. If the calculated feed rate adjustment exceeds the preset maximum adjustment range, a limiting process is required to constrain it within the safe range allowed by the process, avoiding cutting stability issues caused by over-adjustment.

[0060] Simultaneously, force deviation can be converted into cutting depth adjustment according to a proportional relationship. By establishing a quantitative correlation between force deviation and cutting depth correction through offline calibrated cutting depth compensation gain, precise control of the cutting section dimensions can be achieved. If this adjustment exceeds the machining tolerance range, it needs to be truncated to the upper or lower tolerance limit to ensure that the correction range does not exceed the allowable error range of the part's precision requirements, thus guaranteeing the compliance of the machining dimensions.

[0061] Finally, an adjusted feed segment can be inserted at the current CNC trajectory point position. The original feed rate is superimposed with the feed rate adjustment to generate a local feed rate command, and the original depth of cut is superimposed with the depth of cut adjustment to generate a depth of cut increment command. By combining these two types of commands, a complete local feed compensation command can be formed. This command can then be directly sent to the actuator to dynamically correct the feed parameters of the current machining segment, ensuring that the cutting process corrects itself from deviations from the ideal state in a timely manner, ultimately guaranteeing the consistency of accuracy in deep hole machining.

[0062] By accurately quantifying force deviation and dynamically adapting parameters, a closed-loop link from sensing the cutting state to executing corrections was constructed, thereby achieving adaptive control of the deep hole drilling process.

[0063] The process involves executing the local feed compensation command and continuing machining. The machined hole diameter deviation and guide deviation are fed back to the force separation module via an online measurement device to verify the correction effect and adjust the peeling and compensation parameter set until the machining accuracy target is achieved. Specifically, this includes: The local feed compensation command is received and executed, and the diameter D of the machined deep hole is measured by an electromagnetic diameter sensor mounted at the deep hole exit. meas (t), and the offset G of the drill pipe centerline is measured by a laser triangular guide sensor installed at the front end of the drill pipe. meas (t), combined with aperture reference D ref and guiding benchmark G ref Calculate the aperture deviation ΔD(t) = D meas (t)- D ref With guidance deviation ΔG(t) = G meas (t)- G ref : The maximum absolute deviation E is calculated based on the aperture deviation ΔD(t) and the guide deviation ΔG(t). max Obtain the accuracy tolerance R tol If E max ≤R tol If the correction is successful, the correction effect is acceptable; otherwise, the correction effect is unacceptable, and the peeling and compensation parameter set needs to be readjusted to continue the next round of peeling and compensation operations until the processing accuracy target is achieved.

[0064] The maximum absolute deviation Emax The calculation formula is as follows: .

[0065] First, it can receive and execute local feed compensation commands, driving the actuator to continue the deep hole machining process according to the adjusted feed speed and cutting depth. Thus, through precise control of mechanical actions, theoretical parameter compensation can be transformed into actual machining trajectory correction, providing physical machining results for subsequent effect verification.

[0066] Then, high-precision sensing devices can be used to achieve real-time monitoring of machining quality. Specifically, the electromagnetic diameter sensor installed at the deep hole exit can non-contactly measure the actual hole diameter after machining, capturing subtle changes in hole diameter; the laser triangulation sensor installed at the front end of the drill rod is used to monitor the spatial offset of the drill rod centerline, reflecting the straightness deviation of the machining trajectory. The two types of sensors work together to obtain machining result data from both dimensional accuracy and geometrical accuracy dimensions, providing a direct basis for deviation calculation.

[0067] Furthermore, preset aperture and guide references can be retrieved from the process standards. Subtracting the actual measured aperture value from the aperture reference yields the aperture deviation, which characterizes the dimensional deviation. Subtracting the actual measured centerline offset from the guide reference yields the guide deviation, which characterizes the trajectory deviation. These two types of deviations quantify the degree of deviation between the current processing result and the ideal state, and are the core indicators for evaluating the correction effect.

[0068] It should be noted that the maximum absolute deviation can be determined by comprehensively considering the absolute values ​​of the aperture deviation and the guide deviation, taking the maximum value as the overall deviation evaluation index to fully reflect the worst-case scenario for machining accuracy. This index can then be compared with a preset accuracy tolerance, which is the maximum allowable deviation range. If the maximum absolute deviation is within the accuracy tolerance, the current correction effect is considered acceptable, and the machining accuracy meets the expected target. If it exceeds the accuracy tolerance, the correction effect is considered unacceptable, and the parameter adjustment mechanism needs to be activated.

[0069] Specifically, a parameter adjustment algorithm can be invoked to recalibrate the stripping and compensation parameter sets. The stripping parameters are the core parameters used to separate the additional load from the pure cutting force, mainly including the additional load scaling factor α and the nonlinear correction coefficient β in the vibration-force coupling model. Here, α is the offline-calibrated additional load scaling factor, with a value range of [0.1, 0.5]; β is the nonlinear correction coefficient obtained from offline fitting, used to optimize the nonlinear mapping relationship between vibration and additional load. The compensation parameters refer to the feed rate compensation gain K in the feed compensation strategy. f and cutting depth compensation gain K d K fThe value range is [0.01-0.01], and the unit is millimeters per Newton-second, K. d The value range is [0.001-0.01], and the unit is millimeters per Newton. When verification fails, i.e., the maximum absolute deviation E... max Exceeding the accuracy tolerance R tol The specific adjustment logic is as follows: For the peeling parameters, if the pure cutting force is abnormal, increasing α can enhance the peeling force against chip interference; if the separation nonlinearity deviation is large, β can be corrected to optimize the mapping relationship between vibration and additional load, thereby reducing the separation error. For compensation parameters, if the aperture deviation is large, increase K. d To increase the depth of cut correction range; if the guide deviation is large, increase K. f This corrects the deviation in the feed direction path and reduces the guidance error.

[0070] By iteratively adjusting the parameter set, the stripping and compensation process is re-executed until Emax ≤ Rtol. The stripping parameters directly affect the separation accuracy of the additional load and pure cutting force; inaccurate separation will lead to unreliable feed compensation. The compensation parameters determine the mapping relationship between force deviation and feed rate / depth of cut adjustment; parameter mismatch will prevent corrective measures from accurately offsetting the deviation. By specifically adjusting these two types of parameters, the accuracy of load separation and the adaptability of feed compensation can be optimized.

[0071] After parameter adjustments, a new round of iterative verification is required. First, the adjusted set of stripping and compensation parameters can be applied to the next machining cycle. The local feed compensation command is re-executed, and new aperture and guide deviations are simultaneously collected using an online measuring device. The maximum absolute deviation is then recalculated and compared with the accuracy tolerance. This process is repeated until the maximum absolute deviation is controlled within the accuracy tolerance range, and the machining accuracy reaches the preset target. This continuous iterative optimization method can dynamically adapt to complex variables during the machining process, such as material hardness fluctuations and changes in chip accumulation characteristics, ensuring that the correction strategy always matches the actual machining conditions.

[0072] By using the above methods, machining deviations can be continuously eliminated, ensuring that the accuracy of deep hole drilling gradually converges from the deviation state to the qualified range, thus guaranteeing the stability of deep hole machining quality.

[0073] Example 2

[0074] like Figure 2 As shown, a CNC correction system for deep hole drilling based on force feedback is provided, the system comprising: The index extraction module is used to monitor the vibration amplitude and spectral characteristics of the drill pipe surface using a micro vibration sensor array, and compare them with a normal cutting vibration feature library to extract the chip adhesion signal intensity index. The sequence separation module is used to fuse the chip adhesion signal intensity index with the collected resultant force time domain signal to generate a vibration-force coupling data pair, and apply a preset interference stripping algorithm to the vibration-force coupling data pair to separate the additional load time domain sequence and the pure cutting force time domain sequence. The chip cleaning module is used to send a self-cleaning trigger signal if the additional load time domain sequence exceeds a predefined threshold, thereby driving the multi-blade scraper at the tail of the drill pipe and the directional jet cleaning nozzle to work together to peel off and discharge the chips accumulated in the deep hole. The instruction generation module is used to compare the pure cutting force time-domain sequence with the ideal cutting force reference and calculate the force deviation, and generate a local feed compensation instruction based on the force deviation according to a preset feed compensation strategy. The effect verification module is used to execute the local feed compensation command and continue processing. It feeds back the hole diameter deviation and guide deviation after processing to the force separation module through an online measurement device to verify the correction effect and adjust the peeling and compensation parameter set until the processing accuracy target is achieved.

[0075] The module comprises several components: a micro-vibration sensor array, magnetically fixed to the drill pipe surface, with a sampling frequency of 2kHz, for monitoring vibration amplitude and spectral characteristics and comparing them with a normal cutting vibration feature library to extract chip adhesion signal intensity indicators; a sequence separation module equipped with a PLC featuring a preset interference stripping algorithm, receiving vibration and force signals via an EtherCAT bus, fusing them to generate vibration-force coupled data pairs, and separating the time-domain sequences of additional load and pure cutting force; a chip cleaning module including a high-speed steel multi-blade scraper, keyed to the drill pipe tail, and a brass directional jet cleaning nozzle threaded behind the scraper, with a jet pressure of 0.6-0.8MPa and controlled by a solenoid valve; a command generation module based on a Siemens 828D CNC system, generating local feed compensation commands according to a preset feed compensation strategy; and an effect verification module equipped with a laser diameter gauge, feeding back the online measured aperture and guide deviation to the sequence separation module via an RS485 interface, and adjusting the stripping and compensation parameter sets in a closed loop until the target is met.

[0076] Through the above embodiments, this invention, through a force feedback-based CNC correction method and system for deep hole drilling, utilizes a micro-vibration sensor array to monitor the vibration amplitude and spectral characteristics of the drill rod surface, compares them with a normal cutting vibration feature library, and extracts the chip adhesion signal intensity index. The chip adhesion signal intensity index is fused with the collected resultant force time-domain signal to generate a vibration-force coupled data pair. A preset interference stripping algorithm is applied to the vibration-force coupled data pair to separate the additional load time-domain sequence and the pure cutting force time-domain sequence. If the additional load time-domain sequence exceeds a predefined threshold, a self-cleaning trigger signal is issued, driving the multi-blade scraper at the drill rod tail to... The directional jet cleaning nozzles work together to strip and remove accumulated chips from deep holes. The pure cutting force time-domain sequence is compared with the ideal cutting force benchmark, and the force deviation is calculated. According to the preset feed compensation strategy, a local feed compensation command is generated based on the force deviation. The local feed compensation command is executed and machining continues. The machined hole diameter deviation and guide deviation are fed back to the force separation module through an online measurement device to verify the correction effect and adjust the stripping and compensation parameter set until the machining accuracy target is achieved. This allows for accurate identification of chip accumulation and triggering of cleaning, effectively separating the real cutting force from the additional interference load, and dynamically compensating the feed parameters, significantly improving the accuracy and stability of deep hole machining.

[0077] This invention effectively solves the accuracy fluctuation problem in deep hole drilling by constructing a complete dynamic correction system. Through a vibration sensor array, it can accurately capture abnormal vibration characteristics caused by chip accumulation, achieving real-time identification of the chip accumulation state and providing a reliable basis for subsequent intervention. This invention uses a vibration-force coupling model to separate the load signal, eliminating additional interference from chip accumulation and accurately extracting the true cutting force, laying the foundation for feed parameter adjustment. When chip accumulation exceeds the limit, the self-cleaning device of this invention can be triggered in time, clearing the accumulated chips through the coordinated action of scrapers and jets, preventing the continuous accumulation of interference. Based on the feed compensation command generated from the actual cutting force deviation, this invention can dynamically adjust the feed rate and cutting depth, correct the machining trajectory in real time, and reduce hole diameter and guide errors. The closed-loop verification mechanism of this invention continuously optimizes the separation parameters and compensation strategy through online measurement feedback results, ensuring stable and compliant machining accuracy, and significantly improving the accuracy consistency and machining efficiency of deep hole drilling.

[0078] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the steps in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0079] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, including read-only memory (ROM), random access memory (RAM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), one-time programmable read-only memory (OTPROM), electrically-Erasable Programmable Read-Only Memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, disk storage, magnetic tape storage, or any other computer-readable medium capable of carrying or storing data.

[0080] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

Claims

1. A CNC correction method for deep hole drilling based on force feedback, characterized in that, The method includes: The vibration amplitude and spectral characteristics of the drill pipe surface are monitored by a micro vibration sensor array and compared with a normal cutting vibration feature library to extract the chip adhesion signal intensity index. The chip adhesion signal intensity index is fused with the collected resultant force time domain signal to generate a vibration-force coupling data pair, and a preset interference stripping algorithm is applied to the vibration-force coupling data pair to separate the additional load time domain sequence and the pure cutting force time domain sequence. If the additional load time-domain sequence exceeds a predefined threshold, a self-cleaning trigger signal is sent to drive the multi-blade scraper at the tail of the drill pipe and the directional jet cleaning nozzle to work together to remove the chips accumulated in the deep hole. The pure cutting force time-domain sequence is compared with the ideal cutting force reference and the force deviation is calculated. Based on the force deviation, a local feed compensation command is generated according to the preset feed compensation strategy. The local feed compensation command is executed and processing continues. The hole diameter deviation and guide deviation after processing are fed back to the force separation module through the online measurement device to verify the correction effect and adjust the peeling and compensation parameter set until the processing accuracy target is achieved.

2. The CNC correction method for deep hole drilling based on force feedback according to claim 1, characterized in that, The method of using a micro vibration sensor array to monitor the vibration amplitude and spectral characteristics of the drill pipe surface, and comparing them with a normal cutting vibration feature library to extract the chip adhesion signal intensity index, specifically includes: N miniature piezoelectric vibration sensors are arranged at equal intervals along the axial and circumferential directions on the drill pipe shell, and the N miniature piezoelectric vibration sensors are in close contact with the deep hole wall. According to the sampling rate U s With a sampling duration T, the output of each miniature piezoelectric vibration sensor is synchronously sampled using a high-speed analog-to-digital converter to generate a vibration time-domain signal matrix O=[O1(t), ..., O i (t), ..., O I [t], where I represents the total number of vibration frequency domain signals in the vibration time domain signal matrix, i represents the index of the vibration frequency domain signal in the vibration time domain signal matrix, t represents the current time, and the collected vibration time domain signal matrix O is stored in a circular buffer to form a C1×C2 dimensional matrix, C2=U s ×T, U s ≥10kHz; A bandpass filter is used to remove the time-domain signal O of each vibration. i The low-frequency workpiece vibration and high-frequency noise in (t) are used to calculate the filtered vibration time-domain signal O. i The envelope amplitude A of (t) i =max|O i (t)|, for the filtered vibration time-domain signal O i (t) Perform a fast Fourier transform to obtain the amplitude spectrum h i [k], k=1,…,M, where k represents the spectral line index and M represents the total number of spectral lines in the amplitude spectrum. The current spectrum H is generated by aggregating the spectral lines according to their average values. k ; Obtain the normal spectrum from the normal cutting vibration feature library obtained through offline calibration. And calculate the current spectrum H k With the normal spectrum The correlation, i.e., the chip adhesion signal intensity index S; If the decrease in the chip adhesion signal intensity index S exceeds the amplitude threshold, it is determined that there are accumulated chips in the deep hole, and a chip removal request signal is issued.

3. The CNC correction method for deep hole drilling based on force feedback according to claim 2, characterized in that, The current spectrum H k The calculation formula is as follows: In the formula, FFT() represents the Fast Fourier Transform operation; The formula for calculating the chip adhesion signal intensity index S is as follows: 。 4. The CNC correction method for deep hole drilling based on force feedback according to claim 2, characterized in that, The process of fusing the chip adhesion signal intensity index with the acquired resultant force time-domain signal to generate a vibration-force coupled data pair, and applying a preset interference stripping algorithm to the vibration-force coupled data pair to separate the additional load time-domain sequence and the pure cutting force time-domain sequence, specifically includes: The resultant force time-domain signal F is acquired by a force sensor at the drill pipe spindle end. total (t), the bandpass filter is used to remove the resultant force time-domain signal F. total The low-frequency workpiece vibration and high-frequency noise in (t) are determined according to the sampling rate U. s The time-domain signal F of the resultant force after filtering will be obtained from the sampling duration T. total (t) is truncated into an equal-length sequence; The chip adhesion signal intensity index S is updated to the chip adhesion vibration index S(t) according to the actual vibration sampling period; The chip adhesion vibration index S(t) and the resultant force time domain signal F total (t) Perform time synchronization alignment to generate the vibration-force coupling data pair; The preset interference stripping algorithm employs a vibration-force coupling model. The vibration-force coupling data is input into the vibration-force coupling model to calculate the time-domain sequence F of the additional load. int (t), and then combined with the resultant force time domain signal F total (t), the time-domain sequence F of the pure cutting force is calculated. cut (t); If F cut (t) is negative or F cut (t) If the cutting force exceeds the normal cutting force range, then F cut (t) is marked as an outlier and subjected to amplitude limiting.

5. The CNC correction method for deep hole drilling based on force feedback according to claim 4, characterized in that, The additional load time-domain sequence F int The formula for calculating (t) is as follows: In the formula, α represents the additional load scaling factor obtained from offline calibration, β represents the nonlinear correction coefficient obtained from offline fitting, and both α and β are parameters of the vibration-force coupling model; The pure cutting force time-domain sequence F cut The formula for calculating (t) is as follows: 。 6. The CNC correction method for deep hole drilling based on force feedback according to claim 1, characterized in that, The step of comparing the pure cutting force time-domain sequence with the ideal cutting force reference and calculating the force deviation, and generating a local feed compensation command based on the force deviation according to a preset feed compensation strategy, specifically includes: At the current machining position, the cutting time T is recorded. w The pure cutting force time-domain sequence F within cut (t), calculate the cutting time T w The pure cutting force time-domain sequence F within cut The time average of (t), i.e., the average cutting force F avg : In the formula, Represents the integral variable; The ideal cutting force reference F, calibrated offline, is retrieved from the process database for the current machining condition. ref Combined with the average cutting force F avg Calculate the force deviation ΔF: According to the preset feed compensation strategy, the force deviation ΔF is mapped to the feed rate adjustment amount ΔV. f : In the formula, K f This indicates the feed rate compensation gain calibrated offline. If the feed rate adjustment amount ΔV f If the maximum adjustment range is exceeded, the feed rate adjustment amount ΔV is adjusted. f Amplitude limiting is applied; The force deviation ΔF is converted into the cutting depth adjustment amount ΔL according to the proportional relationship: ΔL=-K d ΔF In the formula, K d This indicates the depth of cut compensation gain calibrated offline; If the depth of cut adjustment ΔL exceeds the machining tolerance range, then the depth of cut adjustment ΔL is truncated to the tolerance limit; Insert the original feed segment at the current CNC trajectory point position, and adjust the original feed speed V0 to V0 + ΔV. f The local feed rate command is generated, and the original depth of cut L0 is adjusted to L0+ΔL. The depth of cut increment command is generated, and the local feed rate command and the depth of cut increment command are combined to generate the local feed compensation command.

7. The CNC correction method for deep hole drilling based on force feedback according to claim 1, characterized in that, The process involves executing the local feed compensation command and continuing machining. The machined hole diameter deviation and guide deviation are fed back to the force separation module via an online measurement device to verify the correction effect and adjust the peeling and compensation parameter set until the machining accuracy target is achieved. Specifically, this includes: The local feed compensation command is received and executed, and the diameter D of the machined deep hole is measured by an electromagnetic diameter sensor mounted at the deep hole exit. meas (t), and the offset G of the drill pipe centerline is measured by a laser triangular guide sensor installed at the front end of the drill pipe. meas (t), combined with aperture reference D ref and guiding benchmark G ref Calculate the aperture deviation ΔD(t) = D meas (t)-D ref With guidance deviation ΔG(t) = G meas (t)- G ref : The maximum absolute deviation E is calculated based on the aperture deviation ΔD(t) and the guide deviation ΔG(t). max Obtain the accuracy tolerance R tol If E max ≤R tol If the correction is successful, the correction effect is acceptable; otherwise, the correction effect is unacceptable, and the peeling and compensation parameter set needs to be readjusted to continue the next round of peeling and compensation operations until the processing accuracy target is achieved.

8. The CNC correction method for deep hole drilling based on force feedback according to claim 7, characterized in that, The maximum absolute deviation E max The calculation formula is as follows: 。 9. A force feedback-based CNC correction system for deep hole drilling, applied to the force feedback-based CNC correction method for deep hole drilling as described in any one of claims 1-8, characterized in that, The system includes: The index extraction module is used to monitor the vibration amplitude and spectral characteristics of the drill pipe surface using a micro vibration sensor array, and compare them with a normal cutting vibration feature library to extract the chip adhesion signal intensity index. The sequence separation module is used to fuse the chip adhesion signal intensity index with the collected resultant force time domain signal to generate a vibration-force coupling data pair, and apply a preset interference stripping algorithm to the vibration-force coupling data pair to separate the additional load time domain sequence and the pure cutting force time domain sequence. The chip cleaning module is used to send a self-cleaning trigger signal if the additional load time domain sequence exceeds a predefined threshold, thereby driving the multi-blade scraper at the tail of the drill pipe and the directional jet cleaning nozzle to work together to remove the chips accumulated in the deep hole. The instruction generation module is used to compare the pure cutting force time-domain sequence with the ideal cutting force reference and calculate the force deviation, and generate a local feed compensation instruction based on the force deviation according to a preset feed compensation strategy. The effect verification module is used to execute the local feed compensation command and continue processing. It feeds back the hole diameter deviation and guide deviation after processing to the force separation module through an online measurement device to verify the correction effect and adjust the peeling and compensation parameter set until the processing accuracy target is achieved.

Images