A method for processing a thin-walled profiled cantilever elbow

Abstract

The application provides a method for processing a thin-walled special-shaped cantilever elbow pipe. The method provides basic stiffness and axial support for the workpiece to be processed through a fixing part. Axial pre-stress can eliminate the gap between the workpiece and the cavity, improve the system stiffness and damping, and provide basic conditions for the anti-vibration effect of high and low blades. The double-blade boring cutter actively generates double-frequency non-harmonic excitation through the height difference, disperses the excitation energy, avoids resonance, and blocks regenerative chatter. The fixing part and the double-frequency non-harmonic excitation work together to suppress self-excited vibration from two dimensions of "improving anti-vibration ability" and "reducing excitation strength". Through real-time acquisition of cutting force, vibration and temperature signals, time-frequency domain fusion, combination of machine learning classifier and multi-level threshold judgment method for redundant verification, early identification of chatter precursors is realized, and online adaptive adjustment of response is realized through a piezoelectric ceramic driver.

Description

Technical Field

[0001] This invention belongs to the field of machining technology, specifically relating to a method for processing thin-walled irregular-shaped cantilevered pipes. Background Technology

[0002] Large, thin-walled irregularly shaped parts offer advantages such as high design freedom and significant weight reduction, and are widely used in aerospace, shipbuilding, and energy equipment industries. Among these, thin-walled irregularly shaped parts with lateral cantilevered bends, such as aero-engine recirculation combustion chambers and bending guides, require straight sections of the bends to form a certain angle (typically 30°) with the central axis of the main body due to structural and functional requirements, and the bends are spaced circumferentially along the main body. Since the wall thickness is generally ≤3.5mm and the straight section length of the bend can reach over 450mm, these parts face significant technical challenges in the finishing of the inner surface after bending.

[0003] (a) Deficiencies in workpiece clamping

[0004] Thin-walled irregularly shaped parts have poor rigidity. Existing tooling fixtures for thin-walled irregularly shaped parts, such as patent CN222114316U - a tooling fixture for thin-walled irregularly shaped parts, use an upper cover plate, a lower fixed seat, and a workpiece for cavity positioning; another example is patent CN121290132A - a vertical machining fixture for herringbone-shaped thin-walled workpieces, which uses a tensioning mechanism to press the workpiece towards a support plate. However, the above tooling fixtures all use a unidirectional clamping method, applying clamping force only on one side, resulting in an asymmetrical stress state on the workpiece. In addition, for irregularly shaped parts with lateral cantilever bends, the cantilever ends lack auxiliary support, making them prone to bending deformation and torsional vibration under cutting forces. Existing tooling fixtures cannot provide effective axial support for the cantilever bends, nor can they eliminate the fit clearance between the workpiece and the tooling cavity, resulting in insufficient overall rigidity of the workpiece during machining.

[0005] (ii) Deficiencies in cutting and machining

[0006] During the boring of the inner surface of a cantilevered pipe, chatter (self-excited vibration) is highly likely to occur due to the poor rigidity and low damping of the cantilever structure. This results in chatter marks on the machined surface, with a surface roughness of only about Ra6.3. In severe cases, it can even cause tool wear, rendering the workpiece unusable. Conventional double-edged boring tools aim for equal height and radial force on both edges. However, for weakly rigid workpieces like cantilevered pipes, perfectly symmetrical double-edged cutting generates a continuous single-frequency cutting force excitation. When this excitation frequency approaches a certain natural frequency of the cantilever structure, resonance is easily induced, which in turn exacerbates chatter. Existing patents, such as CN202367223U - Multi-purpose double-edged adjustable boring tool, propose manufacturing two sets of boring bar holders to achieve high and low cutting. However, their purpose is only to improve efficiency through layered cutting, without revealing the anti-chatter principle of high and low edges from the perspective of chatter mechanism. On the other hand, single-edged boring tools only bear force on one side, and the radial component of the cutting force causes unidirectional bending deformation of the cantilevered pipe, resulting in severe tool wear and making it difficult to guarantee machining accuracy.

[0007] (III) Deficiencies in intelligent monitoring and adaptive control

[0008] Patent CN110103076B, "An Intelligent Boring Bar System for Real-Time Monitoring of Deep Hole Boring Machining Status," employs cutting force sensors and vibration sensors for signal acquisition and monitoring. Patent CN109746762B, "An Online Monitoring and Suppression Method for Chatter in Deep Hole Boring Machining," achieves online chatter monitoring by constructing a dynamic model. However, these systems only operate at an open-loop monitoring level of "signal acquisition - abnormal alarm," failing to connect the sensed information with actuator control and thus unable to achieve closed-loop adaptive control of "perception - judgment - adjustment." Specific shortcomings include: lack of pre-machining sensor calibration; lack of a systematic multi-sensor data fusion and chatter status recognition algorithm; and monitoring results used only for alarm prompts, unable to achieve real-time autonomous adjustment.

[0009] (iv) Gaps in the integration of tooling, cutting tools, and monitoring systems

[0010] Currently, there is no systematic solution that organically integrates a special tooling for thin-walled irregular parts, a high-low blade anti-chatter boring tool, and a multi-sensor real-time monitoring system. The tooling solves the problem of "how to clamp the workpiece", the tool solves the problem of "how to cut", and the monitoring solves the problem of "how to judge", but the three are independent and cannot form a closed-loop control link of "clamping provides basic rigidity - tool actively suppresses chatter - sensor real-time monitoring - adaptive adjustment". Summary of the Invention

[0011] To address the technical problems in the prior art of machining large thin-walled irregular parts with lateral cantilevered pipe structures during boring, such as poor surface quality, frequent tool breakage, and low product qualification rate due to insufficient clamping rigidity, severe cutting chatter, and lack of adaptive control, this application provides a method for machining cantilevered pipes of thin-walled irregular parts.

[0012] To achieve the above objectives, the present invention provides the following technical solution:

[0013] This application provides a method for processing thin-walled irregularly shaped cantilevered pipes, including the following steps:

[0014] Step S1: Place the workpiece to be processed in the fixture, which provides axial bidirectional compressive force to the workpiece and obtains the pre-compression stress of the workpiece.

[0015] Step S2: Under no-load conditions, obtain the reference temperature of the inner wall of the double-edged boring bar;

[0016] Step S3: The high and low cutting edges of the double-edged boring bar enter the inner hole of the cantilevered bend of the workpiece in sequence. Cutting force signal, vibration signal, temperature signal and machining parameter signal are collected in real time. The cutting force signal is corrected based on the reference temperature and temperature signal. Based on the corrected chip force, a dual-frequency non-harmonic excitation based on the high-edge excitation frequency and the low-edge excitation frequency is constructed.

[0017] Step S4: Construct time-domain features based on the corrected chip force, and construct frequency-domain features based on dual-frequency harmonic excitation;

[0018] Step S5: Input the frequency domain features and time domain features into the preset classifier, output the first classification result, obtain the second judgment result based on the multi-level threshold judgment method, and obtain the final processing state based on the first judgment result and the second judgment result.

[0019] Step S6: Based on the final machining state, adjust the machining process of the double-edged boring tool, evaluate the adjustment effect, and construct an adjustment sample.

[0020] Furthermore, the workpiece to be processed includes a cylinder and a plurality of cantilever bends evenly distributed on the outer wall of the cylinder. The cantilever bends include a bent section and a straight section, and the angle between the axis of the straight section and the outer wall of the cylinder is °.

[0021] The fastener includes a base, an upper cover plate, and a lower cover plate, as well as multiple locking components disposed between the upper cover plate and the lower cover plate. The locking components include a tension bolt rod and a locking nut. The upper cover plate has a first through hole for the tension bolt rod to pass through, and an annular groove adapted to the end structure of the cylinder is provided on the bottom surface. The lower cover plate has an L-shaped structure. The top abutting surface of the vertical section of the L-shaped structure has a concave spherical cavity and a support platform adapted to the end structure of the cylinder. The horizontal section has a second through hole for the tension bolt rod to pass through. A pressure sensor is provided on the abutting surface of the annular groove and the support platform. The pressure sensor is used to obtain the pre-compression stress of the workpiece to be processed.

[0022] The lower cover plate is detachably connected to the base via a locking device;

[0023] One end of the tension bolt rod is engaged with the second through hole, and the other end is fixed to the first through hole by a lock nut. There are at least two tension bolt rods between adjacent cantilever bends.

[0024] Furthermore, in step S3, the cutting force signal is acquired in real time based on the cutting force sensor. The cutting force signal includes radial cutting force, tangential cutting force, and axial cutting force, wherein the radial cutting force equals the preload stress. times, The angle between the axis of the straight section of the cantilever bend and the outer wall of the cylinder;

[0025] Real-time acquisition of radial vibration acceleration based on a vibration acceleration sensor;

[0026] The temperature signal of the inner wall of the boring bar is collected in real time based on a temperature sensor;

[0027] The reference temperature and temperature signal When correcting the cutting force signal, the following steps are included:

[0028] The corrected radial cutting force In the formula, The radial cutting force is collected in real time; This is the temperature compensation coefficient, with a value ranging from 0.4 N / ℃ to 0.6 N / ℃.

[0029] The corrected tangential cutting force In the formula, The tangential cutting force is collected in real time;

[0030] The corrected axial cutting force In the formula, This refers to the axial cutting force collected in real time.

[0031] Furthermore, in step S3, the high-edge excitation frequency is extracted using a fast Fourier transform based on the corrected cutting force signal. And based on high blade excitation frequency Obtain low blade excitation frequency ,include:

[0032] ;

[0033] in, The radius of rotation of the boring tool; The angle between the cutting edge and the workpiece axis; The height difference between the high and low cutting edges in the axial direction;

[0034] When satisfied And harmonic deviation : At that time, a dual-frequency non-harmonic excitation is generated on the workpiece being processed;

[0035] in, for The rounded integer value.

[0036] Furthermore, the time-domain features constructed in step S4 include: mean radial force. Radial force standard deviation Maximum radial force Mean tangential force Tangential force standard deviation Maximum tangential force , average axial force axial force standard deviation and maximum axial force and the maximum vibration acceleration and mean vibration acceleration ;

[0037] The constructed frequency domain features include: high-edge excitation frequency and low blade excitation frequency and high blade excitation frequency With low blade excitation frequency ratio .

[0038] Furthermore, in step S5, the preset classifier adopts an SVM classifier based on radial basis function kernel function. The SVM classifier is trained offline based on sample training data with known states. The sample training data includes sample frequency domain features and sample time domain features, as well as manually labeled sample state labels. The sample state labels include four categories: stable processing, flutter precursor, flutter aggravation, and severe flutter. During the training phase, the SVM classifier outputs the first classification result through the nonlinear relationship between sample frequency domain features, sample time domain features, and sample state labels.

[0039] The first classification result includes real-time status labels. In the formula, Indicates stable processing; Indicates a precursor to flutter; This indicates that the flutter has intensified; Indicates severe flutter;

[0040] The first classification result also includes real-time probability confidence. .

[0041] Furthermore, the multi-level threshold judgment method in step S5 includes:

[0042] Based on the mean radial force and radial force standard deviation Radial force fluctuation rate is obtained :

[0043] ;

[0044] When R When the threshold judgment result is reached, output multi-level threshold judgment results. =1 indicates that the current processing state is stable processing;

[0045] when R When the threshold judgment result is reached, output multi-level threshold judgment results. =2 indicates that the current processing state is a precursor to chatter;

[0046] when R When the threshold judgment result is reached, output multi-level threshold judgment results. =3 indicates that the current processing condition is characterized by increased chatter.

[0047] when When R is reached, the multi-level threshold judgment result is output. =4 indicates that the current processing condition is severe chatter.

[0048] Furthermore, in step S5, when obtaining the final processing state based on the first and second judgment results, the following steps are included:

[0049] like At that time, the real-time status label is used. Or multi-level threshold judgment results As the final processing state;

[0050] like At that time, and Then use real-time status tags As the final processing state;

[0051] like At that time, and ,but In this case, the multi-level threshold judgment result is adopted. ;

[0052] like At that time, and ,at the same time At the same time, maintain the state from the previous moment to avoid frequent jumps.

[0053] Furthermore, when adjusting the machining process of the double-edged boring tool based on the final machining state in step S6, it includes:

[0054] State 1 indicates that the current processing status is stable, so the current parameters are maintained and monitoring continues;

[0055] State 2: Indicates that the current processing state is a precursor to flutter, so an early warning is issued and minor adjustments are performed;

[0056] State 3: Indicates that the current processing state is flutter intensification, which triggers intervention and executes multi-level active adjustment;

[0057] Status 4: Indicates that the current machining status is severe chatter, which triggers an emergency stop and executes emergency tool retraction protection;

[0058] The micro-adjustment refers to: the piezoelectric ceramic actuator adjusting the low-edge extension amount from ±0.02 mm to ±0.05 mm, breaking the emerging resonance condition by performing a small adjustment, and randomly selecting the direction of extension or retraction, which is used to determine which direction of micro-adjustment is more effective based on the final processing state at the next moment;

[0059] The multi-stage active adjustment includes the following steps: the piezoelectric ceramic actuator adjusts the low blade extension amount in the reverse direction by 0.05 mm to 0.10 mm, reduces the rotational speed by 10% to 20%, and reduces the feed rate by 30% to 50%. The reverse adjustment means that the adjustment direction is opposite to that of the previous moment. The large adjustment changes the dynamic characteristics of the system to suppress the developed chatter.

[0060] The emergency retraction protection includes the following actions in sequence: stop feed, radial retraction of 0.2 mm, spindle stop, and alarm.

[0061] Furthermore, in step S6, the tremor suppression rate is calculated. Evaluate the effectiveness of the adjustments:

[0062] ;

[0063] In the formula, This represents the standard deviation of radial force at the time corresponding to the final machining state. This represents the standard deviation of radial force after the machining process of the double-edged boring tool has been adjusted.

[0064] like This indicates that flutter suppression is effective;

[0065] like If this indicates that chatter suppression is ineffective, then within the corresponding state, change the machining parameters of the double-edged boring tool. If this is done three times consecutively, Still smaller If so, it is determined to be an invalid adjustment.

[0066] Compared with the prior art, the present invention has the following beneficial technical effects:

[0067] This application provides a method for machining thin-walled irregularly shaped cantilevered pipe bends. This method provides basic stiffness and axial support for the workpiece through a fixing component. The axial preload can eliminate the gap between the workpiece and the cavity, improve the system stiffness and damping, and provide the basic conditions for the anti-chatter effect of high and low cutting edges. The double-edged boring tool actively generates dual-frequency non-harmonic excitation through the height difference, which disperses the excitation energy, avoids resonance, and blocks regenerative chatter. The fixing component and the dual-frequency non-harmonic excitation work together to suppress self-excited vibration from two dimensions: "improving vibration resistance" and "reducing excitation intensity". By real-time acquisition of cutting force, vibration and temperature signals, time-frequency domain fusion is performed. Combined with the redundancy verification of machine learning classifier and multi-level threshold judgment method, early identification of chatter precursors is achieved. The response is achieved through online adaptive adjustment by piezoelectric ceramic actuator. The fixing component of this method can increase the system stiffness of the workpiece by more than 30%, and the dual-frequency non-harmonic excitation reduces the chatter amplitude by more than 50%. The surface roughness of the workpiece after processing is reduced from Ra6.3 to below Ra3.2. Through early identification of chatter precursors, the tool-biting phenomenon is basically eliminated, and the product qualification rate is increased from less than 70% to more than 98%. Attached Figure Description

[0068] Figure 1 A schematic diagram of the clamping structure for the fastener and the workpiece being processed, as shown in an embodiment of this disclosure, is illustrated.

[0069] Figure 2 A structural schematic diagram of the fastener in an embodiment of this disclosure is shown;

[0070] Figure 3 A schematic diagram of the structure of the upper cover plate in an embodiment of this disclosure is shown;

[0071] Figure 4 A schematic diagram of the structure of the lower cover plate in an embodiment of this disclosure is shown.

[0072] In the figure: 1-cylinder; 11-cantilever bend; 21-base; 22-upper cover plate; 23-lower cover plate; 24-tension bolt rod; 25-locking nut; 26-annular groove; 27-concave spherical cavity; 28-first through hole; 29-second through hole; 30-support platform. Detailed Implementation

[0073] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0074] This disclosure provides a method for processing thin-walled irregularly shaped cantilevered pipes, including the following steps:

[0075] Step S1: Place the workpiece to be processed in the fixture, which provides axial bidirectional compressive force to the workpiece and obtains the pre-compression stress of the workpiece.

[0076] Specifically, the workpiece to be processed includes a cylinder 1 and a plurality of cantilever bends 11 evenly distributed on the outer wall of the cylinder 1. Each cantilever bend 11 includes a bent portion and a straight section. The angle between the axis of the straight section and the outer wall of the cylinder 1 is 30°. In this embodiment, as shown... Figure 1 As shown, there are four cantilever bends 11.

[0077] like Figures 2 to 4 As shown, the fixing component includes a base 21, an upper cover plate 22, and a lower cover plate 23, as well as multiple locking assemblies disposed between the upper cover plate 22 and the lower cover plate 23. Each locking assembly includes a tension bolt rod 24 and a locking nut 25. The upper cover plate 22 has a first through hole 28 for the tension bolt rod 24, and its bottom surface has an annular groove 26 adapted to the end structure of the cylinder 1. The lower cover plate 23 has an L-shaped structure, and the top abutment surface of the vertical section of the L-shaped structure has a concave spherical cavity 27 adapted to the end structure of the cylinder 1 and a support. The support platform 30 has a second through hole 29 in its horizontal section for passing through the tension bolt rod 24. A pressure sensor is provided on the contact surface between the annular groove 26 and the support platform 30. The pressure sensor is used to obtain the pre-compression stress of the workpiece to be processed. It should be noted that the lower end of the workpiece cylinder 1 is an arc-shaped structure. A stepped surface is provided on the upper side of the arc-shaped structure. The support platform 30 is used to abut against the stepped surface. The structure of the concave spherical cavity 27 is adapted to the arc-shaped structure of the lower end. The concave spherical cavity 27 is used to abut against the arc-shaped structure of the lower end.

[0078] One end of the tension bolt rod 24 is engaged with the second through hole 29, and the other end is fixed to the first through hole 28 by the locking nut 25. There are at least two tension bolt rods 24 between adjacent cantilever bends 11. In this embodiment, there are eight tension bolt rods 24.

[0079] In some embodiments, both the upper cover plate 22 and the lower cover plate 23 are lightweight structures: the upper cover plate 22 is provided with weight-reducing grooves at intervals along the circumference, and the horizontal section of the lower cover plate 23 is also provided with multiple weight-reducing grooves along the circumference.

[0080] In some embodiments, the lower cover plate 23 is detachably connected to the base 21 via a locking member. The locking member includes an L-shaped fixing plate and a fastening bolt. The L-shaped fixing plate includes a vertical plate and a horizontal plate. The vertical plate abuts against the base 21, and the horizontal plate presses against the top edge of the lower cover plate 23. A strip groove is provided on the horizontal plate, and a fastening bolt is threaded through the strip groove. An inverted T-shaped groove is provided on the base 21. During assembly, the nut is first placed in the inverted T-shaped groove in the area to be processed, and the fastening bolt is threaded through the strip groove and engaged with the nut thread and tightened to fix the lower cover plate 23 on the base 21.

[0081] It should be noted that the stable clamping state provided by the fastener is a prerequisite for the sensor to collect a clean and effective signal. If the workpiece to be processed is itself in a shaking state, the sensor signal will be drowned out by noise and will not be able to accurately identify the precursors of chatter.

[0082] Step S2: Under no-load conditions, obtain the reference temperature of the inner wall of the double-edged boring bar;

[0083] Step S3: The high and low cutting edges of the double-edged boring bar enter the inner hole of the cantilevered bend of the workpiece in sequence. The cutting force signal, vibration signal, temperature signal and machining parameter signal of the double-edged boring bar are collected in real time. The cutting force signal is corrected based on the reference temperature and temperature signal. Based on the corrected chip force, a dual-frequency non-harmonic excitation based on the high-edge excitation frequency and the low-edge excitation frequency is constructed.

[0084] Specifically, in step S3, the cutting force signal is acquired in real time based on the cutting force sensor. The cutting force signal includes radial cutting force, tangential cutting force, and axial cutting force, wherein the radial cutting force should be equal to the preload stress. times, The angle between the axis of the straight section of the cantilever bend 11 and the outer wall of the cylinder 1; in this embodiment, Since the angle is 30°, the radial cutting force is 0.5 times the preload stress, which in turn generates a constant reverse component force in the radial direction of the straight section of the cantilever bend 11 that is opposite to the direction of the radial component of the cutting force, in order to counteract the mean component of the radial component of the cutting force in real time.

[0085] Real-time acquisition of radial vibration acceleration based on a vibration acceleration sensor;

[0086] The temperature signal of the inner wall of the boring bar is collected in real time based on a temperature sensor.

[0087] Specifically, in step S3, based on the reference temperature and temperature signal When correcting the cutting force signal, the following steps are included:

[0088] The corrected radial cutting force In the formula, The radial cutting force is collected in real time; This is the temperature compensation coefficient, with a value ranging from 0.4 N / ℃ to 0.6 N / ℃.

[0089] The corrected tangential cutting force In the formula, The tangential cutting force is collected in real time;

[0090] The corrected axial cutting force In the formula, This refers to the axial cutting force collected in real time.

[0091] It's important to note that when cutting high-temperature alloys, the temperature in the cutting zone can reach over 800℃, and the heat is rapidly transferred to the boring bar. Cutting force sensors are typically based on strain gauges; strain gauges expand when heated, generating micro-strains. Even if the tool tip is not under force, the sensor may mistakenly interpret this as a force output—a phenomenon called zero-point drift. For example, when the machine is first turned on, the cutting force sensor may register 0N of cutting force. After machining for a while, thermal deformation might cause the reading to become "+50N". This 50N is a completely false signal and will severely interfere with chatter detection.

[0092] Based on reference temperature and temperature signal Set the temperature compensation coefficient Temperature compensation coefficient It is a proportional value measured through calibration experiments, representing how many Newtons the force reading will drift for every 1°C increase in temperature. When the machine tool stops briefly and then resumes contact, the instantaneous thermal shock current caused by the alternation of hot and cold will be misjudged as a precursor to chatter, causing the piezoelectric ceramic to frequently make meaningless positive and negative adjustments, wasting time, and even damaging the cutting tool due to excessive tool adjustment.

[0093] Specifically, in step S3, the high-edge excitation frequency is extracted using a fast Fourier transform based on the corrected cutting force signal. And based on high blade excitation frequency Obtain low blade excitation frequency ,include:

[0094] ;

[0095] in, The radius of rotation of the boring tool; The angle between the cutting edge and the workpiece axis; The height difference between the high and low cutting edges in the axial direction;

[0096] When satisfied And harmonic deviation : At that time, a dual-frequency non-harmonic excitation is generated on the workpiece being processed;

[0097] in, for The rounded integer value, i.e., the ratio of the two frequencies deviates from any integer ratio by at least 5%, satisfies the non-harmonic condition. For example, simple ratios like 3:2 or 4:3 may still produce periodic superposition, while the high-low blade design aims to create approximate but non-overlapping frequencies, such as 100:101, to disperse the excitation energy and avoid continuous accumulation at a certain point. By adjusting the height difference between the high and low blades, the generated dual-frequency excitation can effectively avoid the system's inherent frequency determined by the boring bar structure. It should be noted that, according to Fourier analysis, the spectral energy of the dual-frequency non-harmonic excitation is dispersed across multiple frequency components, rather than concentrated at a single frequency. Its core anti-chatter mechanism is as follows: the excitation energy is no longer concentrated at a specific frequency, and the energy density of each frequency component is insufficient to form a continuous energy accumulation at any natural frequency of the boring bar structure; the high and low cutting edges cut in sequence, which interrupts the cutting thickness variation law, destroys the phase condition on which regenerative chatter depends, and chatter cannot establish a positive feedback loop; at the same time, when the low cutting edge cuts in later, it generates a small impact on the machined surface, and the impact energy is absorbed by the elastic deformation of the boring bar structure and converted into internal dissipation.

[0098] Step S4: Construct time-domain features based on the corrected chip force, and construct frequency-domain features based on dual-frequency harmonic excitation;

[0099] Specifically, the constructed time-domain features include: mean radial force. Radial force standard deviation Maximum radial force Mean tangential force Tangential force standard deviation Maximum tangential force , average axial force axial force standard deviation and maximum axial force and the maximum vibration acceleration and mean vibration acceleration ;

[0100] The constructed frequency domain features include: high-edge excitation frequency and low blade excitation frequency and high blade excitation frequency With low blade excitation frequency ratio .

[0101] It should be noted that the corrected radial cutting force It is the core indicator for chatter detection, the corrected radial cutting force The corrected radial cutting force is the component of the cutting force perpendicular to the tool axis. The weakest rigid direction acting directly on the cantilever bend is the dominant factor causing bending deformation and chatter in the bend. The corrected radial cutting force... The amplitude of the fluctuation is the most sensitive indicator of flutter. In the early stages of flutter, the standard deviation of the radial force... It will increase first, before the change in tangential force. Maximum radial force. It reflects the degree of instantaneous impact between the tool and the workpiece and is a key indicator for judging the risk of tool breakage. Average radial force. It directly reflects whether the axial support force provided by the 8 sets of tensionable bolt rods 24 effectively counteracts the radial component of the cutting force.

[0102] Step S5: Input the frequency domain features and time domain features into the preset classifier, output the first classification result, obtain the second judgment result based on the multi-level threshold judgment method, and obtain the final processing state based on the first judgment result and the second judgment result.

[0103] Specifically, in step S5, the preset classifier adopts an SVM classifier based on radial basis function kernel function. The SVM classifier is trained offline based on sample training data with known state. The sample training data includes sample frequency domain features and sample time domain features, as well as manually labeled sample state labels. The sample state labels include four categories: stable processing, flutter precursor, flutter aggravation, and severe flutter. During the training phase, the SVM classifier outputs the first classification result through the nonlinear relationship between sample frequency domain features, sample time domain features and sample state labels.

[0104] The first classification result includes real-time status labels. In the formula, Indicates stable processing; Indicates a precursor to flutter; This indicates that the flutter has intensified; Indicates severe flutter;

[0105] The first classification result also includes real-time probability confidence. .

[0106] It should be noted that the SVM classifier based on the radial basis function (RBF) first maps the vibration time / frequency domain features to a high-dimensional space using the RBF kernel, and then finds the optimal classification surface in the high-dimensional space to distinguish between four types of working conditions, including stable processing and flutter precursors. The advantage of the SVM classifier lies in its ability to capture the complex nonlinear relationships between feature parameters, and its ability to identify ambiguous states such as flutter precursors is superior to that of a fixed threshold.

[0107] Specifically, the multi-level threshold judgment method in step S5 includes:

[0108] Based on the mean radial force and radial force standard deviation Radial force fluctuation rate is obtained :

[0109] ;

[0110] The radial force fluctuation rate R, derived from the cutting force sensor, characterizes the macroscopic fluctuation intensity of the cutting force, where the standard deviation of the radial force is... The standard deviation of radial force reflects the degree of fluctuation. With the mean radial force The ratio of radial force fluctuation rate It eliminates the influence of the absolute value of cutting conditions and has good adaptability to working conditions.

[0111] When R When the threshold judgment result is reached, output multi-level threshold judgment results. =1 indicates that the current processing state is stable processing;

[0112] when R When the threshold judgment result is reached, output multi-level threshold judgment results. =2 indicates that the current processing state is a precursor to chatter;

[0113] when R When the threshold judgment result is reached, output multi-level threshold judgment results. =3 indicates that the current processing condition is characterized by increased chatter.

[0114] when When R is reached, the multi-level threshold judgment result is output. =4 indicates that the current processing condition is severe chatter.

[0115] It should be noted that the above thresholds are derived from statistical results of a large number of cutting tests, and correspond to the critical point where surface roughness begins to deteriorate, the point where chatter has developed into stable self-excited vibration requiring active intervention, and the tool-biting safety limit, respectively.

[0116] Specifically, in step S5, when obtaining the final processing state based on the first and second judgment results, the following steps are included:

[0117] like At that time, the real-time status label is used. Or multi-level threshold judgment results As the final processing state;

[0118] like At that time, and Then use real-time status tags As the final processing state;

[0119] like At that time, and ,but In this case, the multi-level threshold judgment result is adopted. ;

[0120] like At that time, and ,at the same time At the same time, maintain the state from the previous moment to avoid frequent jumps.

[0121] Step S6: Based on the final machining state, adjust the machining process of the double-edged boring tool, evaluate the adjustment effect, and construct an adjustment sample.

[0122] Specifically, when adjusting the machining process of the double-edged boring tool based on the final machining state in step S6, it includes:

[0123] State 1 indicates that the current processing status is stable, so the current parameters are maintained and monitoring continues;

[0124] State 2: Indicates that the current processing state is a precursor to flutter, so an early warning is issued and minor adjustments are performed;

[0125] State 3: Indicates that the current processing state is flutter intensification, which triggers intervention and executes multi-level active adjustment;

[0126] Status 4: Indicates that the current machining status is severe chatter, which triggers an emergency stop and executes emergency tool retraction protection;

[0127] The micro-adjustment refers to: the piezoelectric ceramic actuator adjusting the low-edge extension amount from ±0.02 mm to ±0.05 mm, breaking the emerging resonance condition by performing a small adjustment, and randomly selecting the direction of extension or retraction, which is used to determine which direction of micro-adjustment is more effective based on the final processing state at the next moment;

[0128] The multi-stage active adjustment includes the following steps: the piezoelectric ceramic actuator adjusts the low blade extension amount in the reverse direction by 0.05 mm to 0.10 mm, reduces the rotational speed by 10% to 20%, and reduces the feed rate by 30% to 50%. The reverse adjustment means that the adjustment direction is opposite to that of the previous moment. The large adjustment changes the dynamic characteristics of the system to suppress the developed chatter.

[0129] The emergency retraction protection includes the following actions in sequence: stop feed, radial retraction of 0.2 mm, spindle stop, and alarm.

[0130] Specifically, in step S6, the tremor suppression rate is calculated. Evaluate the effectiveness of the adjustments:

[0131] ;

[0132] In the formula, This represents the standard deviation of radial force at the time corresponding to the final machining state. This represents the standard deviation of radial force after the machining process of the double-edged boring tool has been adjusted.

[0133] like This indicates that flutter suppression is effective;

[0134] like If this indicates that chatter suppression is ineffective, then within the corresponding state, change the machining parameters of the double-edged boring tool. If this is done three times consecutively, Still smaller If so, it is determined to be an invalid adjustment.

[0135] In some embodiments, after an invalid adjustment is determined, manual intervention is adopted. The manual intervention process first involves switching the machine tool to manual mode, taking out the workpiece for inspection, and if there are signs of chipping or scratching, the workpiece is determined to be scrapped. The reason for the scrap is recorded and the workpiece is replaced.

[0136] If there are no obvious tool marks: first check the height difference between the high and low edges of the tool. If the current ∆h=0.10mm adjustment is ineffective, try adjusting it to 0.15-0.20mm and changing the excitation frequency.

[0137] Secondly, reduce cutting parameters: reduce spindle speed by 15-20% (e.g., from 900 r / min to 750 r / min), and reduce feed rate by 30-50% (e.g., from 0.08 mm / r to 0.05 mm / r).

[0138] Recheck and adjust the tooling preload: increase the tightening torque of the fixing parts on the workpiece to be processed, and increase the axial preload stress;

[0139] Check that the coolant nozzles are aligned with the cutting area and that the flow rate is sufficient.

[0140] Finally, replace the blade with a new one to eliminate tool wear as a contributing factor.

[0141] After completing the above adjustments, without clamping the workpiece to be processed, run a "dry run" program in manual mode for one machining stroke to confirm that: there is no abnormal interference or noise in the boring bar movement, no abnormal sudden changes in sensor signals, and the coolant lines are unobstructed. Load the test piece, run the program, and if the machining status is normal, you can proceed to normal product machining.

[0142] It should be noted that the adjustment sample is used to assist in the subsequent adjustment of the machining process of the double-edged boring tool. For example, when performing minor adjustments, the system prioritizes extending or retracting the lower edge based on the same or similar adjustment sample, thereby forming a reference-like experience.

[0143] The above embodiments are merely preferred implementation measures of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be covered within the scope of protection of the present invention.

Claims

1. A method for processing thin-walled irregularly shaped cantilevered pipes, characterized in that, Includes the following steps: Step S1: Place the workpiece to be processed in the fixture, which provides axial bidirectional compressive force to the workpiece and obtains the pre-compression stress of the workpiece. Step S2: Under no-load conditions, obtain the reference temperature of the inner wall of the double-edged boring bar; Step S3: The high and low cutting edges of the double-edged boring bar enter the inner hole of the cantilevered bend of the workpiece in sequence. Cutting force signal, vibration signal, temperature signal and machining parameter signal are collected in real time. The cutting force signal is corrected based on the reference temperature and temperature signal. Based on the corrected chip force, a dual-frequency non-harmonic excitation based on the high-edge excitation frequency and the low-edge excitation frequency is constructed. Step S4: Construct time-domain features based on the corrected chip force, and construct frequency-domain features based on dual-frequency harmonic excitation; Step S5: Input the frequency domain features and time domain features into the preset classifier, output the first classification result, obtain the second judgment result based on the multi-level threshold judgment method, and obtain the final processing state based on the first judgment result and the second judgment result. Step S6: Based on the final machining state, adjust the machining process of the double-edged boring tool, evaluate the adjustment effect, and construct an adjustment sample.

2. The method for processing thin-walled irregular-shaped cantilevered pipes according to claim 1, characterized in that, The workpiece to be processed includes a cylinder (1) and a plurality of cantilever bends (11) evenly distributed on the outer wall of the cylinder (1). The cantilever bends (11) include a bent section and a straight section. The angle between the axis of the straight section and the outer wall of the cylinder (1) is 30°. The fasteners include a base (21), an upper cover plate (22) and a lower cover plate (23), and multiple locking components disposed between the upper cover plate (22) and the lower cover plate (23). The locking components include a tension bolt rod (24) and a locking nut (25). The upper cover plate (22) is provided with a first through hole (28) for passing through the tension bolt rod (24), and the bottom surface is provided with an annular groove (26) adapted to the end structure of the cylinder (1). The lower cover plate (23) is an L-shaped structure. The top abutting surface of the vertical section of the L-shaped structure is provided with a concave spherical cavity (27) adapted to the end structure of the cylinder (1) and a support platform (30). The horizontal section is provided with a second through hole (29) for passing through the tension bolt rod (24). The abutting surfaces of the annular groove (26) and the support platform (30) are provided with a pressure sensor. The pressure sensor is used to obtain the pre-compression stress of the workpiece to be processed. The lower cover plate (23) is detachably connected to the base (21) via a locking element; One end of the tension bolt rod (24) is engaged with the second through hole (29), and the other end is fixed to the first through hole (28) by the lock nut (25). There are at least two tension bolt rods (24) between adjacent cantilever bends (11).

3. The method for processing thin-walled irregular-shaped cantilevered pipes according to claim 1, characterized in that, In step S3, the cutting force signal is acquired in real time based on the cutting force sensor. The cutting force signal includes radial cutting force, tangential cutting force, and axial cutting force, wherein the radial cutting force equals the preload stress. times, The angle between the axis of the straight section of the cantilever bend (11) and the outer wall of the cylinder (1); Real-time acquisition of radial vibration acceleration based on a vibration acceleration sensor; The temperature signal of the inner wall of the boring bar is collected in real time based on a temperature sensor; The reference temperature and temperature signal When correcting the cutting force signal, the following steps are included: The corrected radial cutting force In the formula, The radial cutting force is collected in real time; This is the temperature compensation coefficient, with a value ranging from 0.4 N / ℃ to 0.6 N / ℃. The corrected tangential cutting force In the formula, The tangential cutting force is collected in real time; The corrected axial cutting force In the formula, This refers to the axial cutting force collected in real time.

4. The method for processing thin-walled irregular-shaped cantilevered pipes according to claim 3, characterized in that, In step S3, the high-speed cutting force excitation frequency is extracted using a fast Fourier transform based on the corrected cutting force signal. And based on high blade excitation frequency Obtain low blade excitation frequency ,include: ； in, The radius of rotation of the boring tool; The angle between the cutting edge and the workpiece axis; The height difference between the high and low cutting edges in the axial direction; When satisfied And harmonic deviation : At that time, a dual-frequency non-harmonic excitation is generated on the workpiece being processed; in, for The rounded integer value.

5. The method for processing thin-walled irregular-shaped cantilevered pipes according to claim 4, characterized in that, The time-domain features constructed in step S4 include: mean radial force. Radial force standard deviation Maximum radial force Mean tangential force Tangential force standard deviation Maximum tangential force , average axial force axial force standard deviation and maximum axial force and the maximum vibration acceleration and mean vibration acceleration ; The constructed frequency domain features include: high-edge excitation frequency and low blade excitation frequency and high blade excitation frequency With low blade excitation frequency ratio .

6. The method for processing thin-walled irregular-shaped cantilevered pipes according to claim 5, characterized in that, In step S5, the preset classifier is an SVM classifier based on radial basis function kernel function. The SVM classifier is trained offline based on sample training data with known state. The sample training data includes sample frequency domain features and sample time domain features, as well as manually labeled sample state labels. The sample state labels include four categories: stable processing, flutter precursor, flutter aggravation, and severe flutter. During the training phase, the SVM classifier outputs the first classification result through the nonlinear relationship between sample frequency domain features, sample time domain features and sample state labels. The first classification result includes real-time status labels. In the formula, Indicates stable processing; Indicates a precursor to flutter; This indicates that the flutter has intensified; Indicates severe flutter; The first classification result also includes real-time probability confidence. .

7. The method for processing thin-walled irregular-shaped cantilevered pipes according to claim 6, characterized in that, The multi-level threshold judgment method in step S5 includes: Based on the mean radial force and radial force standard deviation Radial force fluctuation rate is obtained : ； When R When the threshold judgment result is reached, output multi-level threshold judgment results. =1 indicates that the current processing state is stable processing; when R When the threshold judgment result is reached, output multi-level threshold judgment results. =2 indicates that the current processing state is a precursor to chatter; when R When the threshold judgment result is reached, output multi-level threshold judgment results. =3 indicates that the current processing condition is characterized by increased chatter. when When R is reached, the multi-level threshold judgment result is output. =4 indicates that the current processing condition is severe chatter.

8. The method for processing thin-walled irregular-shaped cantilevered pipes according to claim 7, characterized in that, Step S5, when obtaining the final processing state based on the first and second judgment results, includes: like At that time, the real-time status label is used. or multi-level threshold judgment results As the final processing state; like At that time, and Then use the real-time status label As the final processing state; like At that time, and ,but In this case, the multi-level threshold judgment result is adopted. ; like At that time, and ,at the same time At the same time, the state of the previous moment remains unchanged to avoid frequent jumps.

9. The method for processing thin-walled irregular-shaped cantilevered pipes according to claim 8, characterized in that, When adjusting the machining process of the double-edged boring tool based on the final machining state in step S6, it includes: State 1 indicates that the current processing status is stable, so the current parameters are maintained and monitoring continues; State 2: Indicates that the current processing state is a precursor to flutter, so an early warning is issued and minor adjustments are performed; State 3: Indicates that the current processing state is flutter intensification, which triggers intervention and executes multi-level active adjustment; Status 4: Indicates that the current machining status is severe chatter, which triggers an emergency stop and executes emergency tool retraction protection; The micro-adjustment refers to: the piezoelectric ceramic actuator adjusting the low-edge extension amount from ±0.02 mm to ±0.05 mm, breaking the emerging resonance condition by performing a small adjustment, and randomly selecting the direction of extension or retraction, which is used to determine which direction of micro-adjustment is more effective based on the final processing state at the next moment; The multi-stage active adjustment includes the following steps: the piezoelectric ceramic actuator adjusts the low blade extension amount in the reverse direction by 0.05mm to 0.10mm, reduces the rotational speed by 10% to 20%, and reduces the feed rate by 30% to 50%. The reverse adjustment means that the adjustment direction is opposite to that of the previous moment. The large adjustment changes the dynamic characteristics of the system to suppress the developed chatter. The emergency retraction protection includes the following actions in sequence: stop feed, radial retraction of 0.2 mm, spindle stop, and alarm.

10. The method for processing thin-walled irregular-shaped cantilevered pipes according to claim 9, characterized in that, In step S6, the tremor suppression rate is calculated. Evaluate the effectiveness of the adjustments: ； In the formula, This represents the standard deviation of radial force at the time corresponding to the final machining state. This represents the standard deviation of radial force after the machining process of the double-edged boring tool has been adjusted. like This indicates that flutter suppression is effective; like If this indicates that chatter suppression is ineffective, then within the corresponding state, change the machining parameters of the double-edged boring tool. If this is done three times consecutively, Still smaller than If so, it is determined to be an invalid adjustment.

Images