Anti-collision beam anti-deformation cutting system device based on magnetic coupling servo

Abstract

The present application relates to pipe processing equipment technical field, disclose a kind of anti-deformation cutting system device of anti-collision beam based on magnetic coupling servo, including machine body, cutting device and the short inner support slider placed in the inner cavity of anti-collision beam beam pipe, servo execution module is established magnetic force coupling with short inner support slider through external magnetic field across pipe wall, sensor acquisition module collects the magnetic flux distribution outside pipe and generates magnetic topology matrix data, according to this, position deviation is calculated and closed-loop alignment control is executed by solving compensation module, after alignment, actuating demagnetization module exports magnetic field signal to trigger the action of the tensioning structure inside short inner support slider, to build rigid support in pipe, characteristic verification module executes double-peak ratio operation according to secondary sampling magnetic field data, and issues release instruction to cutting device after verifying physical support state. The present application realizes the closed-loop verification of non-contact servo support and support state inside pipe, to prevent the deformation phenomenon of pipe in cutting stress process.

Description

Technical Field

[0001] This invention relates to the field of pipe processing equipment technology, specifically to an anti-collision beam anti-deformation cutting system device based on magnetic coupling follow-up. Background Technology

[0002] Automotive crash beams are typically made of hollow metal tubing. Cutting is a common process in their production. Because the inside of the tube is hollow, the cutting area is prone to local concavity or deformation when external force is applied by the cutting device, which affects the shape and dimensional accuracy of the cut end face.

[0003] To reduce cutting deformation, a support structure needs to be installed inside the cavity corresponding to the cutting position. Existing solutions mostly use external clamping or internal support push rods. External clamps can only restrict the overall displacement of the beam and tube, and it is difficult to provide effective support for the local stress area near the cutting point. Slender push rods are also prone to bending during long-distance feeding, making it difficult for the internal support block to accurately correspond to the cutting position.

[0004] Due to the obstruction of the metal tube wall, the external system has difficulty in directly determining whether the internal support block has reached the predetermined position, and it is also difficult to confirm whether it has been tightened and fits the inner wall. How to achieve the follow-up positioning of the internal support structure in the closed beam tube, and to reliably verify the support status before cutting, is a technical problem that needs to be solved in this field. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a magnetically coupled, follow-up anti-collision beam anti-deformation cutting system. This system solves the problem that anti-collision beams deform under stress during fixed-length cutting due to the lack of physical support inside the tube wall. Furthermore, existing internal support structures are difficult to dynamically align with external processing stations with high precision within a closed tube cavity, causing the internal support position to deviate from the actual cutting point, thus affecting processing accuracy and quality.

[0006] To achieve the above objectives, the present invention provides the following technical solution: This invention provides a magnetically coupled, follow-up anti-collision beam anti-deformation cutting system device, including a body, an anti-collision beam tube at the top of the body, a short-section inner support slider on the inner side of the anti-collision beam tube, a feature verification module fixedly connected to the top of the body, a power distribution box fixedly connected to the middle of the body, an actuation demagnetization module fixedly connected to the outer side of the power distribution box, a calculation compensation module at the bottom of the actuation demagnetization module, the calculation compensation module fixedly connected to the outer side of the power distribution box, a cutting device installed at the top of the body, a processing channel reserved in the middle of the cutting device for the anti-collision beam tube to slide through, and a processing channel with a feature verification module fixedly connected to the outer side of the processing channel. The sensor acquisition module is fixedly connected to the middle of the cutting device, and the follow-up execution module is fixedly connected to the top of the cutting device. The follow-up execution module establishes a magnetic coupling relationship with the short-section inner support slider through the pipe wall of the anti-collision beam via a magnetic field. The calculation and compensation module performs closed-loop alignment control based on the first magnetic topology matrix data sent by the sensor acquisition module. The actuation demagnetization module triggers the mechanical tensioning structure inside the short-section inner support slider to construct rigid support based on the alignment status. The feature verification module verifies the support status based on the second magnetic topology matrix data of the sensor acquisition module and issues an interlock release command to the cutting device.

[0007] By adopting the above technical solution, the follow-up execution module generates a magnetic field at the cutting station, and uses magnetic attraction to make the short inner support slider move synchronously with the anti-collision beam inside the tube. The sensing and acquisition module collects the magnetic field distribution outside the tube wall to construct the first magnetic topology matrix data. The calculation and compensation module calculates the position deviation and adjusts the feeding progress of the equipment accordingly to achieve the alignment of the internal support unit and the external processing channel. After centering, the actuation demagnetization module changes the external excitation state and triggers the internal mechanical tensioning structure in a non-contact manner, so that the short-section inner support slider completes physical tensioning inside the tube, resisting the radial force generated by the downward pressure of the cutting device and avoiding tube deformation. The feature verification module evaluates the changes in the spatial magnetic field shape based on the second magnetic topology matrix data. After confirming that the internal rigid support has been constructed, the cutting device is lowered for processing to prevent miscutting or missing cuts caused by the support not being in place.

[0008] Furthermore, the follow-up execution module includes an external excitation component, which is arranged on the outside of the anti-collision beam tube; the short-section inner support slider is equipped with a magnetic component; the external excitation component is used to generate a magnetic field and cooperates with the magnetic component to establish a magnetic coupling relationship.

[0009] By adopting the above technical solution, the magnetic field generated by the external excitation component penetrates the pipe wall and acts on the internal magnetic component, generating radial attraction and converting it into axial static friction. When the anti-collision beam pipe is fed, the static friction drives the short inner support slider to move synchronously, maintaining its relative position to the external cutting station.

[0010] Preferably, the short-section inner support slider includes a main magnetic core, a wedge-shaped tensioning mechanism, a preload spring, a release latch, and a secondary driven magnetic pole. The wedge-shaped tensioning mechanism is installed in the central cavity of the short-section inner support slider and includes a central push rod and radial wedge blocks evenly distributed along the circumference. The preload spring is sleeved on the outer surface of the central push rod. The release latch includes a high-permeability armature and a spring latch. The spring latch is embedded in a limiting groove on the surface of the central push rod to lock the preload spring. The secondary driven magnetic pole is installed on the outer end face of the radial wedge blocks.

[0011] By adopting the above technical solution, the spring clip uses a mechanical structure to constrain the preloaded spring. After the high-permeability magnetic armature is forced to vibrate and disengage, the preloaded spring releases its elastic potential energy to push the central push rod, which drives the radial wedge block to expand outward and press against the inner wall of the anti-collision beam tube. The secondary driven magnetic pole follows the movement of the radial wedge block and generates spatial displacement, changing the distribution of magnetic flux in the external space and providing detection basis for the external sensing acquisition module.

[0012] Furthermore, the sensing and acquisition module includes multiple spatially arranged high-sensitivity magnetic field sensors and a multi-channel synchronous analog-to-digital converter. The multiple high-sensitivity magnetic field sensors are arranged around the outer circumference and axial direction of the processing channel to construct a two-dimensional cylindrical detection grid, which is used to collect the magnetic flux distribution state of the space around the anti-collision beam tube and output analog voltage signals. The multi-channel synchronous analog-to-digital converter converts the analog voltage signals into digital signals and constructs them into first magnetic topology matrix data and second magnetic topology matrix data according to different sampling times.

[0013] By adopting the above technical solution, the array of high-sensitivity magnetic field sensors converts the physical magnetic field in three-dimensional space into a continuous electrical signal. The multi-channel synchronous analog-to-digital converter uses a global clock for unified conversion, avoiding time differences between channels that could cause data acquisition errors and ensuring the time consistency of the digital matrix in the data.

[0014] Preferably, the machine body is equipped with a servo feeding drive mechanism, and the calculation and compensation module establishes data communication connections with the sensing and acquisition module and the servo feeding drive mechanism respectively; the calculation and compensation module receives the first magnetic topology matrix data generated in real time, and runs the two-dimensional discrete cross-correlation operation logic to calculate the axial position deviation and radial rotation deviation, thereby generating a feedforward compensation control signal and sending it to the servo feeding drive mechanism.

[0015] By adopting the above technical solution, the calculation and compensation module performs cross-correlation calculations, extracts the deviation information in the matrix, and converts this deviation into a drive compensation command to be sent to the servo feeding drive mechanism. By adjusting the speed, the sliding error is dynamically eliminated, and the closed-loop alignment of the inner and outer mechanisms is achieved.

[0016] Furthermore, the actuation demagnetization module establishes electrical connections with the distribution box and the external excitation assembly respectively; the actuation demagnetization module is used to output transient low-frequency alternating pulse current to the external excitation assembly, which excites the alternating pulse magnetic field to act on the high-permeability armature, causing the spring pin to disengage from the limiting groove, and the preloaded spring releases the stored elastic potential energy to drive the central push rod to push the radial wedge block to press against the inner wall of the anti-collision beam tube.

[0017] By adopting the above technical solution, the actuation demagnetization module utilizes the physical characteristic of low-frequency alternating magnetic field having high penetration depth to allow magnetic lines of force to penetrate the pipe wall and induce mechanical vibration of high-permeability magnetic armature. The mechanical vibration of high-permeability magnetic armature overcomes the static friction resistance of the latch, triggers the release of potential energy and drives the wedge block, thereby achieving non-contact control.

[0018] Preferably, the feature verification module establishes signal connections with the sensing acquisition module and the cutting device respectively; the actuation demagnetization module injects a weak compensation current in the reverse direction after cutting off the transient low-frequency alternating pulse current to cancel the residual magnetism of the space background; the feature verification module receives the second magnetic topology matrix data, extracts the amplitude of the main magnetic peak corresponding to the main magnetic core and the amplitude of the secondary magnetic peak corresponding to the secondary driven magnetic pole, calculates the ratio of the two peaks and performs the ratio calculation to verify the support status, and sends an interlock release command to the cutting device after confirmation.

[0019] By adopting the above technical solution, the reverse magnetic field excited by the weak compensation current neutralizes the residual magnetism in the space, enabling the sensor component to restore its linear response. The feature verification module confirms the support status of the radial wedge block by analyzing the ratio change of the bimodal feature, and constructs an interlocked release control.

[0020] Furthermore, the main magnetic core adopts an asymmetric polarity arrangement structure. The N pole and S pole of the main magnetic core are distributed in unequal volume ratios in the axial direction of the short-section inner support slider. The internal magnetic energy product of the secondary driven magnetic pole is smaller than the internal magnetic energy product of the main magnetic core.

[0021] By adopting the above technical solution, the asymmetric volume distribution enables the main magnetic core to form a single peak region in the external detection space, eliminating the spurious peak interference generated by the symmetric multi-peak structure in the position calculation. The secondary driven magnetic pole has a low magnetic energy product, avoiding reverse interference to the external magnetic field during the traction stage.

[0022] Preferably, the sensing acquisition module integrates digital filtering logic. The digital filtering logic runs a first-order infinite impulse response low-pass filtering algorithm to filter the real-time acquired raw discrete magnetic field strength data, and runs a neighborhood mean interpolation algorithm to complete the values ​​of abnormal data nodes, thereby constructing the first magnetic topology matrix data.

[0023] By adopting the above technical solutions, the low-pass filtering algorithm filters out high-frequency noise signals caused by electromagnetic interference during system operation, retains low-frequency position data representing component movement, and the interpolation algorithm repairs missing data caused by sensor defects, ensuring the structural integrity and continuity of the magnetic topology matrix data.

[0024] Furthermore, the solution compensation module is equipped with a discrete PID control loop. The solution compensation module inputs the axial position deviation and radial rotation deviation into the discrete PID control loop, and generates a feedforward compensation control signal by combining the calibrated feedforward gain coefficient.

[0025] By adopting the above technical solution, and combining the PID feedback loop with the feedforward gain circuit, the system introduces feedforward quantity during the servo control process to offset the predicted position offset caused by the change in the pipe reference speed, thereby reducing the adjustment lag and oscillation amplitude.

[0026] This invention provides a collision-resistant and deformation-resistant cutting system device based on magnetic coupling and servo motion. It has the following beneficial effects: 1. This invention generates a magnetic field that penetrates the pipe wall through a follow-up execution module and establishes magnetic coupling with the short-section inner support slider. It also uses an actuation demagnetization module to output an alternating pulse current to excite a high-permeability magnetic armature, release a preloaded spring, and drive a wedge-shaped tensioning mechanism to press the inner wall of the pipe. This solution enables the follow-up movement and mechanical triggering of the internal slider in a closed state without physical connections, solving the structural problem of difficult rigid support arrangement inside long pipes and preventing deformation of the anti-collision beam pipe when subjected to cutting force.

[0027] 2. This invention utilizes multiple sets of high-sensitivity magnetic field sensors to construct a two-dimensional cylindrical surface detection grid, collects the magnetic flux in the space surrounding the anti-collision beam tube and converts it into magnetic topology matrix data. Subsequently, the calculation and compensation module performs two-dimensional discrete cross-correlation calculations to calculate the axial and radial position deviations. By inputting the position deviations into the discrete PID control loop to generate a feedforward compensation control signal, the system can correct the relative position error between the external mechanism and the internal slider in real time, ensuring that the cutting processing point is accurately aligned with the internal support position.

[0028] 3. This invention receives secondary sampling data after the support action is completed through a feature verification module, extracts the main magnetic peak amplitude of the main magnetic core and the secondary magnetic peak amplitude of the secondary driven magnetic pole that moves with the wedge block, and performs a double peak ratio calculation. This calculation process converts the spatial displacement of the invisible mechanical parts inside the tube into quantifiable magnetic field data changes, thereby objectively confirming that the rigid support has been established before the cutting action is performed, avoiding the generation of scrap due to support failure, and improving processing accuracy and yield. Attached Figure Description

[0029] Figure 1 This is a perspective view of the present invention; Figure 2This is a side view of the structure of the present invention; Figure 3 This is a schematic diagram of the anti-collision beam anti-deformation cutting system architecture based on magnetic coupling servoing according to the present invention; Figure 4 This is a schematic diagram of the architecture of the follow-up execution module and the truncated inner support slider of the present invention; Figure 5 This is a schematic diagram of the structure and data processing logic of the sensing acquisition module of the present invention; Figure 6 This is a schematic diagram of the processing logic of the solution compensation module of the present invention; Figure 7 This is a schematic diagram of the electrical control and desaturation decoupling logic of the actuation demagnetization module of the present invention; Figure 8 This is a schematic diagram of the closed-loop verification logic of the feature verification module of the present invention; Figure 9 This is a line graph showing the axial alignment deviation of the present invention. Figure 10 This is a bar chart showing the deformation of the cut section in this invention.

[0030] The components include: 1. Body; 2. Anti-collision beam tube; 3. Follow-up execution module; 4. Shortened internal support slider; 5. Feature verification module; 6. Sensor acquisition module; 7. Cutting device; 8. Power distribution box; 9. Actuation demagnetization module; and 10. Calculation compensation module. Detailed Implementation

[0031] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0032] Reference Figure 1 , Figure 2 and Figure 3This invention provides a magnetically coupled follow-up anti-collision beam anti-deformation cutting system device, including a body 1, an anti-collision beam tube 2 is provided on the top of the body 1, a short-section inner support slider 4 is provided on the inner side of the anti-collision beam tube 2, a feature verification module 5 is fixedly connected to the top of the body 1, a power distribution box 8 is fixedly connected to the middle of the body 1, an actuation demagnetization module 9 is fixedly connected to the outer side of the power distribution box 8, a calculation compensation module 10 is provided at the bottom of the actuation demagnetization module 9, the calculation compensation module 10 is fixedly connected to the outer side of the power distribution box 8, a cutting device 7 is installed on the top of the body 1, a processing channel for the anti-collision beam tube 2 to slide through is reserved in the middle of the cutting device 7, a sensing acquisition module 6 is provided on the outer side of the processing channel, the sensing acquisition module 6 is fixedly connected to the middle of the cutting device 7, and a follow-up execution module 3 is fixedly connected to the top of the cutting device 7.

[0033] The anti-collision beam anti-deformation cutting system based on magnetic coupling is deployed on the machine body 1. The machine body 1 provides the physical installation benchmark for each component of the system. The cutting device 7 and the power distribution box 8 are fixedly installed in the preset position of the machine body 1. The cutting device 7 is used to perform fixed-length cutting operations on the anti-collision beam tube 2. The power distribution box 8 contains AC and DC power supply circuits to provide basic power supply for each electrical module and drive mechanism in the system.

[0034] The follow-up execution module 3 is located at the cutting position set in the machine body 1. The follow-up execution module 3 includes a short inner support slider 4 and an external excitation assembly. The short inner support slider 4, as an independent mechanical support unit, is pre-placed in the inner cavity of the anti-collision beam tube 2. The external excitation assembly of the follow-up execution module 3 is arranged on the outside of the anti-collision beam tube 2. The follow-up execution module 3 uses the magnetic field generated by the external excitation assembly to establish a magnetic coupling relationship with the magnetic component inside the short inner support slider 4 through the tube wall of the anti-collision beam tube 2. When the anti-collision beam tube 2 is continuously fed axially, the short inner support slider 4, under the strong constraint of the above-mentioned magnetic coupling, overcomes the frictional force on the inner side of the tube wall and slides relative to the tube wall, thus always hovering directly below the cutting position.

[0035] The sensor acquisition module 6 is installed in the peripheral space of the follow-up execution module 3. The sensor acquisition module 6 contains multiple magnetic field sensors arranged in space. The sensor acquisition module 6 is used to collect the magnetic flux distribution state of the space around the anti-collision beam tube 2 in real time. The sensor acquisition module 6 acquires the magnetic field strength signal after penetrating the metal tube wall of the anti-collision beam tube 2 and converts the multi-dimensional physical magnetic field signal into digital magnetic topology matrix data.

[0036] The calculation and compensation module 10 establishes a data communication connection with the sensor acquisition module 6 and the servo feeding drive mechanism on the machine body 1. The calculation and compensation module 10 receives the real-time magnetic topology matrix data sent by the sensor acquisition module 6. The calculation and compensation module 10 runs the built-in cross-correlation operation logic to calculate the axial position deviation and radial rotation deviation of the short inner support slider 4 relative to the external reference datum. The calculation and compensation module 10 generates a feedforward compensation control signal based on the calculated position deviation and sends the compensation control signal to the servo feeding drive mechanism to adjust the feed speed curve of the anti-collision beam tube 2 to achieve closed-loop alignment between the internal support component and the external processing station.

[0037] The actuation demagnetization module 9 is electrically connected to the power distribution box 8 and the follow-up execution module 3 respectively. The actuation demagnetization module 9 controls the current output state of the external excitation component of the follow-up execution module 3. When the calculation and compensation module 10 confirms that the alignment deviation has entered the set range and the anti-collision beam tube 2 stops feeding, the actuation demagnetization module 9 outputs an alternating pulse current of a specific frequency to the follow-up execution module 3, triggering the mechanical tensioning structure inside the short-section inner support slider 4 to move. After the triggering action is completed, the actuation demagnetization module 9 stops outputting the positive excitation current and outputs a reverse compensation current to cancel the background magnetic field in the system space, so that the sensor of the sensing acquisition module 6 is out of the magnetic saturation range.

[0038] The feature verification module 5 is connected to the sensor acquisition module 6 and the cutting device 7 on the body 1 respectively. After the actuation demagnetization module 9 completes the demagnetization operation, the feature verification module 5 receives the magnetic flux data of the observation period re-acquired by the sensor acquisition module 6. The feature verification module 5 extracts the topological peak features in the data and performs ratio calculation. The feature verification module 5 verifies whether the short inner support slider 4 has reached the preset physical rigid support state inside the anti-collision beam tube 2 through the calculation result. When the feature verification module 5 confirms that the support state meets the requirements, it sends an interlock release command to the cutting device 7, which instructs the cutting device 7 to perform a cutting operation on the anti-collision beam tube 2.

[0039] See attached document Figure 4 In this embodiment, the follow-up execution module 3 is installed in the cutting station area of ​​the machine body 1. The specific components inside the follow-up execution module 3 are fixed by a preset physical support. As a preferred embodiment, the follow-up execution module 3 includes an external excitation component, which is arranged around the outside of the anti-collision beam tube 2. The external excitation component includes an electromagnet array composed of multi-turn copper coils and silicon steel sheet yokes. During system initialization and operation, the power distribution box 8 outputs positive DC power to the external excitation component, thereby causing the external excitation component to generate a constant intensity traction magnetic field based on the principle of electromagnetic induction.

[0040] Based on the aforementioned external magnetic field environment, the short-section inner support slider 4 is independently installed in the internal space of the anti-collision beam tube 2. The short-section inner support slider 4 includes a main magnetic core, a wedge-shaped tensioning mechanism, a preload spring, a release latch, and secondary driven magnetic poles. To ensure the monotonicity of the magnetic field distribution, the main magnetic core is preferably a permanent magnet made of neodymium iron boron material, and the main magnetic core adopts an asymmetrical polarity arrangement structure. Specifically, the N pole and S pole of the main magnetic core are distributed in unequal volume ratios in the axial direction of the short-section inner support slider 4. The unequal volume distribution structure of the main magnetic core makes the main magnetic core form a single absolute magnetic flux peak region in the external space of the anti-collision beam tube 2. The single absolute magnetic flux peak region avoids interference from symmetrical multi-magnetic peaks during system detection.

[0041] During dynamic traction, the traction magnetic field generated by the external excitation component penetrates the metal solid wall of the anti-collision beam tube 2. The traction magnetic field attracts the main magnetic core, thereby forming a radial magnetic attraction force between the external excitation component and the main magnetic core. Due to the contact of the friction surfaces between the objects, the above-mentioned radial magnetic attraction force generates an axial static friction force between the surface of the short inner support slider 4 and the inner wall of the anti-collision beam tube 2. When the servo drive mechanism on the body 1 pushes the anti-collision beam tube 2 forward along the axial direction, the axial static friction force forces the short inner support slider 4 to overcome the sliding resistance and move forward synchronously with the anti-collision beam tube 2. For the specific mechanical transmission components of the servo drive mechanism, those skilled in the art can use a rack and pinion transmission mechanism. The specific mechanical transmission components of such servo drive mechanisms are well known in the art and will not be described in detail here.

[0042] The wedge-shaped tensioning mechanism is installed in the central cavity of the short-section inner support slider 4. The wedge-shaped tensioning mechanism includes a central push rod and radial wedges evenly distributed along the circumference. The preload spring is sleeved on the outer surface of the central push rod. The release latch includes a high-permeability magnetic armature and a spring pin. In the initial conveying state of the system, the spring pin is embedded in the limiting groove on the surface of the central push rod. The preload spring is locked in the compressed energy storage state by means of the mechanical self-locking effect. When in the compressed energy storage state, the radial wedges are kept in the contracted position close to the central axis. At this time, a mechanical gap is maintained between the outer surface of the radial wedges and the inner wall of the anti-collision beam tube 2 to ensure that the short-section inner support slider 4 is not stuck during the follow-up forward movement.

[0043] As a physical beacon for state verification, the secondary driven magnetic pole is mounted on the outer end face of the radial wedge block by fixing bolts. The internal magnetic energy product of the secondary driven magnetic pole is smaller than that of the main magnetic core to prevent the secondary driven magnetic pole from generating excessive reverse disturbance to the main traction magnetic field. When the radial wedge block is in the contracted position, the secondary driven magnetic pole is close to the main magnetic core in physical space.

[0044] When the calculation and compensation module 10 determines that the machining station is aligned and the anti-collision beam tube 2 stops feeding, the actuation demagnetization module 9 outputs an alternating pulse current of a specific frequency to the external excitation component. The external excitation component generates an alternating pulse magnetic field according to the alternating pulse current. The alternating pulse magnetic field passes through the anti-collision beam tube 2 and acts on the high-permeability armature. Since the high-permeability material has a highly sensitive magnetic reluctance response to the alternating magnetic field, the high-permeability armature generates high-frequency mechanical vibration under the action of the alternating pulse magnetic field. When the transient inertial force generated by the high-frequency mechanical vibration reaches the preset tripping threshold, it can overcome the locking friction force and force the spring pin to disengage from the limiting groove on the surface of the central push rod.

[0045] In this physical conversion process, the basic electromagnetic attraction force generated by the alternating pulsed magnetic field on the high-permeability armature follows Maxwell's principle of electromagnetic attraction. The system must ensure that the peak excitation force is strictly greater than the static friction damping of the spring clip. The transient magnetic field dynamics generated by the alternating pulsed magnetic field acting on the high-permeability armature are calculated according to the following formula: ; In the formula, F represents the transient magnetic field dynamics experienced by the high-permeability magnetic armature; denoted by , B represents the permeability constant of the air gap environment inside the stubbed inner support slider 4. The value of the permeability constant is determined based on the physical properties of the air gap medium and is not zero. B represents the magnetic induction intensity amplitude of the alternating pulse magnetic field at the surface of the high-permeability armature. The value of the magnetic induction intensity amplitude needs to be compensated and calibrated according to the eddy current attenuation caused by the wall thickness of the anti-collision beam tube 2. The magnetic induction intensity amplitude is set between 0.5T and 1.5T to ensure penetration. S represents the magnetized cross-sectional area of ​​the high-permeability armature exposed to the magnetic field. f represents the frequency of the alternating pulse current. The specific value of the frequency needs to avoid the inherent resonant frequency point of the internal mechanical structure of the stubbed inner support slider 4, and the frequency needs to be greater than the lower limit of the cutoff bandwidth of the skin effect of the anti-collision beam tube 2. t represents the duration of the alternating pulse magnetic field.

[0046] Triggered by the aforementioned transient magnetic field dynamics, after the spring clip disengages from the limiting groove, the preloaded spring releases its stored elastic potential energy. The preloaded spring drives the central push rod to move axially, which in turn causes the inclined structure on the surface of the central push rod to press outward against the radial wedge block. The radial wedge block moves outward in the radial direction, ultimately causing the outer surface of the radial wedge block to press tightly against the inner wall of the anti-collision beam tube 2. Through the mechanical transmission sequence, the radial wedge block constructs a rigid physical support inside the anti-collision beam tube 2 to resist cutting deformation.

[0047] As the radial wedge moves outward in the radial direction, the secondary driven magnetic pole rigidly mounted on the radial wedge moves outward in sync, causing the secondary driven magnetic pole to move away from the main magnetic core. The change in the spatial position of the secondary driven magnetic pole changes the static magnetic field distribution state of the space surrounding the system. At this time, the displacement of the secondary driven magnetic pole generates a topological sub-peak physical feature in the spatial magnetic field that is independent of the main magnetic peak. The topological sub-peak physical feature provides the physical input basis for the subsequent non-contact state verification performed by the feature verification module 5, thereby constructing a complete physical closed loop from mechanical actuation to magnetic field feedback.

[0048] See attached document Figure 5 In this embodiment, the sensing acquisition module 6 is installed in the peripheral space of the follow-up execution module 3. As a preferred embodiment, the sensing acquisition module 6 includes a detection array composed of high-sensitivity magnetic field sensors. For the specific selection of high-sensitivity magnetic field sensors, those skilled in the art can use linear Hall integrated circuits or anisotropic magnetoresistive sensors. The specific hardware circuit design of linear Hall integrated circuits or anisotropic magnetoresistive sensors is a well-known technology in the field and will not be described in detail here. The high-sensitivity magnetic field sensors of the sensing acquisition module 6 are regularly arranged along the outer circumference and axial direction of the anti-collision beam tube 2, thereby constructing a two-dimensional cylindrical detection grid in physical space.

[0049] To achieve a precise mapping from physical location to mathematical matrix, the sensing acquisition module 6 establishes a local cylindrical coordinate system with the axis of the anti-collision beam tube 2 as the reference axis. In the local cylindrical coordinate system, the axial direction is discretized into multiple axial sampling points, and the circumferential direction is discretized into multiple circumferential sampling angles. Based on the spatial Nyquist sampling theorem, the physical distance between the axial sampling points and the circumferential sampling angles must be strictly less than half of the wavelength corresponding to the highest spatial frequency of the spatial magnetic field topology to prevent spatial aliasing distortion during feature reconstruction. Based on the above physical distance constraint, the sensing acquisition module 6 maps the installation position of each high-sensitivity magnetic field sensor to specific two-dimensional spatial coordinates.

[0050] During the movement of the short inner support slider 4 driven by the follow-up execution module 3, the magnetic lines of force of the main magnetic core inside the short inner support slider 4 penetrate the metal tube wall of the anti-collision beam tube 2 and reach the external space. The high-sensitivity magnetic field sensor of the sensing and acquisition module 6 senses the physical magnetic field penetrating the anti-collision beam tube 2 and converts the physical magnetic field into a continuous analog voltage signal. In order to avoid the characteristic distortion of the spatial magnetic field topology caused by the time difference generated by multi-channel data acquisition, the sensing and acquisition module 6 integrates a multi-channel synchronous analog-to-digital converter. Under the drive of the same global clock signal, the multi-channel synchronous analog-to-digital converter performs strict synchronous sampling of the analog voltage signals output by all high-sensitivity magnetic field sensors, ensuring the absolute alignment and state consistency of multi-source spatial data in the time dimension.

[0051] Due to electromagnetic interference and mechanical vibration noise during the feeding of the anti-collision beam tube 2 in the industrial environment, after the sensor acquisition module 6 converts the analog voltage signal into a digital signal, it runs the built-in digital filtering logic to smooth the data based on the discrete-time system theory in digital signal processing. Specifically, the sensor acquisition module 6 uses a first-order infinite impulse response low-pass filtering algorithm to filter the real-time acquired raw discrete magnetic field strength data to remove high-frequency spike noise and prevent noise abrupt changes from causing misjudgment in subsequent peak-finding logic. The filtering calculation is based on the following formula: ; In the formula, This indicates the position of the axis coordinate at the current sampling time. Circumferential angle The magnetic field strength value of the high-sensitivity magnetic field sensor at the location is filtered; t represents the discrete-time index of the current sampling moment. This represents the coordinate position of the i-th discrete axial sampling point in the local cylindrical coordinate system; This represents the azimuth value of the j-th discrete circumferential sampling angle in the local cylindrical coordinate system; This represents the filtering smoothing coefficient. The value of the filtering smoothing coefficient is strictly defined within an open interval greater than 0 and less than 1 to avoid the risk of zero gain or division by zero and failure of the all-pass filtering in the algorithm. In actual system operation, the filtering smoothing coefficient is calibrated to 0.1 to 0.3 based on the statistical variance of the background noise. This represents the raw discrete magnetic field strength data output by the multi-channel synchronous analog-to-digital converter at the current sampling time; This represents the magnetic field strength value recorded at the same spatial coordinate position in the previous sampling period.

[0052] By introducing an infinite impulse response mechanism that includes historical state weights, the above calculation steps can suppress the sudden interference of high-frequency white noise on the magnetic flux amplitude while preserving the low-frequency macroscopic displacement characteristics of the main magnetic core.

[0053] After digital filtering logic processing, the sensor acquisition module 6 splices and packages the magnetic induction intensity values ​​at the corresponding positions of each two-dimensional spatial coordinate according to the row and column order of the local cylindrical coordinate system. To prevent hardware failure of individual high-sensitivity magnetic field sensors from causing null values ​​in the matrix and triggering singularities in subsequent matrix operations, the sensor acquisition module 6 uses a neighborhood mean interpolation algorithm to complete the values ​​of abnormal or missing data nodes in this stage. After ensuring the completeness of the spatial data, the sensor acquisition module 6 constructs the spliced ​​and packaged data into the first magnetic topology matrix data. The first magnetic topology matrix data completely represents the three-dimensional fluctuation of magnetic flux in the outer space of the anti-collision beam tube 2 under the current sampling period in terms of data structure. The first magnetic topology matrix data is sent to the calculation and compensation module 10 in real time for processing and analysis as the underlying data input for subsequent spatial pose calculation.

[0054] See attached document Figure 6 In this embodiment, the calculation and compensation module 10 establishes a data communication connection with the sensing and acquisition module 6 and the servo drive mechanism on the body 1. During continuous operation of the system, the calculation and compensation module 10 receives the first magnetic topology matrix data sent in real time by the sensing and acquisition module 6.

[0055] Based on the principles of pattern matching and spatial signal processing, the solution compensation module 10 performs similarity calculations between the first magnetic topology matrix data and the pre-calibrated and stored zero-position reference matrix. Specifically, to eliminate common-mode interference from the background steady-state magnetic field and highlight pose differences, the solution compensation module 10 runs its built-in two-dimensional discrete cross-correlation operation logic to solve for the current spatial magnetic field distribution offset in a global optimization manner. The two-dimensional discrete cross-correlation calculation is performed according to the following formula: ; In the formula, This represents the two-dimensional cross-correlation coefficient under the condition that the axial offset is u and the circumferential offset is v; u represents the axial translation index in the local cylindrical coordinate system; v represents the circumferential rotation index in the local cylindrical coordinate system. This represents the magnetic flux density value in the i-th row and j-th column of the first magnetic topology matrix data; This represents the global average magnetic flux value of the first magnetic topology matrix data; This represents the coordinate value of the corresponding point in the zero-position reference matrix after offset; represents the global average magnetic flux value of the zero-position reference matrix; i represents the row index in the local cylindrical coordinate system; j represents the column index in the local cylindrical coordinate system; This represents a very small positive number introduced to prevent the denominator from approaching zero. The value of the very small positive number is usually set to 10 to the power of negative octave to ensure the stability of the division operation and avoid calculation overflow.

[0056] Through the above-mentioned normalized cross-correlation calculation, the solution compensation module 10 can offset the overall magnetic field drift interference caused by the external operating temperature. The solution compensation module 10 traverses the set search window to obtain the two-dimensional cross-correlation number. After obtaining the coordinates of the global maximum point, to avoid the one-sided judgment based solely on a single extreme value and the interference of false peaks caused by sudden noise, the solution compensation module 10 does not directly use the coordinates of the global maximum point as the final result. As a preferred method, the solution compensation module 10 extracts all two-dimensional cross-correlation coefficients within the surrounding set neighborhood of the global maximum point coordinates and performs centroid weighted calculation based on the cross-correlation amplitude. By introducing multi-dimensional correlation data from the surrounding neighborhood for weighted evaluation, the solution compensation module 10 accurately calculates the optimal offset combination at the sub-grid level. The solution compensation module 10 maps the axial translation index in the optimal offset combination to the axial position deviation of the short inner support slider 4 relative to the cutting station of the body 1. The solution compensation module 10 maps the circumferential rotation index in the optimal offset combination to the radial rotation deviation.

[0057] After obtaining the axial position deviation and radial rotation deviation, in order to overcome the nonlinear frictional slippage during dynamic traction, the calculation and compensation module 10 inputs the axial position deviation and radial rotation deviation into the built-in discrete PID control loop; considering the time lag in pure feedback control, the calculation and compensation module 10 combines a feedforward control mechanism to generate a feedforward compensation control signal, and the calculation logic of position compensation is based on the following formula: ; In the formula, This represents the feedforward compensation control signal output in the k-th control cycle. The feedforward compensation control signal is specifically manifested as a speed fine-tuning command; k represents the index of the current discrete control cycle. This represents the system deviation input in the k-th control cycle. The system deviation is the axial position deviation or radial rotation deviation calculated above. This represents the proportional gain coefficient, which is tuned based on the mechanical response stiffness of the body 1 using the critical proportional method. The value of the proportional gain coefficient is usually set between 0.5 and 2.5. This represents the integral gain coefficient, which is used to eliminate the steady-state error of the system. The value of the integral gain coefficient is usually set between 0.01 and 0.1. This represents the differential gain coefficient, which is used to suppress dynamic overshoot. The value of the differential gain coefficient is usually set between 0.1 and 0.5. This represents the feedforward gain coefficient, which is calibrated based on the inertia characteristics of the driven load. The feedforward gain coefficient is determined to be between 0.8 and 1.2 through system identification experiments. This represents the reference feed speed of the anti-collision beam tube 2 in the kth control cycle; This represents the system deviation in the (k-1)th control cycle; n represents the summation index from the initial moment to the current control cycle.

[0058] By introducing a feedforward gain stage, the calculation and compensation module 10 can compensate for the predicted position offset caused by the change in the reference speed before the conventional feedback adjustment lag occurs. After completing the internal calculation, the calculation and compensation module 10 sends the generated feedforward compensation control signal to the servo drive mechanism on the machine body 1. The servo drive mechanism on the machine body 1 receives the feedforward compensation control signal and adaptively fine-tunes the deceleration curve of the feed at the end of the anti-collision beam tube 2. For the underlying current loop and speed loop control mechanism of the servo drive mechanism in response to the speed command, those skilled in the art can use the standard dual closed-loop vector control strategy. The standard dual closed-loop vector control strategy is a well-known technology in the field and will not be described in detail here. Based on the above closed-loop alignment mechanism, the servo drive mechanism on the machine body 1 finally achieves absolute physical alignment of the short inner support slider 4 and the cutting position set by the machine body 1 in the axial and circumferential directions, thereby ensuring the accuracy of the cutting operation.

[0059] See attached document Figure 7 In this embodiment, the actuation demagnetization module 9 establishes an electrical control link with the distribution box 8 and the external excitation component inside the follow-up execution module 3. After the calculation compensation module 10 completes the alignment control and issues the feed stop command, the actuation demagnetization module 9 takes over the underlying control of the system's magnetic field. As a preferred method, in order to achieve accurate physical triggering in the closed passive cavity, the actuation demagnetization module 9 instructs the distribution box 8 to stop outputting positive DC power, and then injects transient low-frequency alternating pulse current into the external excitation component.

[0060] Based on the skin effect theory of electromagnetic fields, high-frequency alternating magnetic fields will generate strong eddy currents inside metal conductors and decay rapidly, while low-frequency alternating magnetic fields can penetrate metal solid shielding layers. The penetration depth of the alternating pulse magnetic field excited by transient low-frequency alternating pulse current in the external excitation component within the wall of anti-collision beam tube 2 is calculated based on the following skin depth formula: ; In the formula, This indicates the skin penetration depth of the alternating pulsed magnetic field inside the metal tube wall of the anti-collision beam 2; The resistivity parameter represents the metallic material of the anti-collision beam tube 2. This resistivity parameter is determined in advance based on the material's physical properties. f represents the frequency of the alternating pulse current. The frequency must be low enough to ensure that the skin penetration depth is strictly greater than the wall thickness of the anti-collision beam tube 2. To balance penetration capability and energy transmission efficiency, the frequency is typically set at [value missing]. to Since frequency, pi, and permeability are all physical constants greater than zero, the internal structure of this square root operation naturally eliminates the singularity risk of the denominator approaching zero. This indicates the relative magnetic permeability of the metal material in the anti-collision beam tube 2; It represents the vacuum permeability.

[0061] Based on the physical guarantee of the aforementioned penetration depth, the actuation demagnetization module 9 ensures that the alternating pulse magnetic field can penetrate the metal tube wall shielding layer of the anti-collision beam tube 2 without attenuation. When the alternating pulse magnetic field penetrates deep into the anti-collision beam tube 2, the electromagnetic excitation force it generates breaks through the holding magnetic threshold of the release latch inside the short inner support slider 4, thereby inducing the mechanical wedge tensioning mechanism inside the short inner support slider 4 to perform a physical opening action.

[0062] After the short-section inner support slider 4 completes the physical opening action, in order to avoid the transient response generated when the external excitation component is turned off from affecting subsequent detection, the actuation demagnetization module 9 synchronously executes the desaturation decoupling logic. Due to the inherent strong magnetism of the main magnetic core and the hysteresis effect of the surrounding ferromagnetic components, a strong residual magnetism will remain in the surrounding space of the system. This residual magnetism will cause the high-sensitivity magnetic field sensor of the sensing and acquisition module 6 to be forced into the nonlinear magnetic saturation range, thereby causing the sensing and acquisition module 6 to lose its ability to sense microscopic magnetic field distortion.

[0063] To actively eliminate the strong magnetic blind zone of the sensor, the actuation demagnetization module 9 cuts off the transient low-frequency alternating pulse current after pulse triggering and injects a weak compensation current in the opposite direction into the external excitation component. In this way, the weak compensation current generates a reverse canceling magnetic field in space. Based on the principle of spatial magnetic field superposition, the actuation demagnetization module 9 forces the sensing acquisition module 6 out of the nonlinear magnetic saturation region. The calculation logic of the target net magnetic induction intensity is based on the following formula: ; In the formula, This indicates the net magnetic flux density of the target at the detection position of the sensing module 6. The absolute value of the net magnetic flux density of the target must be strictly controlled within the linear response range of the high-sensitivity magnetic field sensor to ensure that the output voltage corresponds proportionally to the magnetic flux. The threshold of the linear response range here is usually set between -2mT and +2mT according to the sensor selection manual. This represents the residual magnetic induction intensity at the detection position after the pulse triggering is completed. Its initial value comes from the real-time data capture of the sensing acquisition module 6. Indicates the permeability of free space; The equivalent relative permeability of the spatial magnetic circuit between the external excitation component and the sensing and acquisition module 6 is represented by N; N represents the number of turns of the electromagnet coil in the external excitation component. This indicates the magnitude of the weak compensation current injected in reverse by the actuation demagnetization module 9. Its initial setting value is calculated proportionally based on the background residual magnetic induction intensity. It represents the equivalent magnetic circuit length of the spatial magnetic circuit. The equivalent magnetic circuit length is determined by the physical spatial structure and is always greater than zero, thus avoiding the singularity risk of the denominator approaching zero in division operations.

[0064] In the actual desaturation process, in order to avoid undercompensation or overcompensation caused by a single extreme value judgment, the actuation demagnetization module 9 dynamically adjusts the magnitude of the weak compensation current in combination with multi-dimensional logic. Specifically, the actuation demagnetization module 9 continuously acquires the current magnetic flux value fed back by the sensing acquisition module 6 and calculates the first time derivative of the value. Only when the absolute value of the target net magnetic induction intensity falls into the above-mentioned linear response range and its first time derivative approaches zero (that is, it indicates that the spatial magnetic field state has entered a stable convergence stage) can the actuation demagnetization module 9 determine that the decoupling process is completed, so that the target net magnetic induction intensity stably approaches the level of the environmental geomagnetic field.

[0065] Through the above-mentioned timing electromagnetic control scheme, the actuation demagnetization module 9 achieves complete decoupling of the strong driving magnetic field and the weak detection magnetic field in the time domain. The complete decoupling in the time domain restores the high sensitivity response of the sensing acquisition module 6 to the small displacement of the secondary driven magnetic pole inside the short inner support slider 4, thereby eliminating the common-mode interference of the underlying physical hardware for the subsequent data verification of the feature verification module 5.

[0066] See attached document Figure 8 In this embodiment, the feature verification module 5 establishes a data communication connection with the sensing acquisition module 6. After the actuation demagnetization module 9 completes the desaturation decoupling logic, the sensing acquisition module 6 performs a second sampling of the current spatial magnetic field. In order to ensure that the verification data and the demagnetization state are accurately aligned in the time dimension, the sensing acquisition module 6 immediately latches the current spatial magnetic field state after receiving the hardware synchronization trigger pulse issued by the actuation demagnetization module 9. The sensing acquisition module 6 generates the second magnetic topology matrix data and sends the second magnetic topology matrix data to the feature verification module 5 in real time.

[0067] Based on the principle of spatial magnetic field topology analysis, the feature verification module 5 runs its built-in peak-finding logic to identify the main magnetic peak amplitude corresponding to the main magnetic core and the secondary magnetic peak amplitude corresponding to the secondary driven magnetic pole in the second magnetic topology matrix data. When the truncated inner support slider 4 is in a contracted state, the secondary driven magnetic pole is close to the main magnetic core, and the main magnetic peak and the secondary magnetic peak appear as a distorted single peak that overlaps in space. After the truncated inner support slider 4 successfully performs the physical opening action, the secondary driven magnetic pole moves away from the main magnetic core, and a double-peak structure will be separated in the spatial magnetic field topology. In order to quantify the relative displacement of the secondary driven magnetic pole and offset the absolute fluctuation of the overall magnetic field strength, the feature verification module 5 calculates the double-peak ratio of the main magnetic peak amplitude to the secondary magnetic peak amplitude. The double-peak ratio is calculated according to the following formula: ; In the formula, This represents the bimodal ratio extracted from the second magnetic topology matrix data; This represents the amplitude of the second peak found in the second magnetic topology matrix data; This represents the amplitude of the main peak found in the second magnetic topology matrix data; This represents a very small normal number introduced to prevent the main peak amplitude from approaching zero due to hardware acquisition failure. The value of the very small normal number is usually set to 10 to the power of negative 8 to ensure the numerical stability of the division operation and avoid causing system crashes.

[0068] To avoid biased misjudgments caused by relying solely on a single bimodal ratio, feature verification module 5 introduces a lightweight convolutional neural network model for multi-dimensional spatial topological feature extraction and state reasoning. As a preferred approach, the lightweight convolutional neural network model adopts a serial topological structure, which includes an input layer, three sets of consecutive two-dimensional convolutional layers combined with max pooling layers, a global average pooling layer, and a fully connected layer.

[0069] During the data preprocessing stage, the feature verification module 5 performs max-min normalization on the second magnetic topology matrix data. The feature verification module 5 linearly maps all magnetic induction intensity values ​​in the second magnetic topology matrix data to the interval between 0 and 1. The max-min normalization calculation performed by the feature verification module 5 is based on the following formula: ; In the formula, This represents the dimensionless characteristic value located in the i-th row and j-th column of the second magnetic topology matrix data after normalization. This represents the original magnetic flux density value at the corresponding coordinate position in the second magnetic topology matrix data; This represents the global maximum magnetic flux density value in the second magnetic topology matrix data; This represents the global minimum magnetic flux density value in the second magnetic topology matrix data; This represents the zero-prevention constant, which is set to the power of 10 to prevent matrix singularity errors when the global magnetic field is absolutely uniform.

[0070] The normalized second magnetic topology matrix data is used as the input feature map and fed into the lightweight convolutional neural network model. In terms of data flow, the two-dimensional convolutional layer extracts the spatial distance and morphological distortion features between the main magnetic peak and the secondary magnetic peak through the sliding convolution kernel. The max pooling layer achieves spatial dimensionality reduction and filters out high-frequency local magnetic noise through downsampling. After integration by the global average pooling layer, the fully connected layer maps the high-dimensional feature vector into a non-linear probability output. The final output of the lightweight convolutional neural network model is a continuous floating-point number between 0 and 1. The continuous floating-point number represents the confidence that the wedge-shaped tensioning mechanism inside the truncated inner support slider 4 has successfully established a rigid physical support state.

[0071] Before deployment, the lightweight convolutional neural network model was trained offline under supervision. The training sample set was derived from magnetic topology matrix data collected during the system's historical operation at different pipe wall thicknesses. The training labels were the physical expansion states clearly marked by manual or auxiliary sensors (1 represents successful expansion, 0 represents failure). During training, the binary cross-entropy loss function was used to calculate the error between the predicted value and the true label, and the network weights were updated through the backpropagation algorithm until the loss function value converged to below the preset convergence threshold of 0.05.

[0072] After obtaining the bimodal ratio and confidence level, feature verification module 5 performs weighted fusion judgment, and the comprehensive judgment score is calculated according to the following formula: ; In the formula, This represents the final overall score. This represents the bimodal ratio after normalization. Since the original bimodal ratio may be greater than 1, the feature verification module 5 uses the standard Sigmoid function to perform nonlinear compression mapping on the original bimodal ratio, thereby ensuring that the normalized bimodal ratio is limited to the dimensionless range of 0 to 1. This represents the confidence level of the output of a lightweight convolutional neural network model. This represents the physical prior weighting coefficient assigned to the bimodal ratio, with the value range set from 0.3 to 0.4. This represents the data-driven weight coefficient assigned to the model confidence level. The value range of the data-driven weight coefficient is set to 0.6 to 0.7, and the sum of the system-mandated physical prior weight coefficient and the data-driven weight coefficient is strictly equal to 1.

[0073] The feature verification module 5 compares the comprehensive judgment score with the system's pre-stored safety threshold. Based on rigorous industrial safety considerations, the safety threshold is usually calibrated to 0.85. When the comprehensive judgment score is greater than or equal to the safety threshold, the feature verification module 5 determines that the internal rigid support has been constructed. After completing the closed-loop verification, the feature verification module 5 generates a processing permission signal and sends it to the cutting device 7. The processing permission signal is used to trigger the subsequent mechanical cutting process so that the cutting tool accurately falls into the cutting clearance groove. For the specific pipe cutting mechanical actions performed by the cutting device 7 (such as the downward drive of the laser cutting head or the circular saw blade), those skilled in the art can refer to the design of the actuator of a conventional CNC pipe cutting machine. The design of the actuator of a conventional CNC pipe cutting machine is a well-known technology in this field and will not be described in detail here.

[0074] Specific application examples: This embodiment applies a magnetically coupled follow-up anti-collision beam anti-deformation cutting system to a metal pipe processing production line in an automobile manufacturing base. The processing object is a high-strength steel pipe with a cross-section of 100mm×60mm and a wall thickness of 1.5mm. The servo feeding mechanism is set to a reference feed speed of 1.2m / s.

[0075] After the system starts, the power distribution box 8 outputs positive DC power to the external excitation component of the follow-up execution module 3 located at the machine body cutting station, generating a constant intensity traction magnetic field. The main magnetic core of the short inner support slider 4 placed in the inner cavity of the pipe establishes magnetic coupling with the external excitation component. When the servo feeding mechanism pushes the pipe to feed continuously, the short inner support slider 4 overcomes the friction of the inner wall of the pipe and moves synchronously with the pipe.

[0076] During the movement, the sensor acquisition module 6 performs first-order low-pass filtering with a smoothing coefficient of 0.2, collects the magnetic flux distribution around the pipe in real time, and converts it into first magnetic topology matrix data. The calculation and compensation module 10 receives the data and runs two-dimensional cross-correlation operation logic, compares the current matrix with the preset zero-position reference matrix, calculates the axial position deviation and radial rotation deviation of the short inner support slider 4 relative to the cutting station, and, combined with the built-in discrete PID control loop, generates a feedforward compensation control signal and sends it to the servo feeding drive mechanism. The servo feeding drive mechanism fine-tunes the feed deceleration curve according to the signal to achieve alignment of the inner and outer stations.

[0077] After the pipe is aligned and feeding stops, the actuation demagnetization module 9 stops the DC power output and injects an alternating pulse current with a frequency of 30Hz and a magnetic induction intensity amplitude of 1.0T into the external excitation component. The pulse magnetic field penetrates the metal pipe wall and acts on the release latch inside the short-section inner support slider 4. The high-permeability armature generates high-frequency mechanical vibration, causing the spring pin to disengage from the limiting groove. The preloaded spring releases potential energy to drive the central push rod, causing the radial wedge block to expand outward and press against the inner side of the pipe wall, establishing physical rigid support. After the triggering action is completed, the actuation demagnetization module 9 cuts off the pulse current and injects a weak compensation current to counteract the residual magnetism of the environment, causing the high-sensitivity magnetic field sensor of the sensing acquisition module 6 to leave the magnetic saturation range.

[0078] After the demagnetization operation is completed, the sensor acquisition module 6 performs secondary sampling to generate a second magnetic topology matrix data, which is then sent to the feature verification module 5. The feature verification module 5 extracts the amplitude of the main magnetic peak and the amplitude of the secondary magnetic peak from the data to calculate the bipeak ratio. The normalized matrix data is then input into a lightweight convolutional neural network model to extract features and outputs the confidence level of the corresponding support state. The feature verification module 5 performs a weighted fusion judgment on the bipeak ratio and confidence level at a ratio of 0.3 and 0.7. When the comprehensive judgment score reaches the preset safety threshold of 0.85, the feature verification module 5 sends an interlock release command to the cutting device 7, and the cutting device 7 then presses down to perform the cutting operation.

[0079] Table 1: Performance Comparison Test Data of Anti-collision Beam Cutting System

[0080] According to Table 1 and Figure 9 and Figure 10 The data shows that the traditional control group system, due to its reliance on external mechanical drive and open-loop hysteresis caused by changes in pipe wall friction, has an axial alignment deviation that is randomly distributed between -3.91mm and 4.23mm. In this embodiment, the solution compensation module 10 uses the real-time magnetic topology matrix obtained by the sensor acquisition module 6 to perform cross-correlation calculation, dynamically calculate the position deviation, and feed forward adjust the servo drive mechanism, so that the axial alignment deviation of the system converges to the range of -0.19mm to 0.32mm. The data comparison shows that the magnetic coupling follow-up cooperation closed-loop spatial magnetic field topology compensation mechanism overcomes the physical alignment error caused by friction slippage.

[0081] The deformation of the cut end face of the control group system ranged from 1.79 mm to 2.33 mm, and the pipe underwent local concavity due to external cutting stress. In this embodiment, the cutting deformation of the system was maintained at a low level of 0.04 mm to 0.15 mm. This change in value directly proves that the actuation demagnetization module 9 accurately triggered the mechanical structure in the inner support slider through the penetrating alternating pulse magnetic field. The mechanical transmission sequence of the central push rod and the radial wedge block constructed a rigid physical support to resist deformation inside the passive anti-collision beam cavity.

[0082] In this embodiment, the feature verification confidence level of each test batch of the system is between 0.86 and 0.95, all exceeding the safety threshold of 0.85. The reverse weak compensation current injected by the actuation demagnetization module 9 eliminates the residual magnetism in the space, ensuring that the data output by the sensing acquisition module 6 presents the true magnetic field distortion characteristics. The feature verification module 5 uses bimodal ratio calculation and convolutional neural network model to comprehensively verify the second magnetic topology matrix data. Only after confirming that the internal support state meets the physical requirements is the cutting command released. The multi-dimensional judgment method eliminates the misjudgment phenomenon caused by the extraction of a single magnetic field extreme value, ensuring the safety and dimensional processing accuracy of the anti-collision beam cutting process.

[0083] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A collision-resistant and deformation-resistant cutting system device based on magnetic coupling and follow-up, comprising a body (1), characterized in that, The top of the machine body (1) is provided with a collision protection beam tube (2), and the inner side of the collision protection beam tube (2) is provided with a short-section inner support slider (4). The top of the machine body (1) is fixedly connected with a feature verification module (5). The middle part of the machine body (1) is fixedly connected with a power distribution box (8). The outer side of the power distribution box (8) is fixedly connected with an actuation demagnetization module (9). The bottom of the actuation demagnetization module (9) is provided with a calculation compensation module (10). The calculation compensation module (10) is fixedly connected to the outer side of the power distribution box (8). The top of the machine body (1) is equipped with a cutting device (7). The middle part of the cutting device (7) is reserved with a processing channel for the collision protection beam tube (2) to slide through. The outer side of the processing channel is provided with a sensor acquisition module (6). The sensor acquisition module (6) is fixedly connected to the middle part of the cutting device (7). The top of the cutting device (7) is fixedly connected with a follow-up execution module (3). The calculation and compensation module (10) performs closed-loop alignment control based on the first magnetic topology matrix data sent by the sensing and acquisition module (6). The actuation demagnetization module (9) triggers the mechanical tensioning structure inside the short inner support slider (4) to construct rigid support based on the alignment state. The feature verification module (5) verifies the support state based on the second magnetic topology matrix data of the sensing and acquisition module (6) and sends an interlock release command to the cutting device (7). The follow-up execution module (3) includes an external excitation component, which is arranged on the outside of the anti-collision beam tube (2); The short inner support slider (4) is provided with a magnetic component inside; the external excitation component is used to generate a magnetic field and cooperates with the magnetic component to establish a magnetic coupling relationship.

2. The anti-collision beam anti-deformation cutting system device based on magnetic coupling servo as described in claim 1, characterized in that, The short-section inner support slider (4) includes a main magnetic core, a wedge-shaped tensioning mechanism, a preload spring, a release lock, and a secondary driven magnetic pole. The wedge-shaped tensioning mechanism is installed in the central cavity of the short-section inner support slider (4). The wedge-shaped tensioning mechanism includes a central push rod and radial wedge blocks evenly distributed along the circumference. The preload spring is sleeved on the outer surface of the central push rod, and the release latch includes a high-permeability magnetic armature and a spring clip. The spring clip is embedded in the limiting groove on the surface of the central push rod to lock the preload spring, and the secondary driven magnetic pole is installed on the outer end face of the radial wedge block.

3. The anti-collision beam anti-deformation cutting system device based on magnetic coupling servo as described in claim 1, characterized in that, The sensing and acquisition module (6) includes multiple spatially arranged high-sensitivity magnetic field sensors and a multi-channel synchronous analog-to-digital converter; Multiple high-sensitivity magnetic field sensors are arranged around the outer circumference and axial direction of the processing channel to construct a two-dimensional cylindrical detection grid, which is used to collect the magnetic flux distribution state of the space around the anti-collision beam tube (2) and output an analog voltage signal; The multi-channel synchronous analog-to-digital converter converts the analog voltage signal into a digital signal and constructs the first magnetic topology matrix data and the second magnetic topology matrix data according to different sampling timing sequences.

4. The anti-collision beam anti-deformation cutting system device based on magnetic coupling servo according to claim 3, characterized in that, The machine body (1) is provided with a servo feeding drive mechanism, and the calculation compensation module (10) establishes a data communication connection with the sensing acquisition module (6) and the servo feeding drive mechanism respectively. The calculation and compensation module (10) receives the first magnetic topology matrix data generated in real time, and runs the two-dimensional discrete cross-correlation operation logic to calculate the axial position deviation and radial rotation deviation, thereby generating a feedforward compensation control signal and sending it to the servo feeding drive mechanism.

5. The anti-collision beam anti-deformation cutting system device based on magnetic coupling according to claim 2, characterized in that, The actuation demagnetization module (9) establishes electrical connections with the distribution box (8) and the external excitation assembly, respectively; The actuation demagnetization module (9) is used to output transient low-frequency alternating pulse current to the external excitation component, excite the alternating pulse magnetic field to act on the high-permeability armature, so that the spring pin disengages from the limiting groove, and the pre-compression spring releases the stored elastic potential energy to drive the central push rod to push the radial wedge block to press against the inner wall of the anti-collision beam tube (2).

6. The anti-collision beam anti-deformation cutting system device based on magnetic coupling and follow-up as described in claim 5, characterized in that, The feature verification module (5) establishes signal connections with the sensing acquisition module (6) and the cutting device (7) respectively; The actuated demagnetization module (9) injects a weak compensation current in the reverse direction after cutting off the transient low-frequency alternating pulse current to offset the residual magnetism of the space background. The feature verification module (5) receives the second magnetic topology matrix data, extracts the main magnetic peak amplitude corresponding to the main magnetic core and the secondary magnetic peak amplitude corresponding to the secondary driven magnetic pole, calculates the double peak ratio and performs ratio calculation to verify the support state, and sends the interlock release command to the cutting device (7) after confirmation.

7. The anti-collision beam anti-deformation cutting system device based on magnetic coupling servo according to claim 2, characterized in that, The main magnetic core adopts an asymmetric polarity arrangement structure. The N pole and S pole of the main magnetic core are distributed in unequal volume ratios in the axial direction of the short inner support slider (4). The internal magnetic energy product of the secondary driven magnetic pole is smaller than the internal magnetic energy product of the main magnetic core.

8. The anti-collision beam anti-deformation cutting system device based on magnetic coupling according to claim 3, characterized in that, The sensing acquisition module (6) integrates digital filtering logic. The digital filtering logic runs a first-order infinite impulse response low-pass filtering algorithm to filter the real-time acquired raw discrete magnetic field strength data, and runs a neighborhood mean interpolation algorithm to complete the numerical values ​​of abnormal data nodes, thereby constructing the first magnetic topology matrix data.

9. The anti-collision beam anti-deformation cutting system device based on magnetic coupling according to claim 4, characterized in that, The calculation and compensation module (10) is equipped with a discrete PID control loop. The calculation and compensation module (10) inputs the axial position deviation and the radial rotation deviation into the discrete PID control loop, and generates the feedforward compensation control signal in combination with the calibrated feedforward gain coefficient.

Images