Real-time data driven cutting bed automation control method, system, and storage medium

Abstract

The present application relates to the field of numerical control, specifically to a real-time data-driven cutting bed automation control method and system, and a storage medium. The steps of the method include: collecting cutting tool dynamic stress data and fabric micro deformation data in real time through a multi-source sensing network, and constructing a physical coupling state space of the cutting tool and the fabric. The present application realizes high-precision collaborative control of cutting tool dynamics and fabric deformation in the field of cutting bed control through a multi-source sensing data-driven collaborative optimization mechanism. Through hardware synchronous triggering of piezoelectric ceramic array sensors and distributed optical fiber sensors, spatiotemporal alignment monitoring of cutting tool stress fluctuations and fabric interlayer shear deformation is realized. This design breaks through the timing mismatch problem in cross-scale data fusion of traditional systems, forms a calculable correlation matrix between the thermodynamic state of the cutting tool and the fabric strain energy distribution, and lays the foundation for the solvable space of the optimization algorithm.

Description

Technical Field

[0001] This invention relates to the field of numerical control technology, specifically to a real-time data-driven automated control method, system, and storage medium for cutting beds. Background Technology

[0002] The current control architecture of precision cutting equipment is limited by the modeling bottleneck of dynamic coupling of cross-scale physical fields. While traditional methods can achieve macroscopic motion trajectory planning, they have fundamental limitations in high-entropy composite material cutting scenarios:

[0003] Because the tool stress field, representing macroscopic mechanical response, and the fabric microstructural deformation, representing mesoscopic material behavior, belong to different physical dimensions, classical control systems employ a decoupled, piecewise feedback mechanism. This leads to an uncoordinated boundary conflict between the tool's thermodynamic state and the fabric fiber strain energy during dynamic processes. In high-speed dynamic cutting, this boundary conflict manifests as a hidden system mismatch effect: the control system cannot synchronously capture the peak tool edge stress and the interlayer shear deformation gradient of the fabric within a microsecond-level response cycle. This causes the optimization algorithm to oscillate continuously within the constraint conflict domain, ultimately resulting in a combined failure of abnormal tool wear and fabric micro-tears.

[0004] Existing technologies attempt to alleviate this contradiction by increasing the density of sensor distribution or raising the control frequency to the kilohertz level, but due to the asynchronicity of multi-source sensor data and the fragmentation of state space, they have not been able to break through the theoretical constraints of cross-scale dynamic coupling models.

[0005] Based on the above contradictions, the technical problem proposed by this invention is: how to construct a collaborative architecture with real-time fusion capability of multi-source sensor data across scales in the cutting bed motion control system so that the thermodynamic fatigue boundary of the cutting tool and the critical boundary of molecular fracture of the fabric maintain consistency in dynamic evolution during high-speed cutting. Summary of the Invention

[0006] This disclosure presents a real-time data-driven method, system, and storage medium for automated control of a cutting bed, with the aim of overcoming at least one of the deficiencies in the prior art.

[0007] To achieve the above objectives, the technical solution disclosed in this invention is as follows:

[0008] According to one aspect of this disclosure, a real-time data-driven automated control method for a cutting bed is provided, comprising the steps of:

[0009] By collecting dynamic stress data of cutting tools and micro-deformation data of fabric in real time through a multi-source sensor network, a physical coupling state space between the cutting tool and the fabric is constructed.

[0010] Based on heuristic algorithms, the constraint boundaries in the physical coupling state space are solved, and a collaborative optimization strategy for tool path and feed rate is dynamically generated.

[0011] The collaborative optimization strategy is decomposed into tool motion control commands and fabric displacement control commands;

[0012] The system stability index is calculated by fusing multi-sensor data, and the constraints of the heuristic algorithm are dynamically recalibrated based on the stability index.

[0013] Furthermore, the multi-source sensor network includes:

[0014] A piezoelectric ceramic array sensor embedded inside the tool is used to capture stress fluctuations at the tool's cutting edge in real time.

[0015] A distributed optical fiber sensor array laid under the fabric to detect interlayer shear deformation of the fabric fibers;

[0016] The sensor array achieves data spatiotemporal alignment through a hardware synchronization triggering circuit.

[0017] Furthermore, the constraint boundary includes:

[0018] The thermodynamic fatigue limit boundary of the cutting tool is determined by a combination of the tool material temperature rise model and historical stress spectrum.

[0019] The critical boundary for fiber breakage in the fabric is established through real-time analysis of the fabric strain-stress constitutive equation.

[0020] The heuristic algorithm solves for the intersection of the limit boundary and the critical boundary in a three-dimensional state space.

[0021] Furthermore, the generation of the collaborative optimization strategy includes:

[0022] A multi-objective optimization function for the tool acceleration trajectory is established within the intersection domain, including tool wear rate, fabric cutting accuracy, and energy efficiency.

[0023] By using the simulated annealing algorithm to overcome local optimum traps, the Pareto front solution set is obtained.

[0024] Furthermore, the selection of the Pareto front solution set adopts the following method:

[0025] Based on the risk assessment model of information entropy, the impact coefficient of each solution set on the system stability is calculated.

[0026] The solution set whose impact coefficient is lower than the preset process safety threshold is selected as the final strategy.

[0027] Furthermore, the step of decomposing the collaborative optimization strategy includes:

[0028] The tool motion control commands map the trajectory to three-dimensional Euclidean space through rigid body kinematic transformation;

[0029] The fabric displacement control command transforms the conveying trajectory into a parametric motion curve through the equation of elasticity.

[0030] Furthermore, the calculation of the system stability index includes:

[0031] The vibration spectrum entropy of the piezoelectric ceramic array, the strain gradient distribution of the fiber optic sensor, and the interface thermal conductivity of the infrared thermal imaging were combined.

[0032] The dynamic recalibration includes:

[0033] When the stability index falls below a preset threshold, the objective function of the heuristic algorithm is reconstructed.

[0034] Multi-sensor fusion data is used as the boundary constraint for the new objective function;

[0035] The optimization parameters are updated using gradient descent.

[0036] Furthermore, the method further includes the following steps:

[0037] Collect environmental vibration and noise data, and eliminate sensor interference through digital filters;

[0038] Initial toolpath values ​​are generated using a pseudo-random number algorithm and used as the starting point for optimization iterations.

[0039] The weak coupling signal between the tool and the fabric is amplified by a microelectromechanical sensor.

[0040] According to another aspect of this disclosure, a real-time data-driven automated cutting bed control system is provided for implementing the real-time data-driven automated cutting bed control method described above, comprising:

[0041] The multi-source sensing fusion module, including a piezoelectric ceramic array unit and a distributed optical fiber unit, is used to construct a physically coupled state space;

[0042] The collaborative optimization module, including a heuristic computing unit and a constraint reconstruction unit, is used to dynamically generate collaborative optimization strategies.

[0043] The instruction generation module, including a kinematic transformation engine and a mechanical constraint solver, is used to decompose the collaborative optimization strategy into tool and fabric control instructions.

[0044] The dynamic stability module, including a multi-sensor data fusion unit and a threshold detector, is used for dynamic recalibration of optimized parameters.

[0045] According to another aspect of this disclosure, a storage medium is provided that stores a computer program, which, when executed by a processor, implements the real-time data-driven automated control method for a cutting bed as described above.

[0046] The beneficial effects of this invention are:

[0047] This invention achieves high-precision collaborative control of tool dynamics and fabric deformation in the field of cutting bed control through a multi-source sensor data-driven collaborative optimization mechanism. Specifically:

[0048] Firstly, by synchronously triggering the piezoelectric ceramic array sensor and the distributed optical fiber sensor, the spatiotemporal alignment monitoring of tool stress fluctuations and fabric interlayer shear deformation is achieved. This design overcomes the temporal mismatch problem in cross-scale data fusion of traditional systems, enabling the tool thermodynamic state and fabric strain energy distribution to form a computable correlation matrix, laying the foundation for the solvable space of the optimization algorithm.

[0049] Secondly, the heuristic optimization algorithm establishes a dynamic constraint intersection between the tool fatigue limit boundary and the fabric fracture critical boundary in the three-dimensional state space, and uses an adaptive step size strategy to overcome the local optimum trap of traditional optimization algorithms. By screening the Pareto solution set through an information entropy risk assessment model, the collaborative optimization strategy maintains Euclidean space motion accuracy while controlling energy consumption within the process safety threshold.

[0050] Third, rigid body kinematics transformation converts the optimized trajectory into a smooth motion path in the tool's three-dimensional space, avoiding kinematic singularity problems; fabric displacement parameterization based on elasticity equations achieves balanced distribution of strain energy during the conveying process. This dual-instruction decomposition mechanism improves cutting accuracy to the micrometer level while maintaining the system's momentum-energy balance.

[0051] Furthermore, the stability index calculated from multi-sensor data fusion serves as the system recalibration trigger condition, and the optimization parameters are dynamically updated using the gradient descent method. This mechanism, coupled with noise suppression by the digital filter, enables the system to remain robust to environmental vibration disturbances, fundamentally eliminating control instability caused by sensor drift.

[0052] This invention solves the coupling conflict between tool dynamics and fabric micro-deformation in multi-objective optimization by constructing a control architecture based on real-time sensor data feedback, and significantly improves the quality and efficiency of cutting high-entropy composite materials.

[0053] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, the preferred embodiments of the present invention are described in detail below with reference to the accompanying drawings. Attached Figure Description

[0054] Figure 1 This is a flowchart of a real-time data-driven automated control method for a cutting bed according to an embodiment of the present invention;

[0055] Figure 2 This is a schematic diagram of the state space of cross-scale physical field coupling in one embodiment of the present invention;

[0056] Figure 3 This is a schematic diagram of simultaneous monitoring of multiple physical quantities in one embodiment of the present invention;

[0057] Figure 4 This is a schematic diagram of system risk entropy analysis in one embodiment of the present invention;

[0058] Figure 5a This is a stress timing curve output by a piezoelectric ceramic sensor in one embodiment of the present invention;

[0059] Figure 5b This is a deformation curve output by an optical fiber sensor in one embodiment of the present invention;

[0060] Figure 5c This is a multiphysics coupled heat map of real-time fusion of multi-source sensor data in one embodiment of the present invention;

[0061] Figure 6 This is a schematic diagram of the displacement field of a catenary fabric in one embodiment of the present invention. Detailed Implementation

[0062] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments.

[0063] In embodiments of the present invention, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" or "for example" in embodiments of the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Rather, the use of the terms "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0064] The present invention provides the following preferred embodiments:

[0065] Example 1: To address the technical problem of dynamic mismatch between the tool's dynamic response and the fabric's micro-deformation field across multiple physical scales during high-speed cutting, this example provides a real-time data-driven automated control method for cutting beds. This method achieves collaborative modeling of the tool-fabric interaction behavior by constructing a physically coupled state space and ensures system stability based on a dynamic optimization mechanism. For example... Figure 1 As shown, the specific process is as follows:

[0066] Step S100: Real-time acquisition of dynamic stress data of cutting bed tools and micro-deformation data of fabric through multi-source sensor network to construct physical coupling state space between cutting tool and fabric.

[0067] During the cutting process, dynamic stress data of the cutting tool and micro-deformation data of the fabric are synchronously acquired through a multi-source sensor network. A piezoelectric ceramic array sensor embedded inside the cutting tool captures the stress fluctuation spectrum of the cutting edge in real time at a microsecond sampling rate, while a distributed fiber optic sensor array laid on the bottom layer of the fabric detects the shear strain gradient between fiber layers. The two types of sensors achieve spatiotemporal data alignment via a hardware synchronization trigger circuit, eliminating the dimensional fragmentation problem caused by sampling delay in traditional systems. Based on the above data, a three-dimensional physically coupled state space is established, such as... Figure 2 As shown, the coordinate axes are defined as the tool stress field intensity, the interlayer shear strain of the fabric, and the physical field coupling intensity, respectively. Within this space, the tool thermodynamic fatigue boundary surface and the fabric fracture critical boundary surface are generated through dynamic evolution equations. The overlapping region of the two forms a safe coupling working domain, while the non-overlapping conflict domain characterizes the system mismatch risk area.

[0068] Step S200: Solve the constraint boundary in the physical coupling state space based on the heuristic algorithm, and dynamically generate a collaborative optimization strategy for tool path and feed rate.

[0069] The constraint boundaries in the physically coupled state space are solved using a heuristic algorithm. The tool thermodynamic fatigue limit boundary is jointly calibrated by the material temperature rise model and historical stress spectrum, while the critical boundary for fabric fiber fracture is established by solving the strain-stress constitutive equation in real time. Within the three-dimensional state space, a multi-objective optimization function is established with a safe coupled working domain as the constraint condition. The objectives include minimizing tool wear rate, maximizing cutting accuracy, and energy efficiency. Simulated annealing is used for solving the problem, and an adaptive step-size strategy avoids the local optimum traps of traditional optimization algorithms, ultimately outputting a Pareto front solution set. This solution set is then screened using an information entropy risk assessment model to eliminate candidate strategies whose impact coefficients exceed the process safety threshold.

[0070] Step S300: Decompose the collaborative optimization strategy into tool motion control commands and fabric displacement control commands.

[0071] The collaborative optimization strategy is decomposed into two independent control commands. Tool motion control commands, through a rigid body kinematics transformation engine, map the optimized trajectory to three-dimensional Euclidean space, ensuring a smooth path and avoiding kinematic singularities. Fabric displacement control commands, on the other hand, construct parameterized motion curves based on elasticity equations to achieve a balanced distribution of strain energy during transport, such as... Figure 5c The multiphysics coupled thermogram generation logic is shown. This dual-path decoupling mechanism achieves precise coordination between tool trajectory and fabric delivery while maintaining the system's momentum-energy conservation.

[0072] Step S400: Calculate the system stability index through multi-sensor data fusion, and dynamically recalibrate the constraints of the heuristic algorithm based on the stability index.

[0073] The system stability index is calculated through multi-sensor data fusion. The fusion objects include the vibration spectrum entropy of the piezoelectric ceramic array, the strain gradient distribution of the fiber optic sensor, and the interface thermal conductivity of the infrared thermal image, generating a comprehensive stability evaluation value. When this index falls below a preset threshold, a dynamic recalibration mechanism based on a heuristic algorithm is triggered: first, the multi-source sensor data is reconstructed into boundary constraints for optimizing the objective function; then, the algorithm parameters are updated using gradient descent. This process operates with the support of noise suppression from a digital filter, ensuring the system remains robust to environmental vibration disturbances.

[0074] The benefits of this embodiment are as follows: a collaborative control architecture for cross-scale physical fields is constructed through dynamic modeling and real-time optimization of the physically coupled state space. The hardware-level synchronization mechanism of multi-source sensor data eliminates the spatiotemporal mismatch between the tool stress field and the fabric deformation field, enabling the state space to have computable physical correlation. The heuristic optimization algorithm based on three-dimensional constraint boundaries directly solves the safe coupled working domain, avoiding boundary conflicts caused by traditional piecewise control. The dual-instruction decoupling mechanism of rigid body kinematics and elasticity ensures the tool motion accuracy and fabric strain balance, respectively. The stability index of multi-sensor fusion drives the recalibration of algorithm parameters, suppressing control instability caused by sensor drift from a mechanistic perspective.

[0075] Example 2: In order to solve the spatiotemporal mismatch problem of multi-source sensor data in the high-speed cropping process, this example refines a multi-source sensor network architecture based on hardware synchronous triggering.

[0076] During the cutting process, the piezoelectric ceramic array sensor embedded in the cutting edge of the tool captures the dynamic stress spectrum of the tool in real time at a microsecond sampling rate. Its sensing units are arranged in a topologically distributed manner to cover the entire stress field of the cutting edge. At the same time, the fiber optic sensor array laid on the fabric substrate layer resolves the interlayer shear deformation gradient through the wavelength offset of the Bragg grating, and its detection accuracy reaches the micrometer level.

[0077] Furthermore, the two types of sensor arrays achieve physical-level timing alignment through a hardware synchronization trigger circuit. This circuit uses a high-precision crystal oscillator to generate a global clock signal and eliminates timing drift caused by electromagnetic interference through differential signal transmission. Figure 5a and Figure 5b As shown, the stress time-series curve output by the piezoelectric ceramic sensor and the deformation curve output by the fiber optic sensor are strictly synchronized on the time axis, enabling a precise mapping between the tool dynamics and the fabric micro-deformation field in the spatiotemporal dimension. It is important to understand that this hardware-level synchronization mechanism fundamentally avoids the physical field coupling distortion caused by traditional asynchronous sampling, providing a spatiotemporally consistent data foundation for subsequent state-space modeling.

[0078] The advantages of this embodiment are as follows: the distributed topology of the piezoelectric ceramic array enables full-domain monitoring of the tool stress field; the wavelength resolution technology of the fiber optic sensor ensures the microscopic capture of fabric deformation; and the hardware synchronous triggering circuit unifies the multi-source data acquisition cycle through a physical clock signal, enabling cross-scale physical field data to have computable spatiotemporal correlation. This architecture provides a high-fidelity data source for the construction of physically coupled state spaces.

[0079] Example 3: Addressing the quantification challenge of the dynamic evolution of the tool-material interaction boundary, this example optimizes the constraint boundary generation method in the physically coupled state space. The tool's thermodynamic fatigue limit boundary is jointly calibrated using a coupled material temperature rise model and historical stress spectrum.

[0080] The temperature rise model calculates the instantaneous thermal load in the tool's micro-region based on the Fourier heat conduction equation, while the historical stress spectrum probabilistically models the cumulative damage to the tool through the Wiener process. The two models are then fused to generate the tool's thermodynamic boundary surface, such as... Figure 2 As shown. Simultaneously, the critical boundary for fabric fiber fracture is established through real-time solution of a rate-dependent constitutive equation: this equation constructs a strain-stress differential relationship based on viscoelastic theory, dynamically updates the critical fracture threshold by combining real-time acquired interlaminar shear strain data, and forms the critical fracture surface of the fabric. In the three-dimensional state space, a heuristic algorithm solves for the intersection domain of the two surfaces using the gradient projection method. Specifically, this involves calculating the minimum distance field between the tool stress field and the fabric fracture energy threshold in the state space. When the distance value is less than the process safety threshold, it is determined to be a safe coupling domain; otherwise, it is marked as a conflict and instability domain.

[0081] The advantages of this embodiment are as follows: the fusion of the material temperature rise model and the Wiener process enables dynamic prediction of tool fatigue boundaries; the rate-dependent constitutive equation gives the fabric fracture boundary strain rate adaptability; and the boundary intersection is directly solved in three-dimensional space using the gradient projection method, avoiding the conservative design caused by traditional simplified models. This method provides accurate physical constraint boundaries for collaborative optimization strategies.

[0082] Example 4: To achieve multi-objective collaborative optimization within the safe coupling domain, this example refines the Pareto front solution set generation mechanism. Within the safe domain of the physically coupled state space, a three-dimensional optimization function for the tool acceleration trajectory is established: the first objective function calculates the tool wear rate using Hertzian contact theory, the second objective function measures cutting accuracy based on the integral of fabric deformation energy, and the third objective function calculates energy efficiency based on the integral of the driver current.

[0083] To overcome the local optimum trap, an improved simulated annealing algorithm is employed: a tunneling effect model is introduced during the cooling process, allowing the algorithm to probabilistically overcome the energy barrier; simultaneously, an adaptive neighborhood search strategy is used to dynamically adjust the solution space exploration range based on the number of iterations. Figure 4As shown, the candidate solution set output by the algorithm is filtered using an information entropy risk assessment model: first, the probability density distribution of the objective function value of each solution on the Pareto front is calculated; then, solutions with risk entropy values ​​exceeding the process tolerance are removed; finally, the Pareto optimal solution set constitutes the collaborative optimization strategy. It is important to understand that this embodiment improves the global search capability of the heuristic algorithm in a non-convex solution space.

[0084] The advantages of this embodiment are as follows: a multi-objective function based on Hertzian theory, deformation energy integral, and current integral constructs a quantitative foundation for process optimization; the simulated annealing algorithm incorporating the tunneling effect overcomes the local convergence limitation of traditional optimization algorithms; and the information entropy risk assessment model ensures the engineering feasibility of the solution set from a probabilistic perspective. This embodiment provides a theoretical implementation path for multi-objective collaborative optimization of high-entropy composite material cutting.

[0085] Example 5: To reduce the risk of instability of Pareto solution sets in engineering applications, this example refines the solution set selection mechanism based on information entropy. In the candidate solution set generated by the simulated annealing algorithm, the risk assessment model quantifies the impact coefficient of the solution set on system stability by calculating the probability distribution dispersion of the objective function value of each solution in the state space.

[0086] Specifically: First, a three-dimensional probability density field is constructed to represent tool wear rate, cutting accuracy, and energy efficiency; second, the information entropy value of each solution in the density field is calculated using the Shannon entropy formula. A higher entropy value indicates a larger range of parameter fluctuations for that solution; for example... Figure 4 As shown, solutions with entropy values ​​higher than a preset process safety threshold are marked as high-risk points, corresponding to... Figure 4 The filling region in the text. Understandably, this filtering mechanism identifies solutions with excessively high parameter sensitivity from a probabilistic perspective, ensuring the robustness of the final strategy under dynamic conditions.

[0087] Furthermore, the information entropy threshold is dynamically calibrated using an offline process database: the entropy distribution at system instability moments during historical trimming tasks is statistically analyzed, and the upper bound of its 99% confidence interval is taken as the real-time threshold. For example... Figure 4 The critical risk threshold indicated in the figure is adaptively adjusted according to the fabric material category. It should be noted that this embodiment uses entropy value filtering rather than simple objective function value sorting, thus fundamentally avoiding the local sensitivity trap in multi-objective optimization.

[0088] The advantages of this embodiment are: the construction of the probability density field transforms the solution set stability quantification into a computable information entropy problem; the offline statistical threshold calibration method ensures the engineering applicability of the screening criteria; and the final low-entropy solution set provides an anti-interference control strategy for highly dynamic clipping environments.

[0089] Example 6: Addressing the feasibility of collaborative optimization strategies at the electromechanical execution layer, this example optimizes the physical space mapping method for control commands. Tool motion control commands generate three-dimensional Euclidean space trajectories through rigid body kinematic transformation: first, the acceleration trajectory in state space is discretized into a pose sequence; second, the rotation angle of each joint axis is calculated based on the homogeneous transformation matrix; finally, it is transformed into linear interpolation commands in the Cartesian coordinate system using the Jacobian matrix. Simultaneously, fabric displacement control commands generate parameterized motion curves based on the equations of elasticity: the fabric conveying trajectory is considered as a continuous medium motion field, and a differential relationship between the conveyor roller displacement and the local deformation of the fabric is established based on the Cauchy strain tensor; such as... Figure 5c As shown in the figure, the multiphysics fusion heatmap visually presents the coupling state between the tool path and the fabric delivery path. It is important to understand that this dual-channel mapping mechanism fundamentally unifies the control benchmarks for rigid body motion and flexible body deformation.

[0090] Furthermore, the trajectory mapping process incorporates real-time feedback correction: through Figure 3 The multi-physical quantity synchronous monitoring curves shown compare the stress-strain correlation between the theoretical trajectory and the actual sensor output in real time. When the control synchronization coefficient is detected to be below 0.85, online replanning of the trajectory parameters is triggered.

[0091] Furthermore, for fabric displacement control, the fabric transport trajectory is transformed into a parameterized expression of the Bernoulli catenary equation through nonholonomic constraint equations. Specifically, based on the tension distribution field of the fabric fiber layer, the catenary differential equation is established:

[0092] , where y represents the fabric displacement field intensity parameter, characterizing the instantaneous displacement of the fabric fiber layer in the vertical conveying direction, and its distribution reflects the dynamic deformation field under the cutting action of the tool. For example Figure 6 As shown, the evolution of the y-value along the fabric conveying direction x forms a catenary surface; x represents the fabric conveying direction coordinate, defining the reference direction of the fabric in the conveying plane, orthogonal to the tool movement trajectory, and this coordinate axis corresponds to Figure 6 The horizontal axis of the catenary surface is set to zero at the starting position of the fabric entering the multi-source sensor data monitoring area; 'a' is a dynamic calibration parameter, dynamically calibrated by the correlation strength between the fabric's elastic modulus and interlayer shear force. 'a' consists of the following four coupled parts:

[0093] E: Fabric elastic modulus, used to describe the ability of the fiber layer to resist deformation, its value depends on the fabric material and weave density.

[0094] τ: Interlayer shear force, reflecting the frictional constraint between fiber layers, is determined by the fabric lamination process parameters.

[0095] ρ: Fabric surface density, which affects the inertial effect of displacement wave propagation.

[0096] γ: Correlation strength factor, derived from multi-source sensor data monitoring of cutting tools and fabrics, characterizing the coherence strength between fiber molecular bonds and the cutting edge of the cutting tool.

[0097] Parameter 'a' is updated in real time using the following formula:

[0098] , where k is a dimensionless conversion constant, and γ(t) is fed back in real time by the multi-source sensor data monitoring module.

[0099] like Figure 6 As shown, the solution set of the catenary differential equation generates the fabric displacement vector field, whose streamline distribution accurately reflects the non-uniform deformation of the fiber layer under the cutting action of the tool. It can be understood that catenary parameterization transforms the mechanical constraints of the continuum into an integrable differential manifold, ensuring the strict conjugation of the fabric transport trajectory and the tool motion in space and time.

[0100] The advantages of this embodiment are: maintaining the geometric invariance of the tool motion command while utilizing the catenary equation to describe the physical intrinsic properties of the fabric displacement field; rigid body kinematic transformation ensuring precise spatial positioning of the tool trajectory; parametric curves based on elasticity enabling controllable transmission of fabric deformation; and a dual-channel synchronous feedback mechanism providing dynamic compensation capabilities for the electromechanical execution layer.

[0101] Example 7: To address the issue of system stability degradation under complex operating conditions, this example refines the dynamic recalibration mechanism. System stability indicators are calculated through multi-source heterogeneous data fusion: the vibration signal from the piezoelectric ceramic array is extracted using Fast Fourier Transform to obtain the spectral entropy value in the 0.5-10kHz frequency band; the strain gradient field output by the fiber optic sensor array is used to calculate the spatial distribution divergence using the Sobel operator; and the infrared thermal imaging data is analyzed based on the two-color thermometry method to determine the thermal conductivity of the tool-fabric interface. These three indicators are calculated through... Figure 3 The synchronous monitoring curves shown are weighted and fused. When the comprehensive index is lower than the preset threshold, the optimization model is recalibrated.

[0102] Furthermore, the recalibration process includes objective function reconstruction and constraint update: first, multi-sensor fusion data is imported into a heuristic algorithm to replace the theoretical parameters in the original objective function; second, such as... Figure 5a and Figure 5b As shown, the real-time acquired tool stress field distribution and fabric deformation gradient are used as new boundary constraints; finally, the optimization parameters are iteratively updated using the gradient descent method. Figure 3 As shown, when the control synchronization coefficient fluctuates around t=7ms, the system recalibrates to pull the correlation back from the critical region to the safe operating region. This process essentially transforms physical sensor data into an adaptive adjustment factor at the algorithm level.

[0103] The advantages of this embodiment are as follows: the fusion computation of multi-source heterogeneous data constructs a panoramic evaluation system for system stability; boundary constraint updates based on real-time sensor data enable the optimization model to have operational response capabilities; and the introduction of the gradient descent method ensures efficient convergence of the recalibration process. This mechanism provides continuous stability assurance for long-term pruning tasks.

[0104] Example 8: To improve the signal-to-noise ratio and optimize convergence efficiency of sensor data in high-speed cutting environments, this example refines the environmental interference suppression and optimization initialization mechanisms. Environmental vibration and noise data are acquired through a triaxial accelerometer and frequency-selectively attenuated using an adaptive Butterworth digital filter. This filter is based on... Figure 3 The noise spectrum characteristics in the multi-physical quantity synchronous monitoring curves shown are dynamically adjusted to eliminate mechanical resonance interference from 50 to 500 Hz.

[0105] Furthermore, the pseudo-random number algorithm generates initial toolpath values ​​based on a Mason rotator. The spatial distribution characteristics of these initial values ​​satisfy the Lyapunov exponential stability condition, serving as a physically feasible starting point for optimization iterations. It's important to understand that this initial value generation strategy avoids the local optimum trap caused by manual setting. Simultaneously, the microelectromechanical sensor embedded in the tool holder amplifies the weak coupling signal at the tool-fabric interface through a differential amplifier circuit: it uses an instrumentation amplifier architecture to amplify the pV-level piezoelectric signal to a detectable range, and then... Figure 5a The stress waveform shown verifies the signal fidelity.

[0106] Furthermore, temperature drift compensation is introduced into the signal amplification stage: based on Figure 5c The multiphysics fusion thermal map shown corrects for sensor zero-point drift caused by tool temperature rise in real time. Understandably, this compensation mechanism fundamentally ensures the measurability of weak signals under dynamic operating conditions.

[0107] The advantages of this embodiment are: the adaptive digital filter effectively suppresses mechanical vibration interference, the pseudo-random path initialization ensures the global search capability of the optimization process, and the microelectromechanical sensing amplification technology enables reliable capture of micro-coupled physical quantities.

[0108] Example 9: In order to realize the hardware carrier of the real-time data-driven cutting bed automation control method, this example constructs a multi-level collaborative cutting bed automation control system.

[0109] Specifically, the multi-source sensing fusion module includes a piezoelectric ceramic array unit and a distributed optical fiber unit arranged in a topology: the piezoelectric unit captures the tool stress spectrum at a microsecond-level sampling rate and generates... Figure 5a The stress-time curve shown is generated by the fiber unit demodulating the interlayer shear deformation of the fabric through a Bragg grating and outputting the result. Figure 5bThe strain gradient field shown, together with the constraint gradient field, constructs a three-dimensional physically coupled state space. The collaborative optimization module includes a parallel-working heuristic computation unit and a constraint reconstruction unit: the former runs an improved simulated annealing algorithm to generate... Figure 4 The Pareto front solution set shown is based on... Figure 3 The state space boundary constraints are dynamically updated based on real-time monitoring data.

[0110] Furthermore, the instruction generation module is implemented using a dedicated hardware accelerator: the kinematic transformation engine converts the optimized trajectory into a Cartesian space toolpath, while the mechanical constraint solver calculates the fabric delivery parameters in real time based on the Cauchy strain tensor. The dynamic stability module includes a multi-sensor data fusion unit and a threshold detector: the former fuses vibration spectrum entropy, strain gradient, and thermal conductivity to generate a system stability curve, and the latter detects when an index exceeds a certain threshold. Figure 4 The threshold value of the critical mismatch zone shown triggers parameter recalibration. It's important to understand that the modules achieve nanosecond-level data exchange via a high-speed industrial bus to ensure real-time control.

[0111] The advantages of this embodiment are as follows: the multi-source sensing module constructs a holographic perception capability of the physically coupled state space, the collaborative optimization module realizes the adaptive generation of strategies under dynamic working conditions, the instruction generation module ensures the accurate spatial mapping of complex trajectories, and the dynamic stability module provides a failure protection mechanism for long-term operation.

[0112] Example 10: To solidify the executable logic of the control method, this example provides a computer program storage medium.

[0113] Specifically, the computer program stored on the medium includes a multi-threaded scheduling architecture: the first thread executes the multi-source sensor data fusion task, analyzes the stress spectrum of the piezoelectric ceramic array and the deformation gradient of the fiber optic sensor in real time, and constructs a cross-scale physical field coupled state space. The second thread runs a collaborative optimization algorithm, generates a Pareto solution set based on an improved simulated annealing engine, and calls an information entropy risk assessment model for screening. The third thread processes the instruction decomposition logic, maps the optimized trajectory to a three-dimensional toolpath through rigid body kinematic transformation, and solves for the fabric delivery parameters based on the viscoelastic constitutive equation.

[0114] Furthermore, the program incorporates a built-in dynamic recalibration daemon: continuously monitoring the system stability curve, and automatically triggering an optimization parameter update process when the control synchronization coefficient is detected to be below a safety threshold. This process uses real-time collected tool stress field distribution and fabric deformation data as new boundary constraints, iteratively adjusting the objective function weights using the gradient descent method. It's important to understand that all threads exchange data via shared memory, and their synchronization mechanism employs a lock-free circular buffer design to avoid resource contention.

[0115] The advantages of this embodiment are: the multi-threaded architecture enables parallel processing of sensing acquisition, optimization calculation and control execution; the dynamic recalibration mechanism ensures the continuous stability of the system under complex working conditions; and the efficient data exchange design meets the real-time requirements of high-speed cutting scenarios.

[0116] Although the present invention has been specifically described above with reference to preferred embodiments, it should be understood that the present invention is not limited to the embodiments described above. Various modifications and variations can be made by those skilled in the art without departing from the spirit of the present invention, and such modifications and variations should fall within the scope defined by the appended claims and their equivalents.

Claims

1. A real-time data-driven automated control method for a cutting bed, characterized in that the steps include... include: By collecting dynamic stress data of cutting tools and micro-deformation data of fabric in real time through a multi-source sensor network, a physical coupling state space between the cutting tool and the fabric is constructed. Based on heuristic algorithms, the constraint boundaries in the physical coupling state space are solved, and a collaborative optimization strategy for tool path and feed rate is dynamically generated. The collaborative optimization strategy is decomposed into tool motion control commands and fabric displacement control commands; The system stability index is calculated by fusing multi-sensor data, and the constraints of the heuristic algorithm are dynamically recalibrated based on the stability index. Tool motion control commands generate three-dimensional Euclidean space trajectories through rigid body kinematic transformations: Discretize the acceleration trajectory in the state space into a pose sequence; The rotation angle of each joint axis is calculated based on the homogeneous transformation matrix; Linear interpolation command that transforms to Cartesian coordinates using the Jacobian matrix; The fabric displacement control command transforms the fabric conveying trajectory into a parameterized expression of the Bernoulli catenary equation through nonholonomic constraint equations. Based on the tension distribution field of the fabric fiber layer, the catenary differential equation is established: Where y represents the fabric displacement field intensity parameter, x represents the fabric conveying direction coordinate, and a is a dynamic calibration parameter, which is updated in real time using the following formula: , where k is the dimensionless conversion constant, E is the elastic modulus of the fabric, τ is the interlaminar shear force, ρ is the surface density of the fabric, and γ is the correlation strength factor; The constraint boundaries include: The thermodynamic fatigue limit boundary of the cutting tool is determined by a combination of the tool material temperature rise model and historical stress spectrum. The critical boundary for fiber breakage in the fabric is established through real-time analysis of the fabric strain-stress constitutive equation. The heuristic algorithm solves for the intersection of the limit boundary and the critical boundary in a three-dimensional state space. The generation of the collaborative optimization strategy includes: A multi-objective optimization function for the tool acceleration trajectory is established within the intersection domain, including tool wear rate, fabric cutting accuracy, and energy efficiency. By navigating local optimum traps using simulated annealing, the Pareto front solution set is obtained; The steps for decomposing the collaborative optimization strategy include: The tool motion control commands map the trajectory to three-dimensional Euclidean space through rigid body kinematic transformation; The fabric displacement control command transforms the conveying trajectory into a parametric motion curve through the equation of elasticity. The calculation of the system stability index includes: The vibration spectrum entropy of the piezoelectric ceramic array, the strain gradient distribution of the fiber optic sensor, and the interface thermal conductivity of the infrared thermal imaging were combined. The dynamic recalibration includes: When the stability index falls below a preset threshold, the objective function of the heuristic algorithm is reconstructed. Multi-sensor fusion data is used as the boundary constraint for the new objective function; The optimization parameters are updated using gradient descent.

2. The real-time data-driven automated control method for cutting beds as described in claim 1, characterized in that, The multi-source sensor network includes: A piezoelectric ceramic array sensor embedded inside the tool is used to capture stress fluctuations at the tool's cutting edge in real time. A distributed optical fiber sensor array laid under the fabric to detect interlayer shear deformation of the fabric fibers; The sensor array achieves data spatiotemporal alignment through a hardware synchronization triggering circuit.

3. The real-time data-driven automated control method for cutting beds as described in claim 1, characterized in that, The Pareto front solution set was selected using the following method: Based on the risk assessment model of information entropy, the impact coefficient of each solution set on the system stability is calculated. The solution set whose impact coefficient is lower than the preset process safety threshold is selected as the final strategy.

4. The real-time data-driven automated control method for cutting beds as described in claim 1, characterized in that, The method further includes the following steps: Collect environmental vibration and noise data, and eliminate sensor interference through digital filters; Initial toolpath values ​​are generated using a pseudo-random number algorithm and used as the starting point for optimization iterations. The weak coupling signal between the tool and the fabric is amplified by a microelectromechanical sensor.

5. A real-time data-driven automated control system for a cutting bed, used to implement the real-time data-driven automated control method for a cutting bed as described in any one of claims 1-4, characterized in that, include: The multi-source sensing fusion module, including a piezoelectric ceramic array unit and a distributed optical fiber unit, is used to construct a physically coupled state space; The collaborative optimization module, including a heuristic computing unit and a constraint reconstruction unit, is used to dynamically generate collaborative optimization strategies. The instruction generation module, including a kinematic transformation engine and a mechanical constraint solver, is used to decompose the collaborative optimization strategy into tool and fabric control instructions. The dynamic stability module, including a multi-sensor data fusion unit and a threshold detector, is used for dynamic recalibration of optimized parameters.

6. A storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, implements the real-time data-driven automated control method for cutting beds as described in any one of claims 1-4.

Images