A laser cutting machine cutting control method and system

Abstract

This invention belongs to the field of control and detection technology, specifically relating to a laser cutting machine cutting control method and system, including the following steps: S1, applying sweep frequency excitation at multiple preset grid node positions on the laser cutting machine's worktable and collecting vibration response data of the cutting head; identifying and extracting the dominant poles and zeros of the vibration transfer function of each grid node; and using a non-uniform rational B-spline surface to fit the spatial distribution, establishing a spatial model of continuous mapping between the cutting head position and the vibration transfer function parameters; S2, decomposing the path to be processed into multiple path micro-segments; and calculating the preliminary feedforward compensation command for each path micro-segment based on the spatial model and a zero-phase error tracking control algorithm. This invention reduces the contour error caused by structural vibration and nonlinear factors, improves the processing accuracy and cutting surface quality of the laser cutting machine across its entire working range, and improves the forming effect of complex contours at high speeds.

Description

Technical Field

[0001] This invention belongs to the field of control and detection technology, specifically relating to a laser cutting machine cutting control method and system. Background Technology

[0002] As a high-precision processing equipment, laser cutting machines commonly employ feedforward control technology based on dynamic models to actively compensate for contour errors caused by servo lag and mechanical vibration in order to simultaneously improve processing efficiency and cutting accuracy. Traditional control methods typically simplify the entire laser cutting machine into one or a few linear time-invariant models, and design feedforward compensators such as zero-phase error tracking controllers based on these models. Laser cutting machines often adopt gantry or bridge structures, and the mass distribution and structural stiffness of moving parts change with the position of the cutting head on the worktable, resulting in spatial variability in vibration characteristics.

[0003] When machining complex contours at high speeds, especially along paths with small arcs and sharp turns, the cutting head undergoes drastic acceleration and deceleration, easily triggering high-frequency vibrations in the machine tool structure. This, in turn, induces nonlinear factors such as friction and clearance, resulting in complex coupled vibrations. These high-frequency and nonlinear vibration components are difficult to represent and compensate for using traditional linear models, leading to quality defects such as ripples at the cutting edge and overcutting or undercutting at corners. Existing control strategies often use fixed acceleration or jerk limits to smooth the speed curve, sacrificing machining efficiency on straight or large arc paths to ensure worst-case machining quality. Furthermore, these methods lack online identification and suppression capabilities for real-time vibration states, failing to adjust the control strategy based on path characteristics and actual vibration intensity, thus hindering further improvements in the overall performance of laser cutting machines. Summary of the Invention

[0004] This invention provides a laser cutting machine cutting control method and system to solve the technical problems in the existing technology of laser cutting machine traditional control that make it difficult to compensate for variability and nonlinearity, and difficult to balance efficiency and accuracy.

[0005] In a first aspect, the present invention provides a laser cutting machine cutting control method, comprising the following steps: S1. At multiple preset grid node positions on the laser cutting machine worktable, frequency sweep excitation is applied and vibration response data of the cutting head is collected. By identifying and extracting the dominant poles and zeros of the vibration transfer function of each grid node, the spatial distribution is fitted using a non-uniform rational B-spline surface to establish a spatial model of continuous mapping between the cutting head position and the vibration transfer function parameters. S2 decomposes the path to be processed into multiple path micro-segments, and calculates the preliminary feedforward compensation command for each path micro-segment based on the spatial model and the zero-phase error tracking control algorithm. S3, calculate the predicted contour error sequence of the cutting head within a time window under the initial feedforward compensation command, and perform ensemble empirical mode decomposition on the predicted contour error sequence to identify the intrinsic mode function components whose instantaneous energy exceeds the preset threshold as nonlinear coupled vibration components; calculate the energy integral of all identified nonlinear coupled vibration components to obtain the nonlinear vibration intensity index, and set an error limit based on the nonlinear vibration intensity index; S4. Within the buffer, when a path micro-segment with a path curvature greater than a preset threshold and a predicted contour error exceeding the error limit is detected, the predicted contour error is used as feedback. The acceleration limit parameters of the path micro-segment are iteratively optimized until the predicted contour error is lower than the error limit. The anti-phase signal of the nonlinear coupled vibration component is constructed as a nonlinear compensation signal. The nonlinear compensation signal is superimposed with the preliminary feedforward compensation command after acceleration parameter optimization to generate a cutting control command.

[0006] Furthermore, a sweep frequency excitation is applied and vibration response data of the cutting head is acquired. The dominant poles and zeros of the vibration transfer function of each grid node are identified and extracted, including: At each preset grid node, a linear sweep sine signal from 0.1Hz to 500Hz is generated as the excitation input using a signal generator; at the same time, a triaxial accelerometer mounted on the cutting head is used to collect vibration acceleration response signals in the X, Y, and Z directions; the collected input and output signals are identified using a vector fitting algorithm to obtain the transfer function of each grid node, and two pairs of dominant poles and one pair of dominant zeros are extracted.

[0007] Furthermore, based on the spatial model and the zero-phase error tracking control algorithm, the preliminary feedforward compensation commands for each path micro-segment are calculated, including: For each path micro-segment, the discrete transfer function G(z) corresponding to the center point position is obtained from the spatial model; the transfer function is then applied sequentially to the desired path signal. Non-causal filtering and causal filtering with transfer function G(z) / G(1) are used to generate preliminary feedforward compensation instructions.

[0008] Furthermore, ensemble empirical mode decomposition is performed on the predicted contour error sequence to identify eigenmode function components with instantaneous energy exceeding a preset threshold as nonlinear coupled vibration components, including: Gaussian white noise with a standard deviation 0.2 times that of the predicted profile error sequence is added to the predicted profile error sequence. 100 empirical mode decompositions are performed, and the resulting intrinsic mode function (IMF) components are ensemble-averaged to obtain the IMF components. A Hilbert transform is applied to each IMF component to calculate its instantaneous amplitude, and the square of the instantaneous amplitude is taken as the instantaneous energy. An energy threshold is set, which is three times the average power of all IMF components. IMF components whose instantaneous energy exceeds the energy threshold at any given moment are identified as nonlinear coupled vibration components.

[0009] Furthermore, an error limit is set based on the nonlinear vibration intensity index, including: The error limit is set using the following formula: ,in For error limits, The set basic error limit, It is a nonlinear vibration intensity index. This is a preset adjustment coefficient, with a value range of 0.1-0.5.

[0010] Furthermore, the acceleration constraint parameters of the path micro-segments are iteratively optimized until the predicted contour error is below the error limit, including: If the predicted contour error of the current path micro-segment Exceeding the error limit Then update the acceleration limit parameter as follows: ,in To define the updated acceleration parameters, Use the acceleration constraint parameters from the previous iteration; use the updated acceleration constraint parameters. Recalculate the initial feedforward compensation command and predicted profile error for the path micro-segments, and repeat this process until... .

[0011] Furthermore, constructing the inverse phase signal of the nonlinear coupled vibration component as the nonlinear compensation signal includes: After iterative optimization of the acceleration constraint parameters, the predicted contour error sequence is recalculated based on the optimized preliminary feedforward compensation command. The nonlinear coupled vibration components are then identified through analysis of the predicted contour error sequence. Finally, all identified nonlinear coupled vibration components are summed point-by-point to obtain the total nonlinear vibration signal. ;to the total nonlinear vibration signal Invert each point to generate an inverse phase signal. , as a nonlinear compensation signal.

[0012] Furthermore, in S1, the least squares complex frequency domain estimation algorithm is used to obtain the transfer function of each grid node.

[0013] Furthermore, in S2, the G-code file of the workpiece to be processed is read, and according to the servo control cycle, the straight lines or arcs in it are interpolated into a series of dense path point sequences, with each pair of adjacent points forming a path micro-element segment.

[0014] Secondly, the present invention provides a laser cutting machine cutting control system, including a memory and a processor. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned laser cutting machine cutting control method is implemented.

[0015] The beneficial effects are as follows: By establishing a continuous spatial dynamic model covering the entire processing area, the vibration characteristics of the laser cutting machine as the cutting head position changes are obtained. This invention can identify nonlinear coupled vibrations generated during high-speed processing of complex curves, and based on this, set an error limit that varies with vibration intensity, thereby optimizing the acceleration of local high-curvature path segments where the prediction error exceeds the limit. Furthermore, by generating anti-phase signals of nonlinear vibration components for active compensation, vibration sources that are difficult to model are suppressed. Without sacrificing overall processing efficiency, this invention reduces the contour error caused by both structural vibration and nonlinear factors, improves the processing accuracy and cutting surface quality of the laser cutting machine across its entire working range, and enhances the forming effect of complex contours at high speeds. Attached Figure Description

[0016] Figure 1 This is a flowchart of the laser cutting machine cutting control method. Detailed Implementation

[0017] The technical solutions of 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, not all, of the embodiments of the present invention. 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.

[0018] An embodiment of the laser cutting machine cutting control method provided by the present invention: like Figure 1 As shown, the laser cutting machine cutting control method includes the following steps: S1. At multiple preset grid node positions on the laser cutting machine worktable, frequency sweep excitation is applied and vibration response data of the cutting head is collected. By identifying and extracting the dominant poles and zeros of the vibration transfer function of each grid node, a spatial model of continuous mapping between the cutting head position and vibration transfer function parameters is established by fitting the spatial distribution using a non-uniform rational B-spline surface.

[0019] Specifically, the worktable of a laser cutting machine, for example, with a diameter of 3000mm × 1500mm, is divided into a 10 × 5 grid, forming 50 discrete test nodes. The cutting head is moved to each node position, and a servo motor drives the X-axis or Y-axis to execute a sweeping sinusoidal motion command with a frequency linearly increasing from 1Hz to 200Hz as the excitation signal. Simultaneously, a triaxial accelerometer mounted on the cutting head synchronously collects the actual vibration acceleration of the cutting head as the response signal. Fourier transforms are performed on the excitation and response signals of each node to calculate the frequency response function. A least-squares complex frequency domain estimation algorithm is used to fit the frequency response function to obtain a second-order or higher-order transfer function model. From this model, several dominant poles and zeros with the greatest vibration influence are analyzed. The dominant poles are those with the greatest influence on the overall system response and the longest duration. The coordinates of all node positions are taken as input, and the real and imaginary parts of the corresponding poles or zeros are taken as output. A non-uniform rational B-spline surface fitting algorithm is used to establish a continuous surface function for each parameter. The set of functions constitutes a complete spatial model, which can obtain the transfer function parameters of the position in real time by looking up a table or calculating based on any input X and Y coordinate values.

[0020] In an optional embodiment, a frequency sweep excitation is applied and vibration response data of the cutting head is acquired. The dominant poles and zeros of the vibration transfer function of each grid node are identified and extracted, including: At each preset grid node, a linear sweep sine signal from 0.1Hz to 500Hz is generated as the excitation input using a signal generator; at the same time, a triaxial accelerometer mounted on the cutting head is used to collect vibration acceleration response signals in the X, Y, and Z directions; the collected input and output signals are identified using a vector fitting algorithm to obtain the transfer function of each grid node, and two pairs of dominant poles and one pair of dominant zeros are extracted.

[0021] Specifically, at a specific grid node location, such as point (100, 100) in the worktable coordinate system, a signal generator controls the X-axis servo motor of the machine tool to input a sinusoidal excitation signal with a frequency linearly increasing from 0.1Hz to 500Hz for a duration of 200s. During this period, a triaxial accelerometer fixed to the cutting head continuously records acceleration data in the X, Y, and Z directions at a sampling rate of 2000Hz, obtaining three sets of time-series response signals.

[0022] The input swept frequency signal sequence and the three acquired acceleration response signal sequences are used as input-output data pairs. A vector fitting algorithm is used to process the data, for example, to identify the input-output data in the X direction, resulting in a transfer function model that represents the characteristics of the point. From the mathematical expression of the model, two pairs of dominant poles with significant influence on vibration, such as -2±15j and -5±80j, and one pair of dominant zeros, such as -1±10j, are extracted. This process is repeated on all preset grid nodes to construct a complete spatial characteristic model.

[0023] S2 decomposes the path to be processed into multiple path micro-segments, and calculates the preliminary feedforward compensation command for each path micro-segment based on the spatial model and the zero-phase error tracking control algorithm.

[0024] Specifically, the G-code file of the workpiece to be processed is read. Based on the servo control cycle (e.g., 1ms), the straight lines or arcs are interpolated into a dense sequence of path points, with each pair of adjacent points forming a path micro-segment. For each path micro-segment, the X and Y coordinates of the center point are taken and substituted into the non-uniform rational B-spline surface space model established in the previous step to obtain the local transfer function G(z) corresponding to the path micro-segment. Applying the standard formula of the zero-phase error tracking control algorithm, the feedforward controller F(z) is equal to the inverse of the stable and invertible part of the local transfer function G(z), i.e. The inverse of, where G(z) is decomposed into With the full passage The product of the two. The desired displacement command of the path micro-element is input into the calculated feedforward controller F(z), and the output is the preliminary feedforward compensation torque or current command corresponding to the path micro-element.

[0025] In an optional embodiment, based on the spatial model and the zero-phase error tracking control algorithm, the preliminary feedforward compensation command for each path micro-segment is calculated, including: For each path micro-segment, the discrete transfer function G(z) corresponding to the center point position is obtained from the spatial model; the transfer function is then applied sequentially to the desired path signal. Non-causal filtering and causal filtering with transfer function G(z) / G(1) are used to generate preliminary feedforward compensation instructions.

[0026] Specifically, suppose we want to process a small straight path from point A to point B, with the center point of the path segment having coordinates (250, 300). We search the established spatial model database for the discrete transfer function G(z) of the grid node with coordinates (250, 300). For example, the found discrete transfer function is... Meanwhile, the DC gain G(1) is calculated to be 1.

[0027] The desired position command sequence of the path's infinitesimal segments is taken as input. A non-causal filter is used to process the desired position command sequence; the transfer function of the filter is... ,Right now Phase delay is eliminated by processing the data in reverse time. The filtered result is then passed through a causal filter. After filtering, the resulting signal sequence is the initial feedforward compensation command for the path micro-segment. Theoretically, the initial feedforward compensation command can enable the cutting head to track the desired path with zero phase error.

[0028] S3, calculate the predicted contour error sequence of the cutting head within a time window under the initial feedforward compensation command, and perform ensemble empirical mode decomposition on the predicted contour error sequence to identify the intrinsic mode function components whose instantaneous energy exceeds the preset threshold as nonlinear coupled vibration components; calculate the energy integral of all identified nonlinear coupled vibration components to obtain the nonlinear vibration intensity index, and set an error limit based on the nonlinear vibration intensity index.

[0029] Specifically, a buffer of, for example, 200ms is set up to store the path micro-segments to be executed in the future and the preliminary feedforward compensation instructions. For each path micro-segment in the buffer, the predicted actual position of the cutting head is calculated using the corresponding local transfer function model G(z) and the preliminary feedforward compensation instructions. The normal distance from the predicted actual position point to the original desired path is calculated, which is the predicted profile error of the point. The predicted profile errors of all path micro-segments within the time window are concatenated to form a time series. The ensemble empirical mode decomposition algorithm is performed on the predicted profile error time series to decompose it into a set of intrinsic mode function components (IMFs) arranged from high frequency to low frequency. A Hilbert transform is performed on each IMF component to calculate the instantaneous amplitude, and the square of the instantaneous amplitude is defined as the instantaneous energy. If the instantaneous energy of a certain IMF drastically exceeds an empirically set energy threshold at certain time points, the IMF as a whole is determined to be a nonlinear coupled vibration component caused by factors such as friction and gaps.

[0030] All nonlinear coupled vibration IMFs identified in the previous step are summed point-by-point in the time domain to obtain a total nonlinear vibration error signal. The energy integral of the total signal over the entire time window is calculated, i.e., squared and summed to obtain a scalar value, which is the nonlinear vibration intensity index. A function is used to set the error limit; for example, the error limit is equal to the basic error limit plus a constant k divided by the sum of the nonlinear vibration intensity index and a small positive constant. When the calculated nonlinear vibration intensity index is high, the error limit becomes very small, and vice versa.

[0031] In an optional embodiment, ensemble empirical mode decomposition is performed on the predicted contour error sequence to identify intrinsic mode function components with instantaneous energy exceeding a preset threshold as nonlinear coupled vibration components, including: Gaussian white noise with a standard deviation 0.2 times that of the predicted profile error sequence is added to the predicted profile error sequence. 100 empirical mode decompositions are performed, and the resulting intrinsic mode function (IMF) components are ensemble-averaged to obtain the IMF components. A Hilbert transform is applied to each IMF component to calculate its instantaneous amplitude, and the square of the instantaneous amplitude is taken as the instantaneous energy. An energy threshold is set, which is three times the average power of all IMF components. IMF components whose instantaneous energy exceeds the energy threshold at any given moment are identified as nonlinear coupled vibration components.

[0032] Specifically, the calculated predicted contour error sequence is assumed to have a standard deviation of 0.01 mm. A new sequence with a standard deviation of 0.002 mm is then generated and added to the original error sequence. Empirical mode decomposition (EMD) is performed on the noise-added sequence to obtain a set of intrinsic mode function (EMF) components. This noise addition and decomposition process is repeated 100 times, each time using new random noise. The corresponding EMF components obtained from the 100 decompositions are averaged to obtain a set of stable components with clearly defined physical meanings, such as... , , wait.

[0033] For each eigenmode function component, as A Hilbert transform is performed to obtain the analytic signal, from which the instantaneous amplitude at each time point is calculated. Squaring the instantaneous amplitude yields the instantaneous energy curves of the intrinsic mode function components. Simultaneously, all intrinsic mode function components are calculated. , , The average power is assumed to be 0.0005 units. The energy threshold is set to three times this value, i.e., 0.0015 units. The instantaneous energy curves of each intrinsic mode function component are examined. If any are found... If the instantaneous energy exceeds 0.0015 at certain moments, while other intrinsic mode function components remain below this energy threshold, then... It was identified as a nonlinear coupled vibration component.

[0034] In an optional embodiment, an error limit is set based on a nonlinear vibration intensity index, including: The error limit is set using the following formula: ,in For error limits, The set basic error limit, It is a nonlinear vibration intensity index. This is a preset adjustment coefficient, with a value range of 0.1-0.5.

[0035] Specifically, set the basic parameters. For example, set the basic error limits. A precision of 0.01 mm is the highest achievable accuracy under ideal machining conditions. An adjustment coefficient k is selected empirically, for example, a value of 0.2. This adjustment coefficient determines the sensitivity of the error limit to the intensity of nonlinear vibration.

[0036] When planning a specific path segment, the nonlinear vibration intensity index of the path segment is first calculated. Assuming a high-speed turning section of the road, the calculated... A value of 3.0 indicates a strong tendency for nonlinear vibration. In this case, the error limit is calculated using the formula: ≈0.00625mm. However, on another straight section of road where the vibration was weaker, It might only be 0.5, then The error limit is approximately 0.0091 mm, which is relatively lenient.

[0037] S4. Within the buffer, when a path micro-segment with a path curvature greater than a preset threshold and a predicted contour error exceeding the error limit is detected, the predicted contour error is used as feedback. The acceleration limit parameters of the path micro-segment are iteratively optimized until the predicted contour error is lower than the error limit. The anti-phase signal of the nonlinear coupled vibration component is constructed as a nonlinear compensation signal. The nonlinear compensation signal is superimposed with the preliminary feedforward compensation command after acceleration parameter optimization to generate a cutting control command.

[0038] Specifically, the curvature of each path micro-segment within the buffer is calculated. If the curvature of a path micro-segment exceeds a preset value (e.g., corresponding to an arc with a radius less than 5mm), and the predicted profile error of the path micro-segment also exceeds the error limit calculated in the previous step, the optimization process is triggered. The allowable acceleration value of the path micro-segment is reduced by 10%, and the velocity curve of the path micro-segment is recalculated based on the new acceleration limit. The initial feedforward compensation command is recalculated based on the new velocity curve, and the predicted profile error is simulated again. If the new predicted profile error is still higher than the error limit, the acceleration limit is further reduced in 10% increments, and the above process is repeated until the predicted profile error meets the requirements.

[0039] Summing all the identified nonlinear coupled vibration components (IMFs) and multiplying the resulting composite signal by -1 yields an inverse-phase signal. This inverse-phase signal is morphologically opposite to the predicted nonlinear vibration error. Using a preset gain or a simplified inverse model, the inverse-phase signal in the error domain is converted into a feedforward compensation torque or current signal; this is the nonlinear compensation signal. Adding the nonlinear compensation signal point-by-point to the initial feedforward compensation command optimized by acceleration, the resulting signal is the cutting control command sent to the servo driver.

[0040] In an optional embodiment, the acceleration constraint parameters of the path micro-segment are iteratively optimized until the predicted profile error is below the error limit, including: If the predicted contour error of the current path micro-segment Exceeding the error limit Then update the acceleration limit parameter as follows: ,in To define the updated acceleration parameters, Define the acceleration parameters for the previous iteration; use the updated parameters. Recalculate the initial feedforward compensation command and predicted profile error for the path micro-segments, and repeat this process until... .

[0041] Specifically, for a path infinitesimal segment, an initial acceleration constraint parameter is used. ,for example It generates motion trajectories and preliminary feedforward compensation commands, and calculates the predicted contour error. The error limit is 0.008 mm. Assume the error limit calculated using the aforementioned method is... The value is 0.00625mm. Since the prediction error of 0.008mm > 0.00625mm, iterative optimization is required.

[0042] Calculate the new acceleration limit parameters based on the updated formula. = Using a new, reduced acceleration constraint parameter of 1562.5, the velocity and acceleration curves of the path micro-segments are replanned, and the initial feedforward compensation command and the new predicted profile error are recalculated based on the new motion curves. Assuming the newly calculated... The value is 0.006 mm. At this point, since the new prediction error is 0.006 mm < 0.00625 mm, the iteration process stops. The acceleration constraint parameter for the path micro-element segment is... .

[0043] In an optional embodiment, constructing an inverse phase signal of the nonlinear coupled vibration component as a nonlinear compensation signal includes: After iterative optimization of the acceleration constraint parameters, the predicted contour error sequence is recalculated based on the optimized preliminary feedforward compensation command. The nonlinear coupled vibration components are then identified through analysis of the predicted contour error sequence. Finally, all identified nonlinear coupled vibration components are summed point-by-point to obtain the total nonlinear vibration signal. ;to the total nonlinear vibration signal Invert each point to generate an inverse phase signal. , as a nonlinear compensation signal.

[0044] Specifically, after iterative optimization of acceleration constraint parameters for all micro-segments of the entire machining path, a complete preliminary feedforward compensation command sequence is generated based on the kinematic parameters of the entire path. Using this preliminary feedforward compensation command sequence, contour error prediction is performed again to obtain the predicted contour error time series. Ensemble empirical mode decomposition and energy threshold judgment are then performed on the predicted contour error time series to identify the nonlinear coupled vibration components it contains, such as component... and .

[0045] The values ​​of the two identified components at each time point are added together to form the total nonlinear vibration signal, i.e. The total nonlinear vibration signal represents the profile error that is expected to occur due to nonlinear factors even after acceleration optimization. To compensate for the profile error, an inverse phase signal is generated. The value of t at any time is equal to The negative value. For example, if at t=1.2 seconds, The value is 0.002mm, then The value is -0.002mm. The sequence is a nonlinear compensation signal used to be superimposed on the feedforward compensation command.

[0046] An embodiment of a laser cutting machine cutting control system provided by the present invention: A laser cutting machine cutting control system includes a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement the aforementioned laser cutting machine cutting control method.

[0047] A laser cutting machine cutting control system also includes other components well known to those skilled in the art, such as communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

[0048] In this invention, the aforementioned memory can be any tangible medium containing or storing a program that can be used or combined with an instruction execution system, apparatus, or device. For example, a computer-readable storage medium can be any suitable magnetic or magneto-optical storage medium, such as Resistive Random Access Memory (RRAM), Dynamic Random Access Memory (DRAM), Static Random Access Memory (SRAM), Enhanced Dynamic Random Access Memory (EDRAM), High-Bandwidth Memory (HBM), Hybrid Memory Cube (HMC), etc., or any other medium that can be used to store desired information and can be accessed by an application, module, or both. Any such computer storage medium can be part of a device or accessible to or connected to a device. Any application or module described in this invention can be implemented using computer-readable / executable instructions stored or otherwise maintained by such a computer-readable medium.

[0049] The above are all preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Therefore, all equivalent changes made in accordance with the structure, shape and principle of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A laser cutting machine cutting control method, characterized in that, Includes the following steps: S1. At multiple preset grid node positions on the laser cutting machine worktable, frequency sweep excitation is applied and vibration response data of the cutting head is collected. By identifying and extracting the dominant poles and zeros of the vibration transfer function of each grid node, the spatial distribution is fitted using a non-uniform rational B-spline surface to establish a spatial model of continuous mapping between the cutting head position and the vibration transfer function parameters. S2 decomposes the path to be processed into multiple path micro-segments. Based on the spatial model and zero-phase error tracking control algorithm, it calculates the preliminary feedforward compensation commands for each path micro-segment, including: For each path segment, obtain the discrete transfer function corresponding to the center point position from the spatial model. The transfer function is applied sequentially to the desired path signal. Noncausal filtering and transfer function are The causal filtering generates a preliminary feedforward compensation command. S3, calculate the predicted contour error sequence of the cutting head within a time window under the initial feedforward compensation command, and perform ensemble empirical mode decomposition on the predicted contour error sequence to identify the intrinsic mode function components whose instantaneous energy exceeds a preset threshold as nonlinear coupled vibration components, including: Gaussian white noise with a standard deviation 0.2 times that of the predicted profile error sequence is added to the predicted profile error sequence. 100 empirical mode decompositions are performed, and the resulting intrinsic mode function (IMF) components are ensemble-averaged to obtain the IMF components. A Hilbert transform is applied to each IMF component to calculate its instantaneous amplitude, and the square of the instantaneous amplitude is taken as the instantaneous energy. An energy threshold is set, which is three times the average power of all IMF components. IMF components whose instantaneous energy exceeds the energy threshold at any given moment are identified as nonlinear coupled vibration components. Calculate the energy integral of all identified nonlinear coupled vibration components to obtain the nonlinear vibration intensity index, and set an error limit based on the nonlinear vibration intensity index; S4, within the buffer, when a path micro-segment with a path curvature greater than a preset threshold and a predicted contour error exceeding the error limit is detected, the predicted contour error is used as feedback. The acceleration limit parameters of the path micro-segment are iteratively optimized until the predicted contour error is lower than the error limit. An anti-phase signal of the nonlinear coupled vibration component is constructed as a nonlinear compensation signal. The nonlinear compensation signal is superimposed with the preliminary feedforward compensation command after acceleration parameter optimization to generate a cutting control command. All the nonlinear coupled vibration components IMF identified above are summed, and the resulting synthetic signal is multiplied by -1 to form an anti-phase signal. The anti-phase signal is completely opposite in shape to the predicted nonlinear vibration error. Through a preset gain or a simplified inverse model, the anti-phase signal in the error domain is converted into a feedforward compensation torque or current signal, which is the nonlinear compensation signal.

2. The laser cutting machine cutting control method according to claim 1, characterized in that, A sweep frequency excitation was applied and vibration response data of the cutting head was acquired. The dominant poles and zeros of the vibration transfer function of each grid node were identified and extracted, including: At each preset grid node, a linear sweep sine signal from 0.1Hz to 500Hz is generated as the excitation input using a signal generator; at the same time, a triaxial accelerometer mounted on the cutting head is used to collect vibration acceleration response signals in the X, Y, and Z directions; the collected input and output signals are identified using a vector fitting algorithm to obtain the transfer function of each grid node, and two pairs of dominant poles and one pair of dominant zeros are extracted.

3. The laser cutting machine cutting control method according to claim 1, characterized in that, An error limit is set based on the nonlinear vibration intensity index, including: The error limit is set using the following formula: ,in For error limits, The set basic error limit, It is a nonlinear vibration intensity index. This is a preset adjustment coefficient, with a value range of 0.1-0.

5.

4. The laser cutting machine cutting control method according to claim 1, characterized in that, Iteratively optimize the acceleration constraint parameters of the path's infinitesimal segments until the predicted contour error is below the error limit, including: If the predicted contour error of the current path micro-segment Exceeding the error limit Then update the acceleration limit parameter as follows: ,in To define the updated acceleration parameters, Use the acceleration constraint parameters from the previous iteration; use the updated acceleration constraint parameters. Recalculate the initial feedforward compensation command and predicted profile error for the path micro-segments, and repeat this process until... .

5. The laser cutting machine cutting control method according to claim 1, characterized in that, Constructing the inverse phase signal of the nonlinear coupled vibration component as the nonlinear compensation signal includes: After iterative optimization of the acceleration constraint parameters, the predicted contour error sequence is recalculated based on the optimized preliminary feedforward compensation command. The nonlinear coupled vibration components are then identified through analysis of the predicted contour error sequence. Finally, all identified nonlinear coupled vibration components are summed point-by-point to obtain the total nonlinear vibration signal. The total nonlinear vibration signal Invert each point to generate an inverse phase signal. , as a nonlinear compensation signal.

6. The laser cutting machine cutting control method according to claim 1, characterized in that, In S1, the least squares complex frequency domain estimation algorithm is used to obtain the transfer function of each grid node.

7. The laser cutting machine cutting control method according to claim 1, characterized in that, In S2, the G-code file of the workpiece to be processed is read, and according to the servo control cycle, the straight lines or arcs in it are interpolated into a series of dense path point sequences, with each pair of adjacent points forming a path micro-element segment.

8. A laser cutting machine cutting control system, characterized in that, It includes a memory and a processor, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the laser cutting machine cutting control method according to any one of claims 1-7 is implemented.

Images