High-power laser cutting light path system, laser cutting head and use method of laser cutting head

Abstract

The invention relates to the technical field of laser cutting, and discloses a high-power laser cutting light path system, a laser cutting head and a using method of the laser cutting head. The system comprises a dynamic cutting path accounting module, a multi-dimensional cutting difference analysis module, an abnormal region association positioning module and a self-adaptive cutting strategy generation module. The dynamic cutting path accounting module builds a model based on laser cutting historical data, and collects laser power fluctuation, light beam quality distribution and cutting speed sequence data of current cutting in real time. The multi-dimensional cutting difference analysis module receives the information and performs three-dimensional difference analysis with an actual measurement path of the optical sensor to generate a workpiece-level difference coefficient matrix; and the abnormal region association positioning module inputs the matrix into a space geometric analysis network, generates an abnormal probability thermal distribution diagram of a cutting defect propagation path, and positions an abnormal cutting region. And the adaptive cutting strategy generation module configures parameters according to the thermodynamic diagram, and applies multi-band impedance disturbance to the adjacent node cutting head with the maximum abnormal probability gradient change.

Description

Technical Field

[0001] The present invention relates to the field of laser cutting technology, and in particular to a high-power laser cutting optical path system, a laser cutting head and a method of using the same. Background Art

[0002] In the field of high-power laser cutting, as industrial production demands ever-increasing cutting precision, efficiency, and material adaptability, traditional optical path systems are gradually exposing their limitations. Traditional systems often use preset fixed cutting paths, making it difficult to cope with the effects of laser power fluctuations, changes in beam quality, and differences in material properties during the cutting process. For example, when laser power fluctuates momentarily due to the equipment's operating status, deviations in the cutting depth and width under a fixed path can occur. Especially when processing high-hardness alloys or composite materials, such deviations can lead to defects such as burrs and cracks in the workpiece. Existing cutting discrepancy analysis methods are mostly limited to detecting position deviations within a two-dimensional plane, ignoring the impact of uneven energy distribution in the depth direction. Comparing the measured path captured by an optical sensor with the theoretical path often focuses solely on the offset of planar coordinates, while lacking effective analysis of energy density differences caused by changes in beam focus depth. This makes it difficult to detect cutting defects hidden within the material. The location of abnormal areas relies on manual experience or simple threshold judgments, and it is impossible to establish a correlation between cutting defects and material properties or the position of the cutting head. When a local defect appears on the workpiece, it is difficult to quickly trace the propagation path and cause of the defect, resulting in an increased rework rate and limited production efficiency. In terms of adaptive adjustment, parameter modifications in traditional systems are mostly based on a single feedback signal, such as adjusting the cutting head speed based solely on position deviation. This lacks comprehensive consideration of the influence of multiple coupling factors and makes it difficult to achieve accurate adaptation to complex cutting scenarios. Summary of the Invention

[0003] The object of the present invention is to provide a high-power laser cutting optical path system, a laser cutting head and a method of using the same, so as to solve the problems raised in the above-mentioned background technology.

[0004] To achieve the above objectives, the present invention provides a high-power laser cutting optical path system, the system comprising: The dynamic cutting path calculation module builds a dynamic cutting path calculation model based on the historical laser cutting operation data, collects the actual laser power fluctuation parameters, beam quality distribution indicators and cutting speed sequence data of the current cutting process in real time, and outputs the theoretical cutting path information through the dynamic cutting path calculation model; A multi-dimensional cutting difference analysis module receives the theoretical cutting path information and performs a three-dimensional difference analysis on the cutting path information measured by the optical sensor. The three-dimensional difference analysis includes the cumulative position deviation, depth energy offset and velocity sequence similarity index to generate a workpiece-level difference coefficient matrix; The abnormal area association positioning module inputs the difference coefficient matrix into the spatial geometric analysis network, combines the material property characteristic parameters with the cutting head position coordinate information, and generates an abnormal probability thermal distribution map of the cutting defect propagation path for locating the abnormal cutting physical area; The adaptive cutting strategy generation module configures cutting adjustment parameters according to the abnormal probability thermal distribution map, including applying a multi-band impedance perturbation operation to the cutting head of the adjacent node with the largest abnormal probability gradient change in the abnormal probability thermal distribution map.

[0005] Preferably, the dynamic cutting path calculation module specifically includes a historical cutting feature mining component and a dynamic cutting path calculation model building component; The historical cutting feature mining component processes historical laser cutting operation data and performs multi-dimensional decomposition operations, involving extracting the energy proportions of the steady-state and transient components of the power fluctuation parameters. It establishes a correlation matrix between the beam quality distribution index and the cutting load rate through beam coupling analysis, and uses a dynamic sequence alignment algorithm to adjust the speed sequence mode under different working conditions. The dynamic cutting path calculation model construction component transmits the processed historical operation data to the hybrid prediction network, which has a timing prediction unit based on the cutting head aging curve to generate a basic path prediction value. The fully connected network embedded with the spatial attention mechanism is used to correct the prediction deviation caused by beam distortion, and the sequence feature compensator dynamically adjusts the prediction weight coefficient according to the real-time acquired speed sequence; the heterogeneous data acquisition unit deployed by the optical sensor synchronously captures the zero-crossing distortion rate and phase jitter characteristics of the power fluctuation parameters, the fractional content and angular distribution range of the beam quality distribution index, the amplitude mutation gradient and time interval entropy parameters of the speed sequence; the theoretical path calculation component inputs the real-time acquired data into the dynamic cutting path calculation model to obtain the theoretical cutting path information.

[0006] Preferably, the theoretical cutting path information calculation process includes adaptive noise elimination processing based on the cutting stage to eliminate the influence of measurement noise caused by environmental vibration interference; integrating the power fluctuation parameters, beam quality distribution indicators, and associated characteristic vectors of the speed sequence through a spatiotemporal feature fusion algorithm; and outputting theoretical cutting path information including the normal working condition fluctuation range, and the theoretical cutting path information dynamically updates data as the cutting head aging status changes.

[0007] Preferably, the multi-dimensional cutting difference analysis module specifically includes a position cumulative deviation calculation unit, a depth energy offset detection unit and a velocity sequence similarity evaluation unit; The position cumulative deviation calculation unit performs a sliding comparison between the theoretical cutting path information and the measured cutting path information in a preset distance window, uses a dynamic sequence alignment algorithm to align the asynchronously sampled path sequence data, calculates the cumulative deviation in each window, and generates a position deviation vector; The depth energy offset detection unit performs frequency domain decomposition on the depth data of the theoretical cutting path information and the measured cutting path information, calculates the energy spectrum density ratio, extracts the energy offset index of each depth layer, and constructs a depth offset vector; The speed sequence similarity evaluation unit matches the speed sequence distance distribution of the theoretical cutting path information and the measured cutting path information based on the structural matching algorithm, calculates the phase synchronization error value of the speed mutation point, quantifies the difference index of the speed interval distribution, and generates a speed similarity vector; the difference coefficient matrix generation unit performs a tensor splicing operation on the position deviation vector, the depth offset vector, and the speed similarity vector, eliminates the influence of dimensional differences through normalization processing with weighted feature importance, and outputs a third-order difference coefficient matrix with the dimension of workpiece number multiplied by timestamp multiplied by difference type.

[0008] Preferably, the depth energy offset detection unit first extracts the corresponding depth data from the theoretical cutting path information and the measured cutting path information in the frequency domain analysis stage, and uses the spectrum decomposition method to perform multi-scale frequency band analysis operations on each group of depth signals to extract the energy distribution characteristic parameters within the preset sensitive frequency band interval. The sensitive frequency band interval covers the cutting frequency range of typical metal materials. After the extraction is completed, the energy density of the theoretical cutting path information and the measured cutting path information within the sensitive frequency band interval is quantitatively calculated, and based on the relative offset degree of the two, the offset index data of each order depth layer is extracted, the energy offset results of all orders are summarized, and a depth offset vector is constructed.

[0009] Preferably, the structure matching algorithm in the speed sequence similarity evaluation unit adopts an edit distance algorithm to perform a position matching operation on the mutation point sets in the two sequences to identify the phase synchronization error value.

[0010] Preferably, the abnormal area association positioning module specifically includes a workpiece topology modeling unit, an abnormality propagation simulation unit and a probability distribution generation unit; The workpiece topology modeling unit constructs a topological diagram of the workpiece node connection relationship based on the cutting head position coordinate information, annotates the material impedance parameters between each node, and superimposes the reverse heat conduction constraint conditions of the heat-affected zone access points on the topological diagram to generate a geometric topological model including the impedance matrix and the node admittance matrix; The abnormal propagation simulation unit maps the difference coefficient matrix to the corresponding nodes of the workpiece topology model and performs abnormal propagation deduction operations based on the graph neural network. The abnormal propagation deduction calculation includes calculating the attenuation factor of the abnormal heat flow based on the node impedance parameter, capturing the cross-regional abnormal correlation feature vector through the multi-head attention mechanism, and using the Monte Carlo method to simulate the diffusion path trajectory of the abnormal heat flow in the topological network; The probability distribution generation unit counts the frequency of abnormal heat flow in each path in the simulated propagation, calculates the abnormal heat flow residence probability value in combination with the material impedance parameters, generates an abnormal probability thermal distribution map covering the entire workpiece, and marks the set of suspicious paths whose probability values ​​exceed the preset abnormal residence probability threshold; the physical area positioning unit performs spatial clustering analysis on the abnormal probability thermal distribution map, identifies the abnormal probability clustering area, and delineates the physical boundary range of the abnormal cutting based on the cutting head position coordinates and the topological connection relationship between the workpiece.

[0011] Preferably, the abnormal area association and positioning module further includes outputting suspicious device identification information and abnormal propagation main path information; The suspicious device identification information is based on the workpiece nodes connected to the suspicious path set, and the workpiece nodes are bound to the actual cutting heads to form a suspicious device identification set, indicating the potential abnormal cutting source or affected terminal; The anomaly propagation main path information is obtained by recording the node paths and their sequential sequences experienced in each round of propagation during the anomaly diffusion process of Monte Carlo simulation, counting the occurrence frequency data of each path in all simulated paths, and selecting the path sequence with the highest cumulative occurrence frequency as the anomaly propagation main path. The output anomaly propagation main path sequence is a structured ordered node list, which reflects the main propagation trajectory of the anomaly information in the workpiece.

[0012] Preferably, the present invention also includes a high-power laser cutting head, which is suitable for the high-power laser cutting optical path system as mentioned above, and the laser cutting head includes an integrated optical path transmission component, a real-time feedback component and a dynamic adjustment component; the integrated optical path transmission component includes a collimating lens unit, a focusing lens group unit and a beam shaping galvanometer unit, the collimating lens unit receives the original light beam output by the laser generator and performs collimation processing, the focusing lens group unit focuses the collimated light beam onto the surface of the workpiece to form a cutting spot, and the beam shaping galvanometer unit dynamically adjusts the beam deflection angle according to the cutting adjustment parameters issued by the adaptive cutting strategy generation module; The real-time feedback component is data-connected to the multi-dimensional cutting difference analysis module, and includes an embedded optical sensor array, which captures the actual position coordinates, energy distribution parameters and temperature gradient data of the heat-affected zone of the cutting spot in real time, and feeds the captured data back to the dynamic cutting path calculation module; the dynamic adjustment component is linked to the abnormal area association positioning module, and automatically adjusts the cutting head focal offset and the protective gas pressure value according to the suspicious path set output by the abnormal probability thermal distribution map, wherein the focal offset is compensated for nanometer displacement through a piezoelectric ceramic driver, and the protective gas pressure value is adjusted by a millisecond-level pressure gradient through a proportional valve.

[0013] Preferably, the present invention also includes a method for using a high-power laser cutting optical path system, which is applied to the high-power laser cutting optical path system as described above, and the method includes the following steps: Step S1: Loading the material property characteristic parameters of the target workpiece and the preset cutting trajectory data through the dynamic cutting path calculation module to construct an initial dynamic cutting path calculation model; Step S2: activating the real-time feedback component to capture the laser power fluctuation parameters, beam quality distribution index and cutting speed sequence data during the actual cutting process, and synchronously inputting the data into the multi-dimensional cutting difference analysis module; Step S3: the multi-dimensional cutting difference analysis module performs a three-dimensional difference analysis operation, generates a workpiece-level difference coefficient matrix, and transmits it to the abnormal area association positioning module; Step S4: the abnormal area association positioning module maps the difference coefficient matrix according to the workpiece node connection relationship topology map, and outputs the abnormal probability thermal distribution map and suspicious equipment identification information; Step S5: When it is detected that the abnormal probability value exceeds the preset abnormal residence probability threshold, the adaptive cutting strategy generation module sends a high-frequency laser monitoring instruction and a multi-band impedance disturbance instruction to the laser cutting head; Step S6: The dynamic adjustment component of the laser cutting head responds to the instruction, performs the focal length offset compensation operation and the shielding gas pressure gradient adjustment operation, and feeds back the adjusted data to the dynamic cutting path calculation module to update the theoretical cutting path information.

[0014] Compared with the prior art, the present invention has the following beneficial effects: The dynamic cutting path calculation module builds a calculation model based on historical operation data, collects parameters such as laser power fluctuations, beam quality distribution, and cutting speed sequence in real time, and generates a theoretical cutting path. This process breaks away from the reliance on fixed paths and can dynamically adjust the path planning based on the real-time status of the cutting process, making the theoretical path more closely aligned with actual cutting needs and reducing path deviations caused by changes in environmental and equipment factors. The Multi-Dimensional Cutting Difference Analysis module introduces three-dimensional difference analysis, encompassing cumulative position deviation, depth energy offset, and velocity sequence similarity metrics to generate a workpiece-level difference coefficient matrix. Compared to traditional two-dimensional analysis, this comprehensive comparison method more comprehensively captures the differences between theoretical and measured paths, focusing not only on surface position offsets but also on changes in energy distribution and velocity matching in depth, enabling a more comprehensive and accurate assessment of the cutting process. The Abnormal Area Correlation Location Module inputs the difference coefficient matrix into a spatial geometry analysis network, combining material property characteristic parameters with cutting head position coordinate information to generate an abnormality probability thermal distribution map. This approach breaks the limitations of traditional reliance on manual or simple threshold judgment. By establishing correlations between data, it can clearly demonstrate the propagation path of cutting defects and possible abnormal areas, making positioning more targeted and facilitating the rapid identification of the physical area where the problem occurs. Based on the abnormal probability thermal distribution map, the adaptive cutting strategy generation module applies multi-band impedance perturbations to the cutting heads of adjacent nodes with the largest abnormal probability gradient changes. This adjustment method is not a simple modification of a single parameter, but a precise operation based on a comprehensive analysis of multiple factors. It can timely intervene in abnormal areas, respond to complex cutting state changes through multi-band impedance perturbations, and promptly correct the cutting process when deviation trends appear, improving the stability and adaptability of the entire cutting process. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a working principle diagram of the high-power laser cutting optical path system of the present invention; Figure 2 Flowchart for theoretical cutting path information calculation; Figure 3 Flowchart for depth energy offset detection; Figure 4 Flowchart output for associating and locating abnormal areas; Figure 5 Flowchart of the operation of a high power laser cutting head. DETAILED DESCRIPTION

[0016] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0017] See also Figure 1 The present invention provides a high-power laser cutting optical path system, the system comprising: Dynamic cutting path calculation module, multi-dimensional cutting difference analysis module, abnormal area correlation positioning module and adaptive cutting strategy generation module. The specific implementation is as follows: The dynamic cutting path calculation module builds a dynamic cutting path calculation model based on the historical operation data of laser cutting, collects the actual laser power fluctuation parameters, beam quality distribution indicators and cutting speed sequence data of the current cutting process in real time, and outputs the theoretical cutting path information through the dynamic cutting path calculation model.

[0018] The multi-dimensional cutting difference analysis module receives the theoretical cutting path information and performs three-dimensional difference analysis with the cutting path information measured by the optical sensor, including the cumulative position deviation, depth energy offset and speed sequence similarity index, to generate a workpiece-level difference coefficient matrix.

[0019] The abnormal area association positioning module inputs the difference coefficient matrix into the spatial geometric analysis network, combines the material property characteristic parameters and the cutting head position coordinate information, and generates an abnormal probability thermal distribution map of the cutting defect propagation path, which is used to locate the abnormal cutting physical area.

[0020] The adaptive cutting strategy generation module configures cutting adjustment parameters according to the abnormal probability thermal distribution map, including applying a multi-band impedance perturbation operation to the cutting head of the adjacent node with the largest abnormal probability gradient change in the abnormal probability thermal distribution map.

[0021] Example 1: See Figure 2 ,The dynamic cutting path accounting module realizes the calculation of ,the theoretical cutting path information through the historical cutting feature mining ,component and the dynamic cutting path calculation model building ,component.

[0022] The historical cutting feature mining component performs a multi-dimensional decomposition of historical laser cutting operation data, extracting the energy contributions of the steady-state and transient components of the power fluctuation parameters. The steady-state component reflects the long-term stability trend of laser power, while the transient component captures short-term fluctuation characteristics. Through beam coupling analysis, a correlation matrix between beam quality distribution indicators and cutting load rate is established. This matrix is ​​used to quantify the variation of beam quality under different load conditions. The dynamic sequence alignment algorithm adjusts the speed sequence pattern under different operating conditions, eliminating timing deviations caused by changes in equipment status or material differences, and ensuring the comparability of historical data under different operating conditions.

[0023] The dynamic cutting path calculation model building component transmits the processed historical operation data to the hybrid prediction network. The network contains a time series prediction unit based on the cutting head aging curve, which is used to generate the basic path prediction value. The cutting head aging curve establishes an aging trend model by long-term monitoring of key performance parameters of the cutting head, such as the optical attenuation rate of the focusing lens group and the mechanical wear of the galvanometer system. The fully connected network embedded with the spatial attention mechanism corrects the prediction deviation caused by beam distortion. This mechanism analyzes the spatial distribution characteristics of the beam on the transmission path, identifies the distortion areas that may affect the cutting accuracy, and dynamically adjusts the weights during the prediction process. The sequence feature compensator dynamically adjusts the prediction weight coefficient according to the real-time acquired speed sequence to ensure that the model can adapt to real-time changes in cutting speed.

[0024] The heterogeneous data acquisition units deployed by the optical sensors synchronously capture the zero-crossing distortion rate and phase jitter characteristics of the power fluctuation parameters. The zero-crossing distortion rate reflects the degree of distortion of the laser power waveform, while the phase jitter characteristics are used to evaluate the temporal stability of the laser output. The fractionated content and angular distribution range of the beam quality distribution indicators are measured by the optical sensor array. The fractionated content describes the distribution of the beam energy in different spatial frequency bands, while the angular distribution range reflects the divergence characteristics of the beam. The amplitude mutation gradient and time interval entropy parameters of the velocity sequence are used to quantify the dynamic change characteristics of the cutting speed. The amplitude mutation gradient captures the instantaneous change amplitude of the velocity, while the time interval entropy evaluates the degree of randomness of the velocity sequence.

[0025] The theoretical path calculation component inputs the real-time collected data into the dynamic cutting path calculation model to calculate the theoretical cutting path information. The calculation process first performs adaptive noise elimination processing based on the cutting stage to eliminate the influence of measurement noise caused by environmental vibration interference. This processing adopts a multi-scale filtering algorithm to dynamically adjust the filtering parameters according to the different cutting stages to ensure noise suppression while retaining effective signal characteristics. The spatiotemporal feature fusion algorithm integrates the power fluctuation parameters, beam quality distribution indicators, and the associated feature vectors of the velocity sequence to construct a multi-dimensional feature space. The algorithm extracts the key factors affecting the cutting path by analyzing the correlation between different features in time and space, and establishes a dynamic mapping relationship between features. The final output of the theoretical cutting path information contains the normal working condition fluctuation range, which is dynamically updated with the aging status of the cutting head to ensure that the theoretical path can reflect the actual operating status of the equipment.

[0026] The training process of the dynamic cutting path calculation model utilizes a combined optimization strategy based on historical cutting data and real-time operation data. Historical data is used to establish initial model parameters, while real-time data is used to continuously adjust model weights through an online learning mechanism. During model training, a sliding window mechanism is used to segment historical data to ensure that the model can adapt to data changes at different time scales. The training objective function comprehensively considers path prediction accuracy, computational efficiency, and model stability, balancing various performance indicators through a multi-objective optimization algorithm. After model deployment, a periodic self-check mechanism evaluates prediction performance, triggering a model parameter update process when prediction deviation exceeds a threshold.

[0027] The theoretical cutting path information is output as structured data, including key parameters such as the path coordinate sequence, cutting energy distribution, and velocity variation curve. The path coordinate sequence describes the motion of the cutting head on the workpiece surface, while the cutting energy distribution reflects the dynamic changes in laser energy along the path. The velocity variation curve records the real-time adjustment of cutting speed, providing a reference for subsequent variance analysis. The theoretical path information is transmitted to the multi-dimensional cutting variance analysis module via a high-speed data interface, ensuring data synchronization and real-time performance.

[0028] Example 2: See Figure 3 ,The multi-dimensional cutting difference analysis module realizes the ,three-dimensional difference analysis between the theoretical cutting path information and the ,measured cutting path information through the position cumulative deviation ,calculation unit, the depth energy offset detection unit and the ,speed sequence similarity evaluation unit.

[0029] The cumulative position deviation calculation unit performs a sliding comparison between the theoretical and measured paths within a preset distance window. The window size is dynamically adjusted based on the required cutting accuracy. A dynamic sequence alignment algorithm processes asynchronously sampled path sequence data, eliminating timing misalignments caused by differences in sensor sampling rates or communication delays. After alignment, the cumulative deviation of the path data is calculated within each window to generate a position deviation vector. This vector records the path's offset in three dimensions, including deviation components in the X, Y, and Z directions.

[0030] The depth energy offset detection unit performs frequency domain decomposition on the depth data of the theoretical path and the measured path. The depth data reflects the energy deposition of the laser inside the material, and the energy distribution characteristics of different frequency bands are extracted through spectrum analysis methods. The preset sensitive frequency band covers the cutting frequency range of typical metal materials, such as the molten pool oscillation frequency of low-carbon steel and the thermal conductivity characteristic frequency of aluminum alloy. The energy spectral density ratio calculates the difference in energy distribution between the theoretical path and the measured path in each frequency band, and quantifies the degree of energy offset. The energy offset index of each depth layer constitutes a depth offset vector, which reflects the abnormal distribution of cutting energy in the thickness direction of the material.

[0031] The speed sequence similarity assessment unit uses a structural matching algorithm to analyze the speed sequence differences between the theoretical and measured paths. An edit distance algorithm matches the set of breakpoints in the two sequences to identify key time points for speed changes. The phase synchronization error value calculates the degree of alignment of breakpoints on the time axis, reflecting the synchronization differences in speed adjustments. The speed interval distribution difference index uses statistical methods to quantify the discreteness of speed fluctuations and generate a speed similarity vector. This vector contains the evaluation results for three dimensions: speed matching, phase error, and distribution difference.

[0032] The difference coefficient matrix generation unit performs tensor splicing operations on the position deviation vector, depth offset vector, and velocity similarity vector. Before splicing, all types of vectors are normalized to eliminate the influence of different physical dimensions on the analysis results. The feature importance weighting mechanism dynamically adjusts the weight coefficient of each vector according to the cutting process requirements. For example, the weight of position deviation in high-precision cutting scenarios is higher, while the weight of depth energy offset in thick plate cutting scenarios is larger. The normalized vectors are organized through a three-dimensional tensor structure to form a third-order difference coefficient matrix with three dimensions: workpiece number, timestamp, and difference type. The structural design of this matrix supports fast retrieval and batch analysis, which is convenient for subsequent anomaly detection module processing.

[0033] The position cumulative deviation calculation unit utilizes an adaptive window adjustment strategy to optimize computational efficiency. The initial window size is set based on material thickness and cutting speed, and the window size is dynamically adjusted during the calculation process based on the rate of deviation change. Larger windows are used in areas with gently varying deviations to reduce computational effort, while smaller windows are used in areas with frequent changes to improve analysis accuracy. The window sliding step size is proportional to the cutting head speed to ensure that sampling density matches cutting accuracy requirements. Position deviation vectors are stored in a ring buffer structure, ensuring data integrity while optimizing memory efficiency.

[0034] The frequency domain analysis of the depth energy offset detection unit adopts a multi-resolution processing method. Coarse-grained analysis quickly scans the energy distribution of the entire frequency band to identify the main frequency bands where anomalies may exist; fine-grained analysis performs high-precision decomposition of suspicious frequency bands to extract subtle energy difference characteristics. The frequency band division scheme presets multiple configuration templates based on the material type. For example, the stainless steel template focuses on the characteristic frequency band of the heat-affected zone, and the copper alloy template focuses on the energy rebound frequency band caused by high reflectivity. The calculation of the energy spectral density ratio introduces a sliding average filter to suppress the interference of random noise on the analysis results. The update cycle of the depth offset vector is synchronized with the change of the cutting layer depth to ensure that the energy analysis results are consistent with the actual cutting progress.

[0035] The structural matching algorithm of the velocity sequence similarity assessment unit utilizes a bidirectional comparison strategy. Forward matching identifies mutation points starting from the sequence's starting point, while reverse matching verifies the matching results backtracking from the sequence's ending point. This bidirectional verification mechanism improves the accuracy of mutation point identification and avoids mismatches caused by local fluctuations. A dynamic time warping algorithm is incorporated into the calculation of phase synchronization errors to compensate for the nonlinear time deformation between the theoretical and measured paths. The generation of the velocity similarity vector includes an outlier filtering step to eliminate instantaneous velocity sampling points that significantly exceed a reasonable range, thereby improving the reliability of the assessment results.

[0036] The variance coefficient matrix utilizes a hierarchical storage structure. The bottom layer contains the original sampled data, preserving complete time series information; the middle layer contains windowed statistical results, storing the calculated values ​​of various vectors; and the top layer, the matrix view, provides a data interface for analytical tasks. This structure balances data processing efficiency and flexibility, supporting both real-time monitoring and historical backtracking. The matrix update mechanism utilizes a transactional processing model to ensure data consistency in a multi-threaded environment. The variance coefficient is visualized using a heat map, intuitively displaying the distribution of anomaly levels at different locations and time points.

[0037] The multi-dimensional cutting difference analysis module's real-time processing capabilities are achieved through a pipeline architecture. Data acquisition, preprocessing, feature calculation, and matrix generation form a parallel processing pipeline, connected by a high-speed data bus. A computational task scheduler dynamically allocates computing resources based on system load, prioritizing analysis tasks with high real-time requirements. The module's internal status monitoring system continuously tracks the computational latency and resource utilization of each unit, automatically triggering load balancing strategies when performance bottlenecks are detected.

[0038] The analysis results of the variance coefficient matrix are output in a standardized data format, including a matrix dimension description, a timestamp sequence, and an array of coefficient values. The output interface supports multiple data transmission protocols to meet the access requirements of different downstream modules. The matrix data compression storage algorithm selects the optimal compression strategy based on the data type characteristics, such as differential encoding for position deviation data and frequency domain coefficient truncation for depth energy data. Historical matrix archiving uses a time-sliced ​​storage strategy to balance storage space usage and query efficiency.

[0039] The module's calibration and maintenance process includes sensor calibration and algorithm parameter adjustment. Sensor calibration uses standard test pieces to obtain baseline data and establish a correction relationship between measured and true values. Algorithm parameter adjustment analyzes the difference matrix characteristics under typical operating conditions to optimize key parameters such as window size and frequency band division. Calibration records are stored in association with the difference analysis results, forming a complete quality traceability chain. Maintenance diagnostic tools provide matrix anomaly pattern analysis to help technicians quickly locate the source of system deviations.

[0040] Example 3: The speed sequence similarity evaluation unit uses an improved edit distance algorithm to process the speed sequence comparison problem between the theoretical path and the measured path.

[0041] The algorithm introduces a speed change rate weight factor based on the traditional edit distance calculation, making the algorithm more sensitive to the matching of the key speed segments of the cutting process. For the theoretical speed sequence T and the measured speed sequence R of length n, the structural matching cost function is defined as: in: represents the total matching cost, represents the speed change rate of the i-th sampling point in the theoretical sequence, represents the velocity change rate of the jth point in the measured sequence, represents the speed interval weight factor, represents the time series attenuation coefficient, Indicates the absolute value of sequence position difference The velocity change rate in this formula is obtained by calculating the velocity difference of adjacent sampling points, and the weight factor According to the cutting process specification, the value is larger in the acceleration sensitive area and smaller in the uniform speed section. Control the influence of the distance between matching points on the total cost, so that the algorithm is more inclined to match speed mutation points with similar time.

[0042] The workpiece topology modeling unit of the abnormal area association and positioning module constructs a weighted graph structure based on the cutting head position coordinates. Nodes represent discretized inspection areas on the workpiece surface, and edge weights incorporate material impedance parameters and thermal conductivity characteristics. Impedance parameters are obtained from the material database and include physical properties such as thermal conductivity and specific heat capacity. Heat-affected zone constraints are based on the inverse heat conduction equation, and the main heat flow paths are marked in the topology diagram. The node admittance matrix Y of the geometric topological model is calculated based on node connectivity and material parameters and is used for subsequent anomaly propagation simulation.

[0043] The anomaly propagation simulation unit uses a graph neural network to process the mapping between the difference coefficient matrix and the topological model. Difference coefficients are assigned to corresponding topological nodes based on spatial location, forming the initial anomaly excitation signal. The graph neural network's aggregation function design considers the attenuation characteristics of material impedance, and the information transfer weight between adjacent nodes is proportional to the elements of the admittance matrix. A multi-head attention mechanism captures cross-regional anomaly correlations, with each attention head focusing on propagation modes with different physical characteristics, such as heat conduction-dominated and stress wave-dominated. The attention weight calculation incorporates temperature gradient constraints to ensure that the anomaly propagation direction conforms to the second law of thermodynamics.

[0044] The Monte Carlo simulation process implements a random walk strategy on the topological network. Each walk starts from a node with a large variance coefficient, with the transition probability proportional to the product of the admittance matrix element and the attention weight. The walk path records the sequence of nodes traversed and their dwell time. Termination criteria for a single simulation include reaching the workpiece boundary or cumulative attenuation exceeding a threshold. After repeated large-scale simulations, the frequency of abnormal heat flow arrival at each node is counted, and the probability of abnormal residence is calculated based on the material heat capacity parameters. The walk step size is dynamically adjusted during the simulation, with smaller step sizes used to improve resolution in areas of concentrated anomalies.

[0045] The probability distribution generation unit performs kernel density estimation on the Monte Carlo simulation results to produce a continuous and smooth anomaly probability heat map. The kernel bandwidth is adaptively selected based on the workpiece size and node density, automatically reducing the bandwidth in boundary areas to avoid estimation bias. The heat map's color mapping scheme uses nonlinear transformations to highlight areas with critical probability values. Suspicious paths are extracted based on probability gradient changes, connecting the gradient maxima to form a hypothesis of the main anomaly propagation path.

[0046] The physical area positioning unit applies a modified DBSCAN algorithm to spatially cluster the thermal distribution map. Algorithm parameters are dynamically adjusted based on the workpiece geometry, ensuring that the minimum neighborhood radius is proportional to the cutting spot diameter. The clustering results undergo morphological closing operations to eliminate small areas of noise and fill boundary depressions. The final output represents the physical boundary of the abnormal cut using a polygonal approximation, with vertex coordinates aligned with the coordinate system of the cutting head motion control system.

[0047] The maximum spanning tree algorithm is used to generate information about the main anomaly propagation paths. Nodes in the topological network are used as vertices, and the frequency of abnormal heat flow is used as the edge weight to construct a fully connected graph. The Prim algorithm grows from the node with the highest probability, adding the edge with the largest weight outside the current cut set with each iteration until all significant anomaly nodes are covered. The resulting ordered list of nodes is sorted by access time, and each node is annotated with its material properties and anomaly characteristics, forming a structured description of the propagation path.

[0048] The binding process for suspicious device identification information establishes a mapping between workpiece nodes and cutting head actuators. Each node is associated with the nearest neighboring cutting head position coordinates. When a node is marked as suspicious, the associated cutting head enters detailed diagnostic mode. This identification information includes the cutting head number, optical component status code, and mechanical positioning deviation value, which guides subsequent maintenance operations. This binding relationship is established using a spatial nearest neighbor algorithm, taking into account the cutting head's working radius and kinematic constraints.

[0049] The spatial geometry analysis network utilizes a hierarchical feature extraction architecture. The bottom-level convolution processes local geometric features, the middle-level graph convolution aggregates regional characteristics, and the top-level fully connected network integrates global anomaly indicators. Network training utilizes a semi-supervised learning approach, with a small amount of labeled data used to adjust classification boundaries and a large amount of unlabeled data used to learn feature representations. The loss function incorporates topology-preserving constraints to ensure that the learned feature space retains the original workpiece geometry. The network inference process utilizes a sliding window strategy, segmenting large workpieces into overlapping blocks, processing them separately, and then re-integrating the results.

[0050] The update mechanism for the anomaly probability heat map utilizes an incremental recalculation strategy. Newly collected differential data triggers a resimulation of a local area, updating only the probability values ​​of the affected nodes. The scope of the incremental calculation is determined based on an estimated anomaly propagation speed to ensure inclusion of areas of potential change. Historical probability distributions are cached using a sliding time window, supporting probability trend analysis and short-term forecasting. Version management of the distribution map records the timestamp and modified areas of each update, facilitating tracing back the anomaly's development.

[0051] The parameters for dynamic impedance perturbation are configured based on probabilistic gradient analysis. For pairs of adjacent nodes with the largest gradient changes, the correlation coefficient between the material impedance difference and the probability change is calculated. The perturbation frequency is selected to avoid the material's inherent resonant frequency band, and the amplitude is dynamically adjusted based on the current cutting depth. Multi-band perturbation is achieved by superimposing Gaussian pulses with different center frequencies, with the energy distribution within each band matching the anomaly's characteristic spectrum. The perturbation application sequence is synchronized with the cutting head's motion trajectory to avoid interference with the normal cutting process.

[0052] The physical region boundary coordinate conversion module processes the mapping between the workpiece coordinate system and the machine coordinate system. The coordinate transformation matrix, which includes rotation, translation, and scale parameters, is obtained through a calibration process. Boundary polygon vertices are interpolated and encrypted according to machining accuracy requirements, and the resulting path point sequence is fed into the velocity planning module. The coordinate conversion process accounts for workpiece clamping deviation compensation and corrects positioning errors through reference point alignment. Collision detection is performed on the converted machine coordinates before they are sent to the motion controller to ensure safety during abnormal region processing.

[0053] Anomaly localization results are verified using a multi-sensor data fusion approach. Thermal imaging data verifies the temperature distribution in the anomaly area, acoustic emission signals analyze internal material defects, and visual inspection confirms surface topography anomalies. Data from each sensor is synchronized via timestamps and spatially registered to the same coordinate system. Verification results are fed back into the probabilistic model to adjust parameters, such as transition probabilities in Monte Carlo simulations or attention weights in graph neural networks. This iterative verification process gradually improves localization accuracy until the evidence from each sensor is consistent.

[0054] The maintenance decision support system analyzes the correlation between anomaly location results and equipment status data. It establishes a spatiotemporal correlation model between cutting head maintenance records and the locations of anomaly areas, identifying recurring failure patterns. A decision rule library encompasses multi-dimensional criteria such as material properties, process parameters, and equipment status, generating maintenance priority recommendations. The support system provides interactive causal analysis tools, allowing technicians to explore the predicted effects of different maintenance plans. The results of maintenance actions are recorded and used to optimize decision rules.

[0055] Example 4: See Figure 4 ,The suspicious device identification information generation process of the abnormal area ,association positioning module is realized through the binding operation of the ,workpiece node and the actual cutting head.

[0056] In the example of cutting stainless steel sheets, when node N-17 in the workpiece topology model is marked as the starting point of a suspicious path, the system retrieves the spatial coordinates of this node (x=245mm, y=178mm) and queries the cutting head position log. According to the motion control system's records, cutting head H-03 was within ±2mm of this coordinate at timestamp T-428, and its optical lens temperature sensor showed abnormal temperature rise. The system establishes a binding relationship between the H-03 cutting head and node N-17 and generates a record item in the suspicious device identification set, including fields such as the cutting head serial number, the start time of the abnormality, and the temperature deviation value. For scenarios where multiple cutting heads work together, the binding operation takes into account the weight distribution of overlapping working radius areas. When a coordinate point is at the intersection of the working ranges of multiple cutting heads, the system simultaneously marks all potentially affected cutting heads and annotates them with confidence scores.

[0057] The update mechanism for the suspicious device identification set adopts an event-driven mode. When laser cutting carbon fiber composite materials, the optical sensor detects a sudden increase of 30% in the beam reflectivity at the N-09 node, triggering the real-time binding process. The system immediately scans all cutting head position data within a 5mm radius of the node in the last 3 minutes and finds that the Z-axis coordinate of the focusing mirror of the cutting head H-05 has an abnormal offset of 0.15mm. This information is appended to the suspicious device identification set and displayed in association with the historical maintenance records, indicating that the last calibration of the lens group of the cutting head has been working for 98 hours, approaching the maintenance cycle threshold. In a multi-cutting head system, this type of real-time binding operation takes an average of 12ms to complete, ensuring rapid association of abnormal signals with equipment status.

[0058] The verification process of suspicious equipment identification information introduces multi-dimensional cross-checks. During a stainless steel pipe cutting operation, the system initially identified cutting head H-02 as a suspicious device, but the motion trajectory playback showed that the cutting head was in standby mode during the abnormal period. The system automatically initiated a secondary verification process: it retrieved the thermal imager data to confirm that the H-02 barrel temperature did indeed fluctuate abnormally by 3°C; it checked the work log of the adjacent cutting head H-01 and found that its shielding gas flow rate decreased by 25% during the abnormal period; and it was finally determined that the abnormality of H-01 caused the heat flow to interfere with H-02. The verification process in such complex scenarios usually lasts 45-60 seconds. The system will temporarily lock the suspicious equipment identification status until the verification is completed to avoid false alarms that interfere with production.

[0059] Visualization of the main anomaly propagation path supports spatiotemporal analysis. When cutting multi-layer composite armor plates, the main path identified by the system exhibits periodic fluctuations in the Z-axis, corresponding to interface reflections between different material layers. A 3D visualization tool unfolds the path nodes in a time series, showing that the anomalous heat flux propagates 1.7 times faster in the ceramic layer (Z=2.1mm) than in the metal layer (Z=3.4mm). The color gradient of the nodes on the timeline reflects the temperature change history, such as from orange-red (320°C) at node N-28 to dark red (285°C) at node N-31. Operators can observe the dynamic process of anomaly propagation by sliding the interactive timeline, pausing at any time to view detailed node material properties and cutting parameters.

[0060] The maintenance decision support function for suspicious equipment identification is specifically demonstrated through priority sorting. On a certain automobile chassis cutting line, the system flagged two suspicious items: contamination of the focusing lens of cutting head H-04 and vibration of the guide rail of H-07. Analysis based on the abnormal thermal distribution map showed that the abnormal area associated with H-04 accounted for 18% of the total workpiece area, while H-07 affected only 6%. When the system automatically generated maintenance work orders, H-04 was assigned Priority A (requires immediate attention) and H-07 was assigned Priority B (requires attention within 8 hours). Each work order entry included a screenshot of the associated abnormal path, a table of equipment operating parameter deviations, and recommended maintenance actions. For example, the H-04 work order recommended "clean the focusing lens and inspect the cooling lines." Historical data shows that this prioritization can reduce unplanned downtime by 22%.

[0061] The prediction function of the main path of abnormal propagation is realized through a machine learning model. When continuously cutting 10 aluminum alloy components of the same specifications, the system found that the abnormal path always originated from the upper left corner of the workpiece. When the sixth workpiece began to be cut, the system deployed enhanced monitoring around the predicted path [N-05→N-08→N-11] in advance. Sure enough, the initial abnormal signal of the N-05 node was captured when the cutting was 38%. The prediction model is trained based on the path pattern of the first five workpieces, taking into account more than 30 characteristic parameters such as material batch, ambient temperature and humidity. When the overlap between the actual abnormal path and the predicted path exceeds 75%, the system will automatically optimize the cutting parameters, such as reducing the laser power in the upper left corner by 8% when cutting the seventh workpiece.

[0062] Suspicious device identification information is stored using a hierarchical data structure. Each identification entry comprises three dimensions: a foundational layer (device ID, binding time), an evidence layer (sensor readings, video snapshots), and an analysis layer (anomaly type classification, impact assessment). When a cutting head H-09 is repeatedly flagged as having a loose fiber optic interface, the system automatically creates a unique failure mode profile for that device, recording data such as temperature curves and vibration spectrum characteristics from each anomaly. This structured information is transmitted to the factory's MES system via the OPC UA interface, where it is stored in association with business data such as equipment maintenance history and spare parts replacement records, forming a comprehensive equipment health management database.

[0063] The process optimization application of the main path of abnormal propagation is reflected in the adaptive adjustment of parameters. During a copper alloy cutting operation, the system identified that the main path extended along the rolling direction of the material, and the width of the heat-affected zone at the corresponding node exceeded the standard. The control system immediately initiated a compensation strategy: increasing the cutting speed by 12% at the upstream node of the path and increasing the auxiliary gas pressure by 25 kPa at the downstream node. This local parameter adjustment based on the physical propagation path can maintain the processing efficiency of non-abnormal areas compared to global parameter modifications. After the compensation was implemented, subsequent quality inspections showed that the width of the heat-affected zone was within the process requirement of ±0.05 mm, and the overall cutting time increased by only 3.7%.

[0064] Closed-loop management of suspicious device identification is reflected in the automated verification process. When the system flags a misalignment in the collimator of an H-12 cutting head, it automatically performs a three-point calibration test the next time the device is powered on. First, a standard pattern is cut on a test board. A high-precision camera then measures the deviation between the actual cutting path and the theoretical path. Finally, the calibration results are fed back to the identification system. If the verification passes (deviation < 0.02mm), the suspicious identification is cleared. If not, the alarm is upgraded to a fault and the device is locked. The entire verification process is completed automatically during the device's warm-up phase, taking approximately two minutes and not impacting normal production schedules. This mechanism effectively prevents "over-maintenance" issues. Historical data shows that approximately 31% of suspicious identifications are resolved during the automated verification process.

[0065] The cross-workpiece analysis function of the main path of abnormal propagation supports batch quality traceability. When continuously processing 20 wind turbine bearing seat components, the system found that abnormal paths were concentrated in the area around the flange hole. By retrieving the cutting records of similar workpieces in the past three months, statistics show that 86% of the abnormal paths have similar spatial distribution patterns with the current workpiece. This group analysis result triggers the process review process and traces back to the recently replaced shielding gas supplier. The quality engineer confirmed through path pattern comparison that the insufficient purity of the new gas source caused oxidation of the cut, which was the root cause of the path abnormality, and then adjusted the gas composition to solve the batch problem. The path similarity analysis tool provided by the system supports screening and comparison by multiple conditions such as time, material batch, and equipment combination.

[0066] The hardware architecture of the suspicious device identification system adopts a distributed design. Each cutting head is equipped with an independent edge computing unit, which caches the motion parameters and sensor data of the last 8 hours in real time. When the central system initiates binding, the edge unit immediately uploads detailed operation logs for the specified time period, such as the servo motor current fluctuation curve of the H-15 cutting head during the abnormal period. This design reduces the network transmission burden and keeps the average response time of binding queries within 50ms. The data synchronization mechanism ensures that even if the network is temporarily interrupted, the edge unit can maintain a 72-hour data cache and automatically retransmit key event records after the connection is restored. All transmitted data is encrypted using AES-256 to prevent the leakage of device status information.

[0067] Example 5: See Figure 5 The integrated optical path transmission component of the high-power laser cutting head adopts a modular design architecture, and the collimating lens unit is equipped with an adaptive zoom mechanism, which can automatically adjust the focal length position according to the original divergence angle of the output beam of the laser generator.

[0068] When the cutting system switches between laser sources of different power levels, the servo drive mechanism of the collimating lens completes position fine-tuning within 150 milliseconds, ensuring that the collimation error of the output beam is controlled within 0.05 mrad. The focusing lens unit uses a multi-layer coating process and is specially designed with an anti-reflection film system for laser beams in the 1064nm band. Under continuous operation, the temperature rise of the lens does not exceed 8°C above the ambient temperature. The beam shaping galvanometer unit has a built-in high-speed position feedback sensor, and the galvanometer deflection angle resolution reaches 0.001 degrees. Combined with the control instructions of the dynamic adjustment component, it can achieve precise tracking of complex curve trajectories.

[0069] The embedded optical sensor array of the real-time feedback component adopts a ring-shaped distribution design, with 12 high-sensitivity detection units arranged circumferentially around the cutting spot. Each detection unit contains a visible light CCD, an infrared thermal imager and a spectral analysis module, which synchronously collects the spot position coordinates, energy distribution parameters and temperature gradient data of the heat-affected zone at a sampling frequency of 2000Hz. The position coordinate measurement is based on the principle of double-cross laser positioning. The center position of the spot is identified by an image processing algorithm, and the spatial positioning accuracy reaches ±5μm. The energy distribution parameters are obtained through a beam splitter prism array to measure the asymmetry of the energy distribution of the light beam in the X / Y axis direction. The temperature gradient data of the heat-affected zone adopts a non-contact measurement method, and the infrared sensor array scans the temperature field changes at the cutting front with a spatial resolution of 50μm.

[0070] The dynamic adjustment component realizes real-time compensation of cutting parameters through a multi-axis linkage mechanism. The piezoelectric ceramic driver controls the axial displacement of the focusing lens group. Each nanometer displacement corresponds to an adjustment of the focal position of 0.3μm, and the compensation instruction response time is less than 1 millisecond. The protective gas pressure regulation system consists of a high-speed proportional valve and a micro eddy current sensor. The opening resolution of the proportional valve reaches 0.1%, and it can establish a new pressure balance within 10 milliseconds. The gas nozzle is designed as a rotatable structure, which automatically adjusts the airflow injection angle to the optimal position according to the orientation of the abnormal area indicated by the abnormal probability thermal distribution map. All actuators of the adjustment component are equipped with redundant sensors to monitor the deviation between the actual execution status and the instruction value in real time, and immediately trigger a protective shutdown when the deviation exceeds the safety threshold.

[0071] Data exchange between the laser cutting head and the optical path system utilizes a dual-channel architecture of fiber-optic communication and wireless transmission. Critical control commands are transmitted in real time via gigabit fiber, ensuring transmission delays of adjustment commands are less than 0.5 milliseconds. Condition monitoring data is wirelessly transmitted via the 5G millimeter-wave frequency band. Each cutting head is equipped with an independent network slice, achieving a stable data transmission rate exceeding 800 Mbps. The communication protocol utilizes the Time-Sensitive Networking standard, giving the highest priority to the transmission of critical data packets, ensuring timely delivery of control commands even in network congestion. A data encryption module provides end-to-end encryption for all transmitted commands and status information to prevent external interference and malicious attacks.

[0072] The cutting head's thermal management system utilizes phase-change material cooling technology. An array of micro heat pipes is arranged around the collimating lens unit and focusing lens assembly, leveraging the latent heat absorption properties of phase-change materials to control the operating temperature fluctuations of the optical components to within ±0.5°C. The galvanometer drive motor's heat dissipation system utilizes eddy current cooling technology, using a micro-fan to generate a rotating airflow to remove heat, ensuring that the motor winding temperature does not exceed the rated operating range. All cooling circuits are equipped with flow and temperature monitoring sensors that automatically adjust pumping power or switch to an alternate cooling path when a decrease in cooling efficiency is detected. The thermal management system's status data is uploaded to the central monitoring platform in real time, forming a complete thermal load trend chart.

[0073] The pollution protection system for optical components includes a multi-layer protection mechanism. The outermost layer of air curtain protection adopts an annular airflow design to form a stable air barrier around the optical window, effectively blocking more than 99% of flying particles. The self-cleaning coating of the middle layer has oleophobic and hydrophobic properties, making it difficult for pollutants to adhere to the optical surface. The electrostatic dust removal device on the inner layer is automatically activated during the cutting interval, and absorbs residual tiny particles through a high-voltage electrostatic field. The operating status of the protection system is evaluated in real time by a particle counter. When the transmittance drops by more than 3%, the automatic cleaning program is triggered or manual maintenance is prompted. The records of all protective actions form a complete maintenance log, which is used to analyze the source of pollution and optimize the protection strategy.

[0074] The cutting head's mechanical structure utilizes a lightweight magnesium alloy design, resulting in an overall weight reduction of 35% compared to traditional structures while maintaining sufficient rigidity and stability. The bearing system of the moving parts utilizes magnetic levitation technology to eliminate positioning errors caused by mechanical friction. The quick-release interface design enables optical modules to be replaced within three minutes. The optical path calibration process after module docking is fully automated, requiring no human intervention. The structural health monitoring system uses a vibration sensor array to capture changes in mechanical resonant frequency in real time, prompting preventive maintenance when a decrease in structural stiffness is detected. Three-dimensional models of all mechanical components are stored in the digital twin system for simulation analysis and life prediction.

[0075] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0076] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A high-power laser cutting optical path system, characterized in that: include: The dynamic cutting path calculation module builds a dynamic cutting path calculation model based on the historical laser cutting operation data, collects the actual laser power fluctuation parameters, beam quality distribution indicators and cutting speed sequence data of the current cutting process in real time, and outputs the theoretical cutting path information through the dynamic cutting path calculation model; A multi-dimensional cutting difference analysis module receives the theoretical cutting path information and performs a three-dimensional difference analysis on the cutting path information measured by the optical sensor. The three-dimensional difference analysis includes the cumulative position deviation, depth energy offset and velocity sequence similarity index to generate a workpiece-level difference coefficient matrix; The abnormal area association positioning module inputs the difference coefficient matrix into the spatial geometric analysis network, combines the material property characteristic parameters with the cutting head position coordinate information, and generates an abnormal probability thermal distribution map of the cutting defect propagation path for locating the abnormal cutting physical area; The adaptive cutting strategy generation module configures cutting adjustment parameters according to the abnormal probability thermal distribution map, including applying a multi-band impedance perturbation operation to the cutting head of the adjacent node with the largest abnormal probability gradient change in the abnormal probability thermal distribution map.

2. A high-power laser cutting optical path system according to claim 1, characterized in that: The dynamic cutting path calculation module specifically includes a historical cutting feature mining component and a dynamic cutting path calculation model building component; The historical cutting feature mining component processes historical laser cutting operation data and performs multi-dimensional decomposition operations, involving extracting the energy proportions of the steady-state and transient components of the power fluctuation parameters. It establishes a correlation matrix between the beam quality distribution index and the cutting load rate through beam coupling analysis, and uses a dynamic sequence alignment algorithm to adjust the speed sequence mode under different working conditions. The dynamic cutting path calculation model construction component transmits the processed historical operation data to the hybrid prediction network, which has a timing prediction unit based on the cutting head aging curve to generate a basic path prediction value. The fully connected network embedded with the spatial attention mechanism is used to correct the prediction deviation caused by beam distortion, and the sequence feature compensator dynamically adjusts the prediction weight coefficient according to the real-time acquired speed sequence; the heterogeneous data acquisition unit deployed by the optical sensor synchronously captures the zero-crossing distortion rate and phase jitter characteristics of the power fluctuation parameters, the fractional content and angular distribution range of the beam quality distribution index, the amplitude mutation gradient and time interval entropy parameters of the speed sequence; the theoretical path calculation component inputs the real-time acquired data into the dynamic cutting path calculation model to obtain the theoretical cutting path information.

3. The high-power laser cutting optical path system according to claim 2, characterized in that: The theoretical cutting path information calculation process includes adaptive noise elimination processing based on the cutting stage to eliminate the influence of measurement noise caused by environmental vibration interference; integrating the power fluctuation parameters, beam quality distribution indicators, and associated characteristic vectors of the speed sequence through a spatiotemporal feature fusion algorithm; and outputting theoretical cutting path information including the normal working condition fluctuation range. The theoretical cutting path information is dynamically updated as the cutting head ages.

4. The high-power laser cutting optical path system according to claim 1, characterized in that: The multi-dimensional cutting difference analysis module specifically includes a position cumulative deviation calculation unit, a depth energy offset detection unit and a speed sequence similarity evaluation unit; The position cumulative deviation calculation unit performs a sliding comparison between the theoretical cutting path information and the measured cutting path information in a preset distance window, uses a dynamic sequence alignment algorithm to align the asynchronously sampled path sequence data, calculates the cumulative deviation in each window, and generates a position deviation vector; The depth energy offset detection unit performs frequency domain decomposition on the depth data of the theoretical cutting path information and the measured cutting path information, calculates the energy spectrum density ratio, extracts the energy offset index of each depth layer, and constructs a depth offset vector; The speed sequence similarity evaluation unit matches the speed sequence distance distribution of the theoretical cutting path information and the measured cutting path information based on the structural matching algorithm, calculates the phase synchronization error value of the speed mutation point, quantifies the difference index of the speed interval distribution, and generates a speed similarity vector; the difference coefficient matrix generation unit performs a tensor splicing operation on the position deviation vector, the depth offset vector, and the speed similarity vector, eliminates the influence of dimensional differences through normalization processing with weighted feature importance, and outputs a third-order difference coefficient matrix with the dimension of workpiece number multiplied by timestamp multiplied by difference type.

5. The high-power laser cutting optical system according to claim 4, characterized in that: The depth energy offset detection unit first extracts the corresponding depth data from the theoretical cutting path information and the measured cutting path information in the frequency domain analysis stage, and uses the spectrum decomposition method to perform multi-scale frequency band analysis on each group of depth signals to extract the energy distribution characteristic parameters within the preset sensitive frequency band interval. The sensitive frequency band interval covers the cutting frequency range of typical metal materials. After the extraction is completed, the energy density of the theoretical cutting path information and the measured cutting path information within the sensitive frequency band interval is quantitatively calculated, and based on the relative offset degree of the two, the offset index data of each order depth layer is extracted, and the energy offset results of all orders are summarized to construct a depth offset vector.

6. The high-power laser cutting optical system according to claim 4, characterized in that: The structure matching algorithm in the speed sequence similarity evaluation unit adopts the edit distance algorithm to perform position matching operation on the mutation point sets in the two sequences and identify the phase synchronization error value.

7. The high-power laser cutting optical path system according to claim 1, characterized in that: The abnormal area association positioning module specifically includes a workpiece topology modeling unit, an abnormal propagation simulation unit and a probability distribution generation unit; The workpiece topology modeling unit constructs a topological diagram of the workpiece node connection relationship based on the cutting head position coordinate information, annotates the material impedance parameters between each node, and superimposes the reverse heat conduction constraint conditions of the heat-affected zone access points on the topological diagram to generate a geometric topological model including the impedance matrix and the node admittance matrix; The abnormal propagation simulation unit maps the difference coefficient matrix to the corresponding nodes of the workpiece topology model and performs abnormal propagation deduction operations based on the graph neural network. The abnormal propagation deduction calculation includes calculating the attenuation factor of the abnormal heat flow based on the node impedance parameter, capturing the cross-regional abnormal correlation feature vector through the multi-head attention mechanism, and using the Monte Carlo method to simulate the diffusion path trajectory of the abnormal heat flow in the topological network; The probability distribution generation unit counts the frequency of abnormal heat flow in each path in the simulated propagation, calculates the abnormal heat flow residence probability value in combination with the material impedance parameters, generates an abnormal probability thermal distribution map covering the entire workpiece, and marks the set of suspicious paths whose probability values ​​exceed the preset abnormal residence probability threshold; the physical area positioning unit performs spatial clustering analysis on the abnormal probability thermal distribution map, identifies the abnormal probability clustering area, and delineates the physical boundary range of the abnormal cutting based on the cutting head position coordinates and the topological connection relationship between the workpiece.

8. The high-power laser cutting optical path system according to claim 7, characterized in that: The abnormal area association and positioning module further includes outputting suspicious device identification information and abnormal propagation main path information; The suspicious device identification information is based on the workpiece nodes connected to the suspicious path set, and the workpiece nodes are bound to the actual cutting heads to form a suspicious device identification set, indicating the potential abnormal cutting source or affected terminal; The anomaly propagation main path information is obtained by recording the node paths and their sequential sequences experienced in each round of propagation during the anomaly diffusion process of Monte Carlo simulation, counting the occurrence frequency data of each path in all simulated paths, and selecting the path sequence with the highest cumulative occurrence frequency as the anomaly propagation main path. The output anomaly propagation main path sequence is a structured ordered node list, which reflects the main propagation trajectory of the anomaly information in the workpiece.

9. A high-power laser cutting head, suitable for the high-power laser cutting optical path system according to any one of claims 1 to 8, characterized in that: The laser cutting head includes an integrated optical path transmission component, a real-time feedback component and a dynamic adjustment component; the integrated optical path transmission component includes a collimating lens unit, a focusing lens group unit and a beam shaping galvanometer unit, the collimating lens unit receives the original light beam output by the laser generator and performs collimation processing, the focusing lens group unit focuses the collimated light beam onto the workpiece surface to form a cutting spot, and the beam shaping galvanometer unit dynamically adjusts the beam deflection angle according to the cutting adjustment parameters issued by the adaptive cutting strategy generation module; the real-time feedback component is data-connected to the multi-dimensional cutting difference analysis module, includes an embedded optical sensor array, captures the actual position coordinates, energy distribution parameters and heat-affected zone temperature gradient data of the cutting spot in real time, and feeds the captured data back to the dynamic cutting path calculation module; the dynamic adjustment component is linked to the abnormal area association positioning module, and automatically adjusts the cutting head focal length offset and the shielding gas pressure value according to the suspicious path set output by the abnormal probability thermal distribution map, wherein the focal length offset is compensated for by nanometer displacement through a piezoelectric ceramic driver, and the shielding gas pressure value is adjusted by a millisecond pressure gradient through a proportional valve.

10. A method for using a high-power laser cutting optical path system, applied to the high-power laser cutting optical path system according to claims 1 to 8, characterized in that: The following steps are involved: Step S1: Loading the material property characteristic parameters of the target workpiece and the preset cutting trajectory data through the dynamic cutting path calculation module to construct an initial dynamic cutting path calculation model; Step S2: activating the real-time feedback component to capture the laser power fluctuation parameters, beam quality distribution index and cutting speed sequence data during the actual cutting process, and synchronously inputting the data into the multi-dimensional cutting difference analysis module; Step S3: the multi-dimensional cutting difference analysis module performs a three-dimensional difference analysis operation, generates a workpiece-level difference coefficient matrix, and transmits it to the abnormal area association positioning module; Step S4: the abnormal area association positioning module maps the difference coefficient matrix according to the workpiece node connection relationship topology map, and outputs the abnormal probability thermal distribution map and suspicious equipment identification information; Step S5: When it is detected that the abnormal probability value exceeds the preset abnormal residence probability threshold, the adaptive cutting strategy generation module sends a high-frequency laser monitoring instruction and a multi-band impedance disturbance instruction to the laser cutting head; Step S6: The dynamic adjustment component of the laser cutting head responds to the instruction, performs the focal length offset compensation operation and the shielding gas pressure gradient adjustment operation, and feeds back the adjusted data to the dynamic cutting path calculation module to update the theoretical cutting path information.

Images