A laser cutting machine for sheet metal processing and a control method thereof

Abstract

The present application relates to the technical field of laser cutting, and provides a laser cutting machine for sheet metal machining and a control method thereof, the method comprising: acquiring a current working condition characteristic; acquiring a correlation degree of the current working condition and all historical sample working conditions according to the current working condition characteristic, sorting the historical path samples in descending order according to the correlation degree, and extracting the first N groups as a global optimal candidate path set; acquiring actual cutting paths of the same type of sheet metal workpiece in the recent several times and taking them as real-time reference samples, and comparing the real-time reference samples and the global optimal candidate path set in terms of advantages and disadvantages, and taking the optimal path as the final laser cutting path; and controlling the laser cutting head to perform cutting work according to the optimal laser cutting path. The present application can reuse mature historical high-quality paths, does not need to perform full-amount algorithm iteration every time when planning the cutting path, can quickly screen the optimal path relying on historical sample data, thereby shortening the path planning time consumption and improving the sheet metal cutting machining efficiency.

Description

Technical Field

[0001] This invention relates to the field of laser cutting technology, and more specifically, to a laser cutting machine for sheet metal processing and its control method. Background Technology

[0002] A sheet metal frame, as the name suggests, is a frame manufactured using sheet metal processing techniques. Sheet metal processing is a comprehensive technology for processing thin metal sheets, including cutting, stamping, bending, welding, and other methods. Through these processes, thin metal sheets can be processed into components of various shapes and sizes, and then these components are assembled into a complete frame structure.

[0003] Current path control for sheet metal laser cutting machines primarily employs a single algorithm planning mode, mainly based on genetic algorithms, ant colony algorithms, and simulated annealing algorithms. This approach relies solely on the current sheet metal parameters, cutting pattern, and equipment operating conditions for a single path solution, exhibiting significant technical limitations: path planning lacks accumulated process data, cannot reuse high-quality cutting paths validated over long-term processing, and repeatedly performs ineffective iterative calculations for similar sheet metal workpieces of the same type and specifications, resulting in low processing efficiency. In other words, the processing efficiency of current sheet metal laser cutting machine path control technology has room for further optimization. Summary of the Invention

[0004] Based on this, in order to improve the efficiency of sheet metal laser cutting, the present invention provides a laser cutting machine for sheet metal processing and its control method, the specific technical solution of which is as follows: A laser cutting machine for sheet metal processing includes a parameter acquisition component for collecting multi-dimensional real-time processing parameters and an industrial control center for preprocessing the real-time processing parameters to obtain current working condition characteristics. The industrial control center stores historical sample data of sheet metal laser cutting. The industrial control center includes: The alternative path set acquisition module is used to obtain the correlation between the current working condition and all historical sample working conditions based on the characteristics of the current working condition, filter out historical path samples with a correlation greater than the correlation threshold, sort the historical path samples in descending order according to the correlation, and extract the top N groups as the global optimal alternative path set. The cutting path optimization module is used to obtain the actual cutting paths of the same type of sheet metal workpieces in the most recent times and use them as real-time reference samples. It performs a quantitative comparison of the advantages and disadvantages of the real-time reference samples and the global optimal alternative path set, and uses the optimal path as the final laser cutting path. The laser cutting control module is used to control the laser cutting head to perform cutting operations according to the optimal laser cutting path.

[0005] The laser cutting machine for sheet metal processing obtains the global optimal alternative path set and performs a quantitative comparison of the advantages and disadvantages of the real-time reference sample and the global optimal alternative path set. The optimal path is used as the final laser cutting path. It can reuse mature historical high-quality paths. When planning the cutting path, there is no need to perform a full algorithm iteration every time. The optimal path can be quickly selected based on historical sample data, thereby shortening the path planning time and improving the efficiency of sheet metal cutting processing.

[0006] Preferably, the operating condition center also includes a feature dimension classification module, which is used to classify the current operating condition by feature dimension, dividing it into amplitude-sensitive features and trend-sensitive features. The alternative path set acquisition module includes: The Euclidean distance acquisition unit is used to obtain the Euclidean distance between the current working condition and all historical sample working conditions for amplitude-sensitive features, and to obtain the amplitude matching degree based on the Euclidean distance; The similarity acquisition unit is used to obtain the cosine similarity between the current working condition and all historical sample working conditions for trend-sensitive features, and to obtain the trend matching degree based on the cosine similarity. The correlation degree acquisition unit is used to perform weighted fusion of amplitude matching degree and trend matching degree to obtain the correlation degree.

[0007] Preferably, the cutting path optimization module includes: The path quality acquisition unit is used to select at least one quality evaluation index and perform path quality quantification scoring on the real-time reference sample and the global optimal candidate path set based on the quality evaluation index to obtain the path quality score. The path selection and comparison unit is used to compare the real-time reference samples and the global optimal candidate path set based on the path quality score, and select the path with the highest path quality score as the optimal path and as the final laser cutting path.

[0008] Preferably, the central control center also includes: The baseline trajectory acquisition module is used to obtain a baseline path set based on the union of real-time reference samples and the global optimal candidate path set, and to perform weighted fusion on all paths in the baseline path set to obtain an ideal baseline trajectory. The target trajectory acquisition module is used to acquire a target cutting trajectory that matches the current working condition based on real-time processing parameters. The target cutting trajectory includes trajectory point coordinates, cutting speed, laser output power, and real-time kerf forming parameters. The cutting error acquisition module is used to combine the ideal reference trajectory and the target cutting trajectory to obtain the mechanical clearance position error and the consumable wear and forming error. The trajectory error compensation module is used to compensate and correct the current real-time cutting path based on mechanical clearance position error and consumable wear and forming error. The target cutting trajectory can be either a real-time cutting trajectory or a historical cutting trajectory.

[0009] A control method for a laser cutting machine for sheet metal processing, applied to the laser cutting machine for sheet metal processing, includes the following steps: S1, collect multi-dimensional real-time processing parameters and preprocess the real-time processing parameters to obtain the current working condition characteristics; S2, based on the characteristics of the current working condition, obtain the correlation between the current working condition and all historical sample working conditions, filter out historical path samples with a correlation greater than the correlation threshold, sort the historical path samples in descending order according to the correlation, and extract the top N groups as the global optimal candidate path set. S3: Obtain the actual cutting paths of the same type of sheet metal workpieces in the most recent times and use them as real-time reference samples. Perform a quantitative comparison of the advantages and disadvantages of the real-time reference samples and the global optimal alternative path set, and use the optimal path as the final laser cutting path. S4 controls the laser cutting head to perform the cutting operation according to the optimal laser cutting path.

[0010] Preferably, step S2 specifically includes the following steps: S21, classify the current working condition by feature dimension, dividing it into amplitude-sensitive features and trend-sensitive features, and normalize the amplitude-sensitive features and trend-sensitive features. S22, for amplitude-sensitive features, obtain the Euclidean distance between the current working condition and all historical sample working conditions, and obtain the amplitude matching degree based on the Euclidean distance. At the same time, for trend-sensitive features, obtain the cosine similarity between the current working condition and all historical sample working conditions, and obtain the trend matching degree based on the cosine similarity. S23, weighted fusion of amplitude matching degree and trend matching degree to obtain correlation degree.

[0011] Preferably, step S3 specifically includes the following steps: S31, Select at least one quality evaluation index and perform path quality quantification scoring on the real-time reference sample and the global optimal alternative path set based on the quality evaluation index to obtain the path quality score; S32, based on the path quality score, compares the real-time reference samples and the global optimal candidate path set, and selects the path with the highest path quality score as the optimal path and the final laser cutting path.

[0012] Preferably, the control method for a laser cutting machine for sheet metal processing further includes the following steps: S5: Obtain the baseline path set based on the union of real-time reference samples and the global optimal alternative path set, and perform weighted fusion on all paths in the baseline path set to obtain the ideal baseline trajectory. S6. Obtain the target cutting trajectory that matches the current working condition based on the real-time processing parameters. The target cutting trajectory includes the coordinates of the trajectory points, the cutting speed, the laser output power, and the real-time parameters of the kerf formation. S7, by combining the ideal reference trajectory and the target cutting trajectory, obtains the mechanical clearance position error and the consumable wear and forming error; S8, compensate and correct the current real-time cutting path based on the mechanical clearance position error and the wear and forming error of the consumables; The target cutting trajectory can be either a real-time cutting trajectory or a historical cutting trajectory.

[0013] Preferably, step S7 specifically includes the following steps: S71, extract the real-time XY coordinates of all straight lines, arcs, and corners in the ideal reference trajectory and the target cutting trajectory, calculate the position deviation point by point, generate the global trajectory deviation matrix, remove the random noise of the global trajectory deviation matrix, select the fixed deviations that exist in the same direction, same magnitude, and continuous interval of multiple paths, and perform mean filtering on the fixed deviations to obtain the mechanical clearance position error. S72 compares the cutting forming parameters of the ideal reference trajectory and the target cutting trajectory point by point to obtain the global forming deviation. Based on the multiple features corresponding to nozzle wear and the multiple features corresponding to lens optical path wear, the global forming deviation is coupled and judged. Combined with the time-series cumulative error and multi-path commonality verification, the effective system error of wear is obtained and the effective system error of wear is processed by mean filtering to obtain the consumable wear forming error.

[0014] Preferably, the control method for a laser cutting machine for sheet metal processing further includes the following steps: obtaining the current processing accuracy level and sheet metal processing difficulty, and performing weighted fusion on the normalized current processing accuracy level and sheet metal processing difficulty to obtain the overall complexity of the working condition; In step S2, the Euclidean distance is a weighted Euclidean distance, and the cosine similarity is a weighted cosine similarity. The weights of both the weighted Euclidean distance and the weighted cosine similarity are positively correlated with the overall complexity of the working condition. Attached Figure Description

[0015] The invention will be further understood from the following description taken in conjunction with the accompanying drawings. The components in the drawings are not necessarily drawn to scale, but rather the emphasis is on illustrating the principles of the embodiments. In different views, the same reference numerals designate corresponding parts.

[0016] Figure 1 This is a schematic diagram of the overall three-dimensional structure of a laser cutting machine for sheet metal processing according to one embodiment of the present invention. Figure 1 ; Figure 2This is a schematic diagram of the overall three-dimensional structure of a laser cutting machine for sheet metal processing according to one embodiment of the present invention. Figure 2 ; Figure 3 This is a front view schematic diagram of the overall structure of a laser cutting machine for sheet metal processing according to an embodiment of the present invention; Figure 4 This is a side view of the overall structure of a laser cutting machine for sheet metal processing according to an embodiment of the present invention; Figure 5 This is a schematic diagram of the overall process of a control method for a laser cutting machine for sheet metal processing according to one embodiment of the present invention; Figure 6 This is a schematic diagram of the overall process of a control method for a laser cutting machine for sheet metal processing according to another embodiment of the present invention; Figure 7 This is a flowchart illustrating step S7 in one embodiment of the present invention.

[0017] Explanation of reference numerals in the attached diagram: 1. Industrial control center; 2. Laser cutting head; 3. Moving mechanism; 4. Frame. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to its embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and do not limit the scope of protection of the invention.

[0019] It should be noted that when an element is referred to as being "fixed to" another element, it can be directly attached to the other element or there may be an intervening element. When an element is referred to as being "connected to" another element, it can be directly connected to the other element or there may be an intervening element. The terms "vertical," "horizontal," "left," "right," and similar expressions used herein are for illustrative purposes only and do not represent the only possible implementation.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0021] In this invention, "first" and "second" do not represent a specific quantity or order, but are merely used to distinguish names.

[0022] like Figure 1 , Figure 2 , Figure 3 as well as Figure 4 As shown, an embodiment of the present invention provides a laser cutting machine for sheet metal processing, including a parameter acquisition component for collecting multi-dimensional real-time processing parameters and an industrial control center 1 for preprocessing the real-time processing parameters to obtain current working condition characteristics. The industrial control center stores historical sample data of sheet metal laser cutting. The laser cutting machine also includes a moving mechanism 3 and a laser cutting head 2. The laser cutting head is fixedly mounted on the moving mechanism, which is mounted on a frame 4 and used to drive the laser cutting head to move. The industrial control center is used to control the moving mechanism and the laser cutting head to perform cutting operations.

[0023] The moving mechanism is an XYZ three-axis moving module. The laser cutting head is generally installed at the end of the Z axis. The working center is connected to the moving mechanism and controls the moving mechanism to move along the XYZ axes, driving the laser cutting head to move to the target position and performing laser cutting on the target sheet metal part according to the predetermined laser cutting path.

[0024] The parameter acquisition component collects multi-dimensional real-time processing parameters, including but not limited to sheet metal workpiece parameters, equipment operating condition parameters, environmental parameters, and process parameters. Sheet metal workpiece parameters include material (carbon steel, stainless steel, aluminum alloy, etc.), sheet thickness, workpiece dimensions, cutting pattern complexity, and number of openings. Equipment operating condition parameters include cutting head wear, guide rail running accuracy, servo motor response speed, and continuous equipment operation time. Environmental parameters include temperature and humidity, and dust concentration in the sheet metal processing workshop. Process parameters include cutting speed, focal point position, auxiliary gas pressure, and nozzle type. The real-time processing parameters are normalized to construct a unified feature vector of the current operating condition, which serves as the basis for subsequent correlation determination.

[0025] The industrial control center has established a historical sample database for sheet metal laser cutting. This database stores historical sample data, each representing a mature and valid case of actual processing. Each historical sample includes historical working condition feature vectors, processing quality data, processing efficiency data, and equipment wear data. After each completed sheet metal laser cutting operation, the system automatically organizes the data, including the working conditions, path, and processing effect, into a new sample and adds it to the database after data cleaning. The database samples can be categorized into multiple levels of tags based on sheet metal material, thickness, workpiece type, and processing precision requirements.

[0026] The industrial control center includes a candidate path set acquisition module, a cutting path optimization module, and a laser cutting control module.

[0027] The alternative path set acquisition module is used to obtain the correlation between the current working condition and all historical sample working conditions based on the characteristics of the current working condition, filter out historical path samples with a correlation greater than the correlation threshold, sort the historical path samples in descending order according to the correlation, and extract the top N groups as the global optimal alternative path set.

[0028] Specifically, based on the characteristics of the current working condition, an algorithm combining weighted Euclidean distance and weighted cosine similarity can be used to calculate the comprehensive correlation between the current working condition and all historical working conditions in the sample database, thus selecting the globally optimal alternative path. When calculating the correlation, the characteristics of sheet metal cutting processes can be considered, and differentiated weights can be set for different parameters. Sheet thickness, material, cutting pattern complexity, and laser power stability are the core weight parameters, with appropriate increases in weight (e.g., set to 0.7). Environmental temperature and humidity, and equipment operating time are auxiliary weight parameters, with less weight (e.g., set to 0.3).

[0029] Assuming the correlation range is 0-1, the correlation threshold can be set to 0.85 or 0.9, etc. Historical path samples with a correlation greater than the correlation threshold are selected. The historical path samples are sorted in descending order according to the correlation, and the top 5 groups are extracted as the global optimal candidate path set. If there are fewer than 5 historical path samples with a correlation greater than the correlation threshold, the correlation threshold can be appropriately lowered until there are no fewer than 5 historical path samples with a correlation greater than the correlation threshold.

[0030] As a preferred technical solution, the working condition center also includes a feature dimension classification module. This module is used to classify the current working condition by feature dimension, dividing it into amplitude-sensitive features and trend-sensitive features. Amplitude-sensitive features are generally absolute numerical rigid parameters, and numerical deviations directly determine whether the cutting path is usable. These include sheet metal thickness, laser rated power, auxiliary gas pressure, cutting speed, nozzle orifice diameter, positioning accuracy deviation, etc. Trend-sensitive features are state fluctuation parameters, and the trend and stability of parameter changes determine the consistency of cutting quality. These include equipment power fluctuation coefficient, workshop temperature and humidity, guide rail running stability, sheet flatness deviation, continuous running time, etc.

[0031] Before calculating the corresponding correlation degree through the alternative path set acquisition module, the amplitude-sensitive features and trend-sensitive features are first subjected to [0, 1] interval extreme value normalization processing to eliminate the influence of dimensional differences.

[0032] The alternative path set acquisition module includes an Euclidean distance acquisition unit, a similarity acquisition unit, and a correlation acquisition unit.

[0033] The Euclidean distance acquisition unit is used to obtain the Euclidean distance between the current working condition and all historical sample working conditions for amplitude-sensitive features, and obtain the amplitude matching degree based on the Euclidean distance; the similarity acquisition unit is used to obtain the cosine similarity between the current working condition and all historical sample working conditions for trend-sensitive features, and obtain the trend matching degree based on the cosine similarity; the correlation acquisition unit is used to perform weighted fusion of amplitude matching degree and trend matching degree to obtain the correlation degree.

[0034] Specifically, the Euclidean distance is a weighted Euclidean distance, and the cosine similarity is a weighted cosine similarity. The current processing accuracy level and sheet metal processing difficulty can be obtained, and the normalized current processing accuracy level and sheet metal processing difficulty can be weighted and fused to obtain the overall working condition complexity. Based on the overall working condition complexity, the weighted Euclidean distance and weighted cosine similarity, which are positively correlated with the overall working condition complexity, are obtained.

[0035] When the normalized sheet metal processing difficulty approaches 0, it generally corresponds to easily cut thin carbon steel plates; when it approaches 1, it generally corresponds to thick plates, high-reflectivity materials, aluminum alloys, and workpieces with high flatness requirements. When the normalized processing accuracy level approaches 0, it generally corresponds to ordinary sheet metal blanking and low precision; when it approaches 1, it corresponds to precision sheet metal and high precision dimensional tolerance requirements.

[0036] For the weighting of the weighted Euclidean distance, amplitude-sensitive features can be divided into core amplitude features (such as plate thickness, laser power, and auxiliary gas pressure) and secondary amplitude features (such as cutting speed, nozzle diameter, and positioning deviation). Then, based on the overall complexity of the working conditions, different amplitude gain curves, such as exponential gain, are used for the core and secondary amplitude features. This ensures that as the overall complexity of the working conditions increases, the amplitude gain rate of the core amplitude features is significantly higher than that of the secondary amplitude features. Finally, by obtaining the intermediate weights that are positively correlated with the amplitude gain and normalizing them, the weighted weights corresponding to the core and secondary amplitude features are obtained. The gain coefficient can be used to amplify the basic amplitude weights to obtain the intermediate weight exp (gain coefficient × overall complexity of the working conditions), which is defined as the product of the basic amplitude weights and the amplitude gain. The basic amplitude weights of the amplitude-sensitive features can be set by technicians based on experience, and the sum of multiple basic amplitude weights is 1. This allows for higher processing precision requirements, greater sheet metal processing difficulty, and higher overall working condition complexity, resulting in a higher weighting of core amplitude features.

[0037] Accordingly, for the weighting of weighted cosine similarity, trend-sensitive features can be divided into core trend features (such as power fluctuation coefficient and sheet flatness) and secondary trend features (such as temperature and humidity, guide rail stability, and continuous running time). Then, based on the overall complexity of the operating conditions, different trend gain curves, such as exponential gain, are used for the core and secondary trend features. This ensures that as the overall complexity of the operating conditions increases, the trend gain rate of the core trend features is significantly higher than that of the secondary trend features. Finally, by obtaining the intermediate weights that are positively correlated with the trend gain and normalizing them, the weighted weights corresponding to the core and secondary trend features are obtained. The intermediate weights can be defined as the product of the trend base weight and the trend gain. The trend base weights of trend-sensitive features can be set by technicians based on experience, and the sum of multiple trend base weights is 1. In this way, since trend-type disturbances such as power fluctuations and sheet warping can easily cause batch defects, the weighting of the core trend features is automatically increased when the processing accuracy requirements are higher and the sheet metal processing difficulty is greater, i.e., when the overall complexity of the operating conditions is high. The system can prioritize matching historical samples with stable states.

[0038] The cosine similarity is normalized to obtain the trend matching degree. When weighted fusion of amplitude matching degree and trend matching degree to obtain the correlation, the corresponding weighted fusion weight can be obtained based on the overall complexity of the working conditions, and the fusion ratio of amplitude matching degree and trend matching degree can be dynamically adjusted to adapt to different processing scenarios. For example, the weighted fusion weight is expressed as: Overall complexity of working conditions σ = (Current sheet metal processing difficulty × Current processing accuracy level) / (Maximum sheet metal processing difficulty × Maximum processing accuracy level), where sheet metal processing difficulty and processing accuracy level are both normalized, with a value range of [0,1]. The larger the value, the more complex the overall working conditions and the more stringent the current processing task. Of course, the overall complexity of working conditions can also be defined as the weighted sum of the normalized current sheet metal processing difficulty and the current processing accuracy level.

[0039] Ultimately, the correlation degree is calculated as (1-σ) × amplitude matching degree + σ × trend matching degree. For simple, routine operating conditions, where the overall complexity is relatively low, the correlation degree focuses on weighted Euclidean distance matching to ensure accurate matching of basic cutting parameters. This approach offers advantages such as fast calculation speed and high stability. For complex, high-precision operating conditions, the correlation degree focuses on weighted cosine similarity matching to prioritize matching equipment operational stability and environmental fluctuation trends, thus avoiding cutting defects caused by thermal deformation and vibration. In this way, by splitting laser cutting operating parameters into amplitude and trend features based on their physical attributes and matching them with weighted Euclidean distance and weighted cosine similarity respectively, the industry-wide problem of mixed dimensions and misaligned matching logic in general algorithms can be solved. This enables intelligent adaptation that prioritizes accuracy in simple operating conditions and stability in complex operating conditions, thus meeting the diverse processing needs of sheet metal laser cutting.

[0040] The cutting path optimization module is used to obtain the actual cutting paths of the same type of sheet metal workpieces in the most recent times and use them as real-time reference samples. It performs a quantitative comparison of the advantages and disadvantages of the real-time reference samples and the global optimal alternative path set, and uses the optimal path as the final laser cutting path.

[0041] As a preferred technical solution, the cutting path optimization module includes a path quality acquisition unit and a path optimization comparison unit.

[0042] The path quality acquisition unit is used to select at least one quality evaluation index and perform path quality quantification scoring on the real-time reference sample and the global optimal candidate path set based on the quality evaluation index to obtain the path quality score; the path selection comparison unit is used to perform selection comparison on the real-time reference sample and the global optimal candidate path set according to the path quality score, and select the path with the highest path quality score as the optimal path and as the final laser cutting path.

[0043] Specifically, the actual cutting paths of the two most recent cuts of the same type of sheet metal workpiece can be used as real-time reference samples. Then, the advantages and disadvantages of the real-time reference samples and the globally optimal candidate paths are quantitatively compared from the perspectives of efficiency, quality, and equipment wear. Processing efficiency indicators include normalized total path idle travel length, effective cutting time, and number of start-stop actions; processing quality indicators include normalized path thermal deformation control degree, cutting corner smoothness, and repeatability accuracy; and equipment wear indicators include normalized cutting head motion trajectory stability, servo motor load fluctuation value, and laser output stability. For each path in the real-time reference samples and the globally optimal candidate paths, weighted calculations are performed on the efficiency, quality, and equipment wear indicators to obtain a comprehensive index score. The path with the highest comprehensive index score is selected as the optimal path. If two or more paths have the same comprehensive index score, one path can be randomly selected as the optimal path, or the path with lower equipment wear can be selected as the optimal path.

[0044] The laser cutting control module controls the laser cutting head to perform cutting operations based on the optimal laser cutting path. During the laser cutting process, the module monitors path deviations in real time. If sudden fluctuations occur, minor path adjustments can be made. After sheet metal cutting is completed, the system automatically collects data such as path parameters, processing quality, efficiency, and equipment wear. After data cleaning and feature processing, the data is stored and updated in the historical sample database for continuous iteration and optimization.

[0045] In summary, the laser cutting machine for sheet metal processing obtains the global optimal candidate path set and performs a quantitative comparison of the advantages and disadvantages of the real-time reference sample and the global optimal candidate path set, and uses the optimal path as the final laser cutting path. It can reuse mature historical high-quality paths, and does not need to perform a full algorithm iteration every time when planning the cutting path. It can quickly select the optimal path based on historical sample data, thereby shortening the path planning time and improving the efficiency of sheet metal cutting processing.

[0046] During long-term operation of sheet metal laser cutting machines, mechanical transmission clearance, cutting head assembly clearance, nozzle wear, lens wear, and guide rail micro-movement deviation are the core inherent error sources that generate processing errors. These errors are characterized by slow time-varying nature, cumulative nature, inability to be completely eliminated, and coupling with operating conditions. Most existing laser cutting error compensation technologies rely solely on the deviation between a single preset planning path and the actual processing trajectory for correction. However, a single path has occasional deviations, and these technologies cannot distinguish between instantaneous operating condition fluctuation errors and inherent systematic errors caused by equipment wear / clearance. Consequently, they suffer from low compensation accuracy and a high probability of miscompensation.

[0047] To improve the accuracy of error compensation, in one embodiment, the central control center also includes a reference trajectory acquisition module, a target trajectory acquisition module, a cutting error acquisition module, and a trajectory error compensation module.

[0048] The baseline trajectory acquisition module is used to obtain a baseline path set based on the union of real-time reference samples and the global optimal candidate path set, and to perform weighted fusion on all paths in the baseline path set to obtain an ideal baseline trajectory; the target trajectory acquisition module is used to obtain a target cutting trajectory that matches the current working condition based on real-time processing parameters, wherein the target cutting trajectory includes trajectory point coordinates, cutting speed, laser output power and real-time kerf forming parameters.

[0049] It can be understood that the real-time reference sample equipment represents the actual optimal cutting paths of the two most recent cutting operations under the same working conditions and with the same material, which includes the actual working condition characteristics of the equipment's current clearance and wear; while the global optimal alternative path set represents the ideal cutting trajectory under the current working conditions with no error and standard working conditions. The ideal reference trajectory integrates the common advantages of all reference paths and can serve as a precise theoretical benchmark adapted to the current working conditions.

[0050] The cutting error acquisition module combines the ideal reference trajectory and the target cutting trajectory to obtain the mechanical clearance position error and the consumable wear and forming error. The trajectory error compensation module compensates and corrects the current real-time cutting path based on the mechanical clearance position error and the consumable wear and forming error. The target cutting trajectory can be either the real-time cutting trajectory or a historical cutting trajectory. For the real-time cutting trajectory, during the real-time laser cutting process, the current real-time cutting path trajectory points can be collected at high frequency using a servo encoder, laser position sensor, and real-time vision sampling module to construct the real-time cutting trajectory.

[0051] Mechanical backlash position errors are caused by transmission guide rail clearance, cutting head assembly clearance, and servo return clearance. These manifest as trajectory offset, angular position deviation, and positioning repeatability errors, and belong to the trajectory position system error category. Consumable wear and forming errors are caused by nozzle orifice wear, lens burning loss, and optical path offset. These manifest as kerf width offset, sheet metal edge burning, actual cutting size being too large / too small, and insufficient penetration, and belong to the process forming system error category.

[0052] The cutting error acquisition module, in conjunction with the ideal reference trajectory and the target cutting trajectory, uses a triple criterion of characteristic attribute differences, deviation distribution patterns, and temporal stability to obtain the mechanical clearance position error and the consumable wear forming error. As mentioned above, the mechanical clearance position error often exhibits characteristics of global unidirectionality, trajectory repeatability, and no parameter correlation, while the consumable wear forming error is independent of the equipment transmission clearance and exhibits characteristics of local forming differences, process parameter correlation, and temporal accumulation. Therefore, based on orthogonality, the two types of errors can be decoupled and free from cross-interference.

[0053] Specifically, the process begins by collecting the coordinate deviation of the entire trajectory. Simultaneously, the real-time X and Y coordinates of all straight segments, arc segments, and turning points in the ideal reference trajectory and the target cutting trajectory are extracted. The position deviation vector is calculated point-by-point to generate the entire trajectory deviation matrix. Next, based on this matrix, common deviations across multiple paths are filtered to identify fixed deviations that are in the same direction, of the same magnitude, and existing in continuous intervals, while random jitter errors are eliminated. For example, if three or more consecutive paths exhibit coordinate offsets of the same direction and magnitude, it is determined to be a fixed gap deviation. For fixed deviations, the idle distance deviation at trajectory reversal nodes (positions where positive and negative speeds switch) is extracted separately. The abrupt changes in trajectory displacement before and after the reversal are statistically analyzed to distinguish between steady-state offset of straight segments, steady-state deviation of turning angles, and idle distance deviation of reversals. Finally, the steady-state offset of straight segments, steady-state deviation of turning angles, and idle distance deviation of reversals are subjected to mean filtering and purification processing to obtain the mechanical gap position error.

[0054] For consumable wear forming error, the core cutting forming parameters of the ideal reference trajectory and the target cutting trajectory are first compared, including the standard kerf width and the actual kerf width, the preset laser focus position and the actual focus position, the standard cut flatness and the actual cross-sectional roughness, and the theoretical cutting penetration depth and the actual penetration depth, to obtain the global forming deviation. Then, based on the patterns of multiple different characteristics of consumable wear, the effective systematic error of wear is obtained by multi-feature coupling judgment and combined with time-series cumulative error and multi-path commonality verification. The effective systematic error of wear is then subjected to mean filtering to obtain the consumable wear forming error. For example, if the forming deviation shows a continuous unidirectional cumulative change (the kerf continues to widen and the penetration continues to decrease) when retrieving the forming data of the last 5 processing conditions, it is determined to be an effective systematic error caused by tool wear. Among them, nozzle wear corresponds to characteristics such as uniform kerf widening, auxiliary airflow disturbance, and increased slag on the cross-section, while lens / optical path wear corresponds to characteristics such as laser power density attenuation, focus shift, local incomplete cutting, and edge burning.

[0055] To avoid miscoupling and miscompensation, a dual verification mechanism is implemented. First, spatial position deviations are independent of changes in process forming parameters, and forming deviations are independent of trajectory coordinate offsets, achieving orthogonal isolation and non-interference between the two types of errors. Second, only common, continuous deviations along multiple paths are considered valid systematic errors, while single-point instantaneous deviations and irregular fluctuations are filtered out. This allows for accurate and independent extraction of mechanical clearance position errors and consumable wear forming errors even under complex operating conditions.

[0056] In this way, by constructing a benchmark path set using real-time reference samples and the global optimal alternative path set, and by extracting the inherent error of the system through the common deviation features of multiple paths, random errors in single processing can be filtered out, and mechanical clearance errors and material wear and forming errors can be accurately distinguished.

[0057] To compensate for mechanical backlash errors, a dynamic feedforward correction method for trajectory coordinates can be used. For example, first calculate the average position deviation vector of multiple paths, and statistically analyze the fixed offsets of the X / Y axes and corner positions. Then, before the real-time path interpolation calculation, the planned path coordinates are pre-corrected to obtain: corrected trajectory coordinates = original optimal path coordinates - average position deviation vector of multiple paths; or, based on the historical reversal deviation patterns of multiple paths, compensation displacements are output in advance at the moment of guide rail and servo reversal to offset the return backlash.

[0058] For nozzle wear compensation in consumable wear forming error, the auxiliary gas pressure and cutting speed can be appropriately fine-tuned according to the cutting deviation to compensate for the airflow turbulence and cutting deviation caused by the increase in nozzle orifice diameter; or the laser focus position and laser output power density can be dynamically fine-tuned according to the focusing deviation to compensate for the optical path loss and the cutting incompleteness and edge burning caused by focus deviation, thereby achieving lens / optical path wear compensation.

[0059] An embodiment of the present invention also provides a control method for a laser cutting machine for sheet metal processing, applied to the aforementioned laser cutting machine for sheet metal processing, such as... Figure 5 As shown, it includes the following steps: S1 collects multi-dimensional real-time processing parameters and preprocesses the real-time processing parameters to obtain the current working condition characteristics.

[0060] The parameter acquisition component collects multi-dimensional real-time processing parameters, including but not limited to sheet metal workpiece parameters, equipment operating condition parameters, environmental parameters, and process parameters.

[0061] S2. Based on the characteristics of the current working condition, obtain the correlation between the current working condition and all historical sample working conditions, filter out historical path samples with a correlation greater than the correlation threshold, sort the historical path samples in descending order according to the correlation, and extract the top N groups as the global optimal candidate path set.

[0062] As a preferred technical solution, step S2 specifically includes the following steps: S21 classifies the current working condition by feature dimension into amplitude-sensitive features and trend-sensitive features, and normalizes the amplitude-sensitive features and trend-sensitive features.

[0063] S22, for amplitude-sensitive features, obtain the Euclidean distance between the current working condition and all historical sample working conditions, and obtain the amplitude matching degree based on the Euclidean distance. At the same time, for trend-sensitive features, obtain the cosine similarity between the current working condition and all historical sample working conditions, and obtain the trend matching degree based on the cosine similarity.

[0064] S23, weighted fusion of amplitude matching degree and trend matching degree to obtain correlation degree.

[0065] Specifically, the current processing accuracy level and sheet metal processing difficulty are obtained, and the normalized current processing accuracy level and sheet metal processing difficulty are weighted and fused to obtain the overall working condition complexity. Among them, the Euclidean distance in step S2 is the weighted Euclidean distance, and the cosine similarity is the weighted cosine similarity. The weights of the weighted Euclidean distance and the weighted cosine similarity are positively correlated with the overall working condition complexity.

[0066] S3: Obtain the actual cutting paths of the most recent several sheet metal workpieces of the same type and use them as real-time reference samples. Perform a quantitative comparison of the advantages and disadvantages of the real-time reference samples and the global optimal alternative path set, and use the optimal path as the final laser cutting path.

[0067] As a preferred technical solution, step S3 specifically includes the following steps: S31. Select at least one quality evaluation indicator and perform path quality quantification scoring on the real-time reference sample and the global optimal alternative path set based on the quality evaluation indicator to obtain the path quality score.

[0068] S32, based on the path quality score, compares the real-time reference samples and the global optimal candidate path set, and selects the path with the highest path quality score as the optimal path and the final laser cutting path.

[0069] Specifically, the actual cutting paths of the two most recent cuts of the same type of sheet metal workpiece can be used as real-time reference samples. Then, the advantages and disadvantages of the real-time reference samples and the globally optimal candidate paths are quantitatively compared from the perspectives of efficiency, quality, and equipment wear. For each path in the real-time reference samples and the globally optimal candidate paths, the efficiency, quality, and equipment wear dimensions are weighted and calculated to obtain a comprehensive index score. The path with the highest comprehensive index score is selected as the optimal path. If there are two or more paths with the same comprehensive index score, one path can be randomly selected as the optimal path, or the path with lower equipment wear can be selected as the optimal path.

[0070] S4 controls the laser cutting head to perform the cutting operation according to the optimal laser cutting path.

[0071] The laser cutting control module is used to control the laser cutting head to perform cutting operations based on the optimal laser cutting path. During the laser cutting operation driven by the laser cutting head based on the optimal laser cutting path, the module monitors path deviations in real time throughout the process. If sudden fluctuations in operating conditions occur, minor adjustments to the path can be made as needed.

[0072] In summary, the control method for the laser cutting machine used in sheet metal processing obtains a globally optimal candidate path set and performs a quantitative comparison of the advantages and disadvantages of the real-time reference sample and the globally optimal candidate path set. The optimal path is then used as the final laser cutting path. It can reuse mature historical high-quality paths. When planning the cutting path, it does not need to perform a full algorithm iteration every time. It can quickly select the optimal path based on historical sample data, thereby shortening the path planning time and improving the efficiency of sheet metal cutting processing.

[0073] In one embodiment, such as Figure 6 As shown, the control method for a laser cutting machine for sheet metal processing further includes the following steps: S5 obtains a baseline path set based on the union of real-time reference samples and the global optimal alternative path set, and performs weighted fusion on all paths in the baseline path set to obtain an ideal baseline trajectory.

[0074] Due to subtle differences in the number of trajectory interpolation points and node positions among different optimal paths, trajectory standardization and alignment are performed uniformly before fusion to eliminate interference from structural differences. Specifically, trajectory timing alignment is first performed, unifying the timing of cutting start points, end points, turning points, and process nodes for all paths to ensure a one-to-one correspondence between trajectory points on each path. Next, equal-interval resampling is performed, resampling each path at equal steps to generate a standard trajectory point sequence with consistent quantity and corresponding positions. Each path outputs N trajectory points uniformly, and the dimensional parameters are unified. Path fusion weights are obtained based on the overall working condition complexity, and the real-time reference samples and the globally optimal candidate paths are fused based on these weights. The path fusion weight of the real-time reference samples is positively correlated with the overall working condition complexity, while the path fusion weight of the globally optimal candidate paths is negatively correlated with the overall working condition complexity, and the sum of their fusion weights is 1. Based on the path fusion weights of real-time reference samples, a normalized allocation is performed by combining the correlation between real-time reference samples and the current working conditions to obtain the fusion weight of a single actual cutting path; similarly, based on the path fusion weights of the globally optimal candidate paths, a normalized allocation is performed by combining the correlation between the globally optimal candidate paths and the current working conditions to obtain the fusion weight of a single historical path.

[0075] Based on the obtained fusion weights of a single actual cutting path and a single historical path, a node-by-node coordinate weighted fusion method can be used to perform weighted fusion on all paths in the benchmark path set to obtain an ideal benchmark trajectory.

[0076] S6: Obtain the target cutting trajectory that matches the current working condition based on real-time processing parameters. The target cutting trajectory includes trajectory point coordinates, cutting speed, laser output power, and real-time parameters for kerf formation.

[0077] S7 combines the ideal reference trajectory and the target cutting trajectory to obtain the mechanical clearance position error and the consumable wear forming error.

[0078] As a preferred technical solution, such as Figure 7 As shown, step S7 specifically includes the following steps: S71 extracts the real-time XY coordinates of all straight segments, arc segments, and corner inflection points in the ideal reference trajectory and the target cutting trajectory, calculates the position deviation point by point, generates a global trajectory deviation matrix, removes random noise from the global trajectory deviation matrix, filters out fixed deviations that exist in multiple paths in the same direction, of the same magnitude, and in continuous intervals, and performs mean filtering on the fixed deviations to obtain the mechanical clearance position error.

[0079] Specifically, the global trajectory deviation matrix has a dimension of N×2, where N is the total number of trajectory nodes after resampling and alignment. Each row contains the coordinate position deviation between the ideal reference trajectory and the target cutting track. After filtering the global trajectory deviation matrix for random noise, at least three paths with consistent deviation direction and magnitude within the same trajectory interval are selected as fixed deviations. These fixed deviations are then categorized into global coordinate offsets in the uniform velocity segment and abrupt changes in trajectory displacement before and after the changeover. Mean denoising filtering is applied to both the global coordinate offsets in the uniform velocity segment and the abrupt changes in trajectory displacement before and after the changeover to obtain the mechanical clearance position error.

[0080] S72 compares the cutting forming parameters of the ideal reference trajectory and the target cutting trajectory point by point to obtain the global forming deviation. Based on the multiple features corresponding to nozzle wear and the multiple features corresponding to lens optical path wear, the global forming deviation is coupled and judged. Combined with the time-series cumulative error and multi-path commonality verification, the effective system error of wear is obtained and the effective system error of wear is processed by mean filtering to obtain the consumable wear forming error.

[0081] Generally, laser cutting consumable wear exhibits inherent characteristics of slow time-varying, irreversible, and simultaneous, unidirectional shifts in multiple features, distinguishing it from external random disturbances such as uneven sheet material, instantaneous fluctuations in workshop airflow, and sudden changes in temperature and humidity. For example, nozzle wear accumulates over processing time, with continuous ablation and expansion of the nozzle inner hole and air outlet, constituting permanent physical aperture loss. During this wear process, the kerf width, airflow stability, and slag adhesion on the cross-section often deteriorate simultaneously in a unidirectional manner, without spontaneous recovery. Lens / optical path wear, on the other hand, is caused by the protective lens and focusing lens being burned by metal vapor sputtering and by laser thermal damage, resulting in a continuous decrease in light transmittance and a fixed shift in the optical focus. This wear process often leads to a simultaneous aggravation of four defects: output power density reduction, focus shift, incomplete cutting in certain areas, and edge burning. Therefore, based on the inherent degradation patterns of these two types of wear, multi-feature coupling judgment can be performed. A single abnormal feature is not considered wear; only when multiple wear features of the same category show simultaneous, unidirectional deviations can it be determined as an effective forming deviation caused by consumable wear, thus filtering out external working condition interference.

[0082] The characteristics corresponding to nozzle wear include, but are not limited to, uniform widening of the kerf, increased auxiliary airflow disturbance, and increased slag adhesion on the cross-section. If, within the same processing sampling period, at least two or more nozzle wear characteristics simultaneously exceed their corresponding deviation thresholds and the deviations deteriorate in the same direction, the multi-feature coupling judgment of nozzle wear is considered successful.

[0083] The characteristics corresponding to lens optical path wear include, but are not limited to, laser power density attenuation, continuous focus shift, incomplete cutting of the substrate, and widening of the cut edge due to burning. Only when at least two or more lens optical path wear characteristics simultaneously exceed their corresponding deviation thresholds and the deviations continue to deteriorate within a single sampling period is the multi-feature coupling judgment of lens optical path wear considered successful.

[0084] Cutting parameters include, but are not limited to, kerf width, laser focal point position, material penetration depth, cut surface roughness, and actual effective laser power density. Global forming deviation is obtained by subtracting the cutting parameters of the ideal reference trajectory from the target cutting trajectory dimension by dimension. This includes random deviations such as localized material inhomogeneity in the material, instantaneous airflow disturbances in the workshop, and noise from a single acquisition; short-term operating condition disturbances such as short-term temperature and humidity fluctuations and localized differences in material flatness; and slow time-varying system deviations caused by tool wear.

[0085] The time-series cumulative error is defined as follows: if the deviation of the sampled frames processed under the same working condition for several consecutive times, such as 5 times, shows a unidirectional cumulative change, such as the kerf continuously widening, the effective power continuously decreasing, and the penetration depth continuously decreasing, the time-series cumulative error is determined. The multi-path commonality verification is defined as follows: if the cutting paths of at least 3 reference path sets or historical samples under the same working condition all show the same direction and the same magnitude of forming deviation, it is determined to be a multi-path commonality.

[0086] In the wear and forming error of consumables, each component, such as the deviation of the kerf width, the deviation of the laser focus position, the deviation of the plate penetration depth, and the deviation of the effective laser power density, is the time-series average of the effective systematic error of wear in the corresponding dimension. It represents the forming defects caused by nozzle wear, lens burning, and optical path attenuation.

[0087] S8 compensates and corrects the current real-time cutting path based on the mechanical clearance position error and the wear and forming error of the consumables.

[0088] Based on the mechanical backlash position error, the feedforward coordinate compensation amount is output to the servo interpolator to compensate and correct the current real-time cutting path. Specifically, since sudden changes in backlash reversal can directly cause the contour dimensions to exceed tolerances, instantaneous compensation is adopted for each rotation angle and each reversal. After obtaining the reversal deviation, the compensation vector can be output immediately and the interpolation cycle can be refreshed synchronously; for the global offset component of the uniform speed segment, periodic batch updates are adopted, for example, refreshing once every 100 interpolation frames.

[0089] Based on the wear and forming error of consumables, process parameter adjustment commands are output. Specifically, since nozzle and lens wear is a slow, cumulative change over tens of minutes or even longer, there is no risk of instantaneous abrupt changes, so there is no need to update it on an interpolation cycle basis. After obtaining the wear and forming error of consumables, process compensation parameters such as power, focus, air pressure, and cutting speed can be updated every certain period of time, such as 1 to 3 seconds.

[0090] When the multi-feature coupling judgment of nozzle wear, the time-series cumulative error, and the multi-path commonality verification all pass, the current forming deviation is marked as an effective forming deviation of nozzle wear, and is used as a subset of consumable wear forming error. The dimensional deviations of cutting, airflow, and slag are extracted for the calculation of nozzle compensation components. When the multi-feature coupling judgment of nozzle wear, the time-series cumulative error, and the multi-path commonality verification all pass, the current forming deviation is marked as an effective forming deviation of lens / optical path wear, and is used as a subset of consumable wear forming error. Power, focal point, penetration depth, and burning edge deviation are extracted as lens / optical path compensation components, and process parameter adjustment commands are output.

[0091] The target cutting trajectory can be a real-time cutting trajectory or a historical cutting trajectory. Specifically, the real-time cutting trajectory is the dynamic trajectory and synchronous forming parameters that are synchronously collected by the servo encoder and vision sensor on an interpolation cycle while the workpiece is being cut. It is synchronized with the current material wear, real-time gap status, and sheet metal status. The historical cutting trajectory is the complete historical trajectory and forming dataset of sheet metal workpieces of the same specification and process, which is offline historical data.

[0092] To reduce system hardware costs and real-time computing power consumption, and to address latency issues, the target cutting trajectory can be selected as the most recent historical cutting trajectory. In this scenario, it is suitable for continuous batch production and equipment that experiences no external disturbances for a short period, such as no nozzle / lens replacement, no changes in lead screw lubrication, stable workshop temperature and humidity, and stable auxiliary air pressure, and where the processing time for a single piece is short (e.g., within 5 minutes), and there are no significant incremental changes in mechanical clearance and tool wear.

[0093] For long-term continuous batch sheet metal processing, such as single sheet metal laser cutting processing time exceeding a certain time (e.g., 5 minutes), frequent switching of workpiece type and sheet metal specifications, sheet metal batch fluctuations, unstable workshop environment, etc., the real-time cutting trajectory is used as the target cutting trajectory to filter random noise disturbances in the single sheet metal workpiece cutting process and adapt to the time-series cumulative evolution characteristics of mechanical clearance position error and consumable wear and forming error.

[0094] In summary, by distinguishing between mechanical clearance position error and consumable wear and forming error, this invention can accurately identify and layer-by-layer differentiate mechanical clearance position error and tool wear and forming error, thereby eliminating trajectory offset errors caused by transmission clearance and assembly clearance during long-term operation of the laser cutting machine and improving the accuracy of laser cutting of sheet metal workpieces.

[0095] The technical features of the embodiments described can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0096] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this invention patent should be determined by the appended claims.

Claims

1. A laser cutting machine for sheet metal processing, comprising a parameter acquisition component for acquiring multi-dimensional real-time processing parameters and an industrial control center for preprocessing the real-time processing parameters to obtain current working condition characteristics, wherein the industrial control center stores historical sample data of sheet metal laser cutting, characterized in that, Industrial control centers include: The alternative path set acquisition module is used to obtain the correlation between the current working condition and all historical sample working conditions based on the characteristics of the current working condition, filter out historical path samples with a correlation greater than the correlation threshold, sort the historical path samples in descending order according to the correlation, and extract the top N groups as the global optimal alternative path set. The cutting path optimization module is used to obtain the actual cutting paths of the same type of sheet metal workpieces in the most recent times and use them as real-time reference samples. It performs a quantitative comparison of the advantages and disadvantages of the real-time reference samples and the global optimal alternative path set, and uses the optimal path as the final laser cutting path. The laser cutting control module is used to control the laser cutting head to perform cutting operations according to the optimal laser cutting path.

2. The laser cutting machine for sheet metal processing as described in claim 1, characterized in that, The operating condition center also includes a feature dimension classification module, which classifies the current operating condition by feature dimension, dividing it into amplitude-sensitive features and trend-sensitive features. The alternative path set acquisition module includes: The Euclidean distance acquisition unit is used to obtain the Euclidean distance between the current working condition and all historical sample working conditions for amplitude-sensitive features, and to obtain the amplitude matching degree based on the Euclidean distance; The similarity acquisition unit is used to obtain the cosine similarity between the current working condition and all historical sample working conditions for trend-sensitive features, and to obtain the trend matching degree based on the cosine similarity. The correlation degree acquisition unit is used to perform weighted fusion of amplitude matching degree and trend matching degree to obtain the correlation degree.

3. The laser cutting machine for sheet metal processing as described in claim 2, characterized in that, The cutting path optimization module includes: The path quality acquisition unit is used to select at least one quality evaluation index and perform path quality quantification scoring on the real-time reference sample and the global optimal candidate path set based on the quality evaluation index to obtain the path quality score. The path selection and comparison unit is used to compare the real-time reference samples and the global optimal candidate path set based on the path quality score, and select the path with the highest path quality score as the optimal path and as the final laser cutting path.

4. A laser cutting machine for sheet metal processing as described in claim 3, characterized in that, The central control center also includes: The baseline trajectory acquisition module is used to obtain a baseline path set based on the union of real-time reference samples and the global optimal candidate path set, and to perform weighted fusion on all paths in the baseline path set to obtain an ideal baseline trajectory. The target trajectory acquisition module is used to acquire a target cutting trajectory that matches the current working condition based on real-time processing parameters. The target cutting trajectory includes trajectory point coordinates, cutting speed, laser output power, and real-time kerf forming parameters. The cutting error acquisition module is used to combine the ideal reference trajectory and the target cutting trajectory to obtain the mechanical clearance position error and the consumable wear and forming error. The trajectory error compensation module is used to compensate and correct the current real-time cutting path based on mechanical clearance position error and consumable wear and forming error. The target cutting trajectory can be either a real-time cutting trajectory or a historical cutting trajectory.

5. A control method for a laser cutting machine for sheet metal processing, applied to the laser cutting machine for sheet metal processing as described in any one of claims 1-4, characterized in that, Includes the following steps: S1, collect multi-dimensional real-time processing parameters and preprocess the real-time processing parameters to obtain the current working condition characteristics; S2, based on the characteristics of the current working condition, obtain the correlation between the current working condition and all historical sample working conditions, filter out historical path samples with a correlation greater than the correlation threshold, sort the historical path samples in descending order according to the correlation, and extract the top N groups as the global optimal candidate path set. S3: Obtain the actual cutting paths of the same type of sheet metal workpieces in the most recent times and use them as real-time reference samples. Perform a quantitative comparison of the advantages and disadvantages of the real-time reference samples and the global optimal alternative path set, and use the optimal path as the final laser cutting path. S4 controls the laser cutting head to perform the cutting operation according to the optimal laser cutting path.

6. The control method for a laser cutting machine for sheet metal processing as described in claim 5, characterized in that, Step S2 specifically includes the following steps: S21, classify the current working condition by feature dimension, dividing it into amplitude-sensitive features and trend-sensitive features, and normalize the amplitude-sensitive features and trend-sensitive features. S22, for amplitude-sensitive features, obtain the Euclidean distance between the current working condition and all historical sample working conditions, and obtain the amplitude matching degree based on the Euclidean distance. At the same time, for trend-sensitive features, obtain the cosine similarity between the current working condition and all historical sample working conditions, and obtain the trend matching degree based on the cosine similarity. S23, weighted fusion of amplitude matching degree and trend matching degree to obtain correlation degree.

7. The control method for a laser cutting machine for sheet metal processing as described in claim 6, characterized in that, Step S3 specifically includes the following steps: S31, Select at least one quality evaluation index and perform path quality quantification scoring on the real-time reference sample and the global optimal alternative path set based on the quality evaluation index to obtain the path quality score; S32, based on the path quality score, compares the real-time reference samples and the global optimal candidate path set, and selects the path with the highest path quality score as the optimal path and the final laser cutting path.

8. The control method for a laser cutting machine for sheet metal processing as described in claim 7, characterized in that, It also includes the following steps: S5: Obtain the baseline path set based on the union of real-time reference samples and the global optimal alternative path set, and perform weighted fusion on all paths in the baseline path set to obtain the ideal baseline trajectory. S6. Obtain the target cutting trajectory that matches the current working condition based on the real-time processing parameters. The target cutting trajectory includes the coordinates of the trajectory points, the cutting speed, the laser output power, and the real-time parameters of the kerf formation. S7, by combining the ideal reference trajectory and the target cutting trajectory, obtains the mechanical clearance position error and the consumable wear and forming error; S8, compensate and correct the current real-time cutting path based on the mechanical clearance position error and the wear and forming error of the consumables; The target cutting trajectory can be either a real-time cutting trajectory or a historical cutting trajectory.

9. The control method for a laser cutting machine for sheet metal processing as described in claim 8, characterized in that, Step S7 specifically includes the following steps: S71, extract the real-time XY coordinates of all straight lines, arcs, and corners in the ideal reference trajectory and the target cutting trajectory, calculate the position deviation point by point, generate the global trajectory deviation matrix, remove the random noise of the global trajectory deviation matrix, select the fixed deviations that exist in the same direction, same magnitude, and continuous interval of multiple paths, and perform mean filtering on the fixed deviations to obtain the mechanical clearance position error. S72 compares the cutting forming parameters of the ideal reference trajectory and the target cutting trajectory point by point to obtain the global forming deviation. Based on the multiple features corresponding to nozzle wear and the multiple features corresponding to lens optical path wear, the global forming deviation is coupled and judged. Combined with the time-series cumulative error and multi-path commonality verification, the effective system error of wear is obtained and the effective system error of wear is processed by mean filtering to obtain the consumable wear forming error.

10. The control method for a laser cutting machine for sheet metal processing as described in claim 8, characterized in that, It also includes the following steps: obtaining the current machining accuracy level and sheet metal machining difficulty, and performing a weighted fusion of the normalized current machining accuracy level and sheet metal machining difficulty to obtain the overall complexity of the working condition; In step S2, the Euclidean distance is a weighted Euclidean distance, and the cosine similarity is a weighted cosine similarity. The weights of both the weighted Euclidean distance and the weighted cosine similarity are positively correlated with the overall complexity of the working condition.

Images