Adaptive focal length control method and system for laser cutting machine

Abstract

The application discloses a kind of adaptive focal length control method and system of laser cutting machine, specifically relates to laser processing technical field, for solving the problem of insufficient laser focal point control precision caused by the separation of detection point and processing point and the geometric mutation of curved surface in prior art;It is by continuously acquiring instantaneous distance information in the movement process of cutting head, and according to this, the curvature variation characteristics of workpiece surface are calculated, and then the inertia load and dynamic load risk caused by geometric appearance mutation are comprehensively evaluated, and the path risk section and non-path risk section are intelligently identified.According to different risk sections, adaptive switching conservative prediction mode and conventional prediction mode, accurately predict the normal vector of workpiece surface at laser processing point, then according to the normal vector, the cutting head posture adjustment amount is calculated, finally through multi-axis synchronous interpolation control, realize the real-time, smooth posture adjustment of cutting head in the movement process, ensure that laser beam always maintains perpendicular relationship with workpiece surface.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of laser processing, more particularly, the present application relates to a self-adaptive focal length control method and system of a laser cutting machine. BACKGROUND

[0002] In the field of laser cutting processing, in order to cope with uneven workpiece surface or three-dimensional curved surface cutting, self-adaptive focal length control method is the key technology to improve processing quality. The existing technology usually integrates displacement sensors on the cutting head, detects the distance between the nozzle and the workpiece surface in real time, and drives the control shaft to dynamically adjust the height of the cutting head according to the measurement value, so as to stabilize the laser focal point on the workpiece surface. This method has good control effect for general two-dimensional cutting or gently curved surface with variable curvature, and constitutes the basis of current self-adaptive focal length control.

[0003] However, when the above method is applied to curved surface workpieces with complex geometric features, its control precision faces inherent limitations. Since the sensor needs to be installed in front to measure ahead, the detection point and the actual laser processing point are inevitably separated in space. When cutting complex curved surfaces, the curvature and normal direction of the workpiece surface may change significantly, resulting in differences between the surface geometric information at the detection point and the actual situation at the processing point. The existing technology only adjusts the height according to the distance information of the front detection point, which is difficult to accurately reflect the focal point position required at the processing point, causing the laser focal point to deviate from the ideal state and affecting the final cutting quality. SUMMARY

[0004] In order to overcome the above-mentioned defects of the prior art, the present application provides a self-adaptive focal length control method and system of a laser cutting machine to solve the problems raised in the background art.

[0005] To achieve the above-mentioned purpose, the present application provides the following technical solutions:

[0006] A self-adaptive focal length control method of a laser cutting machine, comprising the following steps:

[0007] S1, during the movement of the cutting head along the preset trajectory, continuously acquiring the instantaneous distance information between the detection point of the cutting head and the workpiece surface through the displacement sensor integrated on the cutting head;

[0008] S2, calculating the workpiece surface curvature variation feature of the position where the detection point is located on the preset trajectory based on the instantaneous distance information;

[0009] S3, based on the workpiece surface curvature variation feature, comprehensively analyzing the inertial load risk generated by geometric morphology mutation to the cutting head motion system and the dynamic load risk generated to the laser energy interaction process, identifying the current preset trajectory segment as a path risk segment or a non-path risk segment;

[0010] S4, starting a conservative prediction mode for the path risk section and a regular prediction mode for the non-path risk section to predict the workpiece surface normal vector at the laser processing point;

[0011] S5, calculating the attitude adjustment amount required for the cutting head to reach the laser processing point according to the workpiece surface normal vector at the laser processing point;

[0012] S6, controlling the cutting head to perform attitude adjustment of the cutting head according to the attitude adjustment amount during the movement to the laser processing point.

[0013] Further, during the movement of the cutting head along the preset trajectory, the instantaneous distance information between the detection point and the workpiece surface is continuously obtained by the displacement sensor integrated on the cutting head, including:

[0014] Obtaining the original electric signal output by the displacement sensor;

[0015] Filtering the original electric signal to eliminate interference noise;

[0016] Convert the filtered electric signal into the corresponding actual distance value;

[0017] Continuously record the actual distance value and its corresponding cutting head position information to form the instantaneous distance information.

[0018] Further, based on the instantaneous distance information, the workpiece surface curvature variation feature of the position of the detection point on the preset trajectory is calculated, including:

[0019] Based on the cutting head position information and the corresponding actual distance value contained in the instantaneous distance information, the three-dimensional coordinates of the current detection point and the adjacent previous detection point on the workpiece surface are determined;

[0020] According to the three-dimensional coordinates of the current detection point and the adjacent previous detection point, the curvature estimation value of the workpiece surface at the current detection point is calculated;

[0021] According to the movement sequence of the cutting head along the preset trajectory, the curvature estimation values at the continuous multiple detection points are differentially operated to obtain the workpiece surface curvature variation feature.

[0022] Further, based on the workpiece surface curvature variation feature, the inertial load risk generated by the geometric morphology mutation to the cutting head motion system and the dynamic load risk generated by the laser energy interaction process are comprehensively analyzed to identify the current preset trajectory section as a path risk section or a non-path risk section, including:

[0023] Based on the workpiece surface curvature variation feature, the curvature variation rate is calculated;

[0024] The rate of change of curvature is compared with a first preset threshold. When the rate of change of curvature exceeds the first preset threshold, it is determined that there is a risk of inertial load due to the lack of coordination between the adjustment of the motion direction and the normal vector.

[0025] The rate of curvature change is compared with a second preset threshold. When the rate of curvature change exceeds the second preset threshold, it is determined that there is a dynamic load risk caused by a sudden change in the energy density of the focal spot.

[0026] When inertial load risk and dynamic load risk coexist, and it is determined that the two risks have a mutually reinforcing coupling effect in the current preset trajectory segment, the current preset trajectory segment is identified as a path risk segment.

[0027] Otherwise, identify the current preset trajectory segment as a non-path risk segment.

[0028] Furthermore, a conservative prediction mode is activated for path-risk segments, and a conventional prediction mode is activated for non-path-risk segments, to predict the workpiece surface normal vector at the laser processing point, including:

[0029] When the current preset trajectory segment is identified as a path risk segment, the number of adjacent detection points used for normal vector calculation is increased and a weighted average algorithm is used to predict the workpiece surface normal vector at the laser processing point.

[0030] When the current preset trajectory segment is identified as a non-path risk segment, the workpiece surface normal vector is directly calculated based on the three-dimensional coordinates of the current detection point and the immediately preceding detection point.

[0031] Furthermore, by increasing the number of adjacent detection points used for normal vector calculation and employing a weighted average algorithm for prediction, the following steps are taken: obtaining the three-dimensional coordinates of the current detection point and at least two previous historical detection points; calculating local normal vectors based on the three-dimensional coordinates of every three consecutive detection points; and performing a weighted average calculation on multiple local normal vectors using the reciprocal of the trajectory distance between the center detection point and the laser processing point corresponding to each local normal vector as the weight, with the final result used as the predicted workpiece surface normal vector.

[0032] Furthermore, the three-dimensional coordinates of the current detection point and the immediately preceding detection point are directly calculated, including: using the three-dimensional coordinates of the current detection point, the immediately preceding detection point, and the third point determined according to the direction of the tangent vector at the current point, two spatial vectors are constructed; by calculating the cross product of the two spatial vectors, the direction of the normal vector of the workpiece surface is obtained.

[0033] Furthermore, based on the workpiece surface normal vector at the laser processing point, the required attitude adjustment amount when the cutting head reaches the laser processing point is calculated, including:

[0034] comparing the predicted workpiece surface normal vector at the laser processing point with the reference axis direction of the laser beam of the cutting head, determining a spatial angle deflection amount of the cutting head required to keep the two consistent;

[0035] According to the spatial angle deflection amount of the cutting head, and in combination with the kinematic relationship between the cutting head and each motion control axis, the position command increment of each motion control axis is calculated, and the position command increment is the attitude adjustment amount.

[0036] Further, the attitude adjustment of the cutting head is performed according to the attitude adjustment amount during the movement of the cutting head to the laser processing point, including:

[0037] The position command increment of each motion control axis contained in the attitude adjustment amount is synchronously interpolated with the main path position command of the cutting head moving along the preset trajectory;

[0038] Based on the control command generated after the synchronous interpolation, each motion control axis is driven to move coordinately, so that the cutting head completes the smooth adjustment of the attitude during the movement to the laser processing point.

[0039] On the other hand, the application provides an adaptive focal length control system of a laser cutting machine, including the following modules:

[0040] An information acquisition module is used to continuously acquire the instantaneous distance information between the detection point and the workpiece surface of the cutting head through the displacement sensor integrated on the cutting head during the movement of the cutting head along the preset trajectory;

[0041] A feature calculation module is used to calculate the workpiece surface curvature variation feature of the position of the detection point on the preset trajectory based on the instantaneous distance information;

[0042] A risk identification module is used to comprehensively analyze the inertial load risk of the cutting head motion system and the dynamic load risk of the laser energy interaction process caused by the geometric morphology mutation based on the workpiece surface curvature variation feature, identify whether the current preset trajectory segment is a path risk segment or a non-path risk segment;

[0043] A vector prediction module is used to start a conservative prediction mode for the path risk segment and a regular prediction mode for the non-path risk segment to predict the workpiece surface normal vector at the laser processing point;

[0044] An adjustment calculation module is used to calculate the attitude adjustment amount required when the cutting head reaches the laser processing point according to the workpiece surface normal vector at the laser processing point;

[0045] An attitude adjustment module is used to control the attitude adjustment of the cutting head according to the attitude adjustment amount during the movement of the cutting head to the laser processing point.

[0046] Compared with the prior art, the application has the following beneficial effects:

[0047] 1. By introducing a dynamic risk identification and prediction mode based on the curvature variation characteristics of the workpiece surface, the focal point control accuracy and process stability of the laser cutting machine in the complex curved surface machining process are significantly improved. Traditional methods only rely on the single distance information of the front detection point for height adjustment, which cannot effectively deal with the focal point prediction error caused by the spatial separation of the detection point and the machining point and the geometric morphology mutation of the curved surface. The present application assesses the comprehensive risk of geometric mutation to the inertia load of the motion system and the interactive dynamic load of laser energy by calculating the curvature variation characteristics on the detection point trajectory, and intelligently distinguishes the machining path into risk segments and non-risk segments according to the risk. For the risk segment, a conservative prediction mode based on the weighted average of multiple point information is used to smooth the noise and enhance the robustness of the normal vector prediction. For the non-risk segment, a conventional prediction mode based on direct calculation of adjacent points is used to ensure calculation efficiency. This differentiated prediction strategy makes the prediction of the workpiece surface normal vector at the laser machining point more accurate and reliable, thereby laying a solid foundation for subsequent high-precision attitude adjustment.

[0048] 2. The control target is improved from maintaining a constant distance to accurately controlling the alignment of the laser beam axis and the workpiece surface normal vector, which is a more essential and accurate focal length control paradigm. The spatial angle deflection amount calculated based on the accurately predicted normal vector is converted into position command increments of each control axis through precise kinematics inverse solution, and finally the attitude is smoothly adjusted through multi-axis synchronous interpolation control. This method ensures that the laser beam of the cutting head can always maintain the ideal vertical relationship with the workpiece surface at the moment of reaching the machining point, fundamentally eliminating the defects of focal point drift and uneven energy density distribution caused by curvature variation. It can effectively ensure the uniformity of the cutting seam width of laser cutting, avoid cutting not transparent or burning phenomenon, and significantly improve the laser cutting quality and processing process reliability of three-dimensional curved surface workpieces, especially those with complex geometric characteristics. BRIEF DESCRIPTION OF DRAWINGS

[0049] Figure 1 The flowchart of the adaptive focal length control method of the laser cutting machine of the present application;

[0050] Figure 2 The structural schematic diagram of the adaptive focal length control system of the laser cutting machine of the present application. DETAILED DESCRIPTION

[0051] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0052] Example 1: Figure 1 An adaptive focal length control method for a laser cutting machine is presented in this invention, which includes the following steps:

[0053] S1. During the movement of the cutting head along the preset trajectory, the instantaneous distance information between the detection point and the workpiece surface is continuously acquired by the displacement sensor integrated on the cutting head.

[0054] S2. Calculate the surface curvature change characteristics of the workpiece at the location of the detection point on the preset trajectory based on instantaneous distance information;

[0055] S3. Based on the curvature change characteristics of the workpiece surface, comprehensively analyze the inertial load risk to the cutting head motion system caused by the sudden change in geometric shape and the dynamic load risk caused by the laser energy interaction process, and identify whether the current preset trajectory segment is a path risk segment or a non-path risk segment.

[0056] S4. For path risk segments, activate the conservative prediction mode; for non-path risk segments, activate the conventional prediction mode to predict the workpiece surface normal vector at the laser processing point.

[0057] S5. Calculate the required attitude adjustment amount when the cutting head reaches the laser processing point based on the workpiece surface normal vector at the laser processing point.

[0058] S6. Control the cutting head to adjust its posture according to the posture adjustment amount during the process of moving to the laser processing point.

[0059] In step S1, the instantaneous distance information between the detection point and the workpiece surface is continuously acquired by the displacement sensor integrated on the cutting head, specifically through the following method:

[0060] The displacement sensor operates continuously as the cutting head moves along a preset trajectory. Its output electrical signal is typically an analog signal in the form of voltage or current, and the magnitude of this signal corresponds precisely to the physical distance between the sensor probe and the workpiece surface. Acquiring this raw electrical signal is accomplished through an analog-to-digital converter. For example, the control system is equipped with a data acquisition card that synchronously acquires the analog output signal of the displacement sensor at a sampling frequency of at least 1000Hz, converting the continuous analog signal into a discrete digital signal sequence, thus completing signal acquisition and laying the foundation for subsequent digital signal processing.

[0061] Filtering the raw electrical signal to eliminate interference noise is necessary because in the actual laser cutting processing environment, power grid fluctuations, driver start-stop and mechanical vibration and other factors will introduce high-frequency noise components, which will be superimposed on the real distance signal, causing the measured value to be distorted. The filtering process is implemented in real time in the processor using a digital filtering algorithm, for example, in this embodiment, a second-order low-pass Butterworth filter with a cutoff frequency of 100 Hz is used to filter the collected digital signal sequence. The type, order and cutoff frequency of the filter are determined according to the maximum motion acceleration of the cutting head, the response characteristics of the displacement sensor and the typical noise frequency range to be suppressed, and the design goal is to maximize the retention of low-frequency effective signals reflecting the real change in distance, while significantly attenuating high-frequency noise interference. The execution process of the filtering algorithm is to calculate the filtered smooth signal value by weighting the digital signal value of the current sampling point and the signal values of the previous multiple sampling points according to a specific difference equation, which ensures the reliability of the data quality.

[0062] Converting the filtered electrical signal to the corresponding actual distance value depends on the parameters of the displacement sensor pre-calibration. Each displacement sensor needs to be calibrated before use. The calibration process changes the distance between the sensor and the standard target plate through a precision displacement platform, while recording the output electrical signal value of the sensor at different known distance points, thereby establishing a mapping relationship from electrical signal to physical distance. This mapping relationship is usually stored in the form of a data table or a fitting formula in the non-volatile memory of the control system. In actual conversion, the processor reads the current filtered electrical signal digital quantity, and through the query of the calibration data table or the substitution of the calibration formula for calculation, the actual distance value between the detection point and the workpiece surface can be accurately obtained. The unit of this distance value is millimeter. For example, for a certain type of laser triangulation displacement sensor, its calibration curve may be approximately a straight line, and the conversion process is a linear calculation.

[0063] The actual distance value and the corresponding cutting head position information are recorded continuously to form complete instantaneous distance information. The cutting head position information is derived from a multi-axis motion controller of a numerical control system, which reads the encoder feedback value of each motion axis servo motor in real time to calculate the precise spatial position coordinates of the cutting head nozzle center point in the workpiece coordinate system. These coordinate axes usually include X-axis, Y-axis, Z-axis linear coordinates and A-axis, C-axis rotation angle coordinates. The core of the recording process is to ensure time synchronization. The control system uses a unified clock source to stamp the same time stamp on each sampled actual distance value and the corresponding cutting head position coordinates, and then stores these pairs of time-related data as a data packet. These data packets are continuously written into a first-in, first-out circular data buffer or a specific log file in chronological order, thereby forming a set of instantaneous distance information sequences that evolve over time and contain distance and position dual information. This set of information sequences not only records the profile height changes of the workpiece surface relative to the probe point, but also accurately associates each height value with the spatial position. It is an indispensable data basis for subsequent analysis of the geometric characteristics of the curved surface. The delay of the entire data acquisition and processing link is accurately calibrated to ensure that the final instantaneous distance information can truly reflect the real-time relative position relationship between the nozzle and the workpiece surface during the cutting head movement.

[0064] In step S2, the curvature variation feature of the workpiece surface at the position of the probe point on the preset trajectory is calculated based on the instantaneous distance information, which is specifically implemented by the following way:

[0065] Firstly, based on the cutting head position information contained in the instantaneous distance information and the corresponding actual distance value, the three-dimensional coordinates of the current detection point and the adjacent previous detection point on the workpiece surface are determined. The cutting head position information is the spatial position coordinates of the cutting head nozzle center point in the preset workpiece coordinate system, which consists of three linear coordinate components of X-axis, Y-axis and Z-axis, and is derived from the real-time feedback of the encoder data of each motion axis of the numerical control system. The actual distance value is the instantaneous linear distance between the detection point of the displacement sensor and the workpiece surface, with the unit of millimeter, which is obtained by filtering and converting the output signal of the displacement sensor. In order to map the measurement value of the detection point to the actual contact point coordinates on the workpiece surface, the installation geometry of the displacement sensor on the cutting head must be combined. The installation geometry is determined in advance through a special calibration process, and the calibration process is as follows: in the state of cutting head static, a standard target plate with known flatness is placed in the detection range of the displacement sensor, the cutting head or the target plate is slowly moved, the displacement sensor scans multiple points on the surface of the target plate, and the cutting head position coordinates and the distance value output by the displacement sensor corresponding to each point are recorded, the detection direction vector of the displacement sensor and the offset distance relative to the center point of the cutting head nozzle are fitted by the least square method, and the spatial transformation relationship between the detection point and the cutting head position is established. In actual calculation, for the current detection point, the cutting head position coordinates at the current time are taken, and a length equal to the actual distance value at the current time is translated in the opposite direction of the calibrated detection direction of the displacement sensor (i.e. the direction pointing to the workpiece surface), and the translated point coordinates are the three-dimensional coordinates of the current detection point in the workpiece coordinate system. The three-dimensional coordinates of the adjacent previous detection point are calculated by the same method, except that the cutting head position information and the actual distance value corresponding to the previous sampling time are indexed from the previously stored instantaneous distance information sequence, and then the same coordinate transformation is performed. Through this series of operations, the time sequence distance measurement sequence is accurately converted into the three-dimensional coordinate sequence of the discrete points on the workpiece surface distributed along the preset trajectory.

[0066] After successfully obtaining the three-dimensional coordinates of the current probe point and adjacent previous probe points (for example, including the current point, the previous point, and the previous two points), the curvature estimate of the workpiece surface at the current probe point is calculated according to the spatial geometric relationship of these points. Curvature is a measure of the local bending degree of a curve or surface in differential geometry, and for a point on a two-dimensional trajectory, it usually refers to the curvature of a curve. When calculating the curvature estimate, the three-dimensional coordinates of at least three non-collinear points are required. The specific calculation process is as follows: select three consecutive points, for example, the current probe point, the immediately preceding probe point, and the immediately following probe point. If the data of the following probe point is temporarily unavailable due to processing delay, the current point, the previous point, and the previous two points can be used. First, two spatial vectors are constructed according to the coordinates of the three points: the first vector is a vector from the previous point to the current point, and the second vector is a vector from the current point to the following point. Then, the included angle between the two vectors is calculated, which can be calculated by the dot product formula, that is, the dot product of the two vectors divided by the product of their lengths, and then taking the inverse cosine function to get the included angle value, with the unit of radian. Next, the Euclidean distance between the current point and the previous point is calculated, as well as the Euclidean distance between the current point and the following point, and the average of the two distances is taken as the characteristic chord length. The curvature estimate can be approximately obtained by dividing the included angle value by the characteristic chord length, and its physical meaning is the rate of change of the tangent direction per unit arc length, with the unit of per millimeter. This is a commonly used geometric method for approximating the curvature of a curve based on a sequence of discrete points, suitable for real-time processing in a numerical control system. In order to improve the robustness of the estimate, if there is a small amount of noise in the coordinates of the points, the coordinates of several consecutive points can be smoothed first, for example, using a three-point moving average method, and then the curvature is calculated. Through the above calculation, a numerical value is assigned to each probe point position, which quantitatively represents the local bending degree of the workpiece surface along the motion trajectory at that position, i.e., the curvature estimate.

[0067] According to the movement sequence of the cutting head along the preset trajectory, the curvature estimation values calculated at the continuous multiple detection points are subjected to difference operation, so as to obtain the curvature variation characteristics of the workpiece surface. The purpose of the difference operation is to quantify the change rate of the curvature value along the movement trajectory, that is, the curvature change rate. In specific implementation, the control system reads the curvature estimation values calculated at the continuous multiple detection points (for example, 5 to 10 points, and the specific number can be adjusted according to the processing capacity of the control system and the required smoothness) in the time sequence of the detection points or more preferably in the path length sequence of the detection points along the preset trajectory (because the path length can better reflect the geometric nature). Then, first-order difference calculation is performed: for two adjacent detection points in the sequence, the difference between the curvature estimation value of the latter point and the curvature estimation value of the former point is calculated to obtain the curvature difference value. At the same time, the path length along the preset trajectory between the two adjacent detection points is calculated, and the path length can be obtained by calculating the Euclidean distance between the three-dimensional coordinates of the two detection points. Finally, the curvature difference value is divided by the corresponding path length to obtain the average curvature change rate in the interval, and the unit is per square millimeter. This sequence of curvature change rate values constitutes the core content of the curvature variation characteristics of the workpiece surface. In order to suppress the fluctuations of the curvature change rate that may be caused by discrete calculation or measurement noise, it is necessary to post-process and smooth filter the initially calculated curvature change rate sequence, for example, using a moving average filtering algorithm: taking the curvature change rate values of the current point and the points before and after it (a total of three points), calculating their arithmetic mean, and using the mean value as the final curvature change feature value of the current point. The size of the filter window (for example, three points) is selected according to the balance between smoothing effect and response speed, and the larger the window, the better the smoothing effect but the response to real sharp changes may be delayed. The final curvature change feature value is a signed scalar, and the absolute value directly reflects the severity of the curvature change of the workpiece surface, and the sign indicates the trend of the curvature, that is, whether it is increasing or decreasing. This series of curvature change feature values distributed along the trajectory provides key and reliable geometric input parameters for dynamically evaluating the processing risk in the subsequent steps. The entire calculation process starts from the original coordinate conversion, and then proceeds through curvature estimation, difference operation and smoothing filtering, to ensure the accuracy and real-time performance of feature extraction, and lays a solid foundation for adaptive control.

[0068] In step S3, based on the curvature variation characteristics of the workpiece surface, the inertial load risk generated by the geometric morphology mutation to the cutting head movement system and the dynamic load risk generated by the laser energy interaction process are comprehensively analyzed, and it is identified whether the current preset trajectory segment is a path risk segment or a non-path risk segment, which is specifically realized by the following methods:

[0069] Firstly, the curvature change rate is calculated based on the workpiece surface curvature variation feature. The workpiece surface curvature variation feature is a series of curvature change rate values along the preset trajectory distribution obtained from step S2. Each value in the series corresponds to a detection point position, representing the change rate of the workpiece surface curvature along the motion trajectory at that point. When calculating the curvature change rate, the value in the workpiece surface curvature variation feature sequence corresponding to the current detection point is directly taken as the curvature change rate of the current point. The curvature change rate is a signed scalar with a unit of per square millimeter. Its absolute value size reflects the severity of curvature change, and the sign indicates the change direction. In order to ensure the stability of the value, the curvature variation feature values of the current point and its adjacent points can be simply averaged, for example, the curvature variation feature values of the current point, the previous point and the next point are averaged, and the average value is taken as the curvature change rate of the current point for risk analysis. This smoothing processing helps to suppress false judgments caused by accidental fluctuations.

[0070] The curvature change rate is compared with the first preset threshold value. When the curvature change rate exceeds the first preset threshold value, it is determined that there is an inertial load risk caused by the uncoordinated adjustment of the motion direction and the normal vector. The setting basis of the first preset threshold value is the dynamic response capability of the cutting head motion system. Specifically, it is determined by the following method: during the machine tool debugging stage, the maximum angular acceleration or maximum normal vector adjustment rate that the cutting head can stably track under multi-axis linkage is tested by experimental method. This maximum capability value is usually limited by the servo motor torque, mechanical structure rigidity and control bandwidth. Then, in the simulation processing or actual processing test, the actual load data of the cutting head motion system under different curvature change rates is collected, for example, the servo motor current or torque command is monitored to indirectly reflect the size of the inertial load, and the curvature change rate critical value corresponding to the load close to the system allowed upper limit is found out by analysis, and the critical value is multiplied by a safety factor, for example, 0.8, to obtain the first preset threshold value. When the absolute value of the real-time calculated curvature change rate exceeds the first preset threshold value, it means that the workpiece surface geometry changes too fast. In order to keep the laser beam perpendicular to the workpiece surface, the motion system needs to complete a large attitude angle adjustment in a very short time. Such urgent motion instructions will cause significant inertial torque, thereby determining that there is an inertial load risk, which specifically manifests as possible servo motor overload, mechanical vibration or increased tracking error.

[0071] The curvature change rate is compared with a second preset threshold value, and when the curvature change rate exceeds the second preset threshold value, it is determined that there is a dynamic load risk caused by a sudden change in the focal spot energy density. The setting basis of the second preset threshold value is the requirement of the laser processing technology on the energy stability, and the second preset threshold value is specifically determined by the following manner: in the process test, cutting tests are performed on work surfaces with different curvatures, and the energy density distribution of the laser actually acting on the workpiece surface is monitored at the same time. The energy density can be indirectly evaluated by measuring the laser power and calculating the actual focal spot area. The change of the focal spot area is directly related to the change of the defocusing amount caused by the change of the normal vector of the workpiece surface. The processing quality, such as the uniformity of the cutting seam width, the slag hanging condition and the like, is analyzed, the critical value of the curvature change rate corresponding to the situation that the energy density fluctuation exceeds the process allowable range and further causes the processing quality to decrease is found out, the critical value is taken as the basis of the second preset threshold value, and a certain process margin is considered to fine-tune. When the absolute value of the real-time calculated curvature change rate exceeds the second preset threshold value, it means that the rapid change of the curvature of the workpiece surface will cause the rapid change of the relative distance between the cutting head and the workpiece. Even if there is a focus tracking system, the position of the laser focal point relative to the workpiece surface will also fluctuate, thereby causing the dramatic change of the laser focal spot area and the energy density. The instability of such energy interaction process is the dynamic load risk, which may be manifested as cutting not transparent, overburning or larger material heat affected zone.

[0072] When the inertial load risk and the dynamic load risk exist simultaneously, that is, the curvature change rate of the current point simultaneously exceeds the first preset threshold and the second preset threshold, it is further determined whether the two risks exist in the current preset track segment. The coupling effect is determined based on the nonlinear influence of risk superposition, and the specific implementation manner is that whether the number of points at which the inertial load risk and the dynamic load risk continuously exist on the current preset track segment (for example, a track segment containing the current point and a plurality of adjacent points before and after the current point) exceeds a critical proportion is checked, for example, whether at least two points of the three consecutive points simultaneously exist the two risks. At the same time, whether the product or weighted sum of the intensities of the two risks exceeds a coupling threshold is evaluated. The coupling threshold is obtained through system identification experiments. The experimental method is to run the cutting head on the track segment with compound risks, and to observe whether the control system appears oscillation, the processing quality is significantly deteriorated, and other unstable phenomena, so as to deduce the critical condition of risk coupling. If the risk continuously exists and the intensity superposition exceeds the threshold, it is determined that the coupling effect of mutual intensification exists, which means that the inertial load of the motion system and the dynamic fluctuation of the laser energy will interfere with each other. For example, rapid posture adjustment may intensify the defocus, and energy mutation may cause thermal deformation feedback to affect the motion accuracy. This coupling will significantly increase the probability of system out-of-control or processing failure. At this time, the current preset track segment is identified as a path risk segment. If the inertial load risk and the dynamic load risk do not exist simultaneously, or although they exist simultaneously, it is determined that there is no mutual intensification coupling effect, for example, the risk points are isolated and discontinuous or the superposition intensity is weak, the current preset track segment is identified as a non-path risk segment. The whole risk identification process ensures the accuracy and reliability of the decision, and provides a clear basis for subsequent selection of different prediction modes.

[0073] In step S4, the conservative prediction mode is started for the path risk segment, and the conventional prediction mode is started for the non-path risk segment to predict the workpiece surface normal vector at the laser processing point. The prediction is realized by the following method:

[0074] When it is determined according to the identification result of step S3 that the current preset track segment is a path risk segment, the conservative prediction mode is started. The core of the mode is to increase the number of adjacent detection points used for normal vector calculation and use a weighted average algorithm for prediction, aiming to smooth the normal vector estimation noise caused by geometric morphology mutation by introducing more historical data, and to enhance the robustness and stability of the prediction result.

[0075] In implementation, firstly, the three-dimensional coordinates of the current detection point and at least two previous historical detection points are obtained, which are derived from the sequence of discrete point coordinates of the workpiece surface calculated and stored in step S2; the number of points obtained can be adjusted according to the severity of the risk, for example, in the area where the curvature changes particularly sharply, the coordinates of the current point and the previous three or four historical detection points can be obtained; after obtaining the coordinates of these points, the local normal vector is calculated based on the three-dimensional coordinates of each three consecutive detection points, and the method for calculating the local normal vector is as follows: for any three consecutive points, denoted as point A, point B and point C, where point B is the center point of the three points, two space vectors are constructed using the coordinates of points A, B and C, that is, vector AB (from point A to point B) and vector BC (from point B to point C), then the cross product of vector AB and vector BC is calculated, and the direction of the vector obtained by the cross product is the approximate normal direction of the plane determined by points A, B and C in the local area, and the cross product vector is normalized to become a unit vector, and the local normal vector based on the three points is obtained; the above calculation is sequentially performed on all combinations of the obtained three consecutive points to obtain multiple local normal vectors.

[0076] Subsequently, the reciprocal of the trajectory distance between the center detection point corresponding to each local normal vector and the laser processing point to be predicted is taken as the weight, and the weighted average calculation is performed on all calculated local normal vectors; the trajectory distance refers to the arc length distance of the center detection point to the laser processing point along the preset trajectory path, which can be approximated by the straight line distance between the two points, or a more accurate path length can be obtained by accumulating the straight line distances of the intermediate points; the setting principle of the weight is that the closer the center detection point to the laser processing point, the greater the weight of the corresponding local normal vector in the weighted average, because the geometric information of the near point is more relevant; in the specific calculation of the weighted average, the three components (X, Y, Z direction components) of each local normal vector are multiplied by the corresponding weight, then all weighted components are summed, and finally the resulting vector is normalized to become a unit vector, and this final unit vector is taken as the workpiece surface normal vector at the laser processing point predicted in the conservative prediction mode. The advantage of this method is that through the fusion of multi-point information, the influence of single-point fluctuation on the prediction result is effectively suppressed.

[0077] When the current preset trajectory segment is determined as a non-path risk segment according to the recognition result of step S3, a normal prediction mode is started. In this mode, the workpiece surface geometry changes gently, and there is no need for complex smoothing processing. A direct calculation method is used to improve the calculation efficiency. The normal vector of the workpiece surface is directly calculated based on the three-dimensional coordinates of the current detection point and the immediately preceding detection point in the normal prediction mode. The specific calculation process includes: first, the three-dimensional coordinates of the current detection point, the immediately preceding detection point, and a third point determined according to the direction of the tangent vector at the current point are used to construct two spatial vectors. The three-dimensional coordinates of the current detection point and the immediately preceding detection point can be directly obtained from the stored sequence. The third point is determined as follows: first, the tangent vector at the current point is calculated, which can be approximated by the vector obtained by subtracting the coordinates of the previous point from the coordinates of the current point, and then the tangent vector is normalized to a unit tangent vector; then, based on the coordinates of the current point, a small step, such as a system-set small distance value (e.g. 0.1 mm), is translated along the direction of the unit tangent vector, thereby theoretically obtaining a virtual point located in the tangent direction of the current point. This virtual point is the required third point.

[0078] In fact, in order to simplify the calculation, the next detection point of the current point (if the data is available) can also be directly used as the third point sometimes, but under the strict definition of the normal prediction mode, it is determined based on the direction of the tangent vector at the current point. After obtaining the coordinates of the three points, two spatial vectors are constructed: the first vector is a vector from the immediately preceding detection point to the current detection point, and the second vector is a vector from the current detection point to the third point. Then, the direction of the workpiece surface normal vector is obtained by calculating the cross product of the two spatial vectors; the calculation rule of the vector cross product is that the direction of the resulting vector is perpendicular to the plane formed by the original two vectors, and the direction conforms to the right-hand rule; the calculated cross product vector represents the normal direction of the plane determined by the three points; finally, the cross product vector is normalized to convert it to a unit vector, and the direction of this unit vector is taken as the workpiece surface normal vector at the laser processing point predicted by the normal prediction mode. This method has small calculation amount and fast response speed, and is suitable for non-path risk segments with smooth surface, which can improve the overall response efficiency of the system while ensuring sufficient accuracy. Through intelligent switching of the two prediction modes, the system realizes adaptive balance between processing accuracy and calculation load.

[0079] In step S5, the attitude adjustment amount required for the cutting head to reach the laser processing point is calculated based on the normal vector of the workpiece surface at the laser processing point, which is realized by the following method:

[0080] First, the predicted workpiece surface normal vector at the laser processing point is compared with the reference axis direction of the laser beam of the cutting head to determine the spatial angle deflection amount required to keep the two consistent. The predicted workpiece surface normal vector at the laser processing point is a unit vector obtained from step S4, and its direction is perpendicular to the workpiece surface tangent plane at the laser processing point. The reference axis direction of the laser beam of the cutting head is a constant direction defined in the cutting head coordinate system in advance, and is usually defined as the propagation direction of the laser beam when the cutting head is in the zero pose, for example, in the commonly used five-axis laser cutting machine, the reference axis direction can be defined as the direction parallel to the machine tool Z axis.

[0081] The purpose of the comparison is to calculate how to rotate the cutting head so that its current laser beam axis direction coincides with the workpiece surface normal vector, thereby ensuring that the laser beam is perpendicular to the workpiece surface at the processing point. The specific calculation process of determining the spatial angle deflection amount is as follows: since both directions are unit vectors, the angle between them can be obtained by calculating the dot product of the two vectors and then taking the inverse cosine function, and this angle is the total angle that needs to be deflected. However, it is not enough to know only the total angle, but also to know the axis direction of the deflection. This deflection axis direction can be obtained by calculating the cross product of the workpiece surface normal vector and the reference axis direction, and the direction of the vector obtained by the cross product is the rotation axis direction of the shortest path for rotating the reference axis to be parallel to the normal vector. Therefore, the spatial angle deflection amount of the cutting head can actually be completely described by a rotation axis vector and a rotation angle, which constitutes a rotation operation of an axis-angle representation. In some control systems, this rotation relationship can also be converted into Euler angles for representation, for example, first rotating around the A axis of the machine tool by an angle, and then rotating around the C axis by an angle, the specific conversion relationship depends on the kinematic structure of the machine tool. This spatial angle deflection amount quantitatively describes the azimuth adjustment that the cutting head needs to make to adapt to the geometric features of the workpiece surface.

[0082] Then, according to the calculated spatial angle deflection amount of the cutting head, combined with the kinematic relationship between the cutting head and each motion control axis, the position command increment of each motion control axis is calculated, which is the final required pose adjustment amount. The kinematic relationship between the cutting head and each motion control axis is determined in advance through the mechanical structure parameters and coordinate transformation relationship of the machine tool, which establishes a mathematical mapping model between the spatial pose (including position and direction) of the cutting head nozzle tip and the physical positions of each motion control axis (such as X axis, Y axis, Z axis three linear axes, and A axis, C axis two rotary axes). For a given spatial angle deflection amount of the cutting head (whether represented by axis-angle or Euler angle), it needs to be converted into a position command for a specific motion control axis. The solving process usually involves inverse kinematics solving.

[0083] For example, if the spatial angle deflection amount is expressed by Euler angles (e.g. tilt angle A and rotation angle C) relative to the machine tool coordinate system, then for a typical double swing head five-axis machine tool, the position command increment of the A-axis and C-axis that need to be moved directly corresponds to the change amount of these two Euler angles, usually in degrees. However, the movement of the rotating axes can cause a slight change in the position of the cutting head nozzle center point (i.e. tool center point offset), so the position commands of the X-axis, Y-axis, Z-axis linear axes need to be fine-tuned at the same time to compensate, so as to ensure that the laser processing point still accurately falls on the preset trajectory after the posture is adjusted. The calculation of this compensation amount is also included in the kinematic relationship, which is usually calculated through a transformation matrix based on the geometric structure of the machine tool.

[0084] Finally, for each motion control axis, the difference between the target position that the next control cycle needs to reach and the current position of the axis is the position command increment of the axis. The position command increments of all these axes together constitute an instruction set, which is the quantitative basis for controlling the posture adjustment of the cutting head, ensuring that the cutting head can smoothly and accurately move to the required posture, so that the laser beam and the workpiece surface maintain a perpendicular relationship at the processing point. The entire calculation process ensures unambiguous conversion from geometric intent to specific execution instructions.

[0085] In step S6, the cutting head is controlled to perform posture adjustment of the cutting head during movement to the laser processing point according to the posture adjustment amount, which is specifically implemented by the following way:

[0086] First, the position command increments of each motion control axis contained in the posture adjustment amount are synchronously interpolated with the main path position commands of the cutting head moving along the preset trajectory. The posture adjustment amount is a set of data calculated from step S5, which contains the displacement or rotation angle of each motion control axis (e.g. X-axis, Y-axis, Z-axis linear axes and A-axis, C-axis rotating axes) that needs to be moved in the next control cycle. The main path position commands are a series of spatial position coordinate sequences that the cutting head nozzle center point needs to pass through in sequence, which are planned by the numerical control system in advance according to the preset machining trajectory (e.g. tool path generated by CAD / CAM software). The purpose of synchronous interpolation is to coordinate the movement commands related to position and posture in time and space, respectively, to ensure that the cutting head moves accurately along the preset trajectory while smoothly adjusting its own posture.

[0087] The synchronous interpolation is performed in each interpolation cycle, which is a fixed time interval of the CNC system, for example, 1 millisecond. At the beginning of each interpolation cycle, the system reads the master path position command target value at the current time, and also reads the position command increments of each motion control axis planned for this cycle. Then, the motion trajectory needs to be planned to decompose the macro movement of the master path and the micro adjustment of the posture into each interpolation cycle. Specifically, for the master path movement, the system calculates the small line segment that the center point of the cutting head nozzle should move in this interpolation cycle according to the preset feed speed, acceleration and jerk limit. For the posture adjustment, it is also necessary to reasonably distribute the total position command increment to a time sequence of multiple interpolation cycles according to the maximum angular velocity and angular acceleration limit of each motion control axis, instead of completing all adjustments in one cycle, so as to achieve smooth posture transition. Finally, at the end of each interpolation cycle, the system generates a set of comprehensive commands containing the accurate target positions of all motion control axes in this cycle, which makes the linear motion and rotary motion proceed synchronously, and the trajectory motion and posture adjustment seamlessly combined.

[0088] Then, based on the control commands generated after synchronous interpolation, the coordinated motion of each motion control axis is driven to complete the smooth adjustment of the posture of the cutting head during the process of reaching the laser processing point. The control commands generated after synchronous interpolation are a sequence of accurate position setting values for all motion control axes at a series of time points. The process of driving the coordinated motion of each motion control axis is performed by the servo driver of the CNC system. Each motion control axis is equipped with a servo driver, a servo motor and a position feedback device (such as an encoder). The servo driver receives the position command from the CNC system, which is the target position that the axis needs to reach at the end of the next control cycle after interpolation calculation. The position loop controller inside the servo driver compares the received target position with the actual position of the motor feedback from the encoder and calculates the position error. Then, the error signal is operated through the proportional-integral-derivative control algorithm to generate a speed command. The speed command is accurately adjusted by the speed loop controller inside the driver, and finally outputs a current command to drive the servo motor to rotate. The motor converts the rotary motion into linear motion of the shaft through transmission mechanisms such as ball screws and gearboxes, or directly drives the rotary shaft to rotate. The key is that the motion of all axes is synchronized, which is achieved by sending a synchronous clock signal to the drivers of all axes or using a real-time communication protocol based on bus by the CNC system, ensuring that all axes start to execute the new position command at the same time and reach the target position at the same time.

[0089] Through this precise cooperative control, the cutting head can not only accurately track the preset trajectory with its nozzle center point, but also continuously and smoothly change its own posture according to the requirements calculated in step S5 within the limited time and space from the current position to the laser processing point. Finally, at the moment when the tool center point reaches the laser processing point, the posture of the cutting head is also adjusted in place, so that the laser beam axis is perfectly aligned with the predicted workpiece surface normal vector, creating the necessary process conditions for high-quality laser processing. The entire execution process ensures the geometric accuracy and motion stability in dynamic processing.

[0090] Embodiment 2: Figure 2 The structure diagram of the adaptive focal length control system of the laser cutting machine is given, and the adaptive focal length control system of the laser cutting machine comprises the following modules:

[0091] The information acquisition module is used to continuously acquire the instantaneous distance information between the detection point and the workpiece surface through the displacement sensor integrated on the cutting head during the movement of the cutting head along the preset trajectory;

[0092] The feature calculation module is used to calculate the workpiece surface curvature variation feature of the position where the detection point is located on the preset trajectory based on the instantaneous distance information;

[0093] The risk identification module is used to comprehensively analyze the inertial load risk generated by the geometric morphology mutation to the cutting head motion system and the dynamic load risk generated to the laser energy interaction process based on the workpiece surface curvature variation feature, and identify whether the current preset trajectory segment is a path risk segment or a non-path risk segment;

[0094] The vector prediction module is used to start the conservative prediction mode for the path risk segment and start the normal prediction mode for the non-path risk segment to predict the workpiece surface normal vector at the laser processing point;

[0095] The adjustment calculation module is used to calculate the posture adjustment amount required when the cutting head reaches the laser processing point according to the workpiece surface normal vector at the laser processing point;

[0096] The posture adjustment module is used to control the cutting head to perform posture adjustment of the cutting head according to the posture adjustment amount during the movement to the laser processing point.

[0097] The calculations involved in the embodiments are all dimensionless numerical calculations, and the preset parameters and threshold values in the calculations are set by a person skilled in the art according to the actual situation.

[0098] The above embodiments can be realized all or partially by software, hardware, firmware or any combination thereof. When realized by software, the above embodiments can be realized in the form of a computer program product all or partially.

[0099] Those skilled in the art can understand that the modules and algorithm steps of each example described in combination with the embodiments disclosed herein can be realized in electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and the constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0100] In addition, each functional module in each embodiment of the present application can be integrated into one processing module, or each module can exist physically independently, or two or more modules can be integrated into one module.

[0101] In several embodiments provided in the present application, it should be understood that the disclosed system, device and method can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the modules is only a logical function division. In actual implementation, another division manner can be used, for example, a plurality of modules or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the displayed or discussed modules can be indirect coupling or communication connection through some interfaces, devices or modules, and can be electrical, mechanical or other forms.

[0102] The above is merely specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or replacements within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

[0103] Finally: the above is merely preferred embodiments of the present application, and is not used to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. A method of adaptive focus control for a laser cutting machine, characterized by, The method comprises the following steps: S1. During the movement of the cutting head along the preset trajectory, the instantaneous distance information between the detection point of the cutting head and the workpiece surface is continuously obtained through the displacement sensor integrated on the cutting head; S2. The curvature variation feature of the workpiece surface at the position of the detection point on the preset trajectory is calculated based on the instantaneous distance information; S3. Based on the curvature variation feature of the workpiece surface, the inertial load risk and the dynamic load risk generated by the geometric morphology mutation to the cutting head movement system are comprehensively analyzed, and it is identified whether the current preset trajectory segment is a path risk segment or a non-path risk segment, comprising: The curvature variation rate is calculated based on the curvature variation feature of the workpiece surface; The curvature variation rate is compared with the first preset threshold value, and when the curvature variation rate exceeds the first preset threshold value, it is determined that there is an inertial load risk caused by the non-coordination of the movement direction and the normal vector adjustment; The curvature variation rate is compared with the second preset threshold value, and when the curvature variation rate exceeds the second preset threshold value, it is determined that there is a dynamic load risk caused by the energy density mutation of the focal spot; When the inertial load risk and the dynamic load risk exist at the same time, and it is determined that the two risks have a mutual intensification coupling effect in the current preset trajectory segment, the current preset trajectory segment is identified as a path risk segment; Otherwise, the current preset trajectory segment is identified as a non-path risk segment; S4. For the path risk segment, the conservative prediction mode is started, and for the non-path risk segment, the conventional prediction mode is started to predict the workpiece surface normal vector at the laser processing point, comprising: When the current preset trajectory segment is identified as a path risk segment, the workpiece surface normal vector at the laser processing point is predicted by increasing the number of adjacent detection points for normal vector calculation and using a weighted average algorithm; When the current preset trajectory segment is identified as a non-path risk segment, the workpiece surface normal vector is directly calculated based on the three-dimensional coordinates of the current detection point and the immediately preceding detection point; S5. According to the workpiece surface normal vector at the laser processing point, the attitude adjustment amount required for the cutting head to reach the laser processing point is calculated; S6. In the process of moving to the laser processing point, the attitude adjustment of the cutting head is performed according to the attitude adjustment amount.

2. The adaptive focus control method of a laser cutting machine according to claim 1, wherein, During the movement of the cutting head along the preset trajectory, the instantaneous distance information between the detection point of the cutting head and the workpiece surface is continuously obtained through the displacement sensor integrated on the cutting head, comprising: The original electrical signal output by the displacement sensor is obtained; The original electrical signal is filtered to eliminate interference noise; The filtered electrical signal is converted into the corresponding actual distance value; The actual distance value and its corresponding cutting head position information are continuously recorded to form the instantaneous distance information.

3. The adaptive focus control method for a laser cutting machine according to claim 1, wherein, Based on the instantaneous distance information, the curvature variation feature of the workpiece surface at the position of the detection point on the preset trajectory is calculated, comprising: Based on the cutting head position information and the corresponding actual distance value contained in the instantaneous distance information, the three-dimensional coordinates of the current detection point and the adjacent previous detection point on the workpiece surface are determined; According to the three-dimensional coordinates of the current detection point and the adjacent previous detection point, the curvature estimation value of the workpiece surface at the current detection point is calculated; According to the movement sequence of the cutting head along the preset trajectory, the curvature estimation values at a plurality of continuous detection points are subjected to difference operation to obtain the curvature variation feature of the workpiece surface.

4. The adaptive focus control method for a laser cutting machine of claim 1, wherein, The prediction is performed by increasing the number of adjacent probe points used for normal vector calculation and using a weighted average algorithm, including: obtaining three-dimensional coordinates of a current probe point and at least two previous historical probe points; calculating local normal vectors based on three-dimensional coordinates of each three consecutive probe points; performing weighted average calculation on the plurality of local normal vectors with the reciprocal of the trajectory distance between the center probe point corresponding to each local normal vector and the laser processing point as the weight, and taking the final result as the predicted workpiece surface normal vector.

5. The adaptive focus control method for a laser cutting machine according to claim 1, wherein, The workpiece surface normal vector is directly calculated based on three-dimensional coordinates of the current probe point and the immediately preceding probe point, including: using three-dimensional coordinates of the current probe point, the immediately preceding probe point and a third point determined according to the direction of the tangent vector at the current point to construct two space vectors; obtaining the direction of the workpiece surface normal vector by calculating the cross product of the two space vectors.

6. The adaptive focus control method for a laser cutting machine of claim 1, wherein, According to the workpiece surface normal vector at the laser processing point, the attitude adjustment amount required when the cutting head reaches the laser processing point is calculated, including: Comparing the predicted workpiece surface normal vector at the laser processing point with the direction of the reference axis of the cutting head laser beam to determine the spatial angle deflection amount of the cutting head required to keep them consistent; According to the spatial angle deflection amount of the cutting head, combined with the kinematic relationship between the cutting head and each motion control axis, the position command increment of each motion control axis is calculated, which is the attitude adjustment amount.

7. The adaptive focus control method for a laser cutting machine of claim 1, wherein, During the movement of the cutting head to the laser processing point, the attitude adjustment of the cutting head is performed according to the attitude adjustment amount, including: Synchronously interpolating the position command increment of each motion control axis contained in the attitude adjustment amount with the main path position command of the cutting head moving along the preset trajectory; Based on the control command generated after synchronous interpolation, driving the coordinated movement of each motion control axis to complete the smooth adjustment of the attitude of the cutting head during the process of reaching the laser processing point.

8. An adaptive focus control system of a laser cutting machine for implementing the adaptive focus control method of any one of claims 1-7, characterized in that, It includes the following modules: An information acquisition module for continuously acquiring instantaneous distance information between the probe point and the workpiece surface of the cutting head through the displacement sensor integrated on the cutting head during the movement of the cutting head along the preset trajectory; A feature calculation module for calculating the workpiece surface curvature variation feature of the position of the probe point on the preset trajectory based on the instantaneous distance information; A risk identification module for comprehensively analyzing the inertial load risk generated by geometric morphology mutation to the cutting head motion system and the dynamic load risk generated by the laser energy interaction process based on the workpiece surface curvature variation feature, identifying the current preset trajectory segment as a path risk segment or a non-path risk segment; A vector prediction module for starting a conservative prediction mode for the path risk segment and a regular prediction mode for the non-path risk segment to predict the workpiece surface normal vector at the laser processing point; An adjustment calculation module for calculating the attitude adjustment amount required when the cutting head reaches the laser processing point according to the workpiece surface normal vector at the laser processing point; An attitude adjustment module for controlling the cutting head to perform attitude adjustment of the cutting head according to the attitude adjustment amount during the movement of the cutting head to the laser processing point.

Images