Laser self-adaptive cutting method and system for heavy steel pipe with abnormal deformation thickness

Abstract

The invention discloses a laser self-adaptive cutting method and system for a heavy special-shaped thickness steel pipe, and relates to the field of heavy special-shaped workpiece cutting technology application, and the laser self-adaptive cutting method comprises the steps that according to a collected surface image of a workpiece, the geometrical shape, surface defects and material texture features of the workpiece are recognized, and thickness distribution information of the workpiece is obtained; obtaining thickness data of different positions of the workpiece based on the thickness distribution information; based on the geometrical shape, the surface defect, the material texture feature and the thickness data, an optimal cutting path is obtained; and based on the optimal cutting path, a laser cutting task is executed, the laser power, the pulse frequency and the cutting speed are automatically adjusted according to the thickness change, and meanwhile, when vibration is detected, vibration is compensated through reverse excitation or dynamic damping adjustment. The key technical problems in heavy special-shaped workpiece cutting are solved, and the heavy special-shaped workpiece cutting device has wide industrial application prospects and extremely high popularization value.

Description

Technical Field

[0001] The present invention relates to the application field of heavy-duty special-shaped workpiece cutting and processing technology, and in particular to a laser adaptive cutting method and system for heavy-duty special-shaped variable-thickness steel pipes. Background Art

[0002] In the processing of heavy and special-shaped workpieces, cutting technology has always faced huge challenges, especially for complex workpieces such as variable thickness steel pipes. Traditional processing methods are difficult to meet the requirements of high efficiency, high precision and stability at the same time. Such workpieces often have the characteristics of large mass, complex geometric shapes, and obvious thickness changes. The existing technology has exposed many deficiencies in practical applications.

[0003] Traditional cutting methods mainly rely on two ideas: (1) Workpiece rotation scheme: This method clamps and fixes the workpiece on a rotating mechanism, allowing the workpiece to rotate to cooperate with the laser to complete the cutting. Although this scheme is suitable for small and medium-sized workpieces with regular shapes, it has obvious limitations in heavy and special-shaped workpieces. First, the weight of a large workpiece may reach several tons or even higher. The center of gravity is difficult to control during the rotation process. A slight imbalance will cause severe machine vibration, which will lead to processing errors and equipment damage. Secondly, the complex geometric structure makes the rotation clamping process time-consuming and labor-intensive. Improper fixation of the workpiece may also cause accidental falling off, which poses a high operational risk. (2) Laser multi-axis movement scheme: In this technical scheme, the workpiece remains stationary, and the laser is moved and adjusted through a multi-axis linkage system to complete the cutting. Although this method can solve the center of gravity problem caused by workpiece rotation, the flexibility and response speed of existing lasers are limited. Especially when facing variable thickness steel pipes, the laser power and cutting trajectory are difficult to achieve real-time dynamic adjustment, which is prone to overcutting or incomplete cutting. In addition, due to the limited moving range of the laser, the cutting area of some complex workpieces is difficult to cover, which further limits the scope of application in actual production.

[0004] Some studies have attempted to improve on traditional technologies, such as: (1) Adaptive control technology: By adding a thickness detection module, the system can adjust the laser power according to the thickness change of the workpiece. However, this type of technology is often based on offline detection data and cannot achieve real-time adjustment, and is still insufficient in terms of processing speed and accuracy. (2) High-rigidity machine tool design: In order to solve the problem of machine tool vibration, some solutions increase the rigidity and mass of the machine tool to improve stability, but this also brings about the problem of bulky equipment and high cost, and cannot fundamentally solve the dynamic imbalance caused by the center of gravity offset. (3) Local clamping and shock absorption technology: A shock-absorbing clamping device is used to reduce vibration when fixing the workpiece, but for the complex geometric characteristics of special-shaped workpieces, this device has poor adaptability and limited actual effect.

[0005] Through comprehensive analysis, it can be seen that although the existing technical solutions have solved some problems to a certain extent, there are still the following technical bottlenecks that are difficult to overcome: (1) Unstable center of gravity: The center of gravity deviation caused by the rotation of heavy workpieces cannot be completely eliminated, and even if the rigidity of the machine tool is strengthened, it is impossible to effectively avoid the problem of machining vibration. (2) Poor adaptability to thickness changes: When the existing laser cutting technology processes workpieces with variable thickness, the adjustment reaction speed of the laser power and trajectory is relatively slow, resulting in low machining accuracy and even possible damage to the workpiece. (3) Difficult to cover complex workpieces: The moving range and flexibility of the laser are limited. When facing shaped workpieces, it is difficult to complete high-quality cutting in some areas. (4) Inefficiency and cost mismatch: The existing improvement solutions often sacrifice machining efficiency while improving machining accuracy, and the high equipment cost is difficult to meet the needs of industrial large-scale production. Summary of the Invention

[0006] To solve the above problems, the object of the present invention is to provide a variable-thickness steel pipe adaptive laser cutting technology for heavy-shaped workpieces, aiming to completely solve the problems of unstable center of gravity, low cutting accuracy and poor thickness adaptability in the processing of heavy-shaped workpieces by combining the multi-degree-of-freedom rotation of the laser, adaptive thickness detection and center-of-gravity correction technology, and provide a comprehensive solution for the efficient cutting of variable-thickness steel pipes and similar complex workpieces.

[0007] To achieve the above technical object, the present application provides a laser adaptive cutting method for heavy-shaped variable-thickness steel pipes, including the following steps: Based on the surface image of the workpiece collected, identify the geometric shape, surface defects and material texture features of the workpiece, and obtain the thickness distribution information of the workpiece; Based on the thickness distribution information, obtain the thickness data of different positions of the workpiece; Based on the geometric shape, surface defects, material texture features and thickness data, obtain the optimal cutting path; Based on the optimal cutting path, execute the laser cutting task, and automatically adjust the laser power, pulse frequency and cutting speed according to the thickness change. At the same time, when vibration is detected, compensate for the vibration through reverse excitation or dynamic damping adjustment.

[0008] Preferably, when collecting the surface image of the workpiece, by dynamically adjusting the spectrum, intensity and angle of the light source, collect the surface image on workpieces with different materials, colors and surface roughnesses. Among them, a multi-band light source combination is used to enhance the contrast of texture features and obtain material texture features.

[0009] Preferably, when collecting the surface image, set the image acquisition frame rate according to the cutting speed and accuracy requirements.

[0010] Preferably, when obtaining the thickness distribution information of the workpiece, project a pattern onto the surface of the workpiece, acquire the deformed patterns of the workpiece at different angles, and utilize the principle of triangulation. By calculating the geometric relationship between the acquisition points of the deformed patterns and the pattern projection points and the degree of pattern deformation, obtain the three-dimensional coordinates of each point on the workpiece surface and generate the thickness distribution information of the workpiece.

[0011] Preferably, when obtaining the thickness data at different positions of the workpiece, based on the thickness distribution information, utilize laser-excited ultrasonic waves to propagate inside the workpiece, and measure the thickness at different positions of the workpiece by receiving and analyzing the ultrasonic echo signals.

[0012] Preferably, when obtaining the optimal cutting path, based on the geometric shape, surface defects, material texture characteristics, and thickness data, establish a digital twin model of the cutting process, perform a rehearsal and optimization of the cutting path, and aim to ensure the cutting efficiency and accuracy to obtain the optimal cutting path.

[0013] Preferably, when performing the laser cutting task, automatically adjust the laser power and cutting speed according to the real-time thickness, material properties, and cutting process requirements of the workpiece. Among them, the fuzzy control algorithm is used to establish a fuzzy control rule base according to the real-time thickness, material properties, and preset cutting process rules of the workpiece; perform reasoning in the fuzzy control rule base according to the fuzzified input to obtain the fuzzy control output; through defuzzification processing, convert the fuzzy output into a control quantity.

[0014] Preferably, during the process of performing the laser cutting task, when encountering surface defects, thickness mutations, or minor deviations during the cutting process, dynamically adjust the cutting path based on the real-time feedback of three-dimensional topography data and the cutting force sensor signals, bypass the defect area or adjust the cutting direction. Among them, for the case where the defect is small and does not affect the overall structural strength, select to fine-tune the cutting path to avoid the defect area with the laser beam; for the case where the defect is large, re-plan a brand-new cutting path; for the case of thickness mutation, adjust the cutting speed and laser power according to the amplitude and position of the mutation, combined with the capabilities of laser cutting, and re-plan the cutting path at the same time.

[0015] Preferably, when planning the cutting path, based on the three-dimensional model of the workpiece and the cutting process requirements, a path planning method combining genetic algorithm and neural network is adopted to plan the cutting path. Among them, in the genetic algorithm, the cutting path is encoded as a chromosome, the fitness value of each chromosome is calculated, and the roulette wheel selection method is used in the selection operation, so that the probability of each chromosome being selected is proportional to its fitness value; and partially matched crossover is adopted, randomly select and exchange some gene segments of two chromosomes, and handle the gene conflict problem; the neural network is used to predict the best cutting path and parameters under different conditions by learning the complex relationship between the workpiece shape, thickness, material and cutting equipment performance and the optimal cutting parameters.

[0016] The present invention discloses a laser adaptive cutting system for heavy-duty deformed and variable-thickness steel pipes, which is used to implement the above-mentioned laser adaptive cutting method for heavy-duty deformed and variable-thickness steel pipes, including: A data acquisition and analysis module, which is used to identify the geometric shape, surface defects and material texture features of the workpiece according to the collected surface image of the workpiece, and obtain the thickness distribution information of the workpiece; A thickness detection module, which is used to obtain the thickness data of different positions of the workpiece based on the thickness distribution information; A path planning module, which is used to obtain the optimal cutting path based on the geometric shape, surface defects, material texture features and thickness data; A cutting task execution module, which is used to execute the laser cutting task based on the optimal cutting path, and automatically adjust the laser power, pulse frequency and cutting speed according to the thickness change; A vibration compensation module, which is used to compensate for vibration by reverse excitation or dynamic damping adjustment when vibration is detected; A quality evaluation module, which is used to evaluate the cutting quality by obtaining the cutting temperature, cutting force and smoke concentration when executing the laser cutting task; A fault diagnosis module, which is used to quickly locate the cause of the fault by using a diagnosis method based on fault tree analysis FTA and neural network after detecting that the equipment fails.

[0017] The present invention discloses the following technical effects: The present invention can flexibly select the type of laser source according to the workpiece material and cutting requirements. Common equipment may only be adapted to a single or a few types of laser sources, and it is difficult to meet the diverse processing requirements.

[0018] The multi-degree-of-freedom laser head designed by the present invention integrates a high-precision motor drive system, combined with a built-in attitude sensor and a feedback control system, can accurately adjust the angle and position of the laser head to ensure that the laser beam is always on the best cutting path.

[0019] The beam transmission system of the present invention, which consists of a high-precision reflector and an optical fiber component, can better ensure that the beam quality is not greatly affected during transmission.

[0020] An efficient collaborative working mechanism is formed among the laser source, the beam transmission system, the multi-degree-of-freedom laser head, the attitude sensor, and the feedback control system mentioned in the present invention.

[0021] Based on the real-time information of the workpiece, the present invention uses the data monitored by the attitude sensor and adjusts the parameters such as the attitude of the laser head and the power of the laser source in real time through the feedback control system, making the cutting process more intelligent. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required to be used in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0023] Figure 1 It is a schematic diagram of the multi-degree-of-freedom fixture described in the present invention; Figure 2 It is a schematic diagram of the laser clamping and cutting movement described in the present invention; Figure 3 It is a schematic diagram of the laser cutting operation described in the present invention; Figure 4 It is a schematic diagram of the working principle of machine vision described in the present invention; Figure 5 It is a schematic diagram of the workpiece information recognition process based on deep learning described in the present invention; Figure 6 It is a schematic diagram of the thickness detection and feedback control operation described in the present invention; Figure 7 It is a schematic diagram of the working process of the path planning algorithm described in the present invention; Figure 8 It is a schematic diagram of the actual machining and virtual model operation described in the present invention; Figure 9 It is a schematic diagram of the working principle of the vibration compensation mechanism described in the present invention; Figure 10 It is a schematic diagram of the multi-sensor information acquisition working process described in the present invention; Figure 11 It is a schematic diagram of the cutting parameter and path adjustment working process described in the present invention; Figure 12It is a variable-thickness steel pipe adaptive laser cutting device for heavy-duty special-shaped workpieces according to the present invention. Among them, 1. Multi-degree-of-freedom laser fixture, 2. Gantry track, 3. Laser, 4. Workpiece, 5. Industrial camera, 6. Strip light source, 7. Industrial lens, 8. Workpiece, 9. Ring light source, 61. Honeycomb board, 62. Connecting rib plate, 63. High-precision motor, 64. Industrial camera. Detailed implementation manners

[0024] To make the objectives, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application usually described and illustrated in the accompanying drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the present application to be protected, but only represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative efforts belong to the scope of protection of the present application.

[0025] As Figures 1-12 shown, the present invention innovatively proposes a cutting solution for heavy-duty special-shaped workpieces integrating multiple cutting-edge technologies, mainly including advanced machine vision technology, intelligent adaptive laser control technology, high-precision thickness detection technology, intelligent path planning algorithm, and efficient vibration compensation mechanism, etc., aiming to overcome the long-existing problems in the field of cutting heavy-duty and special-shaped workpieces, such as unstable precision, poor cutting quality, and low processing efficiency, and provide an efficient, accurate, and stable cutting solution for related industrial production.

[0026] The adaptive laser cutting system of the present invention consists of multiple closely coordinated core modules, including a high-performance laser cutting device, an advanced machine vision system, a precise thickness detection and feedback control system, an intelligent adaptive path planning and control algorithm, an efficient vibration compensation mechanism, etc. Each module cooperates with each other to achieve high-quality cutting of heavy-duty special-shaped workpieces.

[0027] As Figure 1 shown, the fixture with multiple degrees of freedom can hold the industrial camera 64 or the laser, and through programming, control the high-precision motor 63 to act, so that the end of the fixture (holding the industrial camera 64 or the laser, etc.) is in a suitable position in space, so as to achieve the purpose of image information acquisition and cutting. Further explanation, the honeycomb board 61 is fixed at a suitable position on the machine tool, and the connecting rib plate 62 is connected to the honeycomb board 61. The high-precision motor 63 acts, and the mechanism acts to make the end of the fixture in a suitable working state.

[0028] As shown Figure 2 in the figure, the laser 3 is clamped by a multi-degree-of-freedom fixture 1. The multi-degree-of-freedom fixture 1 can move on the gantry rail 2, and the workpiece 4 is located directly below the laser. Further explanation, this laser integrates a high-precision motor drive system, which can quickly and accurately adjust the angle and position in space according to the complex shape and thickness change of the workpiece. Through the built-in attitude sensor and feedback control system, it ensures that the laser beam is always perpendicular to the workpiece surface or travels along the theoretical optimal cutting path, effectively avoiding cutting deviation caused by unstable workpiece center of gravity and irregular shape, and guaranteeing the cutting accuracy. For example, when cutting large-sized special-shaped steel structure parts, the laser can be adjusted in real time to ensure that the laser beam acts stably on the workpiece surface and improve the cutting quality.

[0029] As shown Figure 3 in the figure, the core of the laser cutting device is a high-power laser source, and its type is selected according to the workpiece material and cutting requirements. For example, a fiber laser can be selected for metal workpieces, and the output power is several kilowatts or even higher to meet the cutting requirements for thick plate materials. After the laser beam is emitted from the laser source, it is transmitted through the beam transmission system. This system includes a series of high-precision mirrors and fiber optic components. The mirrors are used to adjust the beam direction, and the fiber optic is responsible for efficiently transmitting the beam and ensuring that the beam quality is not greatly affected during the transmission process.

[0030] As shown Figure 3 in the figure, the high-precision motor drive system of the multi-degree-of-freedom laser works in cooperation with multiple servo motors. These servo motors respectively control the linear motion of the laser in the X, Y, and Z axis directions and the rotational motion around the X, Y, and Z axes, realizing the precise attitude adjustment of the laser in three-dimensional space. The built-in attitude sensor uses a high-precision inertial measurement unit (IMU), which can real-time monitor the acceleration and angular velocity information of the laser, so as to accurately obtain the attitude of the laser.

[0031] The feedback control system transmits the data collected by the attitude sensor to the control system. The control system, based on the received workpiece data (such as CAD model data, real-time measured workpiece contour and thickness data), uses a preset control algorithm (such as an algorithm based on model predictive control) to calculate the ideal position and angle of the laser. The control system sends control commands to the motor drive system according to the calculated control quantity, driving the servo motors to act, so that the laser is adjusted to the target position and angle, ensuring that the laser beam acts on the workpiece surface in the best state.

[0032] As shown Figure 4As shown, the machine vision system, as one of the core technologies, utilizes a high-resolution, large field-of-view industrial camera 5, in conjunction with multi-spectral bar light sources 6 and ring light source 9 illumination technologies, to collect real-time surface images of the workpiece 8 from multiple angles. Using deep learning image processing algorithms, the system can automatically identify features such as the geometric shape, surface defects, and material texture of the workpiece. At the same time, by combining structured light three-dimensional measurement technology, it can accurately obtain the thickness distribution information of the workpiece. Further, the control system can, according to the existing working conditions, dynamically adjust the spectrum, intensity, and angle of the light source to ensure clear and accurate image data can be obtained for workpieces with different materials, colors, and surface roughnesses, providing a reliable basis for subsequent cutting control.

[0033] As Figure 4 shown, the multi-spectral light source illumination technology integrates light sources of multiple different wavelengths, such as visible light (400 - 700nm), near-infrared light (700 - 2500nm), etc. For workpieces of different materials, the system will automatically select an appropriate light source combination. Further, the decision-making basis of the system comes from the judgment of data on different materials. For example, for metal workpieces with strong surface reflection, the proportion of near-infrared light can be increased to reduce the impact of reflection on image acquisition; for non-metal workpieces with complex surface textures, a multi-band light source combination can be used to enhance the contrast of texture features.

[0034] Further, the industrial camera is equipped with a high-resolution image sensor and, in conjunction with a wide-angle lens, can obtain surface images of the workpiece with a large field of view. The frame rate of the camera is set according to the cutting speed and accuracy requirements. During high-speed cutting, the frame rate can be increased to ensure real-time images; during high-precision cutting, the frame rate can be appropriately reduced to improve image quality.

[0035] As Figure 5 shown, the deep learning image processing algorithm is constructed based on the convolutional neural network (CNN) architecture. During the training phase, a large amount of image data containing different workpiece shapes, defects, and material textures is used for training. The network automatically extracts features in the image through components such as convolutional layers, pooling layers, and fully connected layers. The structured light three-dimensional measurement technology projects specific structured light patterns, such as Gray code patterns, sine stripe patterns, etc. When the pattern is projected onto the workpiece surface, due to the height changes on the workpiece surface, the pattern will be deformed. The industrial camera takes pictures of the deformed pattern from different angles. Using the principle of triangulation, by calculating the geometric relationship between the camera and the projector and the degree of pattern deformation, the three-dimensional coordinates of each point on the workpiece surface can be accurately calculated, and then the thickness distribution information of the workpiece can be obtained.

[0036] As Figure 6As shown, the thickness detection and feedback control system is equipped with a high-precision laser ultrasonic thickness detection device to solve the problem of uneven thickness of special-shaped workpieces. The system uses lasers to excite ultrasonic waves to propagate inside the workpiece. By receiving and analyzing the ultrasonic echo signals, the thickness of different positions of the workpiece can be accurately measured. The detected data is transmitted to the control unit in real time. The control unit uses an adaptive control algorithm to automatically adjust the laser power, pulse frequency, and cutting speed according to the thickness change. For example, when cutting a pressure vessel plate with variable thickness, the system can dynamically adjust the laser parameters according to the thickness change to ensure the consistency of cutting quality and avoid cutting defects.

[0037] Further explanation: The working principle of the laser ultrasonic thickness detection device is based on the thermoelastic excitation effect. The laser emitter emits short-pulse lasers, which are focused on a very small area on the surface of the workpiece. This area rapidly heats up and expands under the action of laser energy, thus generating ultrasonic waves. The ultrasonic waves propagate inside the workpiece and will be reflected and refracted when encountering different medium interfaces (such as the interface between different thickness regions). The ultrasonic receiver is installed at a suitable position to receive the reflected echo signals and convert them into electrical signals.

[0038] Further explanation: The signal processing unit amplifies, filters, and digitizes the electrical signals transmitted from the ultrasonic receiver. By accurately measuring the propagation time of the echo signals and combining the known propagation speed of ultrasonic waves in the workpiece material, the thickness of the workpiece is calculated using the formula (where d is the thickness, v is the sound speed, and t is the propagation time). The adaptive control algorithm adopts the model predictive control (MPC) strategy. This algorithm first establishes a prediction model based on the material of the workpiece, the initial thickness, and the preset cutting quality target. During the cutting process, the algorithm predicts the optimal adjustment values of the laser power, pulse frequency, and cutting speed in the next period of time according to the real-time measured thickness data and the prediction model to ensure the consistency of cutting quality.

[0039] As Figure 7 shown, the adaptive path planning and control algorithm, based on deep learning and intelligent optimization algorithms, the system deeply analyzes the obtained workpiece geometry, surface features, and thickness data to calculate the globally optimal cutting path. Further explanation: The algorithm not only considers the static geometric shape of the workpiece but also combines dynamic factors in the machining process in real time, such as cutting heat influence, workpiece micro-deformation, etc., to dynamically adjust the cutting path. By establishing a digital twin model of the cutting process, the cutting path is pre-played and optimized to ensure that the laser cutting head always moves along the optimal path, improving cutting efficiency and accuracy.

[0040] Furthermore, the deep learning algorithm first extracts features from the geometric shape, surface features, and thickness data of the workpiece. The convolutional neural network is used to process the two-dimensional image data of the workpiece to extract features such as edges and textures in the image; for the thickness data, a dedicated deep neural network structure is used for feature learning. These feature data are input into an intelligent optimization algorithm, such as the particle swarm optimization (PSO) algorithm or the simulated annealing algorithm. During the machining process, dynamic factors such as the temperature distribution in the cutting heat-affected area and the micro-deformation of the workpiece are monitored in real time by sensors. When a change in the dynamic factors is detected, the optimization algorithm is restarted to dynamically adjust the cutting path. For example, if it is found that the temperature in the cutting heat-affected area is too high, which may cause deformation of the workpiece, the algorithm will adjust the cutting path to avoid the area with too high temperature, or adjust the cutting speed and laser power to reduce the heat impact. As Figure 8 shown, the actual machining process corresponds one-to-one with the virtual model machining process, which is conducive to the monitoring and control of the actual machining situation. The digital twin model constructs a model in the virtual space that is exactly the same as the actual cutting process through the real-time collected device operation data (such as laser power, cutting speed, laser position, etc.) and machining data (such as cutting depth, cutting quality detection data, etc.). By simulating the digital twin model, the cutting results under different cutting paths and parameter settings can be predicted, providing a reference basis for path planning and parameter adjustment.

[0041] As Figure 9 shown, for the vibration compensation mechanism, aiming at the vibration problem generated during the machining of heavy workpieces, high-precision vibration sensors are installed in the system to monitor the vibration states of the machine tool, workpiece, and laser in real time. The active vibration control technology is adopted. When vibration is detected, the system quickly adjusts the motion trajectory, laser power, and cutting speed of the laser, and effectively compensates for the impact of vibration on the cutting quality through reverse excitation or dynamic damping adjustment. At the same time, the structural design and dynamic performance of the machine tool are optimized to improve the anti-vibration ability of the machine tool and ensure the stability of the cutting process.

[0042] Furthermore, the high-precision vibration sensor adopts a piezoelectric vibration sensor or an acceleration sensor, which is installed at key parts of the machine tool, such as the workbench, spindle, and machine body, as well as on the laser and the workpiece. These sensors can monitor information such as the amplitude, frequency, and direction of vibration in real time. The active vibration control technology adopts an electromagnetic or piezoelectric actuator. When the vibration sensor detects a vibration signal, the control system sends a control signal to the actuator according to the characteristics of the vibration. For example, the electromagnetic actuator generates an electromagnetic force opposite to the vibration direction to offset the influence of vibration; the piezoelectric actuator uses the inverse piezoelectric effect of piezoelectric materials to generate a displacement opposite to the vibration direction, thereby compensating for the vibration.

[0043] Further explanation: the machine vision system plays a key role in the entire cutting process, providing accurate data support for cutting control through image acquisition, processing and analysis. Industrial cameras and multi-spectral light sources are used to collect images of the workpiece surface under different lighting conditions. Optical image stabilization and autofocus technology are used to ensure image clarity and stability. The collected original images are pre-processed by filtering, denoising, grayscale conversion and other pre-processing operations to remove noise interference and enhance image contrast. Image enhancement algorithms are used to highlight the edges and feature information of the workpiece, providing high-quality image data for subsequent shape and thickness recognition.

[0044] Identification of geometric shape and thickness: The geometric shape and contour features of the workpiece are identified through edge detection, contour extraction and shape matching algorithms. Combined with multi-view image fusion technology, the three-dimensional shape information of the workpiece is obtained. Using the principle of structured light measurement, the thickness distribution of the workpiece is calculated based on the deformation of the structured light pattern projected on the surface of the workpiece. For complex and special-shaped workpieces, a semantic segmentation algorithm based on deep learning is used to accurately identify the various parts of the workpiece and the thickness change area, providing an accurate basis for cutting path planning and parameter adjustment.

[0045] Further explanation: in edge detection, the Canny edge detection algorithm is used, which extracts edges through the following steps: first, the image is denoised by Gaussian filtering; then the gradient amplitude and direction of the image are calculated; then non-maximum suppression is performed to refine the edges; finally, the edges are detected and connected by double thresholds to obtain the final edge image. In contour extraction, a contour search algorithm is used to extract the contour of the workpiece based on the edge image. For the semantic segmentation algorithm based on deep learning, taking the U-Net network as an example, its network structure consists of an encoder and a decoder. The encoder gradually extracts the features of the image through convolutional layers and pooling layers, and the decoder restores the low-resolution feature map to a segmentation result of the same size as the original image through upsampling and convolution operations.

[0046] Real-time feedback and control transmits the identified geometric shape and thickness data to the control system in real time. Based on these data and the cutting process requirements, the control system automatically calculates and adjusts the various parameters of laser cutting, such as laser power, cutting speed, focal position, etc. During the cutting process, the machine vision system continuously monitors the workpiece status and provides real-time feedback. The control system dynamically adjusts the cutting parameters based on the feedback information to ensure the accuracy and stability of the cutting process. Assume that based on the geometric shape and thickness data of the workpiece, the relationship between the laser power P and the cutting speed v calculated by the empirical formula or machine learning model is: Among them, k1, k2, and k3 are coefficients obtained by fitting experimental data, and d is the thickness of the workpiece. The control system calculates the required laser power P based on the real-time monitored workpiece thickness d and the preset cutting speed v, and sends control instructions to adjust the parameters of the laser cutting device.

[0047] Adaptive control feedback mechanism. The adaptive control system of the present invention realizes the dynamic adjustment and optimization of the cutting process through multi-source data fusion and intelligent algorithms. According to the real-time thickness, material properties, and cutting process requirements of the workpiece, the system automatically adjusts the laser power and cutting speed using a fuzzy control algorithm. For areas with a larger thickness or difficult-to-cut materials, the laser power is appropriately increased and the cutting speed is reduced to ensure the cutting depth and quality; for thinner areas, the laser power is reduced and the cutting speed is increased to avoid over-cutting. At the same time, considering the heat accumulation and heat-affected zone during the cutting process, the laser power and cutting speed are dynamically adjusted to reduce thermal deformation and thermal influence. The fuzzy control algorithm first establishes a fuzzy control rule base according to the real-time thickness of the workpiece, material properties (such as the hardness and thermal conductivity of metals, etc.), and preset cutting process rules. For example, if the workpiece has a large thickness and the material is a high-hardness metal, the rule base may set rules to increase the laser power and reduce the cutting speed. During the actual cutting process, the real-time thickness d and material information (represented by parameters such as hardness H) of the workpiece are obtained through sensors, and these accurate data are fuzzified. The thickness data is divided into fuzzy sets such as "thin", "medium", and "thick", and the material hardness information is divided into fuzzy sets such as "soft", "medium", and "hard". For example, for the thickness d, the fuzzification function can be defined as: where, is a threshold set according to the actual situation). Similarly, the material hardness is fuzzified. Then, based on the fuzzified inputs, reasoning is performed in the fuzzy control rule base to obtain a fuzzy control output, such as "increase power more", "reduce speed a little", etc. Finally, through defuzzification, the fuzzy output is converted into an accurate control quantity, such as the specific increased value of the laser power and the reduced value of the cutting speed, so as to achieve the accurate adjustment of the laser power and cutting speed. At the same time, the system monitors the heat accumulation and heat-affected zone situation during the cutting process through a temperature sensor. If the temperature of the heat-affected zone is too high, indicating a large heat accumulation, the system will appropriately reduce the laser power or increase the cutting speed to reduce thermal deformation and thermal influence.

[0048] Dynamic adjustment of cutting path. The adaptive path planning algorithm dynamically adjusts the cutting path according to real-time feedback data, combining the real-time shape changes of the workpiece and the cutting state. When encountering surface defects, thickness mutations or minor deviations during the cutting process, based on the three-dimensional topography data and cutting force sensor signals of the real-time feedback, the algorithm re-plans the cutting path to bypass the defective area or adjust the cutting direction to ensure the cutting quality. By real-time monitoring and predicting the thermal deformation and stress distribution during the cutting process, the cutting path is adjusted in advance to reduce deformation and stress concentration. During the operation of the adaptive path planning algorithm, it continuously receives real-time feedback data from the machine vision system, thickness detection system and various sensors during the cutting process.

[0049] Furthermore, when encountering surface defects of the workpiece, such as detecting cracks, sand holes, etc., the algorithm first analyzes information such as the location, size and shape of the defects to judge the degree of influence on the cutting quality. If the defect is small and does not affect the overall structural strength, the algorithm may choose to fine-tune the cutting path to make the laser beam avoid the defective area; if the defect is large, a completely new cutting path is re-planned to ensure the quality of the cut workpiece. For the case of thickness mutation, the algorithm will adjust the cutting speed and laser power according to the amplitude and location of the mutation, combined with the laser cutting ability, and re-plan the cutting path at the same time to ensure stable and high-quality cutting in the thickness change area. For example, when it is detected that the workpiece thickness suddenly increases, appropriately increase the laser power and reduce the cutting speed, and adjust the cutting path so that the laser beam acts on the workpiece at a more appropriate angle and trajectory.

[0050] Adjustment of laser angle and position. Using vibration sensors and attitude sensors, the vibration, tilt and other states of the machine tool and the workpiece are monitored in real time. When an abnormality is detected, the system automatically adjusts the angle and position of the laser through a high-precision servo control system. Using intelligent control algorithms, the optimal adjustment amount of the laser is calculated based on the vibration and tilt data to ensure that the laser beam is always perpendicular to the workpiece surface or irradiates along the optimal cutting path, compensating for errors caused by workpiece imbalance or vibration.

[0051] Further explanation: After receiving the sensor data, the high-precision servo control system activates the intelligent control algorithm. Based on the vibration and tilt data, the intelligent control algorithm uses mathematical models and algorithms to calculate the angles and positions that the laser needs to be adjusted. For example, the adaptive PID control algorithm is adopted to dynamically adjust the parameters of the PID controller according to the current vibration and tilt conditions to achieve precise control of the laser. After calculating the optimal adjustment amount, the servo control system sends control commands to the drive motor of the laser, and the drive motor precisely adjusts the angles and positions of the laser. During the adjustment process, the system continuously feeds back the actual position and angle information of the laser to ensure the accuracy of the adjustment. Through such a closed-loop control, it is ensured that the laser beam is always perpendicular to the workpiece surface or irradiates along the optimal cutting path, effectively compensating for errors caused by workpiece imbalance or vibration and improving the cutting accuracy.

[0052] Scanning of the workpiece and acquisition of geometric information: The workpiece is scanned in all directions using a lidar and an industrial camera to obtain high-precision point cloud data. Through point cloud processing algorithms, characteristic information such as the outer contour, surface defects, and cracks of the workpiece is extracted. Combining laser interferometry technology, the thickness of different positions of the workpiece is accurately measured to generate a three-dimensional digital model of the workpiece. During the scanning process, the clamping position and posture of the workpiece are automatically identified to provide accurate geometric data for subsequent path planning and cutting control. The lidar measures the distance information of each point on the workpiece surface by emitting laser beams and receiving reflected light based on the Time-of-Flight (ToF) principle.

[0053] Path planning and dynamic adjustment of cutting parameters: Based on the three-dimensional model of the workpiece and the cutting process requirements, a path planning method combining genetic algorithms and neural networks is adopted to automatically generate the optimal cutting path. During the path planning process, the shape, thickness, material of the workpiece and the performance of the cutting equipment are fully considered to optimize the cutting sequence and direction. According to the thickness data and cutting status collected in real time, the cutting parameters such as laser power, cutting speed, and focus position are dynamically adjusted to ensure the consistency and stability of the cutting quality.

[0054] Further explanation: In the genetic algorithm, the cutting path is encoded as a chromosome. For example, the integer encoding method is adopted, and each gene represents a point passed by the cutting path. The initial population is composed of randomly generated chromosomes. Calculate the fitness value of each chromosome (i.e., the cutting path plan), and the fitness function can be comprehensively determined according to factors such as cutting path length, cutting time, and cutting quality. For example, the fitness function Fitness can be defined as: where L is the cutting path length, T is the cutting time, and Q is the cutting quality evaluation value (which can be obtained by calculating the cutting surface roughness, perpendicularity, etc.). is the weight coefficient, which is adjusted according to actual requirements. The roulette wheel selection method is used for the selection operation, and the probability of each chromosome being selected is proportional to its fitness value. The partially matched crossover (PMX) is used for the crossover operation, and partial gene segments of two randomly selected chromosomes are exchanged, and the gene conflict problem is processed. The mutation operation randomly changes the gene values in the chromosome with a certain probability. The neural network is used to learn the complex relationship between the shape, thickness, material of the workpiece and the performance of the cutting equipment and the optimal cutting parameters. The neural network is trained with a large amount of sample data so that it can accurately predict the best cutting path and parameters under different conditions. During the cutting process, the thickness data d of the workpiece and the cutting state information (such as cutting temperature T, cutting force F, etc.) are collected in real time. According to these real-time data, combined with the preset cutting process rules and models, the cutting parameters are dynamically adjusted. For example, when the cutting temperature is too high, the cutting speed v is adjusted using the following formula: where is the initial cutting speed, is the normal cutting temperature, is the adjustment coefficient. When the thickness of the workpiece changes, the laser power P is adjusted according to the formula: where is the initial laser power, is the preset thickness, is the coefficient related to the material of the workpiece.

[0055] Automatic adjustment and vibration compensation of the laser. The laser is equipped with a high-precision automatic adjustment mechanism, which automatically adjusts the angle, position and focal length of the laser according to the real-time state of the workpiece and the cutting path. The magnetic levitation drive technology is adopted to improve the response speed and positioning accuracy of the laser. The vibration of the machine tool and the workpiece is monitored in real time through vibration sensors, and a method combining active vibration damping and passive vibration damping is used to effectively compensate for the influence of vibration on the cutting accuracy. For example, when cutting large ship components, the system can adjust the laser in real time to ensure that the cutting accuracy is not affected by vibration. The automatic adjustment mechanism of the laser consists of a motor, a guide rail and a transmission device.

[0056] Real-time monitoring and feedback system during the cutting process. A variety of sensors are installed in the cutting area to collect data such as cutting temperature, cutting force, and smoke concentration in real time. Through data analysis and machine learning algorithms, a state monitoring model of the cutting process is established to evaluate the cutting quality and the operating state of the equipment in real time. When an abnormal situation is detected, the system automatically issues an alarm, and adjusts the cutting parameters or pauses the cutting according to the preset strategy to avoid producing defective products and equipment damage. At the same time, the data during the cutting process is stored and analyzed to provide a basis for subsequent process optimization and equipment improvement.

[0057] Further explanation: A thermocouple sensor is used to measure the cutting temperature T. By measuring the thermoelectromotive force and calculating the temperature according to the Seebeck effect. The cutting force sensor uses a strain gauge sensor to convert the change in force into a change in resistance value, and then measures the cutting force F. The smoke concentration sensor uses the optical principle to measure the smoke concentration C based on the scattering or absorption characteristics of light. A cutting quality evaluation model is established using the Support Vector Machine (SVM) algorithm.

[0058] System automation and automatic control: The system realizes fully automated operation. The operator only needs to input the basic information and processing requirements of the workpiece, and the system can automatically complete the whole process from scanning, path planning to cutting. By using artificial intelligence technology, the system can intelligently adjust the cutting parameters and path planning strategy according to historical data and real-time feedback, and continuously optimize the cutting process. With the functions of fault diagnosis and self-repair, the system can automatically detect equipment failures, and through remote monitoring and intelligent diagnosis technology, guide the operator to carry out repairs or automatically repair, improving the reliability and stability of the equipment. The system's automated operation is achieved through a Human-Machine Interface (HMI). The operator inputs the material M, dimensions , shape description S, and processing requirements (such as cutting accuracy A, surface quality Q, etc.) of the workpiece on the HMI. After receiving this information, the system automatically starts the scanning program, and uses lidar and industrial cameras to scan the workpiece to obtain geometric information. Based on the obtained information, the system automatically performs path planning and cutting parameter settings.

[0059] Further explanation: The path planning adopts a method combining genetic algorithm and neural network, as described above. The cutting parameter settings are obtained by querying the process database or using a machine learning model prediction according to the workpiece material and processing requirements. For example, for a certain material M and cutting accuracy A, the prediction formula for the laser power P is: where g is the function relationship obtained through training. During the cutting process, the system uses deep learning algorithms, such as Convolutional Neural Network (CNN) or Recurrent Neural Network (RNN), to intelligently adjust the cutting parameters and path planning strategy according to historical cutting data and real-time feedback of cutting state information. For example, RNN processes time series data, learns the dynamic change rules during the cutting process, predicts the future cutting state, and thus adjusts the parameters in advance. The system has the functions of fault diagnosis and self-repair. By installing sensors at key parts of the equipment, the running state of the equipment is monitored in real time, such as the rotational speed n and temperature of the motor, and the power of the laser source , etc.

[0060] Further explanation, when a device failure is detected, a diagnostic method based on Fault Tree Analysis (FTA) and neural network is used to quickly locate the cause of the failure. For example, if the motor temperature is too high and the rotational speed is abnormal or , through fault tree analysis, it is judged that it may be a motor heat dissipation problem or an overload problem, and then the neural network is used to further determine the specific cause of the failure. For some simple faults, the system can automatically repair them, such as by adjusting parameters or restarting relevant components; for complex faults, the system sends fault information to the operator through a remote monitoring system and provides maintenance guidance to help the operator quickly repair the device.

[0061] By adopting a processing mode of workpiece fixation and laser rotation, the present invention can avoid the problem of center of gravity offset caused by workpiece rotation. The laser realizes spatial positioning through a multi-degree-of-freedom rotation mechanism, and is equipped with a dynamic center of gravity correction system and an efficient vibration absorption device to ensure the stability and accuracy of the cutting process.

[0062] The present invention introduces an advanced real-time thickness detection system, combined with laser power adaptive adjustment technology, to achieve dynamic response to workpieces with variable thickness. The thickness information of the workpiece is obtained through sensors, and the laser cutting parameters are adjusted in real time using an efficient feedback algorithm to ensure the consistency of cutting depth and width, and avoid the phenomena of over-cutting and under-cutting in traditional technologies.

[0063] The laser mentioned in the present invention adopts a multi-degree-of-freedom rotation mechanism, which can flexibly adjust the angle and position to achieve full coverage of special-shaped workpieces. Whether it is a complex curved surface, an edge area or a special structure, high-quality cutting can be achieved through the adaptive movement of the laser, greatly improving the adaptability and processing flexibility of the device.

[0064] To balance processing efficiency and cost, the present invention adopts a modular structure in the design, reduces manufacturing and maintenance costs by simplifying mechanical components. At the same time, the laser cutting path planning and control algorithm are optimized, greatly improving the processing efficiency. In addition, many innovative functions in the device are optimized and integrated based on existing mature technologies, further reducing the economic pressure of the overall equipment development and application.

[0065] The design of the present invention has high versatility, and is especially suitable for processing heavy special-shaped workpieces in the fields of oil pipelines, shipbuilding industry and aerospace. Its high adaptability and high performance can meet the diverse needs of different industries for the processing of complex workpieces, bringing significant technological progress and economic benefits to the industrial manufacturing field.

[0066] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, as well as the combination of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or means for implementing the functions specified in multiple blocks.

[0067] In the description of the present invention, it should be understood that the terms "first" and "second" are used for descriptive purposes only and cannot be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the present invention, "a plurality of" means two or more, unless otherwise specifically defined.

[0068] Obviously, those skilled in the art can make various modifications and variations to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these modifications and variations.

Claims

1. A laser adaptive cutting method for heavy-duty deformed steel pipes with variable wall thickness, characterized in that, Including the following steps: Based on the surface image of the workpiece collected, identify the geometric shape, surface defects and material texture features of the workpiece, and obtain the thickness distribution information of the workpiece; Based on the thickness distribution information, obtain the thickness data at different positions of the workpiece; Based on the geometric shape, the surface defects, the material texture features and the thickness data, generate a globally optimal cutting path through an intelligent algorithm; Based on the optimal cutting path, perform a laser cutting task, and automatically adjust the laser power, pulse frequency and cutting speed according to the thickness change. At the same time, when vibration is detected, compensate for the vibration through reverse excitation or dynamic damping adjustment.

2. The laser adaptive cutting method for a heavy-duty deformed-thickness steel pipe according to claim 1, wherein: When collecting the surface image of the workpiece, by dynamically adjusting the spectrum, intensity and angle of the light source, collect the surface image on workpieces with different materials, colors and surface roughnesses. Among them, a multi-band light source combination is adopted to enhance the contrast of texture features and obtain the material texture features.

3. The laser adaptive cutting method for a heavy-duty deformed-thickness steel pipe according to claim 2, wherein: When collecting the surface image, set the image acquisition frame rate according to the cutting speed and accuracy requirements.

4. The laser adaptive cutting method for a heavy-duty deformed-thickness steel pipe according to claim 3, wherein: When obtaining the thickness distribution information of the workpiece, project a pattern onto the surface of the workpiece, obtain the deformed patterns of the workpiece at different angles, and use the principle of triangulation. By calculating the geometric relationship between the collected points of the deformed pattern and the pattern projection points and the degree of pattern deformation, obtain the three-dimensional coordinates of each point on the surface of the workpiece and generate the thickness distribution information of the workpiece.

5. The laser adaptive cutting method for a heavy-duty deformed-thickness steel pipe according to claim 4, wherein: When obtaining the thickness data at different positions of the workpiece, based on the thickness distribution information, use laser to excite ultrasonic waves to propagate inside the workpiece, and measure the thickness of the workpiece at different positions by receiving and analyzing the ultrasonic echo signals.

6. The laser adaptive cutting method for a heavy-duty deformed-thickness steel pipe according to claim 5, wherein: When obtaining the optimal cutting path, based on the geometric shape, the surface defects, the material texture features and the thickness data, establish a digital twin model of the cutting process, pre-play and optimize the cutting path, and aim to ensure the cutting efficiency and accuracy to obtain the optimal cutting path.

7. The laser adaptive cutting method for a heavy-duty deformed-thickness steel pipe according to claim 6, wherein: When performing a laser cutting task, according to the real-time thickness, material characteristics of the workpiece and the requirements of the cutting process, the laser power and cutting speed are automatically adjusted by using a fuzzy control algorithm. Among them, the fuzzy control algorithm is used to establish a fuzzy control rule base according to the real-time thickness, material characteristics of the workpiece and the preset cutting process rules; reason in the fuzzy control rule base according to the fuzzified input to obtain a fuzzy control output; through defuzzification processing, convert the fuzzy output into a control quantity.

8. The laser adaptive cutting method for heavy-duty deformed-thickness steel pipes according to claim 7, characterized in that: During the process of performing the laser cutting task, when encountering surface defects, thickness mutations or minor deviations during the cutting process of the workpiece, based on the real-time feedback of three-dimensional topography data and the signals of the cutting force sensors, the cutting path is dynamically adjusted to bypass the defective area or adjust the cutting direction. Among them, for the case where the defect is small and does not affect the overall structural strength, the cutting path is finely adjusted to avoid the defective area with the laser beam; for the case where the defect is large, a completely new cutting path is re-planned; for the case of thickness mutation, according to the amplitude and position of the mutation, combined with the capabilities of laser cutting, the cutting speed and laser power are adjusted, and the cutting path is re-planned at the same time.

9. The laser adaptive cutting method for heavy-duty deformed-thickness steel pipes according to claim 8, characterized in that: When planning the cutting path, based on the three-dimensional model of the workpiece and the requirements of the cutting process, a path planning method combining genetic algorithm and neural network is used to plan the cutting path. Among them, in the genetic algorithm, the cutting path is encoded as a chromosome, the fitness value of each chromosome is calculated, and the roulette wheel selection method is used for the selection operation, so that the probability of each chromosome being selected is proportional to its fitness value; and partially matched crossover is adopted, and some gene segments of two randomly selected chromosomes are exchanged, and the gene conflict problem is processed; the neural network is used to predict the best cutting path and parameters under different conditions by learning the complex relationship between the shape, thickness, material of the workpiece and the performance of the cutting equipment and the optimal cutting parameters.

10. A laser adaptive cutting system for heavy-duty deformed steel pipes with variable wall thickness, characterized in that, Including: A data acquisition and analysis module, used to identify the geometric shape, surface defects and material texture features of the workpiece according to the collected surface image of the workpiece, and obtain the thickness distribution information of the workpiece; A thickness detection module, used to obtain the thickness data of different positions of the workpiece based on the thickness distribution information; A path planning module, used to generate a globally optimal cutting path through an intelligent algorithm based on the geometric shape, the surface defects, the material texture features and the thickness data; A cutting task execution module, used to perform a laser cutting task based on the optimal cutting path, and automatically adjust the laser power, pulse frequency and cutting speed according to the thickness change; A vibration compensation module, used to compensate for the vibration by reverse excitation or dynamic damping adjustment when vibration is detected; A quality evaluation module, used to evaluate the cutting quality by obtaining the cutting temperature, cutting force and smoke concentration when performing the laser cutting task; A fault diagnosis module, which is used to quickly locate the cause of a fault by using a diagnosis method based on Fault Tree Analysis (FTA) and neural network after detecting a fault in the device.

Images