A method and system for laser additive path control of a steel rail

Abstract

The present application relates to the technical field of laser additive manufacturing, and provides a rail laser additive path control method and system, the method comprising: monitoring and acquiring a real-time topographic image of a molten pool; processing the real-time topographic image to obtain an asymmetric feature of the molten pool; calculating an asymmetric distortion degree of the molten pool according to the asymmetric feature; controlling the laser beam to perform asymmetric micro-motion according to the asymmetric distortion degree; the asymmetric micro-motion comprises adjusting the scanning speed or the residence time of the laser beam in different regions of the molten pool to reconstruct the energy distribution inside the molten pool; adjusting the parameters of the asymmetric micro-motion of the laser beam according to the real-time topographic image corresponding to the molten pool after the energy distribution inside the molten pool is reconstructed to realize rail laser additive path control. The present application has the effect of improving the quality, geometric accuracy and mechanical properties of the laser additive repair layer of the rail.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of laser additive manufacturing, and in particular to a rail laser additive path control method and system. BACKGROUND

[0002] In the railway transportation system, the rail is a key infrastructure for carrying the train operation. Its long-term service will be damaged due to wear and tear, especially in the high-load area such as the curve. Laser additive manufacturing technology is widely used in the online repair of rails due to its accuracy and efficiency, to restore its geometric shape and prolong its service life. However, in the actual repair process, due to the unevenness of the rail surface pretreatment and the introduction of specific operations in the laser additive process, the molten pool shape may change asymmetrically, thereby affecting the quality of the repair layer and even causing the accumulation of defects.

[0003] Towards the side of the laser head tilt direction, due to the existence of the protrusion, the actual point of the laser beam will become closer to the laser focus, causing the local energy density to be abnormally concentrated, thereby causing the "over-melting" phenomenon, that is, the molten pool on this side becomes abnormally wide or deeply sunk. At the same time, the other side away from the laser head tilt direction may be blocked by the protrusion or cause the distance between the laser beam and the workpiece surface to suddenly increase, causing the energy density to drop sharply, and the powder feeding efficiency may also be affected, thereby appearing "unfused" defects, that is, the new material cannot be fully combined with the lower solidified layer or substrate. This phenomenon is caused by the local unevenness of laser energy and powder delivery, as well as the complex response of the internal fluid dynamics of the molten pool under asymmetric energy input.

[0004] This initial small defect caused by the unevenness of the common pretreatment is enlarged layer by layer under the standard repair process with an inclination angle, and finally evolves into a specific serious forming defect: excessive accumulation on one side of the additive layer, and an unfused area on the other side. The characteristics of this defect are its asymmetry and accumulation. Traditional quality control methods, such as monitoring only a single parameter such as the overall width, length or surface temperature of the molten pool, cannot accurately capture and correct this complex, localized molten pool shape distortion. Because even if the overall width or temperature looks normal, the asymmetry inside the molten pool may already exist.

[0005] The prior art needs to be improved in view of the above problems. SUMMARY

[0006] The present application discloses a rail laser additive path control method and system, which aims to solve the problem of asymmetric distortion of the molten pool caused by uneven rail surface pretreatment and laser inclined cladding during the rail laser additive repair process, thereby causing the quality of the repair layer to decline and defects to accumulate.

[0007] The technical solutions of the present application are as follows:

[0008] In a first aspect, the present application discloses a steel rail laser additive path control method, comprising:

[0009] Monitoring and acquiring a real-time topographic image of the molten pool;

[0010] Processing the real-time topographic image to obtain an asymmetric feature of the molten pool;

[0011] According to the asymmetric feature, the asymmetric distortion degree of the molten pool is calculated;

[0012] According to the asymmetric distortion degree, the laser beam is controlled to perform asymmetric micro-motion; the asymmetric micro-motion includes adjusting the scanning speed or residence time of the laser beam in different regions of the molten pool to reconstruct the energy distribution inside the molten pool;

[0013] According to the real-time topographic image corresponding to the molten pool after the energy distribution inside the molten pool is reconstructed, the parameters of the asymmetric micro-motion of the laser beam are adjusted to realize the steel rail laser additive path control.

[0014] Through the technical solutions, the asymmetric topography of the molten pool can be monitored in real time, and the micro-motion of the laser beam is dynamically adjusted according to the asymmetric distortion degree, so as to reconstruct the energy distribution inside the molten pool, effectively correct the asymmetric distortion of the molten pool, avoid the accumulation of defects, and significantly improve the quality and stability of the steel rail laser additive repair.

[0015] Further, in the above steel rail laser additive path control method, according to the asymmetric distortion degree, the laser beam is controlled to perform asymmetric micro-motion; the step of the asymmetric micro-motion including adjusting the scanning speed or residence time of the laser beam in different regions of the molten pool to reconstruct the energy distribution inside the molten pool includes:

[0016] According to the asymmetric distortion degree, an asymmetric composite swing trajectory is generated;

[0017] The laser beam is controlled to execute the asymmetric composite swing trajectory; the asymmetric composite swing trajectory has higher path density or longer laser residence time in the direction with higher energy demand;

[0018] According to the real-time topographic image corresponding to the molten pool after the energy distribution inside the molten pool is reconstructed, the parameters of the asymmetric micro-motion of the laser beam are adjusted to realize the steel rail laser additive path control.

[0019] According to the real-time topographic image corresponding to the molten pool after the energy distribution inside the molten pool is reconstructed, the geometric parameters of the asymmetric composite swing trajectory are adjusted.

[0020] By the technical scheme, the asymmetric distortion degree of the molten pool can be converted into specific laser swing trajectory parameters, the energy distribution inside the molten pool is accurately reconstructed by adjusting the path density or the residence time, thereby the asymmetric distortion is more effectively corrected, and the repair precision is improved.

[0021] On this basis, after the step of generating the asymmetric composite swing trajectory according to the asymmetric distortion degree, the method further comprises:

[0022] Obtaining rail surface curvature information or molten pool bottom topography information;

[0023] According to the rail surface curvature information or the molten pool bottom topography information, performing three-dimensional space correction on the asymmetric composite swing trajectory; the three-dimensional space correction comprises adjusting the defocusing amount change of the asymmetric composite swing trajectory in the direction perpendicular to the rail surface, and adjusting the projection shape and density of the asymmetric composite swing trajectory in the direction parallel to the rail surface, so as to complete the generation and correction of the asymmetric composite swing trajectory perpendicular to the advancing direction.

[0024] By the technical scheme, the laser swing trajectory can be finely corrected in three-dimensional space in combination with the actual curvature of the rail and the bottom topography of the molten pool, the uniform action of the laser energy on the complex surface is ensured, and the forming quality of the additive layer and the bonding strength with the base body are further improved.

[0025] Further, in the above rail laser additive path control method, the step of generating the asymmetric composite swing trajectory according to the asymmetric distortion degree comprises:

[0026] Obtaining local fluid characteristic parameters or transient micro-geometric change information inside the molten pool;

[0027] According to the local fluid characteristic parameters or the transient micro-geometric change information, generating the asymmetric composite swing trajectory and dynamically adjusting the geometric parameters of the asymmetric composite swing trajectory.

[0028] By the technical scheme, the fluid dynamics and the micro-geometric change inside the molten pool can be deeply analyzed, thereby the laser swing trajectory can be more accurately generated and dynamically adjusted, the real-time and fine control of the energy distribution inside the molten pool is realized, and the complex and changeable additive process is effectively coped with.

[0029] In some preferred embodiments, in the above rail laser additive path control method, the step of calculating the asymmetric distortion degree of the molten pool according to the asymmetric feature comprises:

[0030] Dividing a real-time topographic image of the molten pool into a first region and a second region located on both sides of a center line along the advancing direction of the laser beam;

[0031] Calculate the average brightness value of all valid pixels in the first region and the second region respectively to obtain the first average brightness and the second average brightness.

[0032] Calculate the asymmetry index data based on the first and second average brightness.

[0033] The degree of asymmetric distortion is judged based on asymmetric index data, and the result of the degree of asymmetric distortion is obtained.

[0034] When the asymmetric distortion degree judgment result indicates that the absolute value of the asymmetric index data is greater than the preset threshold, it is determined that there is asymmetric distortion in the molten pool, and the direction of energy compensation is determined according to the positive or negative sign of the asymmetric index data.

[0035] The magnitude of the asymmetric exponent data is used as a quantitative basis for adjusting the asymmetric microscopic motion amplitude and scanning speed of the laser beam.

[0036] Through this technical solution, the degree of asymmetric distortion of the molten pool can be quantified by image processing and brightness analysis, and the direction and magnitude of energy compensation can be determined. This provides a precise quantitative basis for the subsequent asymmetric microscopic motion of the laser beam, thereby achieving accurate correction of the distortion.

[0037] Based on the above, this application further proposes that, in the above-mentioned laser additive manufacturing path control method for steel rails, the steps for obtaining local fluid characteristic parameters or transient micro-geometric change information inside the molten pool include:

[0038] Obtain radiation images of the molten pool surface area at different wavelengths;

[0039] The spatial distribution differences and dynamic changes of molten pool radiance at different wavelengths were analyzed to obtain the fusion analysis results;

[0040] Based on the fusion analysis results, all radiation images at different wavelengths are fused to obtain local fluid characteristic parameters or transient micro-geometric change information inside the molten pool.

[0041] Through this technical solution, multi-wavelength radiation image analysis technology can be used to obtain deeper physical information inside the molten pool, such as local fluid characteristics and transient micro-geometric changes, thereby providing more comprehensive and accurate input for laser path control and further improving the precision and effectiveness of control.

[0042] More specifically, in some implementation schemes, the steps in the above-mentioned rail laser additive manufacturing path control method to analyze the spatial distribution differences and dynamic changes of molten pool radiance at different wavelengths and obtain the fusion analysis results include:

[0043] Obtain radiation images of the molten pool at several preset wavelengths; the preset wavelengths include wavelengths sensitive to temperature, wavelengths sensitive to component segregation, and wavelengths sensitive to bubbles or solidification fronts.

[0044] By dividing the radiation image into regions, we can obtain the characteristic regions corresponding to various physical phenomena inside the molten pool.

[0045] The variation trend and spatial correlation of radiance in the characteristic region under different preset wavelengths are analyzed to decouple the combined effects of various physical phenomena on radiance and obtain radiance characteristics.

[0046] Based on the radiance characteristics, the radiation characteristics corresponding to the preset physical phenomena are identified, and based on the radiation characteristics, the local fluid characteristic parameters or transient micro-geometric change information inside the molten pool are analyzed.

[0047] Through technical solutions, complex physical phenomena inside the molten pool, such as temperature field, component segregation, bubbles, and solidification front, can be identified by fine analysis and decoupling of multi-wavelength images. This allows for a more comprehensive understanding of the molten pool state and provides a deeper basis for intelligent adjustment of the laser path.

[0048] Preferably, in the above-mentioned laser additive path control method for rails, the step of analyzing the variation trend and spatial correlation of radiance in the characteristic region under different preset wavelengths to decouple the combined influence of various physical phenomena on radiance and obtain radiance characteristics includes:

[0049] Acquire radiance data of the feature region at different preset wavelengths;

[0050] The radiance data were normalized.

[0051] Calculate the correlation matrix of the normalized radiance data in the spatial dimension;

[0052] Calculate the rate of change of the normalized radiance data over time;

[0053] The correlation matrix and rate of change are used as radiance features;

[0054] The steps for identifying the radiation characteristics corresponding to a preset physical phenomenon based on radiation brightness characteristics include:

[0055] Principal component analysis was performed on the correlation matrix and rate of change in the radiance characteristics to extract independent feature components characterizing different physical phenomena.

[0056] Based on independent feature components, identify the radiation characteristics corresponding to the preset physical phenomena.

[0057] Through technical solutions, independent characteristic components representing different physical phenomena can be extracted from multi-wavelength radiation data by data processing and principal component analysis, thereby more accurately identifying the complex physical state inside the molten pool and providing more reliable data support for the precise control of the laser path.

[0058] Based on the above, this application further proposes that, in the above-mentioned rail laser additive path control method, the step of identifying the radiation characteristics corresponding to a preset physical phenomenon based on independent feature components includes:

[0059] Obtain the material system parameters, laser process parameters, and environmental parameters of the current laser additive manufacturing process;

[0060] Real-time monitoring of fluctuation ranges in material system parameters, laser process parameters, and environmental parameters;

[0061] Based on the real-time fluctuation range of independent feature components, material system parameters, laser process parameters, and environmental parameters, the mapping weights between independent feature components and preset physical phenomena are dynamically adjusted.

[0062] When the real-time fluctuation range of material system parameters, laser process parameters, and environmental parameters exceeds the preset range, the corresponding mapping weight is reduced.

[0063] Based on the adjusted mapping weights, the radiation characteristics corresponding to the preset physical phenomena are identified.

[0064] Through technical solutions, the mapping relationship between physical phenomena and radiation characteristics can be dynamically adjusted by combining the real-time fluctuations of actual process parameters and environmental parameters, thereby improving the robustness and accuracy of identifying physical phenomena inside the molten pool and ensuring precise control even under complex working conditions.

[0065] Secondly, this application also discloses a rail laser additive path control system for performing rail laser additive path control, including:

[0066] The morphology image acquisition module is used to monitor and acquire real-time morphology images of the molten pool;

[0067] The molten pool feature acquisition module is used to process real-time topography images to obtain the asymmetric features of the molten pool;

[0068] The distortion degree calculation module is used to calculate the degree of asymmetric distortion of the molten pool based on the asymmetric characteristics.

[0069] The laser motion control module is used to control the laser beam to perform asymmetric micro-motion based on the degree of asymmetric distortion. The asymmetric micro-motion includes adjusting the scanning speed or dwell time of the laser beam in different regions of the molten pool to reconstruct the energy distribution inside the molten pool.

[0070] The laser parameter adjustment module is used to adjust the parameters of the asymmetric micro-motion of the laser beam based on the real-time morphology image of the molten pool after the energy distribution inside the molten pool is reconstructed, so as to realize the path control of laser additive manufacturing of rails.

[0071] The technical solution provides a complete hardware and software system to automate and intelligently control the laser additive manufacturing process of rails, effectively solving the problem of asymmetric distortion of the molten pool, improving repair efficiency and quality, and reducing the need for manual intervention.

[0072] Beneficial effects

[0073] The laser additive manufacturing path control method for rails disclosed in this application monitors and acquires the morphological image of the molten pool in real time, then processes it to obtain the asymmetric features of the molten pool and calculates the degree of asymmetric distortion. Based on this degree of asymmetric distortion, this method can intelligently control the laser beam to perform asymmetric micro-motions, including adjusting the scanning speed or dwell time of the laser beam in different regions of the molten pool to accurately reconstruct the energy distribution inside the molten pool. After completing the energy distribution reconstruction, the system further adjusts the asymmetric micro-motion parameters of the laser beam according to the real-time morphological image of the molten pool, thereby achieving precise control of the laser additive manufacturing path for rails.

[0074] This method effectively solves the problem of asymmetric distortion of the molten pool caused by uneven pretreatment of the rail surface and tilted cladding in existing technologies. Traditional methods struggle to accurately capture and correct such localized and cumulative defects. However, this application, through quantitative analysis of the asymmetric characteristics of the molten pool and based on this, asymmetric micro-motion control of the laser beam, can compensate for energy imbalances within the molten pool in real time and dynamically, avoiding the layer-by-layer accumulation of defects such as "over-melting" and "lack of fusion." By reconstructing the energy distribution within the molten pool, this method ensures that the additive layer can form a uniform and dense structure even under complex curved surfaces and tilted cladding conditions. This significantly improves the quality, geometric accuracy, and mechanical properties of the laser additive repair layer for rails, overcoming the shortcomings of existing technologies in controlling molten pool morphological distortion, thereby extending the service life of the rails and reducing maintenance costs. Attached Figure Description

[0075] Figure 1 This is a flowchart of a method for controlling the path of laser additive manufacturing of steel rails in one embodiment of the present invention;

[0076] Figure 2 This is one of the flowcharts of a method for controlling the path of laser additive manufacturing of steel rails according to another embodiment of the present invention;

[0077] Figure 3 This is a second flowchart of a method for controlling the path of laser additive manufacturing of steel rails according to another embodiment of the present invention;

[0078] Figure 4This is a system block diagram of a rail laser additive manufacturing path control system according to another embodiment of the present invention;

[0079] Explanation of reference numerals in the attached figures:

[0080] 1. Rail laser additive manufacturing path control system; 11. Shape image acquisition module; 12. Melt pool feature acquisition module; 13. Distortion degree calculation module; 14. Laser motion control module; 15. Laser parameter adjustment module. Detailed Implementation

[0081] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0082] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0083] This application proposes a laser additive manufacturing path control method for steel rails, combining... Figure 1 As shown, it includes:

[0084] S1, monitor and acquire real-time morphological images of the molten pool;

[0085] S2, process the real-time topography image to obtain the asymmetric features of the molten pool;

[0086] S3, based on the asymmetric characteristics, calculate the degree of asymmetric distortion of the molten pool;

[0087] S4, based on the degree of asymmetric distortion, control the laser beam to perform asymmetric micro-motion; the asymmetric micro-motion includes adjusting the scanning speed or dwell time of the laser beam in different regions of the molten pool to reconstruct the energy distribution inside the molten pool;

[0088] S5, based on the real-time morphology image of the molten pool after the energy distribution inside the molten pool is reconstructed, adjusts the parameters of the asymmetric micro-motion of the laser beam to achieve path control of laser additive manufacturing of steel rails.

[0089] This application achieves precise control of the laser additive manufacturing path for rails by real-time monitoring of the molten pool morphology and identification of its asymmetric features, thereby calculating the degree of asymmetric distortion and dynamically adjusting the microscopic motion of the laser beam based on this.

[0090] The term "molten pool" as used in this application refers to the liquid metal region formed when a laser beam melts the surface of a material during laser additive manufacturing. The morphology, size, temperature distribution, and internal fluid dynamics of the molten pool directly affect the quality of the final formed part. "Real-time morphology images" refer to images of the molten pool surface continuously captured during the laser additive manufacturing process using devices such as vision sensors or infrared cameras. These images contain information such as the geometry, size, and surface fluctuations of the molten pool. "Asymmetric features" refer to the asymmetry of the geometric or physical properties exhibited by the molten pool in a certain direction (e.g., perpendicular to the laser scanning direction), such as inconsistent width, depth, or surface tension distribution on both sides of the molten pool. "Degree of asymmetric distortion" is a quantitative indicator of this asymmetry, used to assess the degree to which the molten pool deviates from an ideal symmetric state. Furthermore, "asymmetric micro-motion" refers to the refined, asymmetric scanning or dwell time adjustment of the laser beam in a localized area of ​​the molten pool, aimed at locally altering the energy input to correct the asymmetric distortion of the molten pool.

[0091] In the path control method for laser additive manufacturing of steel rails, the first step is to monitor and acquire real-time images of the molten pool's morphology. This can be achieved in several ways. For example, a high-speed camera with an appropriate illumination system can be used to directly capture visible light images of the molten pool. Another approach is to use an infrared thermal imager to acquire images of the temperature distribution on the molten pool surface, as temperature distribution is closely related to the molten pool's morphology. Alternatively, three-dimensional vision techniques such as laser triangulation or structured light projection can be used to acquire three-dimensional point cloud data of the molten pool surface, thereby reconstructing its precise morphology. These images or data streams are transmitted to the processing unit in real time, providing a foundation for subsequent analysis.

[0092] After acquiring the real-time topographic image, it needs to be processed to obtain the asymmetric features of the molten pool. One approach is to preprocess the image, such as denoising and contrast enhancement, and then use image segmentation algorithms (such as thresholding, edge detection, or deep learning segmentation models) to separate the molten pool region from the background. Next, the geometric center or centroid of the molten pool can be calculated and used as a reference to analyze the contour symmetry of the molten pool in different directions. For example, the width, area, or average gray value on both sides of the molten pool can be measured along the vertical axis of the laser beam's propagation direction to identify any significant differences. Another approach is to analyze the spatial distribution of the molten pool surface brightness to identify any locally overly bright or dark areas. These areas may indicate non-uniformity in energy input, thus reflecting asymmetric features.

[0093] Based on the obtained asymmetric features, the degree of asymmetric distortion of the molten pool needs to be calculated. This can be achieved by establishing a mathematical model. For example, an asymmetry index can be defined, which is calculated based on the difference in width, area, or brightness between the two sides of the molten pool. When the widths on both sides of the molten pool are W1 and W2, respectively, the degree of asymmetric distortion can be expressed as the absolute value of (W1-W2) / (W1+W2). The magnitude of this index reflects the severity of asymmetry, while its sign indicates the direction of asymmetry. Alternatively, a machine learning model can be used, taking various extracted asymmetric features as input, to train the model and output a quantified value of the degree of asymmetric distortion.

[0094] After calculating the degree of asymmetric distortion in the molten pool, the laser beam needs to be controlled to perform asymmetric micro-movements based on this distortion level. These asymmetric micro-movements include adjusting the scanning speed or dwell time of the laser beam in different regions of the molten pool to reconstruct the energy distribution within the pool. Specifically, if "over-melting" occurs on one side of the molten pool, resulting in excessive width, the scanning speed of the laser beam in that region can be reduced, or its dwell time increased, thereby reducing local energy input and causing the molten pool to shrink. Conversely, if "unfusion" occurs on one side, resulting in excessive narrowness, the scanning speed of the laser beam in that region can be increased, or its dwell time reduced, thereby increasing local energy input and causing the molten pool to expand. This adjustment can be periodic, for example, by changing the oscillation trajectory of the laser galvanometer to achieve higher path density or slower scanning speed in the area requiring compensation. It can also be transient, for example, by rapidly adjusting the laser power or focal position to achieve precise local energy control.

[0095] After reconstructing the energy distribution within the molten pool, the parameters of the asymmetric micro-motion of the laser beam need to be adjusted based on the real-time morphology image of the molten pool to achieve path control for laser additive manufacturing of the rail. This means that this method is a closed-loop control process. After the laser beam performs asymmetric micro-motion, the system acquires the real-time morphology image of the molten pool again and re-evaluates its asymmetric characteristics and distortion degree. If the molten pool still has asymmetric distortion, or the distortion degree does not meet expectations, the asymmetric micro-motion parameters of the laser beam (such as the adjustment range of the scanning speed, the increment of the dwell time, and the geometric parameters of the oscillation trajectory) will be further adjusted based on the new evaluation results. This iterative process continues until the molten pool morphology reaches an ideal symmetrical state, thereby ensuring the uniformity and quality of the additive layer.

[0096] Optional, combined Figure 2 As shown, S4 controls the laser beam to perform asymmetric micro-motion based on the degree of asymmetric distortion; the asymmetric micro-motion includes the steps of adjusting the scanning speed or dwell time of the laser beam in different regions of the molten pool to reconstruct the energy distribution inside the molten pool, including:

[0097] S41, based on the degree of asymmetric distortion, generate an asymmetric composite oscillation trajectory;

[0098] S42 controls the laser beam to execute an asymmetric composite oscillation trajectory; the asymmetric composite oscillation trajectory has a higher path density or a longer laser dwell time in the direction with higher energy demand.

[0099] The steps for adjusting the parameters of the asymmetric micro-motion of the laser beam to achieve path control in laser additive manufacturing of steel rails, based on the real-time topographic image of the molten pool after energy distribution reconstruction, include:

[0100] Based on the real-time topographic image of the molten pool after the energy distribution inside the molten pool is reconstructed, the geometric parameters of the asymmetric composite oscillation trajectory are adjusted.

[0101] Specifically, generating an asymmetric composite oscillation trajectory refers to designing and generating a laser beam motion path based on the degree of asymmetric distortion in the molten pool. This path is not a simple straight line or circular scan, but a composite trajectory combining multiple motion modes (e.g., straight lines, curves, periodic oscillations, etc.), whose shape and density are designed to specifically compensate for energy unevenness within the molten pool. The "asymmetry" is reflected in the fact that the trajectory's geometric characteristics (such as amplitude, frequency, shape, or direction) are adjusted according to the direction and degree of asymmetric distortion in the molten pool to ensure that energy is precisely replenished to areas with insufficient energy.

[0102] Furthermore, controlling the laser beam to execute an asymmetric composite oscillation trajectory means that the laser beam will scan strictly according to the preset asymmetric composite oscillation trajectory. The asymmetric composite oscillation trajectory has a higher path density or a longer laser dwell time in directions with higher energy demand. Path density can be understood as the length or number of laser beam scanning paths per unit area; a higher path density means that the area is irradiated by the laser beam more times or for a longer time, thus obtaining more energy input. Laser dwell time refers to the time the laser beam maintains irradiation on a specific area or point; the longer the dwell time, the more energy that area receives. In this way, precise control of the energy distribution within the molten pool can be achieved, concentrating more energy in areas requiring compensation.

[0103] Furthermore, based on the real-time topographic image of the molten pool after energy distribution reconstruction, the geometric parameters of the asymmetric composite oscillation trajectory are adjusted. This means that the method possesses real-time feedback and adaptive adjustment capabilities. After the laser beam executes the asymmetric composite oscillation trajectory and completes the initial energy reconstruction, the system acquires another real-time topographic image of the molten pool and evaluates the effect of energy reconstruction based on the new topographic image. If residual asymmetric distortion still exists in the molten pool or new distortions occur, the geometric parameters of the asymmetric composite oscillation trajectory, such as oscillation amplitude, frequency, shape, scanning speed, or path density, are dynamically adjusted to further optimize the energy distribution until the molten pool topography reaches an ideal symmetrical state.

[0104] This application's solution effectively addresses the challenge of achieving refined and dynamic energy reconstruction in traditional methods when faced with complex or rapidly changing asymmetric distortions in the molten pool by introducing an asymmetric composite oscillation trajectory. Specifically, when asymmetric distortion occurs in the molten pool, the system can generate a customized asymmetric composite oscillation trajectory based on the degree of distortion. The core design of this trajectory lies in specifically increasing the energy input in the direction of higher energy demand within the molten pool by increasing the laser beam path density or extending the laser dwell time. This proactive and precise energy compensation mechanism allows for effective control of the temperature and flow fields within the molten pool, thereby promoting the restoration of the molten pool morphology to a symmetrical state. Simultaneously, real-time monitoring and feedback of the molten pool morphology after energy reconstruction allows for dynamic adjustment of the oscillation trajectory's geometric parameters, ensuring the adaptability and robustness of the entire control process, thus achieving precise control of the laser additive manufacturing path for steel rails.

[0105] In some preferred embodiments, a specific example is given below. Suppose that during laser additive manufacturing of steel rails, a real-time topographic image of the molten pool shows that the left side of its path exhibits a depression or lower temperature due to insufficient energy caused by uneven heat dissipation or material flow characteristics. In this case, the system generates an asymmetric composite oscillation trajectory based on the detected degree of asymmetric distortion. For example, this trajectory could be an oscillation pattern with higher path density (e.g., denser zigzag scans or shorter oscillation periods) or longer laser dwell time (e.g., brief pauses or deceleration in the left side). As the laser beam executes this trajectory, it inputs more energy into the left side of the molten pool, thereby increasing the temperature in that area and promoting material flow, gradually restoring its symmetrical topography. After completing one energy reconstruction, the system acquires the molten pool topographic image again. If a slight depression is still found in the left side, the system dynamically adjusts the geometric parameters of the asymmetric composite oscillation trajectory based on the new topographic information, for example, further increasing the oscillation amplitude or path density in the left side until the molten pool topography reaches the ideal symmetrical state. In this way, precise and dynamic compensation for asymmetric distortion of the molten pool can be achieved.

[0106] Optionally, after the step of generating the asymmetric composite oscillation trajectory based on the degree of asymmetric distortion, this application further includes:

[0107] Obtain information on the surface curvature of the rail or the bottom morphology of the molten pool;

[0108] Based on the surface curvature information of the rail or the bottom morphology information of the molten pool, the asymmetric composite oscillation trajectory is corrected in three dimensions. The three-dimensional spatial correction includes adjusting the defocusing amount of the asymmetric composite oscillation trajectory in the direction perpendicular to the rail surface, and adjusting the projection shape and density of the asymmetric composite oscillation trajectory in the direction parallel to the rail surface, so as to complete the generation and correction of the asymmetric composite oscillation trajectory perpendicular to the forward direction.

[0109] Specifically, various methods can be used to obtain information on the curvature of the rail surface or the morphology of the molten pool bottom. For example, the curvature of the rail surface can be obtained by pre-scanning a three-dimensional geometric model of the rail, or the transient morphology of the molten pool bottom can be obtained during additive manufacturing using real-time three-dimensional sensing technologies such as laser triangulation, confocal microscopy, and structured light scanning. This information provides the true three-dimensional geometric features of the local area where the molten pool is located. Three-dimensional spatial correction of the asymmetric composite oscillation trajectory refers to dynamically adjusting the scanning path of the laser beam based on the obtained rail surface curvature or molten pool bottom morphology information, enabling it to better adapt to complex three-dimensional geometric environments. Specifically, adjusting the defocusing amount of the asymmetric composite oscillation trajectory in the direction perpendicular to the rail surface aims to ensure that the laser beam remains near the optimal focal plane throughout the scanning process, ensuring effective focusing and absorption of laser energy even if there are undulations on the rail surface or changes in molten pool depth. For example, when there is a depression at the bottom of the molten pool, the defocusing amount can be appropriately reduced to lower the laser focus; when there is a convexity at the bottom of the molten pool, the defocusing amount can be appropriately increased to raise the laser focus. Simultaneously, adjusting the projection shape and density of the asymmetric composite oscillation trajectory in the direction parallel to the rail surface is crucial for precisely controlling the lateral distribution of laser energy in three-dimensional space. For example, when the rail surface has significant curvature, the original two-dimensional oscillation trajectory may be distorted when projected onto the three-dimensional surface, leading to uneven energy distribution. Adjusting the projection shape can compensate for this distortion, ensuring that the energy density distribution of the laser in the actual action area meets expectations. Adjusting the projection density allows for denser or sparser scanning paths in three-dimensional space based on the energy requirements of different areas, thereby achieving more precise energy control.

[0110] In some preferred embodiments, a specific example is illustrated below. Suppose that when performing laser additive repair on a section of rail with significant curvature, traditional two-dimensional path control methods may not be effective in handling the geometric changes in the curved area. First, after generating the asymmetric composite oscillation trajectory, the system uses a three-dimensional scanning module integrated into the laser head to acquire the rail surface curvature information of the current additive area in real time. For example, this module could be a sensor based on structured light or lidar, capable of reconstructing the local three-dimensional topography of the rail surface with high precision. Next, based on the acquired real-time curvature data, the control system dynamically adjusts the defocusing amount of the laser beam. When the laser beam scans an upwardly convex area on the rail surface, the defocusing amount is appropriately increased to ensure that the laser focus remains within the molten pool; conversely, when scanning a downwardly concave area, the defocusing amount is decreased. Simultaneously, to compensate for the influence of the curved surface on the projection of the laser scanning trajectory, the control system adjusts the projection shape and density of the asymmetric composite oscillation trajectory in the direction parallel to the rail surface. For example, on a convex curved surface, the lateral spacing of the trajectories may be appropriately stretched to maintain a uniform energy density; while on a concave curved surface, the lateral spacing of the trajectories may be appropriately compressed. In this way, even on a three-dimensional curved surface, laser energy can be precisely applied to the molten pool according to the expected asymmetric distribution pattern, thereby effectively reconstructing the energy distribution inside the molten pool and ensuring additive manufacturing quality.

[0111] Optionally, the steps for generating an asymmetric composite oscillation trajectory, based on the degree of asymmetric distortion, include:

[0112] To obtain local fluid characteristic parameters or transient micro-geometric changes within the molten pool;

[0113] Based on local fluid characteristic parameters or transient micro-geometric changes, an asymmetric composite oscillation trajectory is generated, and the geometric parameters of the asymmetric composite oscillation trajectory are dynamically adjusted.

[0114] Acquiring local fluid characteristic parameters or transient micro-geometric changes within the molten pool refers to real-time monitoring and acquisition of microscopic physical quantities such as fluid flow state, temperature gradient, component distribution, bubble formation, and solidification front position within the molten pool using specific sensing or analytical methods. These parameters directly reflect the energy distribution and mass transport within the molten pool, providing a deeper basis for the refined control of the laser beam. For example, local fluid characteristic parameters may include flow velocity, eddy current intensity, and temperature field distribution within the molten pool; transient micro-geometric changes may include surface fluctuations, keyhole depth, and solidification line position. Specifically, generating an asymmetric composite oscillation trajectory based on local fluid characteristic parameters or transient micro-geometric changes, and dynamically adjusting the geometric parameters of the asymmetric composite oscillation trajectory, means that the laser beam scanning path is no longer solely based on the macroscopic asymmetric distortion and external morphology of the molten pool, but rather incorporates real-time dynamic information from within the molten pool. For example, when fluid stagnation or insufficient energy is detected in a certain area of ​​the molten pool, the laser path density in that area can be increased or the laser dwell time extended accordingly; when an area is detected to be too hot or at risk of overheating, the energy input in that area can be reduced. The geometric parameters of the asymmetric composite oscillation trajectory, such as oscillation amplitude, frequency, shape, and scanning direction, can be adjusted in real time and adaptively based on this internal information to ensure that energy is accurately transferred to the area inside the molten pool that needs compensation and to avoid overheating or defects.

[0115] In some preferred embodiments, a specific example is given below. Assume that during the laser additive manufacturing process of steel rails, sensors such as high-speed cameras and spectrometers acquire real-time information on the temperature field distribution and fluid velocity vector within the molten pool. When the temperature on one side of the molten pool is significantly lower than that on the other side, and a fluid stagnation region exists, this indicates insufficient energy input or excessive heat dissipation in that region. At this point, based on these local fluid characteristic parameters and transient micro-geometric changes, the system dynamically adjusts the asymmetric composite oscillation trajectory. For example, the scanning path density of the laser beam in the low-temperature, fluid-stagnant region can be increased, or the dwell time of the laser beam in that region can be extended, while potentially slightly reducing the energy input to the opposite side to avoid overheating. This dynamic adjustment allows for precise reconstruction of the energy distribution within the molten pool, ensuring that the entire molten pool achieves a uniform temperature and fluid state before solidification, thereby effectively suppressing the generation of solidification defects and improving the density and microstructure uniformity of the additive layer.

[0116] Optional, combined Figure 3 As shown, the step S3, which calculates the degree of asymmetric distortion of the molten pool based on asymmetric characteristics, may include the following process:

[0117] S31, the real-time topographic image of the molten pool is divided into a first region and a second region located on both sides of the center line along the direction of laser beam propagation;

[0118] S32, calculate the average brightness value of all effective pixels in the first region and the second region respectively, and obtain the first average brightness and the second average brightness.

[0119] S33, calculate the asymmetry index data based on the first average brightness and the second average brightness;

[0120] S34, Based on the asymmetric exponential data, the degree of asymmetric distortion is judged, and the result of the degree of asymmetric distortion is obtained;

[0121] S35, when the asymmetric distortion degree judgment result indicates that the absolute value of the asymmetric index data is greater than the preset threshold, it is determined that there is asymmetric distortion in the molten pool, and the direction of energy compensation is determined according to the positive or negative sign of the asymmetric index data.

[0122] S36 uses the magnitude of the asymmetric exponent data as a quantitative basis for adjusting the asymmetric microscopic motion amplitude and scanning speed of the laser beam.

[0123] The median line typically refers to the central axis of the laser beam in its forward direction. This division aims to perform symmetry analysis of the molten pool morphology image in the transverse direction (perpendicular to the laser beam's forward direction). This allows for the separate examination of the morphological features of the molten pool on both sides, providing foundational data for subsequent asymmetry quantification. Effective pixels refer to the pixels in the molten pool morphology image that represent the actual area where the molten pool exists, usually distinguished from the background through image segmentation or thresholding. Brightness values ​​reflect the temperature or energy density of different regions of the molten pool; therefore, by comparing the average brightness on both sides, the asymmetry of the molten pool's energy distribution can be preliminarily determined. Asymmetry index data can be obtained by calculating the difference, ratio, or normalized difference between the first and second average brightness values. Its magnitude and sign directly reflect the degree and direction of molten pool asymmetry. For example, when the first average brightness is greater than the second average brightness, the asymmetry index data may be positive, and vice versa. Preset thresholds are set based on experience or experimental data to determine whether the molten pool asymmetry reaches a level requiring intervention. The sign of the asymmetry index directly indicates the direction of energy deviation. For example, a positive value may indicate that the energy on the left side of the molten pool is too high, requiring energy compensation to the right, and vice versa. This means that the larger the absolute value of the asymmetry index, the more severe the asymmetric distortion of the molten pool, and the greater the amount of asymmetric micro-motion or scanning speed adjustment required by the laser beam to achieve a stronger energy compensation effect.

[0124] Optionally, the step of obtaining local fluid characteristic parameters or transient micro-geometric change information inside the molten pool can be implemented in the following way:

[0125] Obtain radiation images of the molten pool surface area at different wavelengths;

[0126] The spatial distribution differences and dynamic changes of molten pool radiance at different wavelengths were analyzed to obtain the fusion analysis results;

[0127] Based on the fusion analysis results, all radiation images at different wavelengths are fused to obtain local fluid characteristic parameters or transient micro-geometric change information inside the molten pool.

[0128] Acquiring radiation images of the molten pool surface region at different wavelengths refers to simultaneously or quasi-simultaneously capturing the radiation intensity distribution of the molten pool within a specific wavelength range by configuring a multispectral imaging system or multiple narrowband filter cameras. For example, wavelengths sensitive to temperature (such as near-infrared), wavelengths sensitive to the evaporation or segregation of specific elements (such as certain spectral lines in visible light), and wavelengths sensitive to geometric features such as bubbles or keyholes can be selected. These radiation images can reflect the energy distribution, temperature gradient, and material state of the molten pool surface and near the surface.

[0129] Furthermore, analyzing the spatial distribution differences and dynamic changes in molten pool radiance at different wavelengths to obtain fusion analysis results involves image processing and data analysis of the acquired radiation images at each wavelength. Spatial distribution differences can reveal convection patterns, hotspot regions, or material enrichment areas within the molten pool; dynamic changes can reflect transient processes such as molten pool oscillations, keyhole formation and collapse, and bubble generation and escape. Through comprehensive analysis of these differences and changes, the combined effects of different physical phenomena on radiance can be preliminarily decoupled, laying the foundation for subsequent parameter extraction.

[0130] Therefore, based on the results of fusion analysis, fusing all radiation images at different wavelengths to obtain local fluid characteristic parameters or transient micro-geometric changes within the molten pool refers to the use of advanced data fusion algorithms to deeply integrate and analyze radiation data from different wavelengths and time points. For example, machine learning models or physical models can be used to extract, from the fusion analysis results, local fluid characteristic parameters within the molten pool that are difficult to measure directly, such as local flow velocity, surface tension gradient, and viscosity changes, as well as transient micro-geometric changes, such as the real-time depth, shape, and stability of keyholes, and the size, number, and trajectory of bubbles. These parameters and information are crucial for accurately understanding and controlling the behavior of the molten pool during additive manufacturing.

[0131] Optionally, the steps for analyzing the spatial distribution differences and dynamic changes of molten pool radiance at different wavelengths to obtain fusion analysis results may include:

[0132] Obtain radiation images of the molten pool at several preset wavelengths; the preset wavelengths include wavelengths sensitive to temperature, wavelengths sensitive to component segregation, and wavelengths sensitive to bubbles or solidification fronts.

[0133] By dividing the radiation image into regions, we can obtain the characteristic regions corresponding to various physical phenomena inside the molten pool.

[0134] The variation trend and spatial correlation of radiance in the characteristic region under different preset wavelengths are analyzed to decouple the combined effects of various physical phenomena on radiance and obtain radiance characteristics.

[0135] Based on the radiance characteristics, the radiation characteristics corresponding to the preset physical phenomena are identified, and based on the radiation characteristics, the local fluid characteristic parameters or transient micro-geometric change information inside the molten pool are analyzed.

[0136] The acquisition of radiation images of the molten pool at several preset wavelengths aims to capture the unique spectral responses exhibited by the molten pool under different physical states through multispectral imaging technology. Specifically, selecting temperature-sensitive wavelengths, such as the infrared band, can accurately reflect the instantaneous temperature distribution of the molten pool; selecting wavelengths sensitive to component segregation, such as the characteristic spectral line wavelengths of certain elements, helps to monitor the uniformity or segregation of elements within the molten pool; and selecting wavelengths sensitive to bubbles or solidification fronts can reveal the formation of defects within the molten pool or the dynamic changes in the solidification process. Acquiring these multi-wavelength images provides a rich data foundation for subsequent refined analysis.

[0137] Furthermore, regional segmentation of the radiation image to obtain characteristic regions corresponding to various physical phenomena within the molten pool refers to dividing the acquired multi-wavelength radiation image into several meaningful sub-regions based on the physical characteristics or geometric structure of the molten pool. For example, the molten pool can be divided into a central region, an edge region, a solidification front region, etc., or adaptive segmentation can be performed based on features such as radiation intensity gradient and texture. Each characteristic region may correspond to different thermodynamic states, fluid flow patterns, or solidification behaviors. Through region segmentation, the physical phenomena in different regions can be analyzed in a targeted manner.

[0138] Therefore, analyzing the variation trend and spatial correlation of radiance in characteristic regions at different preset wavelengths to decouple the combined effects of various physical phenomena on radiance and obtain radiance characteristics is the core of this scheme. Multiple physical phenomena within the molten pool (such as temperature changes, component segregation, bubble formation, and fluid flow) often occur simultaneously and influence each other, resulting in complex composite characteristics of radiance. By analyzing the temporal variation trend of radiance in each characteristic region at different wavelengths and the spatial correlation between different regions, signal processing, pattern recognition, or machine learning methods can be used to effectively decouple these composite effects. For example, an increase in temperature leads to a general increase in radiance across all wavelengths, while segregation of specific elements may only cause significant changes at specific wavelengths. Through this decoupling analysis, the independent radiative characteristics of each physical phenomenon can be extracted, thereby providing a more accurate understanding of the true state within the molten pool.

[0139] Finally, based on the radiance characteristics, the radiation features corresponding to preset physical phenomena are identified, and based on these radiation features, the local fluid characteristic parameters or transient micro-geometric changes within the molten pool are analyzed. This means that after obtaining the decoupled radiance characteristics, they can be compared with pre-established physical models or empirical databases to identify the specific physical phenomena currently occurring within the molten pool, such as the presence of Marangoni convection, bubble entrainment, or segregation of specific elements. Based on these identified physical phenomena and their corresponding radiation features, further quantitative analysis can be performed to obtain local fluid characteristic parameters or transient micro-geometric changes within the molten pool, such as local fluid velocity, temperature gradient, component concentration distribution, bubble size and location, and the shape and velocity of the solidification front. This information is crucial for the precise control of the laser additive manufacturing process.

[0140] Optionally, the steps of analyzing the variation trend and spatial correlation of radiance in the characteristic region at different preset wavelengths to decouple the combined effects of various physical phenomena on radiance and obtain radiance characteristics include:

[0141] Acquire radiance data of the feature region at different preset wavelengths;

[0142] The radiance data were normalized.

[0143] Calculate the correlation matrix of the normalized radiance data in the spatial dimension;

[0144] Calculate the rate of change of the normalized radiance data over time;

[0145] The correlation matrix and rate of change are used as radiance features;

[0146] The steps described above for identifying the radiation characteristics corresponding to a preset physical phenomenon based on radiation brightness characteristics include:

[0147] Principal component analysis was performed on the correlation matrix and rate of change in the radiance characteristics to extract independent feature components characterizing different physical phenomena.

[0148] Based on independent feature components, identify the radiation characteristics corresponding to the preset physical phenomena.

[0149] Specifically, acquiring radiance data of feature regions at different preset wavelengths refers to extracting pixel brightness or intensity values ​​corresponding to various physical phenomena within the molten pool from radiation images acquired at several preset wavelengths. This data forms the basis for subsequent analysis and reflects the energy radiation characteristics of the molten pool under different physical states. Normalization of the radiance data aims to eliminate the influence of factors such as differences in sensor response at different wavelengths, changes in ambient light, or inconsistencies in image acquisition parameters on the absolute value of the radiance data. This ensures comparability of data from different wavelengths, regions, or time points, thereby more accurately reflecting the relative changes caused by the physical phenomena themselves. Normalization can employ various methods, such as min-max normalization and Z-score normalization.

[0150] In practical applications, calculating the correlation matrix of normalized radiance data in the spatial dimension aims to quantify the degree of correlation between radiance changes at different spatial locations within the molten pool. For example, if the radiance change in one part of the molten pool is highly positively correlated with that in another part, it may indicate that they are affected by the same or similar physical phenomena, such as overall Marangoni flow. This correlation matrix can reveal spatial patterns of energy distribution or mass transport within the molten pool. Simultaneously, calculating the rate of change of normalized radiance data in the temporal dimension aims to capture the dynamic evolution characteristics of physical phenomena within the molten pool. For example, transient processes such as rapid expansion or contraction of the molten pool, bubble formation and collapse, or the advancement of the solidification front will all exhibit specific temporal rates of change in radiance. By analyzing these rates of change, transient and rapidly changing physical events can be identified. Using the correlation matrix and rate of change as radiance features aims to transform the complex radiance image information of the molten pool into a more representative and quantifiable feature vector. These features can comprehensively reflect the spatial correlation and temporal dynamics of the molten pool, providing rich and structured input for subsequent identification of physical phenomena.

[0151] Furthermore, principal component analysis (PCA) is performed on the correlation matrix and rate of change in the radiance characteristics. The aim is to extract the most representative independent feature components that best represent the data variability from the high-dimensional radiance characteristics. PCA is a statistical method that transforms original, potentially interrelated features (such as spatial correlation and temporal rate of change) into a set of linearly uncorrelated new features, i.e., independent feature components. These independent feature components effectively decouple the combined effects of various physical phenomena within the molten pool on radiance, allowing each physical phenomenon (e.g., Marangoni flow, keyhole instability, spatter formation, solidification front fluctuations, etc.) to correspond to one or several main independent feature components, thus avoiding confusion between features of different physical phenomena. Therefore, based on the independent feature components, the radiance characteristics corresponding to preset physical phenomena can be identified. This means that by analyzing these decoupled independent feature components, it is possible to more accurately determine what physical phenomenon is currently occurring in the molten pool. For example, one independent feature component may be mainly related to the Marangoni flow on the molten pool surface, while another independent feature component may be mainly related to the depth and stability of the keyhole. In this way, accurate identification of complex physical phenomena within the molten pool can be achieved.

[0152] In some preferred embodiments, a specific example is given below. Suppose that during the laser additive manufacturing process of steel rails, it is necessary to identify whether there is significant Marangoni flow and keyhole instability in the molten pool. First, real-time radiation images of the molten pool are acquired using a multi-wavelength camera at preset wavelengths, such as 900 nm (temperature-sensitive) and 1550 nm (keyhole depth-sensitive). Then, these images are divided into feature regions, for example, dividing the molten pool surface into a central region, edge regions, etc. Next, radiance data is extracted from these feature regions and normalized. Subsequently, the correlation matrix of the normalized radiance data in the spatial dimension is calculated. For example, if the radiance changes in the central region and the edge region of the molten pool show a high positive correlation, it may indicate the presence of strong Marangoni flow, causing the temperature distribution on the molten pool surface to tend to be uniform. Simultaneously, the rate of change of the radiance data in the time dimension is calculated. For example, if the radiance of a certain region fluctuates drastically in a short period of time, it may indicate periodic keyhole collapse or bubble rupture. These spatial correlation matrices and time rates of change are used as radiance features and input into the principal component analysis model. Principal component analysis (PCA) can extract several independent characteristic components. For example, the first independent characteristic component may primarily reflect the overall temperature gradient and Marangoni flow intensity of the molten pool, while the second independent characteristic component may be mainly related to the keyhole depth fluctuation and stability. By analyzing the magnitude and variation trend of these independent characteristic components, the intensity of the Marangoni flow in the current molten pool and the stability of the keyhole can be accurately identified. For example, a high value of the independent characteristic component related to Marangoni flow indicates a strong Marangoni flow; when the independent characteristic component related to keyhole stability exhibits periodic large fluctuations, it indicates instability in the keyhole. Based on this accurately identified physical phenomenon information, the asymmetric micro-motion parameters of the laser beam can be further adjusted, such as adjusting the scanning speed or dwell time, to effectively suppress keyhole instability or optimize the Marangoni flow, thereby achieving accurate reconstruction of the molten pool energy distribution and optimized control of the rail additive manufacturing path.

[0153] Optionally, the step of identifying the radiation characteristics corresponding to a preset physical phenomenon based on independent feature components includes:

[0154] Obtain the material system parameters, laser process parameters, and environmental parameters of the current laser additive manufacturing process;

[0155] Real-time monitoring of fluctuation ranges in material system parameters, laser process parameters, and environmental parameters;

[0156] Based on the real-time fluctuation range of independent feature components, material system parameters, laser process parameters, and environmental parameters, the mapping weights between independent feature components and preset physical phenomena are dynamically adjusted.

[0157] When the real-time fluctuation range of material system parameters, laser process parameters, and environmental parameters exceeds the preset range, the corresponding mapping weight is reduced.

[0158] Based on the adjusted mapping weights, the radiation characteristics corresponding to the preset physical phenomena are identified.

[0159] Specifically, acquiring the material system parameters, laser process parameters, and environmental parameters of the current laser additive manufacturing process refers to obtaining key parameters related to the additive manufacturing process in real time through sensors, process control systems, or pre-set databases. Material system parameters may include the composition, purity, and preheating temperature of the material to be additively manufactured; laser process parameters may include laser power, scanning speed, defocusing amount, and spot pattern; and environmental parameters may include ambient temperature, humidity, and atmosphere composition. Real-time acquisition of these parameters is crucial for a comprehensive understanding of the current state of the additive manufacturing process.

[0160] Furthermore, real-time fluctuation range monitoring is performed on the aforementioned material system parameters, laser process parameters, and environmental parameters. This aims to continuously track the actual values ​​of these parameters and compare them with preset normal operating ranges or thresholds. For example, upper and lower limits can be set for each parameter; once the real-time monitored value exceeds these ranges, it is considered a fluctuation or anomaly. This monitoring process can be implemented using a data acquisition system and real-time analysis algorithms.

[0161] Based on this, the mapping weights between the aforementioned independent feature components and preset physical phenomena are dynamically adjusted according to the real-time fluctuation range of the independent feature components, material system parameters, laser process parameters, and environmental parameters. The mapping weights can be understood as the relative importance or confidence level of different independent feature components when identifying a specific physical phenomenon. When external parameters fluctuate, some independent feature components may become unreliable or their correlation with the physical phenomenon may change. In this case, dynamically adjusting their mapping weights can reduce their impact on the final identification result. For example, machine learning models or adaptive algorithms can be used to update the weights in real time based on parameter fluctuations.

[0162] Specifically, when the real-time fluctuation range of the aforementioned material system parameters, laser process parameters, and environmental parameters exceeds a preset range, the corresponding mapping weight is reduced. This means that when one or a group of external parameters deviates from normal operating conditions, the contribution of independent feature components highly correlated with the fluctuations of these parameters in identifying physical phenomena will be actively reduced. For example, if excessively high ambient temperatures cause interference with radiation signals, the weight of independent components related to temperature-sensitive radiation characteristics can be reduced to avoid misjudgment.

[0163] Finally, based on the adjusted mapping weights, the radiation characteristics corresponding to the preset physical phenomena are identified. In this way, even under fluctuating external conditions, the system can more accurately and robustly identify local fluid characteristic parameters or transient micro-geometric changes within the molten pool, thus providing a more reliable basis for subsequent asymmetric micro-motion control of the laser beam.

[0164] In some preferred embodiments, a specific example is given below. Suppose that during the laser additive manufacturing process of steel rails, a sudden increase in ambient temperature leads to an increase in background noise in the radiation signal from the molten pool surface, or a slight fluctuation in laser power output causes a subtle change in the energy distribution within the molten pool. Without the solution of this application, the system might still identify the fluid characteristics within the molten pool according to a fixed mapping weight, potentially leading to deviations.

[0165] This application also discloses a rail laser additive path control system for performing rail laser additive path control, combined with... Figure 4 As shown, the rail laser additive manufacturing path control system 1 includes:

[0166] The morphology image acquisition module 11 is used to monitor and acquire real-time morphology images of the molten pool;

[0167] The molten pool feature acquisition module 12 is used to process the real-time topography image to obtain the asymmetric features of the molten pool;

[0168] The distortion degree calculation module 13 is used to calculate the degree of asymmetric distortion of the molten pool based on the asymmetric characteristics.

[0169] The laser motion control module 14 is used to control the laser beam to perform asymmetric micro-motion according to the degree of asymmetric distortion; the asymmetric micro-motion includes adjusting the scanning speed or dwell time of the laser beam in different regions of the molten pool to reconstruct the energy distribution inside the molten pool.

[0170] The laser parameter adjustment module 15 is used to adjust the parameters of the asymmetric micro-motion of the laser beam based on the real-time morphology image of the molten pool after the energy distribution inside the molten pool is reconstructed, so as to realize the path control of laser additive manufacturing of rails.

[0171] The topographic image acquisition module may include a high-speed camera, an infrared thermal imager, a laser triangulation device, or a structured light scanning device, etc., for continuously acquiring visible light images, temperature distribution images, or 3D point cloud data of the molten pool during the laser additive manufacturing process of rails. This module is equipped with synchronous triggering and data caching functions to ensure that the acquired images correspond to the laser scanning position in real time. In a preferred embodiment, this module has multispectral acquisition capabilities, enabling it to simultaneously acquire molten pool radiation images at multiple preset wavelengths for subsequent multi-band fusion analysis.

[0172] The molten pool feature acquisition module includes an image preprocessing unit, a region segmentation unit, and a feature calculation unit. The preprocessing unit performs operations such as denoising, brightness normalization, and contrast enhancement; the region segmentation unit uses threshold segmentation, edge detection, or machine vision-based segmentation algorithms to separate the molten pool region from the background; the feature calculation unit extracts asymmetric features based on the segmentation results, including width difference, area difference, brightness difference, and temperature gradient difference.

[0173] The distortion calculation module divides the molten pool into two regions, left and right, along the centerline of the laser beam's direction of travel. It calculates the average brightness value or geometric dimensions of each region separately, obtaining the first and second feature values. The module then uses a normalized difference formula to calculate the asymmetry index. This module also supports a multi-feature fusion-based distortion calculation method, combining width difference, brightness difference, and area difference to generate a comprehensive asymmetric distortion value and determine whether it exceeds an intervention threshold.

[0174] The laser motion control module generates an asymmetric composite oscillation trajectory based on the degree of distortion, controlling the scanning speed and dwell time of the laser beam in different regions of the molten pool. This trajectory has a higher path density or longer dwell time in directions with high energy demand, and can be combined with information on the curvature of the rail surface or the topography of the molten pool for three-dimensional correction, adjusting the defocus amount and the shape of the projected path.

[0175] The laser motion control module generates an asymmetric composite oscillation trajectory based on the degree of distortion, controlling the scanning speed and dwell time of the laser beam in different regions of the molten pool. This trajectory has a higher path density or longer dwell time in directions with high energy demand, and can be combined with information on the curvature of the rail surface or the topography of the molten pool for three-dimensional correction, adjusting the defocus amount and the shape of the projected path.

[0176] During the laser additive repair of rails, the topography image acquisition module works simultaneously with a high-speed camera and a near-infrared thermal imager. The high-speed camera acquires visible light images at 5000 frames per second, while the thermal imager acquires temperature distribution images. The molten pool feature acquisition module performs noise reduction and brightness normalization on the acquired images, uses a segmentation algorithm to identify the molten pool region, and calculates the average brightness difference and width difference on both sides of the centerline along the laser beam's propagation direction. The distortion degree calculation module substitutes these differences into the asymmetric exponential formula, obtaining a distortion degree value of 0.25, which is positive, indicating insufficient energy on the left side of the molten pool.

[0177] Based on this, the laser motion control module generates an asymmetric composite oscillating trajectory with a higher path density and longer dwell time on the left side. It then performs three-dimensional correction using rail surface curvature data, slightly shifting the laser focus downwards in the left region to compensate for localized energy deficiency. After the laser beam scans along this trajectory, the laser parameter adjustment module detects in the next frame of the topography image that the distortion has decreased to 0.05, but the left side still shows a slight concavity. Therefore, it further increases the left-side scanning path density by 5%, ultimately restoring the molten pool to its ideal symmetrical state.

[0178] Through modular division of labor and closed-loop feedback control, this system achieves precise energy distribution reconstruction on complex geometric rail surfaces and under rapidly changing molten pool conditions, significantly improving the quality and efficiency of additive repair.

[0179] The above are merely embodiments of this application and are not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for controlling the path of laser additive manufacturing of steel rails, characterized in that, include: Monitor and acquire real-time morphological images of the molten pool; The real-time topography image is processed to obtain the asymmetric features of the molten pool; Based on the aforementioned asymmetric characteristics, the degree of asymmetric distortion of the molten pool is calculated; Based on the degree of asymmetric distortion, the laser beam is controlled to perform asymmetric micro-motion; the asymmetric micro-motion includes adjusting the scanning speed or dwell time of the laser beam in different regions of the molten pool to reconstruct the energy distribution inside the molten pool. Based on the real-time topographic image of the molten pool after the energy distribution inside the molten pool is reconstructed, the parameters of the asymmetric micro-motion of the laser beam are adjusted to achieve path control of laser additive manufacturing of steel rails. The step of calculating the degree of asymmetric distortion of the molten pool based on the asymmetric features includes: dividing the real-time topography image of the molten pool into a first region and a second region located on both sides of the center line along the direction of laser beam propagation; calculating the average brightness value of all effective pixels in the first region and the second region respectively to obtain a first average brightness and a second average brightness; calculating asymmetric index data based on the first average brightness and the second average brightness; judging the degree of asymmetric distortion based on the asymmetric index data to obtain a judgment result of the degree of asymmetric distortion; when the judgment result of the degree of asymmetric distortion indicates that the absolute value of the asymmetric index data is greater than a preset threshold, determining that there is asymmetric distortion in the molten pool, and determining the direction of energy compensation based on the positive or negative sign of the asymmetric index data; using the value of the asymmetric index data as a quantitative basis for adjusting the asymmetric micro-motion amplitude and scanning speed of the laser beam; The step of controlling the laser beam to perform asymmetric micro-motion based on the degree of asymmetric distortion, wherein the asymmetric micro-motion includes adjusting the scanning speed or dwell time of the laser beam in different regions of the molten pool to reconstruct the energy distribution inside the molten pool, includes: generating an asymmetric composite oscillation trajectory based on the degree of asymmetric distortion; controlling the laser beam to execute the asymmetric composite oscillation trajectory; wherein the asymmetric composite oscillation trajectory has a higher path density or a longer laser dwell time in directions with higher energy demand; the step of adjusting the parameters of the asymmetric micro-motion of the laser beam based on the real-time topographic image of the molten pool after the energy distribution inside the molten pool is reconstructed to achieve path control of laser additive manufacturing of rails includes: adjusting the geometric parameters of the asymmetric composite oscillation trajectory based on the real-time topographic image of the molten pool after the energy distribution inside the molten pool is reconstructed.

2. The method for controlling the path of laser additive manufacturing of steel rails according to claim 1, characterized in that, After the step of generating the asymmetric composite oscillation trajectory based on the degree of asymmetric distortion, the method further includes: Obtain information on the surface curvature of the rail or the bottom morphology of the molten pool; Based on the rail surface curvature information or the molten pool bottom morphology information, the asymmetric composite oscillation trajectory is corrected in three-dimensional space. The three-dimensional space correction includes adjusting the defocusing amount of the asymmetric composite oscillation trajectory in the direction perpendicular to the rail surface, and adjusting the projection shape and density of the asymmetric composite oscillation trajectory in the direction parallel to the rail surface, so as to complete the generation and correction of the asymmetric composite oscillation trajectory perpendicular to the forward direction.

3. The method for controlling the path of laser additive manufacturing of steel rails according to claim 2, characterized in that, The step of generating an asymmetric composite oscillation trajectory based on the degree of asymmetric distortion includes: To obtain local fluid characteristic parameters or transient micro-geometric changes within the molten pool; Based on the local fluid characteristic parameters or transient micro-geometric change information, an asymmetric composite oscillation trajectory is generated, and the geometric parameters of the asymmetric composite oscillation trajectory are dynamically adjusted.

4. The method for controlling the path of laser additive manufacturing of steel rails according to claim 3, characterized in that, The steps for obtaining local fluid characteristic parameters or transient micro-geometric change information inside the molten pool include: Obtain radiation images of the molten pool surface area at different wavelengths; The spatial distribution differences and dynamic changes of molten pool radiance at different wavelengths were analyzed to obtain the fusion analysis results; Based on the fusion analysis results, all radiation images at different wavelengths are fused to obtain local fluid characteristic parameters or transient micro-geometric change information inside the molten pool.

5. The method for controlling the path of laser additive manufacturing of steel rails according to claim 4, characterized in that, The steps for analyzing the spatial distribution differences and dynamic changes of molten pool radiance at different wavelengths to obtain fusion analysis results include: Obtain radiation images of the molten pool at several preset wavelengths; the preset wavelengths include wavelengths sensitive to temperature, wavelengths sensitive to component segregation, and wavelengths sensitive to bubbles or solidification fronts. The radiation image is divided into regions to obtain feature regions corresponding to various physical phenomena inside the molten pool; The variation trend and spatial correlation of the radiance of the feature region under different preset wavelengths are analyzed in order to decouple the combined influence of various physical phenomena on the radiance and obtain the radiance characteristics. Based on the radiance characteristics, the radiation characteristics corresponding to the preset physical phenomena are identified, and based on the radiation characteristics, the local fluid characteristic parameters or transient micro-geometric change information inside the molten pool are analyzed.

6. The method for controlling the path of laser additive manufacturing of steel rails according to claim 5, characterized in that, The steps of analyzing the variation trend and spatial correlation of the radiance of the characteristic region under different preset wavelengths, in order to decouple the combined influence of various physical phenomena on the radiance and obtain the radiance characteristics, include: Obtain the radiance data of the feature region at different preset wavelengths; The radiance data is normalized. Calculate the correlation matrix of the normalized radiance data in the spatial dimension; Calculate the rate of change of the normalized radiance data over time; The correlation matrix and the rate of change are used as radiance features; The step of identifying the radiation characteristics corresponding to a preset physical phenomenon based on the radiation brightness characteristics includes: Principal component analysis was performed on the correlation matrix and rate of change in the radiance characteristics to extract independent feature components characterizing different physical phenomena. Based on the independent feature components, the radiation characteristics corresponding to the preset physical phenomena are identified.

7. The method for controlling the path of laser additive manufacturing of steel rails according to claim 6, characterized in that, The step of identifying the radiation characteristics corresponding to a preset physical phenomenon based on the independent feature components includes: Obtain the material system parameters, laser process parameters, and environmental parameters of the current laser additive manufacturing process; Real-time fluctuation range monitoring of the material system parameters, laser process parameters, and environmental parameters; Based on the real-time fluctuation range of the independent feature components, the material system parameters, the laser process parameters, and the environmental parameters, the mapping weight between the independent feature components and the preset physical phenomena is dynamically adjusted. When the real-time fluctuation range of the material system parameters, laser process parameters, and environmental parameters exceeds the preset range, the corresponding mapping weight is reduced. Based on the adjusted mapping weights, the radiation characteristics corresponding to the preset physical phenomena are identified.

8. A rail laser additive path control system, used to execute the rail laser additive path control method according to any one of claims 1-7, characterized in that, include: The morphology image acquisition module is used to monitor and acquire real-time morphology images of the molten pool; The molten pool feature acquisition module is used to process the real-time topography image to obtain the asymmetric features of the molten pool; The distortion degree calculation module is used to calculate the degree of asymmetric distortion of the molten pool based on the asymmetric features. The laser motion control module is used to control the laser beam to perform asymmetric micro-motion according to the degree of asymmetric distortion; the asymmetric micro-motion includes adjusting the scanning speed or dwell time of the laser beam in different regions of the molten pool to reconstruct the energy distribution inside the molten pool; The laser parameter adjustment module is used to adjust the parameters of the asymmetric micro-motion of the laser beam based on the real-time morphology image of the molten pool after the energy distribution inside the molten pool is reconstructed, so as to realize the path control of laser additive manufacturing of rails.

Images