A molten pool closed-loop intelligent cladding system for laser rapid forming

Abstract

The application discloses a molten pool closed-loop intelligent cladding system for laser rapid forming, relates to the technical field of laser cladding forming, and is used for solving the problems of molten pool fluctuation and unstable cladding layer forming in the laser rapid forming process; the application drives a laser beam, a powder feeding mechanism and a moving mechanism to perform cladding according to a forming track through a cladding execution module, divides the forming track into different cladding sections, collects molten pool images, molten pool temperatures, cladding heights and moving positions through a molten pool collection module, extracts molten pool boundaries, molten pool widths, molten pool brightness changes, spatter changes and cladding height deviations, generates a molten pool fusion state based on an extended Kalman filtering algorithm through a state fusion module, calls an abnormal type and a parameter influence relation database according to an abnormal tendency grade through a rule backstepping adjustment module, backstepping adjustment amounts of laser power, scanning speed, powder feeding amount and spot position, corrects rules and parameter relations according to adjustment results through a feedback correction module, and intelligent molten pool cladding control is realized.

Description

Technical Field

[0001] This invention relates to the field of laser cladding technology, and more specifically, to a closed-loop intelligent cladding system for laser rapid prototyping. Background Technology

[0002] In the fields of laser rapid prototyping and laser cladding, high-energy laser beams are used to locally melt the surface of a substrate and simultaneously fed metal powder. A motion mechanism continuously scans the surface along a forming trajectory, causing the molten material to solidify and form a cladding layer or multi-layered structure. This technology is widely used in parts repair, surface strengthening, wear- and corrosion-resistant coating preparation, and rapid prototyping of complex metal components. Such systems typically consist of a laser, a powder feeding mechanism, a motion mechanism, and a controller working in concert. Cladding is achieved by pre-setting laser power, scanning speed, powder feed rate, spot position, and forming trajectory. Some solutions also acquire images of the molten pool, temperature, or cladding height to aid in judging the cladding quality.

[0003] However, the state of the molten pool during laser rapid prototyping is affected by factors such as substrate heat accumulation, powder feeding fluctuations, path corners, molten pool overlap, interlayer continuation, and spatter obstruction. Existing systems mostly perform cladding according to uniform preset parameters, lacking differentiated processing for the initiation section, straight section, corner section, overlap section, melting end section, and interlayer continuation section, making it difficult to adapt to local heat input and changes in the molten pool boundary in a timely manner. At the same time, the sources of molten pool images, molten pool temperatures, cladding height, and movement positions are different. If judgment is made directly based on a single temperature, single frame image, or local height, it is easy to cause misjudgments such as undermelting, overmelting, collapse, overlap misalignment, or interlayer heat accumulation due to image jitter, spatter obstruction, instantaneous temperature fluctuations, and height sampling offset. Existing adjustment methods also mostly involve adjusting laser power, scanning speed, or powder feeding amount individually, lacking collaborative back-reasoning and feedback correction of the relationship between anomaly type, anomaly level, cladding segment type, and parameter influence, resulting in adjustment lag, overcorrection, or repeated occurrence of the same anomaly. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, the following solution is proposed to solve the problems of molten pool state fluctuation, unstable cladding layer formation and parameter adjustment lag in the laser rapid prototyping process mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A closed-loop intelligent cladding system for laser rapid prototyping includes a cladding execution module, a molten pool acquisition module, a state fusion module, a rule-based adjustment module, and a feedback correction module.

[0007] The cladding execution module is used to drive the laser beam, powder feeding mechanism and motion mechanism to perform cladding according to the forming trajectory. It divides the forming trajectory into the melting start section, straight section, corner section, overlapping section, melting end section and interlayer continuation section, and binds the laser power, scanning speed, powder feeding amount and spot position to each cladding section.

[0008] The molten pool acquisition module is used to acquire molten pool images, molten pool temperature, cladding height and movement position, extract molten pool boundary, molten pool width, molten pool brightness change, spatter change and cladding height deviation, and generate observation features of the cladding segment.

[0009] The state fusion module is used to predict, update, and correct the observed features of the cladding section based on the extended Kalman filter algorithm, and generate the fusion state of the molten pool, which includes the heat input state, powder absorption state, boundary stability state, and forming height state.

[0010] The rule-based adjustment module is used to match the fusion state of the molten pool, the type of the cladding section, and the rules for judging cladding anomalies, generate an anomaly tendency level, and call the anomaly type and parameter influence relationship library based on the anomaly tendency level to deduce the coordinated adjustment amount of laser power, scanning speed, powder feed amount, and spot position.

[0011] The feedback correction module is used to compare the fusion state of the molten pool, the cladding height, and the level of abnormality before and after adjustment, correct the cladding anomaly judgment rules, the abnormality type and parameter influence relationship library, and the adjustment step size, and output the closed-loop cladding control results.

[0012] Furthermore, the cladding execution module is used to read the forming trajectory of the part to be formed, determine the cladding segment boundary based on the starting position, ending position, path direction change, overlapping area of ​​adjacent melt channels and interlayer connection area of ​​the forming trajectory, mark the forming trajectory between the cladding segment boundaries as melting start segment, straight segment, corner segment, overlapping segment, melting end segment or interlayer continuation segment, and bind each cladding segment with laser power, scanning speed, powder feeding amount, spot position and execution time to generate cladding segment execution record, so that the molten pool observation characteristics can be mapped to specific cladding segments and their process parameters.

[0013] Furthermore, the molten pool acquisition module is used to simultaneously acquire molten pool images, molten pool temperature, molten pool height and movement position during the cladding process, assign the molten pool images, molten pool temperature and molten pool height to the corresponding cladding segment according to the movement position, and extract the molten pool boundary, molten pool width, molten pool brightness change, spatter change and cladding height deviation from the acquired data after assignment to generate cladding segment observation features;

[0014] The molten pool boundary and molten pool width are used to characterize the molten pool forming range, the molten pool brightness variation and molten pool temperature are used to characterize the heat input variation, the spatter variation is used to characterize the molten pool disturbance, and the cladding height deviation is used to characterize the cladding layer forming deviation.

[0015] Furthermore, the state fusion module is used to generate the predicted state of the molten pool of the current cladding segment based on the extended Kalman filter algorithm, using the fusion state of the molten pool of the previous cladding segment, the laser power, scanning speed, powder feeding amount and spot position of the current cladding segment as prediction inputs, and to perform observation correction on the predicted state of the molten pool using the observation characteristics of the current cladding segment as observation inputs, eliminating abnormal observations caused by molten pool image jitter, spatter occlusion, instantaneous temperature fluctuations and height sampling offsets, and generating the fusion state of the molten pool of the current cladding segment.

[0016] Furthermore, the molten pool fusion state includes a heat input state, a powder absorption state, a boundary stability state, and a forming height state;

[0017] The heat input state is determined by the molten pool temperature, molten pool brightness change and laser power; the powder absorption state is determined by the powder feed rate, molten pool width and cladding height; the boundary stability state is determined by the molten pool boundary, spatter change and spot position; and the forming height state is determined by the cladding height, cladding height deviation and scanning speed.

[0018] The state fusion module associates the above states with the corresponding cladding segment types to generate a molten pool state record for determining the level of abnormality.

[0019] Furthermore, the rule-back-adjustment module is used to call the corresponding cladding anomaly judgment rule according to the cladding segment type, and match the heat input state, powder absorption state, boundary stability state and forming height state with the corresponding judgment boundary respectively;

[0020] When the heat input is lower than the forming requirements, the powder absorption is insufficient, and the forming height decreases, there is a tendency for under-melting anomalies to occur.

[0021] When the heat input is higher than the forming requirements, the boundary stability is reduced, and the spatter variation is increased, an over-melting anomaly is generated.

[0022] When the forming height decreases locally and the continuity of the molten pool boundary is interrupted, a collapse anomaly tends to occur.

[0023] When the boundary of the molten pool of the overlapping section deviates from the overlapping area of ​​the adjacent molten channel, an abnormal tendency of overlapping offset is generated.

[0024] When the heat input state of the interlayer joint continues to increase and the cooling recovery is insufficient, an abnormal tendency for interlayer heat accumulation is generated.

[0025] Furthermore, the rule-based adjustment module is used to determine the level of abnormality based on the number of state deviations, the magnitude of state deviations, the number of consecutive cladding segments, and the type of cladding segments corresponding to the abnormality tendency;

[0026] When the fusion state of the molten pool does not exceed the corresponding judgment boundary, a normal level is generated;

[0027] When a single state exceeds the corresponding judgment boundary without causing an increase in the cladding height deviation, a level of concern is generated;

[0028] An adjustment level is generated when at least two states exceed the corresponding judgment boundaries and cause the cladding height deviation to increase.

[0029] When the abnormal tendency persists within a continuous cladding segment and does not diminish after coordinated adjustment, a restriction level is generated, and the abnormal tendency level is associated with and stored in relation to the corresponding cladding segment, abnormal type, and fusion state of the molten pool.

[0030] Furthermore, the database of anomaly types and parameter influence relationships includes anomaly type, cladding segment type, main control parameter, cooperative parameter, adjustment direction, adjustment step size, and parameter boundary;

[0031] The rule-based adjustment module is specifically used to call the database of abnormality types and parameter influences based on the level of abnormality tendency to determine the coordinated adjustment amount of laser power, scanning speed, powder feed amount and spot position.

[0032] The tendency of under-melting corresponds to increasing laser power, reducing scanning speed, and stabilizing powder feed rate; the tendency of over-melting corresponds to reducing laser power, increasing scanning speed, and shortening local thermal action time; the tendency of collapse corresponds to reducing powder feed rate and increasing scanning speed; the tendency of overlap offset corresponds to correcting the spot position; and the tendency of interlayer heat accumulation corresponds to reducing the heat input of the cladding section.

[0033] Furthermore, the feedback correction module is used to obtain the fusion state, cladding height and abnormal tendency level of the molten pool after the coordinated adjustment is executed, and compare them with the fusion state, cladding height and abnormal tendency level of the molten pool before adjustment.

[0034] When the abnormality tendency level decreases and the cladding height deviation decreases, the adjustment direction and adjustment step size in the corresponding abnormality type and parameter influence relationship database are retained.

[0035] When the abnormality tendency level does not decrease or the cladding height deviation increases, correct the main parameters, coordinating parameters and adjustment step size in the corresponding abnormality type and parameter influence relationship library;

[0036] When the same abnormal trend occurs continuously in the same cladding segment type, the judgment boundary in the corresponding cladding anomaly judgment rule is corrected, and the correction result is used for the closed-loop adjustment of subsequent cladding segments.

[0037] Furthermore, the feedback correction module is also used to generate protection control results when the abnormal tendency level is the limit level;

[0038] When the limiting level is caused by an undermelting anomaly, the limiting scan speed continues to increase and the heat input of the subsequent cladding section is increased;

[0039] When the limiting level is caused by an abnormal tendency of over-melting or an abnormal tendency of interlayer heat accumulation, the laser power is limited to continue to increase and the heat input of the subsequent cladding section is reduced.

[0040] When the restriction level is caused by an abnormal tendency of collapse, the powder feed rate should be increased and the scanning speed of the subsequent cladding section should be adjusted.

[0041] When the restriction level is caused by an abnormal tendency of overlap offset, stop using the original spot position and generate a spot position compensation amount;

[0042] The feedback correction module writes the protection control results back to the abnormality type and parameter influence relationship library, and outputs the closed-loop cladding control results of the corresponding cladding section.

[0043] The technical effects and advantages of the present invention, a closed-loop intelligent cladding system for laser rapid prototyping:

[0044] This invention constructs a closed-loop cladding control link, which includes cladding segment division, molten pool observation feature generation, molten pool state fusion, anomaly tendency determination, parameter back-calculation adjustment, and feedback correction. This allows the forming trajectory, process parameters, molten pool changes, and cladding results during laser rapid prototyping to be processed in a related manner within the same cladding segment. This avoids the problem of existing systems relying solely on preset parameters for cladding, which makes it difficult to respond promptly to molten pool fluctuations. The system divides the forming trajectory into the initiation segment, straight segment, corner segment, overlap segment, melting termination segment, and interlayer continuation segment, and binds laser power, scanning speed, powder feed amount, and spot position to each cladding segment, enabling differentiated monitoring and adjustment strategies for different forming positions.

[0045] Based on this, the present invention acquires images of the molten pool, molten pool temperature, cladding height and movement position, extracts the molten pool boundary, molten pool width, molten pool brightness change, spatter change and cladding height deviation, and performs prediction update and observation correction based on the extended Kalman filter algorithm to reduce the impact of image jitter, spatter occlusion, instantaneous temperature fluctuation and height sampling offset on the judgment results, and generates more stable heat input state, powder absorption state, boundary stability state and forming height state;

[0046] Furthermore, based on the fusion state of the molten pool and the type of cladding segment, abnormal tendency levels such as undermelting, overmelting, collapse, overlap offset, and interlayer heat accumulation are generated. The abnormality type and parameter influence relationship library are used to deduce the coordinated adjustment amount of laser power, scanning speed, powder feed amount, and spot position, avoiding overcorrection or adjustment lag caused by adjusting a single parameter. At the same time, the feedback correction module compares the molten pool state, cladding height, and abnormal tendency level before and after adjustment, and continuously corrects the judgment rules, parameter influence relationship library, and adjustment step size, thereby improving the continuity of the cladding layer, the consistency of the molten channel overlap, and the stability of multi-layer forming. Attached Figure Description

[0047] Figure 1 This is a schematic diagram of the structure of a closed-loop intelligent cladding system for laser rapid prototyping according to the present invention.

[0048] Figure 2 This is a flowchart illustrating the closed-loop cladding control process of a closed-loop intelligent cladding system for laser rapid prototyping according to the present invention. Detailed Implementation

[0049] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0050] In order to achieve the above objectives, Figure 1 A schematic diagram of the structure of a closed-loop intelligent cladding system for laser rapid prototyping is given, which specifically includes a cladding execution module, a molten pool acquisition module, a state fusion module, a rule-based adjustment module, and a feedback correction module.

[0051] The cladding execution module is used to drive the laser beam, powder feeding mechanism and motion mechanism to perform cladding according to the forming trajectory. It divides the forming trajectory into the melting start section, straight section, corner section, overlapping section, melting end section and interlayer continuation section, and binds the laser power, scanning speed, powder feeding amount and spot position to each cladding section.

[0052] The molten pool acquisition module is used to acquire molten pool images, molten pool temperature, cladding height and movement position, extract molten pool boundary, molten pool width, molten pool brightness change, spatter change and cladding height deviation, and generate observation features of the cladding segment.

[0053] The state fusion module is used to predict, update, and correct the observed features of the cladding section based on the extended Kalman filter algorithm, and generate the fusion state of the molten pool, which includes the heat input state, powder absorption state, boundary stability state, and forming height state.

[0054] The rule-based adjustment module is used to match the fusion state of the molten pool, the type of the cladding section, and the rules for judging cladding anomalies, generate an anomaly tendency level, and call the anomaly type and parameter influence relationship library based on the anomaly tendency level to deduce the coordinated adjustment amount of laser power, scanning speed, powder feed amount, and spot position.

[0055] The feedback correction module is used to compare the fusion state of the molten pool, the cladding height, and the level of abnormality before and after adjustment, correct the cladding anomaly judgment rules, the abnormality type and parameter influence relationship library, and the adjustment step size, and output the closed-loop cladding control results.

[0056] Laser cladding forming equipment includes a laser, a powder feeding mechanism, a motion mechanism, a molten pool image acquisition device, a temperature acquisition device, a height acquisition device, and a controller. The laser is used to output a laser beam to the surface of the part to be formed or the substrate. The powder feeding mechanism is used to transport cladding powder to the laser action area. The motion mechanism is used to drive the laser beam, the powder feeding mechanism, or the part to be formed to move relative to each other according to the forming trajectory, so that the powder material forms a molten pool under the action of the laser and solidifies into a cladding layer.

[0057] The controller is equipped with a cladding execution module, a molten pool acquisition module, a state fusion module, a rule-based adjustment module, and a feedback correction module. The cladding execution module drives the laser beam, powder feeding mechanism, and motion mechanism to perform cladding according to the forming trajectory and generates cladding segment execution records. The molten pool acquisition module acquires molten pool images, molten pool temperature, cladding height, and motion position during the cladding process and generates cladding segment observation features. The state fusion module fuses the cladding segment observation features to generate the molten pool fusion state. The rule-based adjustment module determines the abnormal tendency level based on the molten pool fusion state and deduces the collaborative adjustment amount. The feedback correction module corrects the judgment rules, parameter influence relationship library, and adjustment step size based on the changes in molten pool state and cladding height before and after adjustment, thereby forming a closed-loop intelligent cladding control for the molten pool.

[0058] Closed-loop cladding control process as follows Figure 2 As shown, the cladding execution module reads the forming trajectory of the part to be formed before the laser rapid prototyping operation begins. The forming trajectory can be generated by slicing the three-dimensional forming model or imported from a preset cladding path file. The forming trajectory includes at least the path start position, path end position, path direction, distance between adjacent melt channels, interlayer connection position, and motion execution sequence. The cladding execution module determines the cladding segment boundary based on the above trajectory information and divides the continuous forming trajectory into multiple cladding segments so that the subsequent melt pool acquisition results can correspond to the specific forming area.

[0059] Specifically, the cladding execution module uses the starting position of the forming trajectory as the boundary of the initiation melting section and the ending position of the forming trajectory as the boundary of the termination melting section.

[0060] When the forming trajectory changes direction at a certain position, the trajectory area before and after that position is defined as the corner segment;

[0061] When there is an overlapping area between adjacent melt channels, the corresponding area is defined as the overlap section;

[0062] When there is a connection between different forming layers, the corresponding area is defined as the interlayer connection segment;

[0063] In addition to the areas mentioned above, trajectory areas with consistent path direction and no overlapping or interlayer continuity are identified as straight segments. Through the above processing, the cladding execution module marks the forming trajectory as the starting melting segment, straight segment, corner segment, overlapping segment, ending melting segment, or interlayer continuity segment.

[0064] After the cladding segments are divided, the cladding execution module establishes a cladding segment execution record for each segment. The cladding segment execution record includes at least the cladding segment identifier, cladding segment type, start position, end position, execution time, laser power, scanning speed, powder feeding amount, and spot position. Specifically, the laser power limits the energy input of the laser beam acting on the powder and substrate surface; the scanning speed limits the relative speed between the laser beam and the part to be formed; the powder feeding amount limits the amount of powder supplied to the laser action area per unit time; and the spot position limits the positional relationship between the laser action area and the target forming trajectory. After the cladding segment execution record is generated, the controller drives the laser beam, powder feeding mechanism, and motion mechanism to perform the cladding according to the record.

[0065] During the cladding process, the molten pool acquisition module simultaneously acquires molten pool images, molten pool temperature, cladding height, and movement position. The molten pool image reflects the molten pool outline, molten pool brightness, and spatter distribution; the molten pool temperature reflects the heat input changes in the molten pool area; the cladding height reflects the forming height after the powder melts and solidifies; and the movement position determines the cladding segment corresponding to the currently acquired data.

[0066] The molten pool acquisition module assigns the molten pool image, molten pool temperature, and cladding height acquired within the same sampling time window or the same location to the corresponding cladding segment, so that the acquired data and the cladding segment execution record form a corresponding relationship.

[0067] The sampling time window and location assignment window can be preset according to the scanning speed of the motion mechanism, the sampling frequency of the acquisition device, and the length of the cladding section. When there are multiple acquisition results within the same sampling time window, the molten pool acquisition module determines the cladding section assignment of the acquisition result based on the correspondence between the motion position and the boundary of the cladding section. When the acquisition result crosses the boundary of an adjacent cladding section, the molten pool acquisition module determines the assignment object according to the cladding section where the acquisition position is located, so that the molten pool image, molten pool temperature, and cladding height can be mapped to the same cladding section.

[0068] The molten pool acquisition module extracts features from the acquired data after attribution to generate observation features of the cladding section. Specifically, the molten pool acquisition module determines the molten pool boundary based on the brightness difference between the molten pool area and the surrounding area in the molten pool image, and calculates the molten pool width from the molten pool boundary.

[0069] The brightness variation of the molten pool is determined based on continuously acquired images of the molten pool, which is used to characterize the trend of heat input variation in the molten pool; the spatter variation is determined based on the bright particles or instantaneous jet areas that deviate from the main body of the molten pool in the images, which is used to characterize the disturbance of the molten pool.

[0070] The cladding height deviation is determined based on the difference between the cladding height and the target forming height, and is used to characterize the deviation of the cladding layer forming.

[0071] Therefore, the observation features of the cladding section generated by the molten pool acquisition module include at least the molten pool boundary, molten pool width, molten pool brightness variation, spatter variation, and cladding height deviation. Among them, the molten pool boundary and molten pool width are used to characterize the molten pool forming range, the molten pool brightness variation and molten pool temperature are used to characterize the heat input variation, the spatter variation is used to characterize the molten pool disturbance, and the cladding height deviation is used to characterize the deviation of the cladding layer forming. After the cladding section observation features are stored in correspondence with the cladding section execution record, they are used as input for the state fusion module to generate the molten pool fusion state.

[0072] The specific implementation details of the state fusion module include:

[0073] The state fusion module is used to fuse the observation features of the cladding section to reduce the judgment error caused by the influence of single acquisition data on molten pool image jitter, spatter obstruction, instantaneous temperature fluctuation and height sampling offset. The state fusion module uses the extended Kalman filter algorithm to predict, update and correct the observation features of the cladding section. The extended Kalman filter algorithm is not used to predict the shape of the part alone, but to fuse the dynamic correlation between molten pool image, molten pool temperature, cladding height and process parameters during laser cladding.

[0074] In this embodiment, the state variables in the extended Kalman filter algorithm include thermal input state, powder absorption state, boundary stability state, and forming height state; the observation variables include molten pool temperature, molten pool brightness change, molten pool width, spatter change, and cladding height deviation; the control variables include laser power, scanning speed, powder feed rate, and spot position. The state fusion module performs prediction updates based on the state variables of the previous cladding segment and the control variables of the current cladding segment, and performs observation correction based on the observation variables of the current cladding segment, thereby obtaining the molten pool fusion state of the current cladding segment.

[0075] Specifically, the state fusion module uses the fusion state of the molten pool of the previous cladding segment as a reference for the continuous forming of the current cladding segment, and uses the laser power, scanning speed, powder feed rate, and spot position bound to the current cladding segment as prediction inputs. Since the laser rapid forming process has continuous cladding characteristics, the heat accumulation, molten channel width, cladding height, and boundary stability of the previous cladding segment will affect the molten pool changes of the subsequent cladding segment. Therefore, the state fusion module first generates the predicted molten pool state of the current cladding segment based on the heat input state, powder absorption state, boundary stability state, and forming height state of the previous cladding segment, combined with the laser power, scanning speed, powder feed rate, and spot position of the current cladding segment.

[0076] The predicted state of the molten pool is used to represent the trend of molten pool change that should be formed in the current cladding section under the current process parameters, specifically including:

[0077] The predicted results of heat input determined by laser power and scanning speed, the predicted results of powder absorption determined by powder feed rate and molten pool forming range, the predicted results of boundary stability determined by spot position and molten pool boundary change, and the predicted results of forming height determined by cladding height and target forming height.

[0078] Through the above prediction processing, the state fusion module can form a predicted state of the molten pool corresponding to the cladding section type and process parameters before the current observation data enters.

[0079] After obtaining the predicted state of the molten pool, the state fusion module uses the observed features of the current cladding segment as the observation input to perform observation correction on the predicted state of the molten pool. The observed features of the current cladding segment include the molten pool boundary, molten pool width, molten pool brightness change, spatter change, and cladding height deviation, and can also be combined with the molten pool temperature to participate in the correction. The state fusion module compares the observed features with the predicted state of the molten pool. When the deviation between the observed features and the predicted state is within the allowable range, the observed features are used as valid observations to participate in the generation of the molten pool fusion state.

[0080] When an observed feature exhibits a short-term abrupt change and is inconsistent with other observed features within the same cladding section, the observed feature is marked as an anomalous observation.

[0081] In this embodiment, abnormal observations include instantaneous jumps in the molten pool boundary caused by molten pool image jitter, short-term loss of molten pool width caused by spatter obstruction, instantaneous increase or decrease in molten pool temperature caused by temperature acquisition interference, and sudden changes in cladding height caused by height sampling position offset.

[0082] For abnormal observations, the state fusion module reduces its participation in observation correction, or uses effective observations from adjacent positions within the cladding section for alternative correction.

[0083] For observation deviations that occur consecutively and are consistent with other state changes, the state fusion module retains them as valid state changes for subsequent anomaly tendency level determination.

[0084] After prediction updates and observation corrections, the state fusion module generates the fusion state of the molten pool for the current cladding segment. This fusion state includes the heat input state, powder absorption state, boundary stability state, and forming height state. The heat input state is determined by the molten pool temperature, molten pool brightness changes, and laser power, and is used to characterize whether the energy input of the current cladding segment meets the forming requirements. The powder absorption state is determined by the powder feed rate, molten pool width, and cladding height, and is used to characterize whether the fed powder is stably melted and forms a cladding layer. The boundary stability state is determined by the molten pool boundary, spatter changes, and laser spot position, and is used to characterize whether the molten pool contour is continuous and whether the laser spot's area of ​​action deviates from the target trajectory. The forming height state is determined by the cladding height, cladding height deviation, and scanning speed, and is used to characterize whether the forming height of the current cladding segment deviates from the target requirements.

[0085] Furthermore, the state fusion module associates the heat input state, powder absorption state, boundary stability state, and forming height state with the corresponding cladding segment type to generate a molten pool state record.

[0086] The molten pool status record includes at least the cladding segment identifier, cladding segment type, heat input status, powder absorption status, boundary stability status, forming height status, and effective observation markers. By associating the molten pool fusion status with the cladding segment type, the subsequent rule back-reasoning adjustment module can use different anomaly judgment rules for the initiation segment, straight segment, corner segment, overlap segment, melting end segment, and interlayer continuation segment, avoiding misjudgment caused by using the same judgment conditions for different cladding segments.

[0087] Through the above processing, the state fusion module does not simply make independent judgments on the molten pool temperature, molten pool image or cladding height, but uses the correlation between the previous cladding segment state, the current process parameters and the current observation features to generate a fusion result that can reflect the true forming state of the molten pool. This fusion result not only retains the heat effect continuously transmitted by the cladding segment during the laser rapid forming process, but also eliminates short-term interference caused by splashing, obstruction and sampling jitter.

[0088] After the state fusion module generates the molten pool state record, the rule back-propagation adjustment module determines the abnormal tendency of the current cladding segment based on the molten pool state record. The molten pool state record includes the cladding segment identifier, cladding segment type, heat input state, powder absorption state, boundary stability state, forming height state, and effective observation mark. The rule back-propagation adjustment module first reads the cladding segment type, and then calls the cladding anomaly judgment rule corresponding to the cladding segment type, so that the initiation segment, straight segment, corner segment, overlap segment, melting end segment, and interlayer continuation segment are judged according to their corresponding forming characteristics.

[0089] In one embodiment, the cladding anomaly determination rule includes a heat input determination boundary, a powder absorption determination boundary, a boundary stability determination boundary, and a forming height determination boundary. The determination boundaries can be pre-configured according to the substrate material, cladding powder material, target cladding layer width, target cladding layer height, melt channel overlap width, and interlayer forming requirements. They can also be updated after the feedback correction module runs. The rule back-propagation adjustment module matches the heat input state, powder absorption state, boundary stability state, and forming height state with the corresponding determination boundaries, and generates anomaly tendencies based on the matching results.

[0090] Among them, the heat input judgment boundary is used to limit the allowable energy input range of the current cladding section, the powder absorption judgment boundary is used to limit the matching range between the powder feeding amount and the width of the molten pool and the cladding height, the boundary stability judgment boundary is used to limit the allowable range of the continuity of the molten pool boundary, the spatter variation and the spot position offset, the forming height judgment boundary is used to limit the allowable deviation range between the cladding height and the target forming height, and the adjustment step size and parameter boundary are determined according to the allowable power range of the laser, the allowable speed range of the motion mechanism, the allowable powder feeding range of the powder feeding mechanism, the allowable compensation range of the spot position and the process window corresponding to the qualified cladding sample.

[0091] Specifically, when the heat input is lower than the forming requirement, the powder absorption is insufficient, and the forming height decreases, the rule-based adjustment module generates an under-melting anomaly tendency. This anomaly tendency indicates that there is a risk that the powder is not fully melted, the cladding layer height is insufficient, or the melt channel continuity is insufficient in the current cladding section.

[0092] When the heat input is higher than the forming requirements, the boundary stability is reduced and the spatter changes are increased, the rule back-deduction adjustment module generates an over-melting abnormality tendency. This abnormality tendency indicates that there is a risk of excessively wide melt pool, enhanced spatter, excessive dilution, or expansion of local thermal effects in the current cladding section.

[0093] When the forming height decreases locally and the continuity of the molten pool boundary is interrupted, the rule-based adjustment module generates a collapse anomaly tendency. This anomaly tendency indicates that the cladding layer height of the current cladding segment has partially collapsed or that the molten pool support is insufficient.

[0094] For overlapping sections, the rule-based back-calculation adjustment module focuses on determining the positional relationship between the molten pool boundary and the overlapping area of ​​adjacent molten channels. When the molten pool boundary of the overlapping section deviates from the overlapping area of ​​adjacent molten channels, an overlap offset anomaly is generated. This anomaly indicates that there may be insufficient overlap, excessive overlap, or discontinuity at the edge of the cladding layer between adjacent molten channels. For interlayer continuation sections, the rule-based back-calculation adjustment module focuses on determining the thermal impact of the previous layer or previous cladding section on the current cladding section.

[0095] When the heat input state of the interlayer continuation section continues to increase and the cooling recovery is insufficient, an abnormal tendency of interlayer heat accumulation is generated. This abnormal tendency indicates that the subsequent cladding section may experience over-melting, collapse, uneven structure, or fluctuation in forming height.

[0096] After generating the abnormal tendency, the rule-based adjustment module further determines the abnormal tendency level. The abnormal tendency level is determined based on the number of state deviations, the magnitude of state deviations, the number of consecutive cladding segments, and the type of cladding segment. The number of state deviations represents the number of states that exceed the judgment boundary among the heat input state, powder absorption state, boundary stability state, and forming height state; the magnitude of state deviations represents the degree to which the corresponding state deviates from the judgment boundary; the number of consecutive cladding segments represents the number of times the same abnormal tendency occurs consecutively in adjacent cladding segments; and the cladding segment type is used to determine the degree of influence of the abnormal tendency in the current cladding segment.

[0097] When the fusion state of the molten pool does not exceed the corresponding judgment boundary, the rule back-propagation adjustment module generates a normal level, indicating that no parameter adjustment is required for the current cladding segment. When a single state exceeds the corresponding judgment boundary and does not cause an increase in the cladding height deviation, a concern level is generated, indicating that there is a slight deviation in the current cladding segment, but the current process parameters can be maintained and the subsequent molten pool state can continue to be collected.

[0098] When at least two states exceed the corresponding judgment boundary and cause the cladding height deviation to increase, an adjustment level is generated, indicating that the current cladding segment needs to perform coordinated parameter adjustment;

[0099] When an abnormal tendency persists within a continuous cladding section and does not diminish after coordinated adjustment, a restriction level is generated, indicating that the current abnormality has affected the continuity of cladding or the forming quality, and it is necessary to enter the protection control or restriction execution process.

[0100] After determining the level of anomaly tendency, the rule-based adjustment module calls the anomaly type and parameter influence relationship library. This library records the anomaly type, cladding segment type, main adjustment parameter, cooperating parameter, adjustment direction, adjustment step size, and parameter boundaries. The anomaly type includes under-melting anomaly tendency, over-melting anomaly tendency, collapse anomaly tendency, overlap offset anomaly tendency, and interlayer heat accumulation anomaly tendency. The cladding segment type includes the initiation segment, straight segment, corner segment, overlap segment, melt-out segment, and interlayer continuation segment. The main adjustment parameter represents the process parameter that has a major impact on the current anomaly. The cooperating parameter represents the process parameter that needs to be adjusted synchronously with the main adjustment parameter. The adjustment direction indicates increasing, decreasing, stabilizing, or compensating. The adjustment step size represents the magnitude of each adjustment. The parameter boundaries represent the allowable adjustment range of laser power, scanning speed, powder feed rate, and spot position.

[0101] The rule-based reverse adjustment module determines the adjustment intensity based on the level of abnormality and determines the coordinated adjustment amount based on the relationship between abnormality type and parameter influence. For under-melting abnormality, the rule-based reverse adjustment module uses laser power as the main adjustment parameter and scanning speed and powder feeding amount as coordinated parameters to generate coordinated adjustment amounts that increase laser power, decrease scanning speed and stabilize powder feeding amount.

[0102] For the tendency of over-melting, the rule-based back-deduction adjustment module uses laser power and scanning speed as the main adjustment parameters to generate a coordinated adjustment amount that reduces laser power, increases scanning speed, and shortens local thermal action time.

[0103] For the tendency of collapse, the rule-based adjustment module uses the powder feed rate and scanning speed as the main adjustment parameters to generate a coordinated adjustment amount that reduces the powder feed rate and increases the scanning speed.

[0104] For abnormal overlap offset tendency, the rule back-induction adjustment module uses the spot position as the main adjustment parameter to generate spot position compensation amount;

[0105] In response to the abnormal tendency of interlayer heat accumulation, the rule-based adjustment module generates a coordinated adjustment amount to reduce the heat input of the subsequent cladding section.

[0106] In one specific embodiment, when the current cladding segment is an overlapping segment, the rule back-pushing adjustment module reads the molten pool status record output by the status fusion module. If the molten pool status record shows that the boundary stability state has decreased and the molten pool boundary has deviated from the overlapping area of ​​the adjacent molten channel, and the cladding height deviation has increased, the rule back-pushing adjustment module generates an overlapping offset abnormal tendency.

[0107] If the abnormal tendency is accompanied by an increase in thermal input, the abnormal tendency level can be determined as the adjustment level. Subsequently, the rule-based adjustment module calls the abnormality type and parameter influence relationship library to determine the spot position as the primary adjustment parameter. When the overlap offset abnormal tendency is accompanied by an increase in thermal input, the scanning speed is determined as the cooperative parameter, and an adjustment amount to shorten the local thermal action time is generated. When the overlap offset abnormal tendency is accompanied by a decrease in thermal input, the laser power is determined as the cooperative parameter, and an adjustment amount to compensate for the thermal input of the overlap area is generated. The rule-based adjustment module generates the spot position compensation amount and the corresponding cooperative adjustment amount based on the above primary and cooperative parameters.

[0108] In another specific embodiment, when the current cladding segment is an interlayer continuation segment, if the heat input state output by the state fusion module continues to increase, and the cooling recovery within the continuous cladding segment is insufficient, while the forming height state fluctuates, the rule back-propagation adjustment module generates an abnormal tendency for interlayer heat accumulation. If this abnormal tendency persists within the continuous cladding segment and has not weakened after the previous collaborative adjustment, the abnormal tendency level is determined to be a restriction level. At this time, the rule back-propagation adjustment module no longer increases the laser power independently, but generates a collaborative adjustment amount to reduce the heat input of subsequent cladding segments based on the abnormality type and parameter influence relationship library, so that the subsequent feedback correction module can further perform protection control and rule write-back.

[0109] Through the above processing, the rule-based reverse adjustment module does not directly adjust the laser power based on a single molten pool temperature or a single image result. Instead, it combines the molten pool fusion state, cladding segment type, anomaly judgment rules, anomaly tendency level, and anomaly type with the parameter influence relationship library to reverse-calculate the coordinated adjustment amount of laser power, scanning speed, powder feed amount, and spot position. This processing method can make differentiated adjustments for the forming characteristics of different cladding segments and transmit the adjustment results to the feedback correction module, which then judges whether the adjustment is effective, thus forming a correction process.

[0110] After the rule-based adjustment module generates the coordinated adjustment amount, the feedback correction module receives the coordinated adjustment amount and re-acquires the fusion state, cladding height, and abnormal tendency level of the corresponding cladding segment after the coordinated adjustment amount is executed. The coordinated adjustment amount includes the result of adjusting at least one parameter among laser power, scanning speed, powder feed amount, and spot position, and may also include the result of adjusting the heat input or spot position compensation amount of the subsequent cladding segment. The feedback correction module determines whether the current adjustment has a weakening effect on the current abnormal tendency by comparing the fusion state, cladding height, and abnormal tendency level of the fusion pool before and after the adjustment.

[0111] Specifically, the feedback correction module first reads the molten pool state record and abnormal tendency level before adjustment, then reads the molten pool state record and abnormal tendency level after adjustment, and compares the heat input state, powder absorption state, boundary stability state, forming height state and cladding height deviation accordingly.

[0112] When the level of abnormality decreases after adjustment and the deviation of cladding height decreases, the feedback correction module determines that the current coordinated adjustment is effective, retains the corresponding adjustment direction and adjustment step size in the abnormality type and parameter influence relationship library, so that the adjustment relationship can continue to be called when the same cladding segment type and the same abnormality type occur in the future.

[0113] When the abnormality tendency level does not decrease after adjustment, or the cladding height deviation increases, the feedback correction module determines that the current coordinated adjustment has not achieved the expected effect, and corrects the relationship between abnormality type and parameter influence. The correction includes adjusting the main adjustment parameter, coordinated parameter and adjustment step size under the corresponding abnormality type. For example, for the under-melting abnormality tendency, if the thermal input state increases after increasing the laser power but the cladding height deviation still increases, the feedback correction module can reduce the adjustment priority of continuing to increase the laser power and increase the coordination adjustment priority of reducing the scanning speed or stabilizing the powder feeding amount.

[0114] If the boundary stability is not restored after reducing the laser power, the feedback correction module can prioritize increasing the scanning speed adjustment or shortening the local thermal action time in the case of an overmelting abnormality.

[0115] When the same abnormal tendency appears continuously in the same cladding section type, the feedback correction module further corrects the judgment boundary in the corresponding cladding abnormality judgment rule. The judgment boundary includes the heat input judgment boundary, powder absorption judgment boundary, boundary stability judgment boundary and forming height judgment boundary.

[0116] If an abnormal tendency repeatedly occurs in multiple consecutive cladding segments, it indicates that the original judgment boundary or parameter influence relationship library has failed to fully reflect the cladding state under the current material, trajectory, or heat accumulation conditions. The feedback correction module will tighten or loosen the corresponding judgment boundary according to the changes in the fusion state of the molten pool in the consecutive cladding segments, and use the correction results for the closed-loop adjustment of subsequent cladding segments.

[0117] In one embodiment, if the same type of overlap segment continuously exhibits an abnormal tendency of overlap offset, and the level of abnormal tendency does not decrease after spot position compensation, the feedback correction module corrects the boundary stability judgment boundary in the overlap offset anomaly judgment rule, and simultaneously corrects the adjustment step size of the spot position compensation amount in the anomaly type and parameter influence relationship library; if the abnormal tendency of overlap offset decreases after adjustment, and the cladding height deviation decreases, the feedback correction module retains the spot position compensation direction and adjustment step size, and uses it as the priority adjustment relationship for subsequent overlap segments.

[0118] Furthermore, when the abnormal tendency level is the limit level, the feedback correction module generates a protection control result. The limit level indicates that the current abnormal tendency has been maintained in the continuous cladding section, or has not been weakened after coordinated adjustment. Continuing to execute according to the original adjustment method may cause a decrease in the continuity of the cladding layer, an increase in the deviation of the cladding height, or instability of the molten pool. Therefore, the feedback correction module generates a protection control result according to the abnormal type corresponding to the limit level, and writes the protection control result back to the abnormal type and parameter influence relationship database.

[0119] When the limiting level is caused by the tendency of under-melting, the feedback correction module limits the scanning speed from continuing to increase and increases the heat input of the subsequent cladding stage to avoid insufficient powder melting due to excessive scanning speed or insufficient heat input.

[0120] When the limit level is caused by an abnormal tendency of over-melting or an abnormal tendency of interlayer heat accumulation, the feedback correction module limits the laser power from continuing to increase and reduces the heat input of the subsequent cladding section to avoid excessively wide molten pool, increased spatter, aggravated collapse, or continued accumulation of interlayer heat effects.

[0121] When the limit level is caused by an abnormal tendency to collapse, the feedback correction module limits the powder feeding amount from continuing to increase and adjusts the scanning speed of the subsequent cladding section to avoid excessive powder entering the molten pool and causing insufficient local support or high instability.

[0122] When the limitation level is caused by an abnormal tendency of overlap offset, the feedback correction module stops using the original spot position and generates a spot position compensation amount so that the laser action area corresponds to the overlapping area of ​​the adjacent melt channel.

[0123] In one specific embodiment, the current cladding segment is an interlayer continuation segment. The rule-based adjustment module has generated an abnormal tendency for interlayer heat accumulation and determined it to be a restriction level. After the feedback correction module obtains the adjusted molten pool fusion state, if the heat input state continues to rise and the forming height state fluctuates, the feedback correction module generates a protection control result to reduce the heat input of the subsequent cladding segment and restricts the laser power from continuing to rise.

[0124] Meanwhile, the feedback correction module writes the thermal input judgment boundary and adjustment step size corresponding to the inter-layer continuation segment into the abnormal type and parameter influence relationship library, so that subsequent inter-layer continuation segments will preferentially adopt control strategies that reduce thermal input or adjust scanning speed.

[0125] In another specific embodiment, the current cladding segment is an overlapping segment. The rule-based adjustment module generates an abnormal tendency of overlapping offset and outputs the spot position compensation amount. After the compensation is executed, the feedback correction module re-compares the positional relationship between the boundary of the molten pool and the overlapping area of ​​the adjacent molten channel. If the boundary stability is restored and the cladding height deviation is reduced, the corresponding spot position compensation amount is retained. If the boundary stability is still not restored, the spot position compensation step size is corrected and the original spot position is marked as a restricted use state to prevent subsequent overlapping segments from continuing to use the original spot position.

[0126] Through the above feedback correction and protection control, the system can continuously correct the cladding anomaly judgment rules, anomaly type and parameter influence relationship library and adjustment step size according to the actual cladding results after collaborative adjustment, so that the adjustment effect of the previous cladding segment can be applied to the subsequent cladding segment. Thus, the molten pool state identification, anomaly tendency judgment, parameter back-propagation adjustment and protection control in the laser rapid prototyping process form a closed loop, improving the continuity of the cladding layer, the consistency of the molten channel overlap and the stability of multi-layer forming.

[0127] In a specific operational example, the part to be formed is a metal component that requires multi-layer laser cladding. After the controller reads the forming trajectory of the part to be formed, the cladding execution module divides the forming trajectory into the starting section, straight section, corner section, overlapping section, ending section and interlayer continuation section according to the starting position of the path, the change of the path direction, the overlapping area of ​​adjacent melt channels and the interlayer connection area. It also binds the laser power, scanning speed, powder feeding amount and spot position to each cladding section and generates a cladding section execution record.

[0128] When the system performs cladding on a certain overlap segment, the cladding execution module drives the laser beam, powder feeding mechanism, and motion mechanism to operate according to the execution record corresponding to the overlap segment. The molten pool acquisition module simultaneously acquires the molten pool image, molten pool temperature, cladding height, and motion position of the overlap segment, and assigns the acquired data to the overlap segment according to the motion position. Subsequently, the molten pool acquisition module extracts the molten pool boundary, molten pool width, molten pool brightness change, and spatter change from the molten pool image, and generates the cladding height deviation based on the difference between the cladding height and the target forming height, thereby forming the cladding segment observation characteristics of the overlap segment.

[0129] The state fusion module uses the fusion state of the molten pool of the previous cladding segment, as well as the laser power, scanning speed, powder feeding amount and spot position bound to the overlapping segment as prediction inputs to generate the predicted state of the molten pool of the overlapping segment. Then, it uses the observation characteristics of the cladding segment of the overlapping segment as observation inputs to perform observation correction on the predicted state of the molten pool.

[0130] If a short-term boundary is missing in the molten pool image due to spatter occlusion, the state fusion module marks the short-term boundary missing as an abnormal observation and reduces its impact on the molten pool fusion state.

[0131] If the boundary of the molten pool continues to deviate from the overlapping area of ​​adjacent molten channels, and the deviation in cladding height increases synchronously, the state fusion module retains this change as a valid state change, generating a molten pool fusion state that includes the heat input state, powder absorption state, boundary stability state, and forming height state.

[0132] The rule-back-induction adjustment module reads the fusion state of the molten pool and the type of the cladding segment corresponding to the overlapping segment, and calls the cladding anomaly judgment rule corresponding to the overlapping segment. When the boundary stability state decreases and the molten pool boundary deviates from the overlapping area of ​​the adjacent molten channel, the rule-back-induction adjustment module generates an overlapping offset anomaly tendency.

[0133] When the abnormal tendency of the overlap offset is accompanied by an increase in the cladding height deviation, the rule-based adjustment module determines the abnormal tendency level as the adjustment level based on the number of state deviations, the magnitude of state deviations, and the type of cladding segment.

[0134] After determining the level of abnormal tendency, the rule-based adjustment module calls the database of abnormality types and parameter impact relationships, specifically including:

[0135] For abnormal tendencies in overlap offset, the rule-based back-calculation adjustment module determines the spot position as the main tuning parameter, the laser power or scanning speed as the cooperative parameter, and determines the adjustment step size according to the abnormal tendency level of the overlap segment, generating the spot position compensation amount and the corresponding cooperative adjustment amount.

[0136] After receiving the coordinated adjustment amount, the cladding execution module compensates for the position of the spot on the overlapping section or the subsequent adjacent overlapping section, so that the laser action area corresponds to the overlapping area of ​​the adjacent cladding channel again.

[0137] After the coordinated adjustment is executed, the feedback correction module re-acquires the adjusted molten pool fusion state, cladding height and abnormal tendency level, and compares them with the molten pool fusion state, cladding height and abnormal tendency level before adjustment;

[0138] When the boundary stability is restored after adjustment, the cladding height deviation decreases, and the abnormal tendency level is reduced from the adjustment level to the attention level or the normal level, the feedback correction module determines that the spot position compensation is effective and retains the corresponding adjustment direction and adjustment step size in the abnormal type and parameter influence relationship library.

[0139] When the adjusted molten pool boundary still deviates from the overlapping area of ​​adjacent molten channels, or the cladding height deviation continues to expand, the feedback correction module corrects the spot position compensation step size in the abnormality type and parameter influence relationship library, and corrects the boundary stability judgment boundary in the overlap offset abnormality judgment rule, so that the overlap section can adopt the updated judgment rule and adjustment relationship.

[0140] In another specific operational example, when the system performs interlayer cladding, the fusion state of the molten pool generated by the state fusion module shows a continuous increase in heat input and fluctuations in forming height. The rule-based adjustment module generates an abnormal tendency for interlayer heat accumulation based on the cladding anomaly judgment rules corresponding to the interlayer cladding.

[0141] If the abnormal tendency persists within the continuous cladding section and does not weaken after coordinated adjustment, the rule-based adjustment module determines the abnormal tendency level as the limit level. The feedback correction module generates protection control results under the limit level, restricting the laser power from continuing to increase and reducing the heat input of subsequent cladding sections. At the same time, the protection control results are written back to the abnormality type and parameter influence relationship database.

[0142] Through the above specific operation process, the cladding execution module, molten pool acquisition module, state fusion module, rule back-calculation adjustment module, and feedback correction module sequentially complete the execution of the cladding segment, generation of molten pool observation features, generation of molten pool fusion state, determination of abnormal tendency level, back-calculation of coordinated adjustment amount, and feedback correction. This transforms the molten pool control in the laser rapid prototyping process from a single parameter adjustment to a closed-loop control process based on the cladding segment, molten pool fusion state, and adjustment effect.

[0143] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0144] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0145] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0146] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0147] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A closed-loop intelligent cladding system for laser rapid prototyping, characterized in that: It includes a cladding execution module, a melt pool acquisition module, a state fusion module, a rule-based adjustment module, and a feedback correction module; The cladding execution module is used to drive the laser beam, powder feeding mechanism and motion mechanism to perform cladding according to the forming trajectory. It divides the forming trajectory into the melting start section, straight section, corner section, overlapping section, melting end section and interlayer continuation section, and binds the laser power, scanning speed, powder feeding amount and spot position to each cladding section. The molten pool acquisition module is used to acquire molten pool images, molten pool temperature, cladding height and movement position, extract molten pool boundary, molten pool width, molten pool brightness change, spatter change and cladding height deviation, and generate observation features of the cladding segment. The state fusion module is used to predict, update, and correct the observed features of the cladding section based on the extended Kalman filter algorithm, and generate the fusion state of the molten pool, which includes the heat input state, powder absorption state, boundary stability state, and forming height state. The rule-based adjustment module is used to match the fusion state of the molten pool, the type of the cladding section, and the rules for judging cladding anomalies, generate an anomaly tendency level, and call the anomaly type and parameter influence relationship library based on the anomaly tendency level to deduce the coordinated adjustment amount of laser power, scanning speed, powder feed amount, and spot position. The feedback correction module is used to compare the fusion state of the molten pool, the cladding height, and the level of abnormality before and after adjustment, correct the cladding anomaly judgment rules, the abnormality type and parameter influence relationship library, and the adjustment step size, and output the closed-loop cladding control results.

2. The intelligent cladding system for closed-loop molten pool in laser rapid prototyping according to claim 1, characterized in that: The cladding execution module is used to read the forming trajectory of the part to be formed, determine the cladding segment boundary based on the starting position, ending position, path direction change, overlapping area of ​​adjacent melt channels and interlayer connection area of ​​the forming trajectory, and mark the forming trajectory between the cladding segment boundaries as the starting melt segment, straight segment, corner segment, overlapping segment, melting end segment or interlayer continuation segment, and bind each cladding segment with laser power, scanning speed, powder feeding amount, spot position and execution time to generate a cladding segment execution record, so that the molten pool observation characteristics can be mapped to specific cladding segments and their process parameters.

3. The closed-loop intelligent cladding system for laser rapid prototyping according to claim 2, characterized in that: The molten pool acquisition module is used to simultaneously acquire molten pool images, molten pool temperature, molten pool height and movement position during the cladding process. The molten pool images, molten pool temperature and molten pool height are assigned to the corresponding cladding segment according to the movement position. The molten pool boundary, molten pool width, molten pool brightness change, spatter change and cladding height deviation are extracted from the acquired data after assignment to generate cladding segment observation features. The molten pool boundary and molten pool width are used to characterize the molten pool forming range, the molten pool brightness variation and molten pool temperature are used to characterize the heat input variation, the spatter variation is used to characterize the molten pool disturbance, and the cladding height deviation is used to characterize the cladding layer forming deviation.

4. The closed-loop intelligent cladding system for laser rapid prototyping according to claim 3, characterized in that: The state fusion module is used to generate the predicted state of the molten pool for the current cladding segment based on the extended Kalman filter algorithm, using the fusion state of the molten pool of the previous cladding segment, the laser power, scanning speed, powder feeding amount, and spot position of the current cladding segment as prediction inputs. It then uses the observation characteristics of the current cladding segment as observation inputs to perform observation correction on the predicted state of the molten pool, eliminating abnormal observations caused by molten pool image jitter, spatter obstruction, instantaneous temperature fluctuations, and height sampling offsets, thus generating the fusion state of the molten pool for the current cladding segment.

5. The closed-loop intelligent cladding system for laser rapid prototyping according to claim 4, characterized in that: The fusion state of the molten pool includes the heat input state, the powder absorption state, the boundary stability state, and the forming height state; The heat input state is determined by the molten pool temperature, molten pool brightness change and laser power; the powder absorption state is determined by the powder feed rate, molten pool width and cladding height; the boundary stability state is determined by the molten pool boundary, spatter change and spot position; and the forming height state is determined by the cladding height, cladding height deviation and scanning speed. The state fusion module associates the above states with the corresponding cladding segment types to generate a molten pool state record for determining the level of abnormality.

6. The closed-loop intelligent cladding system for laser rapid prototyping according to claim 5, characterized in that: The rule-back adjustment module is used to call the corresponding cladding anomaly judgment rule according to the cladding segment type, and match the heat input state, powder absorption state, boundary stability state and forming height state with the corresponding judgment boundary respectively; When the heat input is lower than the forming requirements, the powder absorption is insufficient, and the forming height decreases, there is a tendency for under-melting anomalies to occur. When the heat input is higher than the forming requirements, the boundary stability is reduced, and the spatter variation is increased, an over-melting anomaly is generated. When the forming height decreases locally and the continuity of the molten pool boundary is interrupted, a collapse anomaly tendency is generated; When the boundary of the molten pool of the overlapping section deviates from the overlapping area of ​​the adjacent molten channel, an abnormal tendency of overlapping offset is generated. When the heat input state of the interlayer joint continues to increase and the cooling recovery is insufficient, an abnormal tendency for interlayer heat accumulation is generated.

7. The closed-loop intelligent cladding system for laser rapid prototyping according to claim 6, characterized in that: The rule-based adjustment module is used to determine the level of abnormality based on the number of state deviations, the magnitude of state deviations, the number of consecutive cladding segments, and the type of cladding segments corresponding to the abnormality tendency. When the fusion state of the molten pool does not exceed the corresponding judgment boundary, a normal level is generated; When a single state exceeds the corresponding judgment boundary without causing an increase in the cladding height deviation, a level of concern is generated; An adjustment level is generated when at least two states exceed the corresponding judgment boundaries and cause the cladding height deviation to increase. When the abnormal tendency persists within a continuous cladding segment and does not diminish after coordinated adjustment, a restriction level is generated, and the abnormal tendency level is associated with and stored in relation to the corresponding cladding segment, abnormal type, and fusion state of the molten pool.

8. The closed-loop intelligent cladding system for laser rapid prototyping according to claim 7, characterized in that: The database of anomaly types and parameter influences includes anomaly type, cladding segment type, main control parameter, cooperative parameter, control direction, control step size, and parameter boundary. The rule-based adjustment module is specifically used to call the database of abnormality types and parameter influences based on the level of abnormality tendency to determine the coordinated adjustment amount of laser power, scanning speed, powder feed amount and spot position. The tendency of under-melting corresponds to increasing laser power, reducing scanning speed, and stabilizing powder feed rate; the tendency of over-melting corresponds to reducing laser power, increasing scanning speed, and shortening local thermal action time; the tendency of collapse corresponds to reducing powder feed rate and increasing scanning speed; the tendency of overlap offset corresponds to correcting the spot position; and the tendency of interlayer heat accumulation corresponds to reducing the heat input of the cladding section.

9. A closed-loop intelligent cladding system for laser rapid prototyping according to claim 8, characterized in that: The feedback correction module is used to obtain the fusion state, cladding height and abnormal tendency level of the molten pool after the coordinated adjustment is executed, and compare them with the fusion state, cladding height and abnormal tendency level of the molten pool before adjustment. When the abnormality tendency level decreases and the cladding height deviation decreases, the adjustment direction and adjustment step size in the corresponding abnormality type and parameter influence relationship database are retained. When the abnormality tendency level does not decrease or the cladding height deviation increases, correct the main parameters, coordinating parameters and adjustment step size in the corresponding abnormality type and parameter influence relationship library; When the same abnormal trend occurs continuously in the same cladding segment type, the judgment boundary in the corresponding cladding anomaly judgment rule is corrected, and the correction result is used for the closed-loop adjustment of subsequent cladding segments.

10. A closed-loop intelligent cladding system for laser rapid prototyping according to claim 9, characterized in that: The feedback correction module is also used to generate protection control results when the abnormal tendency level is the limit level. When the limiting level is caused by an undermelting anomaly, the limiting scan speed continues to increase and the heat input of the subsequent cladding section is increased; When the limiting level is caused by an abnormal tendency of over-melting or an abnormal tendency of interlayer heat accumulation, the laser power is limited to continue to increase and the heat input of the subsequent cladding section is reduced. When the restriction level is caused by an abnormal tendency of collapse, the powder feed rate should be increased and the scanning speed of the subsequent cladding section should be adjusted. When the restriction level is caused by an abnormal tendency of overlap offset, stop using the original spot position and generate a spot position compensation amount; The feedback correction module writes the protection control results back to the abnormality type and parameter influence relationship library, and outputs the closed-loop cladding control results of the corresponding cladding section.

Images