Cross-scale near-net shape manufacturing method for core component of conductor equipment

Abstract

The invention discloses a conductor equipment core component cross-scale near-net shape manufacturing method, and relates to the technical field of additive manufacturing, and the conductor equipment core component cross-scale near-net shape manufacturing method comprises the following steps: S1, in a forming process, synchronously executing macro-scale and micro-scale monitoring, and associating two types of data through coordinate calibration to obtain cross-scale associated data; and S2, based on the cross-scale associated data, performing fusion analysis with real-time process and environmental parameters, and outputting a process parameter adjustment direction according to conductor exclusive analysis logic. According to the cross-scale near-net-shape manufacturing method for the core component of the conductor equipment, the overall manufacturing quality and production efficiency of the core component of the conductor equipment can be remarkably improved, the deviation of the size and the organization structure is recognized in real time in the forming process through synchronous monitoring and data fusion of the macroscale and the microscale, and the manufacturing quality of the core component of the conductor equipment is improved. And intelligent analysis is carried out based on the characteristics of the conductor material, and dynamic optimization and closed-loop regulation and control of process parameters are realized.

Description

Technical Field

[0001] This invention relates to the field of additive manufacturing technology, specifically to a method for cross-scale near-net-shape manufacturing of core components for conductor equipment. Background Technology

[0002] In the field of high-end equipment manufacturing, the manufacturing quality of core components of conductor equipment directly affects the operational stability and service life of the equipment. Near-net-shape additive manufacturing and precision forming technologies have become the core development direction for the production of such components. Currently, the industry has put forward more stringent requirements for the macroscopic dimensional accuracy, microstructure, and comprehensive performance such as conductivity and mechanical properties of these core components. Related manufacturing technologies are advancing towards precision and efficiency, striving to improve the consistency of component forming quality and performance by optimizing process flows and monitoring methods.

[0003] Existing technologies have established a quality control model centered on performance prediction and post-forming inspection in the precision manufacturing of core components for conductor equipment, providing fundamental support for component quality assurance. However, in practical applications, some aspects still need improvement: Firstly, existing methods largely rely on final inspection results to judge component performance. If defects or substandard performance are found, remedial measures such as rework or scrapping are often the only options, failing to achieve real-time and precise intervention during the forming process, making it difficult to effectively reduce material and time costs. Secondly, the performance of conductor components is determined by both macroscopic dimensions and microstructure, while existing monitoring methods are often limited to a single scale, making it difficult to establish a real-time correlation between macroscopic deformation, microscopic evolution, and process parameters, affecting the synergy of control. Simultaneously, general-purpose analytical control models do not fully incorporate the unique properties of conductor materials, resulting in a need to improve the adaptability of control recommendations to the actual service requirements of components. To address these issues, we propose a cross-scale near-net-shape manufacturing method for core components of conductor equipment. Summary of the Invention

[0004] To address the aforementioned technical problems, a cross-scale near-net-shape manufacturing method for core components of conductor equipment is provided. This technical solution solves the problems mentioned above, such as the reliance on post-forming inspection in the manufacturing of core components of conductor equipment, the lack of real-time process intervention, the limitation of monitoring to a single scale, the difficulty in establishing a real-time correlation between macroscopic deformation, microscopic evolution and process parameters, and the fact that the general analysis and control model is not fully adapted to the characteristics of conductor materials and its adaptability to the actual service requirements of components needs to be improved.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A method for near-net-shape manufacturing of core components of a conductor device across scales includes the following steps: S1. During the forming process, macro-scale and micro-scale monitoring are performed simultaneously, and the two types of data are correlated through coordinate calibration to obtain cross-scale correlated data; S2. Based on the cross-scale correlation data, perform fusion analysis with real-time process and environmental parameters, and output the process parameter adjustment direction according to the conductor-specific analysis logic. S3. Based on the adjustment direction, determine the type of deviation and adjust the macroscopic process parameters or field auxiliary parameters in a hierarchical manner. After linkage verification, a monitoring-analysis-adjustment closed loop is formed. S4. In the closed-loop cycle, an adaptive control step is embedded for the conductor's conductivity and complex structural characteristics to optimize the molding quality in real time.

[0006] Preferably, in step S1, performing macroscopic-scale monitoring specifically involves: A laser scanning array and a vision sensor are used to collect surface geometric data of the forming area, and the three-dimensional point cloud coordinate sequence and two-dimensional texture image sequence of the deposition layer or the forming body are obtained in real time over the entire area. Outlier filtering and smoothing are performed on the three-dimensional point cloud coordinate sequence to extract the contour edge point set. Based on the two-dimensional texture image sequence, the contour boundary is corrected with the help of edge detection algorithm and fused to generate a high-precision surface contour topology mesh. Based on the preset component design model, the surface contour topology mesh is registered with the corresponding theoretical model in real time, the deviation between the actual shape and position dimensions of each monitoring point and the theoretical value is calculated, and macroscopic morphological deviation data and dynamic deformation gradient distribution map are output.

[0007] Preferably, in step S1, performing microscale monitoring specifically involves: An online electron backscatter diffraction probe and a Raman spectrometer are used to synchronously trigger acquisition in a preset key area of ​​the molded component, thereby acquiring the backscatter electron diffraction pattern sequence and Raman spectral signal sequence of the selected micro-area in real time. The backscattered electron diffraction pattern sequence is automatically calibrated and matched with the zone axis, the crystal orientation of each acquisition point is analyzed, an orientation mapping map is generated, and the grain boundary position and type data are extracted; the Raman spectral signal sequence is filtered and denoised and baseline subtracted, characteristic Raman peaks are identified by peak position fitting, and the phase composition and relative content are quantitatively analyzed. Based on the orientation mapping, the grain size distribution is statistically analyzed, and the texture intensity index and orientation difference angle distribution function are calculated. Simultaneously, based on the phase composition data, the type, morphology, and spatial distribution density of the second phase in the micro-region are identified. Combined with the peak position shift of the Raman spectrum, the local residual stress tensor is calculated. Integrating crystal orientation characteristics, second-phase distribution characteristics, and residual stress data, it outputs a set of microscale microstructure state parameters and real-time evolution curves.

[0008] Preferably, in step S1, obtaining cross-scale correlation data specifically involves: Using the base coordinate system of the molding equipment and the preset reference points on the surface of the component as a common spatial reference, the global coordinates of each vertex in the macroscopic monitoring topology grid are obtained, and the real-time robotic arm pose data of the probe center of the microscopic monitoring equipment is obtained simultaneously. The microscopic probe pose data is converted to the base coordinate system, and the three-dimensional spatial domain of the probe detection spot or electron beam spot acting on the surface of the component at the current moment is calculated. This domain is defined as the source space region for microscopic data acquisition. Based on the spatial nearest neighbor matching algorithm, the macro data point that is closest to the vertex set of the micro source spatial region is found in the macro monitoring topology grid, and the initial correspondence between the micro source region and the macro nearest neighbor point set is established. An affine transformation model is used to iteratively optimize the initial correspondence and solve for the optimal spatial coordinate transformation matrix, so as to accurately map the set of microscopic tissue structure state parameters from the probe coordinate system to the corresponding spatial position in the macroscopic base coordinate system. Based on the mapping results, each micro-analysis unit is bound to its corresponding macro-morphological deviation data and spatial coordinates, generating an integrated cross-scale associated data list indexed by spatial location. The micro-analysis unit refers to the basic data unit that characterizes the organizational structure state of a specific micro-region after being processed and analyzed in the micro-monitoring step.

[0009] Preferably, in step S2, the fusion analysis specifically involves: Based on cross-scale correlation data, real-time process parameter sequences and environmental parameter sequences are obtained. The process parameter sequences include: energy input, feed rate and path planning coordinates. The environmental parameter sequences include: temperature distribution in the molding chamber and concentration of atmosphere components. Timestamp alignment and spatial coordinate unification are performed on cross-scale correlated data, process parameter sequences, and environmental parameter sequences to construct a multi-source data cube; Extract subsets of macroscopic morphological geometric features, microscopic organizational structure features, process state features, and environmental disturbance features from the multi-source data cube; The macroscopic morphological geometric feature subset is coupled and correlated with the process state feature subset to generate process and morphological response feature vectors; the microscopic structure feature subset is coupled and correlated with the environmental disturbance feature subset to generate environmental and structural evolution feature vectors. Based on the process and morphology response feature vectors and the environment and structural evolution feature vectors, a cross-scale multi-physics joint feature matrix is ​​constructed as the input data for the analysis logic.

[0010] Preferably, in step S2, the adjustment direction of the output process parameters based on the conductor-specific analysis logic is specifically as follows: Based on the aforementioned cross-scale multiphysics joint feature matrix, the conductor material-structure-performance association knowledge base and dynamic threshold table are invoked. The knowledge base stores the mapping relationship between the macroscopic morphology characteristics, microstructure characteristics and final conductivity and mechanical properties of different conductor materials under different process histories. Multidimensional feature decoupling is performed on the joint feature matrix to identify the macroscopic morphology-dominated deviation mode and the microstructure-dominated evolution mode in the current forming state, and the three-dimensional region where each mode occurs is located based on the spatial location index. Based on the aforementioned conductor material-structure-performance correlation knowledge base, the performance impact weight of the identified deviations and evolution patterns is evaluated, and their potential impact on the target conductivity and mechanical properties of the component is calculated. Based on the performance impact weight evaluation results and the tolerance range preset for the current material and structure in the dynamic threshold table, a multi-objective decision analysis is performed. The generated adjustment vector is input into the process parameter-field parameter coupling influence model for simulation verification. The chain effect of implementing the adjustment on non-target performance dimensions and other related regions is evaluated. The magnitude and timing of the adjustment vector are adjusted according to the verification results. The output includes a set of process parameter adjustment instructions, which includes the parameter adjustment category, direction, magnitude, area of ​​effect, and expected control target, and serves as the input for hierarchical dynamic feedback.

[0011] Preferably, the multi-objective decision analysis specifically includes: If the macroscopic dimensional deviation exceeds the tolerance and has a high weight, a first-class adjustment vector is generated with the correction of geometric accuracy as its core, pointing to the macroscopic process parameters of energy input and feed rate; if the microstructure evolution is detrimental to the target performance and has a high weight, a second-class adjustment vector is generated with the optimization of the microstructure as its core, pointing to the electromagnetic field and ultrasonic field auxiliary parameters; if the two are intertwined and have equal weight, a composite adjustment vector is generated and the order of primary and secondary control is determined.

[0012] Preferably, in step S3, the determination of the deviation type is specifically hierarchically as follows: Receive the set of process parameter adjustment instructions and parse the parameter adjustment categories and their effective areas; Based on the parameter adjustment category, it is determined whether the parameter to be adjusted belongs to a macroscopic process parameter or a field auxiliary parameter. If macroscopic process parameters need to be adjusted, the energy input and feed rate should be adjusted according to the adjustment direction and magnitude in the instruction set. If field auxiliary parameters need to be adjusted, adjust the electromagnetic field strength and frequency or the ultrasonic field power and frequency according to the adjustment direction and amplitude in the instruction set. If it is necessary to adjust both macroscopic process parameters and field auxiliary parameters simultaneously, the corresponding parameters should be adjusted sequentially according to the primary and secondary control order specified in the instruction set.

[0013] Preferably, in step S3, the monitoring-analysis-adjustment closed-loop cycle formed after the linkage verification specifically includes: Before performing parameter adjustments, the macroscopic process parameters and field auxiliary parameters to be adjusted are input into the preset process coupling influence model. Run the process coupling effect model to predict the chain effect of the parameter adjustment on the current forming area and the adjacent unforming area; If the predicted chain reaction does not exceed the preset safety threshold, the parameter adjustment is confirmed to be executed; if the predicted chain reaction exceeds the safety threshold, the adjustment range is reduced or the adjustment timing is postponed, and the modified parameter adjustment instruction is generated and executed. After the parameter adjustment is completed, the next round of cross-scale collaborative monitoring of the adjusted area is triggered, and the newly collected cross-scale correlation data is re-input into the fusion analysis and analysis logic steps. Adjust the direction based on the newly output process parameters, repeat the hierarchical adjustment and linkage verification to form a closed loop until the entire component is formed.

[0014] Preferably, in step S4, the step of embedding adaptive control over the conductor's conductivity and complex structural features specifically includes: In the closed-loop cycle, the grain boundary density and second phase distribution data output by microscale monitoring are acquired in real time; Based on the conductivity model of the conductor material and the data, the predicted conductivity of the current forming area is calculated in real time. The predicted conductivity is compared with the target conductivity. If the deviation exceeds the set range, the field-aided parameter adjustment for optimizing the microstructure is triggered first in the hierarchical adjustment step. When the forming path enters the variable cross-section or thin-walled region of the component, the spatial sampling interval between macroscopic and microscopic monitoring is reduced, and the frequency of monitoring data acquisition is increased; For the variable cross-section or thin-walled region, a dedicated feature extraction algorithm for the region with abrupt structural changes is enabled in the fusion analysis step, and a dynamic threshold table preset for this type of region is called in the analysis logic step.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: The cross-scale near-net-shape manufacturing method for core components of conductor equipment proposed in this invention can significantly improve the overall manufacturing quality and production efficiency of core components of conductor equipment. Through synchronous monitoring and data fusion at the macro and micro scales, deviations in size and microstructure are identified in real time during the forming process. Based on the characteristics of conductor materials, intelligent analysis is performed to achieve dynamic optimization and closed-loop control of process parameters. Quality control is moved from traditional post-production inspection to the manufacturing process, which greatly reduces rework and material loss caused by defects. At the same time, cross-scale collaboration ensures the consistency of components in geometric accuracy and microstructure, thereby ensuring the reliability of the final product in terms of conductivity, mechanical properties, etc. It has strong adaptability and learning capabilities and can make targeted adjustments for different conductor materials, complex structural features, and key performance indicators, enhancing the adaptability and stability of the process. Attached Figure Description

[0016] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0017] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.

[0018] Reference Figure 1 As shown, a method for near-net-shape manufacturing of core components of a conductor device across scales includes the following steps: S1. During the forming process, macro-scale and micro-scale monitoring are performed simultaneously, and the two types of data are correlated through coordinate calibration to obtain cross-scale correlated data; In S1, the macro-scale monitoring is specifically performed as follows: A laser scanning array and a vision sensor are used to collaboratively acquire surface geometric data of the molding area. The synchronous triggering of the two is achieved based on an FPGA hardware synchronization clock, with the synchronization clock frequency set to 100MHz. The scanning frequency of the laser scanning array and the frame frequency of the vision sensor are strictly matched to 50Hz. At the same time, parallel data transmission and synchronous verification are achieved through gigabit Ethernet to ensure that the acquisition timing deviation is less than 1μs. The three-dimensional point cloud coordinate sequence and two-dimensional texture image sequence of the deposition layer or the molded body in the whole area are acquired in real time. Outlier removal was performed on the 3D point cloud coordinate sequence using a statistical filtering algorithm. The number of neighboring points was set to 15, and the standard deviation threshold was set to 2.5. Discrete points that deviated from the average distance of the neighboring points by more than 2.5 times the standard deviation were removed. Then, the moving least squares method was used for smoothing, and a quadratic basis function was selected. The influence domain radius was set to 1.2 times the average spacing of the point cloud. The Canny edge detection algorithm was used to assist in correcting the contour boundary. The algorithm was selected because it has the characteristic of suppressing false edges with a double threshold. The high threshold was set to 80, and the low threshold was set to 40. The fusion rule adopted a weighted average strategy. The weight coefficient of the edge points of the 3D point cloud contour was 0.7, and the weight coefficient of the edge points of the 2D texture image was 0.3. A high-precision surface contour topology mesh was generated through weighted fusion. Based on the preset component design model, the surface contour topological mesh is registered with the corresponding theoretical model in real time. The registration algorithm adopts an improved iterative nearest-point algorithm. The specific implementation steps are as follows: First, initial registration is performed based on the reference marker points to obtain the initial transformation matrix; then, the Euclidean distance between corresponding point pairs in the current point cloud and the target point cloud is calculated, and a distance threshold of 0.1 mm is set to remove abnormal corresponding points exceeding the threshold; the optimal transformation matrix is ​​solved using the least squares method, and the point cloud position is iteratively updated; the iteration termination condition is set to the root mean square error of two adjacent iterations being less than 1 mm. Alternatively, after 50 iterations and convergence, the deviation between the actual shape and position dimensions of each monitoring point and the theoretical value is calculated. The shape deviation data is calculated using the following formula: in Due to morphological deviation, These are the actual geometric dimensions. The theoretical dimensions are used; the dynamic deformation gradient distribution diagram uses the formula. Calculation, where For the deformation gradient tensor, As a gradient operator, it outputs macroscopic topographic deviation data and dynamic deformation gradient distribution map.

[0019] In S1, the microscale monitoring is specifically performed as follows: An online electron backscatter diffraction probe and a Raman spectrometer are used to synchronously trigger acquisition in a preset key area of ​​the molded component. The synchronous control logic is based on the closed-loop feedback of the robotic arm's posture. When the robotic arm drives the probe to the preset key area coordinates, a synchronous acquisition signal is triggered. The signal delay is controlled within 5ms. At the same time, the rising edge of the trigger signal is used to achieve precise alignment of the acquisition timing of the two, and the backscatter electron diffraction pattern sequence and Raman spectral signal sequence of the selected micro-area are obtained in real time. The backscattered electron diffraction pattern sequence was automatically calibrated using Hough transform. A standard crystallography database (containing standard unit cell parameters, interplanar spacing, and other data of the target conductor material) was called. The zone axis matching adopted the angle matching method, and the matching threshold was set to 1°. That is, when the angle between the calculated zone axis and the standard zone axis is less than 1°, the matching is considered successful. The crystal orientation of each collection point was analyzed, an orientation mapping map was generated, and the position and type data of the grain boundary were extracted. The grain boundary extraction criteria were that the orientation difference angle between adjacent grains was greater than 15° and was considered a large-angle grain boundary, and less than 15° and was considered a small-angle grain boundary. The Raman spectral signal sequence was denoised using the Savitzky-Golay filtering algorithm. A polynomial order of 3 and a window width of 11 data points were selected. A polynomial fitting baseline subtraction method was used, with the fitting order set to 5. Characteristic Raman peaks were identified using a Lorentz-Gaussian mixed peak fitting model. The expression of the mixed model is as follows: in For the fitting strength, , These are the amplitudes of the Gauss peak and the Lorentz peak, respectively. , These are the peak positions of the two peaks, The half-height width of Gauss Peak, The half-height and width of Lorentz Peak The baseline offset is used to quantitatively analyze the phase composition and relative content through peak position fitting. Based on the orientation mapping, the grain size distribution was statistically analyzed using the intercept method. The test line length was set to 100 μm, and each test line passed through at least 10 grains. The texture intensity index and orientation difference angular distribution function were calculated. The texture intensity index was calculated using the following formula: in The texture strength index, This represents the actual orientation distribution function value. The distribution function value is given under random orientation. Simultaneously, based on the phase composition data, the type, morphology, and spatial distribution density of the second phase within the micro-region are identified. The spatial distribution density is characterized by the number of second-phase particles per unit area. Combining the peak shift of the Raman spectrum, the local residual stress tensor is calculated using the following formula: in For residual stress, This represents the Raman peak shift. This is the stress coefficient, which is obtained through standard specimen calibration experiments. Specifically, a known gradient stress is applied to the standard specimen, the corresponding peak position shift is measured, and linear fitting is used to obtain the coefficient. value; Integrating crystal orientation characteristics, second-phase distribution characteristics, and residual stress data, it outputs a set of microscale microstructure state parameters and real-time evolution curves.

[0020] In S1, the cross-scale correlation data is obtained as follows: Using the base coordinate system of the molding equipment and the preset reference marks on the surface of the component as a common spatial reference, the base coordinate system of the molding equipment is defined as a rectangular coordinate system with the center of the equipment worktable as the origin, the X-axis along the length of the worktable, the Y-axis along the width of the worktable, and the Z-axis perpendicular to the plane of the worktable. The reference markers adopt a circular pit structure with a diameter of 2mm and a depth of 0.5mm. The arrangement rule is to distribute them evenly on the surface of the component, with no less than 4 markers. The distance between adjacent markers is no less than 50mm. The coordinate measurement is carried out with a laser tracker for precise measurement with a measurement accuracy of ±5μm. The global coordinates of each vertex in the macroscopic monitoring topology grid are obtained, and the real-time robotic arm pose data of the probe center of the microscopic monitoring device is obtained simultaneously. The microprobe pose data is transformed to the base coordinate system using a homogeneous coordinate transformation matrix, the expression of which is: ,in It is a 3×3 rotation matrix, representing the rotation relationship between the probe coordinate system and the base coordinate system. This is a 3×1 translation vector, representing the amount of translation between the origin of the probe coordinate system and the origin of the base coordinate system. and The three-dimensional spatial domain of the probe detection spot or electron beam spot acting on the surface of the component at the current moment is obtained by using the forward kinematics solution of the robotic arm. This domain is defined as the source space region for microscopic data acquisition. Based on the spatial nearest neighbor matching algorithm, the KD tree algorithm is selected to achieve fast nearest neighbor search. The selection is based on its efficient search performance in high-dimensional space. In the macroscopic monitoring topology grid, the macroscopic data point that is closest to the vertex set of the microscopic source space region is found, and the initial correspondence between the microscopic source region and the macroscopic nearest neighbor point set is established. Using the affine transformation model, the mathematical expression for the affine transformation is: in , which is the coordinate vector in the macroscopic base coordinate system. The coordinate vector in the probe coordinate system. It is a 3×3 linear transformation matrix. Given a 3×1 translation vector, the initial correspondence is iteratively optimized. The optimal spatial coordinate transformation matrix is ​​solved using the least squares method, with the root mean square error of the corresponding point pairs as the objective function: The iteration termination condition is that the root mean square error is less than 100%. Or, if the number of iterations reaches 30, the set of microscopic tissue structure state parameters can be accurately mapped from the probe coordinate system to the corresponding spatial position in the macroscopic base coordinate system; Based on the mapping results, each micro-analysis unit is bound to its corresponding macro-morphological deviation data and spatial coordinates, generating an integrated cross-scale correlation data list indexed by spatial location. The micro-analysis unit refers to the basic data unit that characterizes the organizational structure state of a specific micro-region after parsing and processing in the micro-monitoring step. The integrated cross-scale correlation data list adopts a tree index structure, with the macro-spatial coordinate interval as the first-level index and the micro-analysis unit number as the second-level index. The storage format adopts JSON format to facilitate fast data reading and parsing.

[0021] S2. Based on the cross-scale correlation data, perform fusion analysis with real-time process and environmental parameters, and output the process parameter adjustment direction according to the conductor-specific analysis logic. In S2, the fusion analysis specifically involves: Based on cross-scale correlation data, real-time process parameter sequences and environmental parameter sequences are obtained. The process parameter sequences include: energy input, feed rate and path planning coordinates. The environmental parameter sequences include: temperature distribution in the molding chamber and concentration of atmosphere components. Cross-scale correlated data, process parameter sequences, and environmental parameter sequences are processed for timestamp alignment and spatial coordinate unification. Timestamp alignment adopts a synchronization mechanism that combines hardware synchronization clock with software interpolation completion. Based on the master clock of the molding equipment, the timestamps of each data source are converted into a unified master clock time. For missing timestamp data, linear interpolation is used to complete it. Spatial coordinate unification adopts a mapping rule based on the base coordinate system. The spatial coordinates of all data sources are transformed to the base coordinate system of the molding equipment to construct a multi-source data cube. The dimensions of the data cube are defined as time dimension, spatial X dimension, spatial Y dimension, spatial Z dimension, and feature dimension. From the multi-source data cube, subsets of macroscopic morphological geometric features, microscopic microstructure features, process state features, and environmental disturbance features are extracted. Specific extraction indicators for the macroscopic morphological geometric feature subset include contour curvature, surface roughness, peak value of morphological deviation, valley value of morphological deviation, and the trace of the deformation gradient tensor. Contour curvature is calculated using the following formula: In the formula, The equation for the profile curve is... The x-axis represents the curve, and the surface roughness is expressed as the arithmetic mean deviation of the profile. ,Right now: In the formula, For length measurement; specific extraction indicators for the microstructure feature subset include average grain size, standard deviation of grain size, proportion of large-angle grain boundaries, texture intensity index, volume fraction of second phase, and peak residual stress. The grain size measurement standard adopts the equivalent circle diameter method, that is, the equivalent circle diameter is calculated from the grain area. In the formula, The area is the grain size. A subset of macroscopic morphological geometric features is coupled with a subset of process state features. A feature weighted summation algorithm is used to generate process and morphological response feature vectors. The weight coefficients are determined by the analytic hierarchy process (AHP). The weight coefficient for energy input is 0.4, the weight coefficient for feed rate is 0.3, and the weight coefficient for path planning coordinates is 0.3. After standardization of each feature, the feature vector is obtained by weighted summation. A subset of microscopic organizational structure features is coupled with a subset of environmental disturbance features. Principal component analysis is used for dimensionality reduction. The first three principal components are extracted, and their cumulative contribution rate is not less than 90%, generating environmental and structural evolution feature vectors. Based on the feature vectors of process and morphological response and the feature vectors of environment and structural evolution, a cross-scale multiphysics joint feature matrix is ​​constructed. The matrix dimension is defined as the number of samples × (the dimension of process and morphological response features + the dimension of environment and structural evolution features). Feature normalization adopts the min-max normalization method, i.e.: in These are the original eigenvalues. The minimum value of the characteristic. To find the maximum eigenvalue, the normalized eigenvectors are arranged in rows to construct a joint feature matrix, which serves as the input data for the analysis logic.

[0022] In step S2, the adjustment direction of the output process parameters based on the conductor-specific analysis logic is specifically as follows: Based on the aforementioned cross-scale multiphysics joint feature matrix, the conductor material-structure-performance association knowledge base and dynamic threshold table are invoked. The knowledge base stores the mapping relationship between the macroscopic morphology, microstructure, and final conductivity and mechanical properties of different conductor materials under different process histories. The experimental basis for the mapping relationship is obtained through systematic process experiments. Samples are prepared for different combinations of process parameters, and the macroscopic morphology, microstructure, conductivity, and mechanical properties of each sample are tested to establish an experimental database. At the same time, the experimental data are supplemented and verified by the finite element simulation model, which is constructed based on the thermo-mechanical-electric multiphysics coupling theory. The principle for setting the dynamic threshold table is based on the performance standards of conductor materials, referring to relevant industry specifications, and combining the actual working requirements of the components to determine the tolerance range. For example, for copper conductor components, the tolerance range of conductivity is set to ±3% of the target value, and the tolerance range of tensile strength in mechanical properties is set to ±5% of the target value. Multidimensional feature decoupling is performed on the joint feature matrix. Principal component analysis is used to achieve feature decoupling. Principal components are extracted as independent feature dimensions to identify the macroscopic morphology-dominated deviation mode and the microstructure-dominated evolution mode in the current forming state. The deviation and evolution mode identification adopts the support vector machine classification algorithm. The classification model is trained through training set data. The kernel function of the model is selected as the radial basis function, the penalty coefficient is set to 10, and the gamma parameter is set to 0.1. Based on the spatial location index, the three-dimensional region where each mode occurs is located. Based on the aforementioned conductor material-structure-performance correlation knowledge base, the performance impact weights of the identified deviations and evolution patterns are evaluated. A judgment matrix is ​​constructed using the analytic hierarchy process (AHP) to determine the impact weights of each deviation and evolution pattern on conductivity and mechanical properties. The weight coefficients are calculated by decomposing the eigenvalues ​​of the judgment matrix to obtain the eigenvector corresponding to the largest eigenvalue, normalizing it, and using this normalized vector as the weight coefficient. The potential impact values ​​on the target conductivity and mechanical properties of the component are then calculated, and these potential impact values ​​are obtained through weighted summation. ,in These are the weighting coefficients. This refers to the performance deviation. Based on the performance impact weight evaluation results and the preset tolerance ranges for the current material and structure in the dynamic threshold table, a multi-objective decision analysis is performed, and the optimized objective function is: in Due to conductivity deviation, For mechanical performance deviations, the constraint is that the adjustment range of process parameters shall not exceed the rated operating range of the equipment; The generated adjustment vector is input into the process parameter-field parameter coupling influence model for simulation verification. The process parameter-field parameter coupling influence model adopts a finite element-based physical simulation model. The governing equations of the model include the heat conduction equation, stress balance equation, and electromagnetic field equation. The input is the adjustment amount of process parameters and field parameters, and the output is the temperature field, stress field, and microstructure evolution results of the forming area. The chain effect of implementing this adjustment on non-target performance dimensions and other related areas is evaluated. The verification standard is that the deviation of non-target performance dimensions after adjustment does not exceed the tolerance range, and the performance fluctuation of related areas is less than 2%. The magnitude and timing of the adjustment vector are adjusted according to the verification results. The output includes a set of process parameter adjustment instructions, which includes the parameter adjustment category, direction, magnitude, area of ​​effect, and expected control target, and serves as the input for hierarchical dynamic feedback.

[0023] The multi-objective decision analysis specifically refers to: If the macroscopic dimensional deviation exceeds the tolerance and has a high weight, a first-class adjustment vector is generated with the correction of geometric accuracy as its core, pointing to the macroscopic process parameters of energy input and feed rate; if the microstructure evolution is detrimental to the target performance and has a high weight, a second-class adjustment vector is generated with the optimization of the microstructure as its core, pointing to the electromagnetic field and ultrasonic field auxiliary parameters; if the two are intertwined and have equal weight, a composite adjustment vector is generated and the order of primary and secondary control is determined.

[0024] S3. Based on the adjustment direction, determine the type of deviation and adjust the macroscopic process parameters or field auxiliary parameters in a hierarchical manner. After linkage verification, a monitoring-analysis-adjustment closed loop is formed. In S3, the deviation type is determined by hierarchical classification as follows: Receive the set of process parameter adjustment instructions and parse the parameter adjustment categories and their effective areas; Based on the parameter adjustment category, it is determined whether the parameter to be adjusted belongs to macroscopic process parameter or field auxiliary parameter. The specific rules for the deviation type judgment are based on the threshold division standard of deviation amplitude. The macroscopic size deviation threshold is set to 0.2mm. When the macroscopic size deviation is greater than 0.2mm, it is determined to be macroscopic deviation-dominant. The microstructure performance deviation threshold is set to 5%. When the performance deviation corresponding to the microstructure is greater than 5%, it is determined to be microdeviation-dominant. If macroscopic process parameters need to be adjusted, the energy input and feed rate should be adjusted according to the adjustment direction and magnitude in the instruction set. The adjustment step size for energy input is set to 50W, and the adjustment step size for feed rate is set to 0.1mm / s. The adjustment magnitude should not exceed ±20% of the current value. If field auxiliary parameters need to be adjusted, adjust the electromagnetic field strength and frequency or the ultrasonic field power and frequency according to the adjustment direction and amplitude in the instruction set. The adjustment range of electromagnetic field strength is 0-5T with an adjustment step of 0.1T, and the adjustment range of frequency is 10-100kHz with an adjustment step of 1kHz. The adjustment range of ultrasonic field power is 0-1000W with an adjustment step of 50W, and the adjustment range of frequency is 20-80kHz with an adjustment step of 5kHz. If it is necessary to adjust both macroscopic process parameters and field auxiliary parameters at the same time, then according to the primary and secondary control sequence specified in the instruction set, first adjust the parameter corresponding to the primary deviation, stabilize for 3 seconds after adjustment, and then adjust the parameter corresponding to the secondary deviation, and so on.

[0025] In S3, the monitoring-analysis-adjustment closed-loop cycle formed after the linkage verification is specifically as follows: Before performing parameter adjustments, the macroscopic process parameters and field auxiliary parameters to be adjusted are input into the preset process coupling influence model. The process coupling effect model is run. The process coupling effect model adopts a response surface model. Sample data is obtained through a central composite design experiment. The response surface equation between process parameters and molding quality indicators is constructed. The chain effect of parameter adjustment on the current molding area and adjacent unmolded areas is predicted. The prediction algorithm details are: substituting the adjusted parameters into the response surface equation to calculate the corresponding change in molding quality indicators. If the predicted chain reaction does not exceed the preset safety threshold, which is set to 80% of the tolerance range of the molding quality index, then the parameter adjustment is confirmed to be executed; if the predicted chain reaction exceeds the safety threshold, then the adjustment range is reduced by the ratio of the excess to the safety threshold, or the adjustment sequence is postponed by 5 seconds, and the modified parameter adjustment instruction is generated and executed. After the parameter adjustment is completed, the next round of cross-scale collaborative monitoring of the adjusted area is triggered. The triggering method adopts an event-driven mechanism, that is, after the parameter adjustment is completed, a trigger signal is sent to trigger the next round of monitoring. After the monitoring data is collected, the newly collected cross-scale correlation data is re-input into the fusion analysis and analysis logic steps. Adjust the direction according to the newly output process parameters, repeat the hierarchical adjustment and linkage verification to form a closed loop. The criterion for the termination of the loop is that the overall part is formed and the morphological deviation and performance deviation of all monitoring points are within the tolerance range. It should be noted that the criterion for the overall part being formed is that the forming path is completed and the deposition layer thickness meets the design requirements.

[0026] S4. In the closed-loop cycle, an adaptive control step is embedded for the conductor's conductivity and complex structural characteristics to optimize the molding quality in real time.

[0027] In step S4, the adaptive control step for the conductor's conductivity and complex structural features is specifically as follows: In the closed-loop cycle, the grain boundary density and second phase distribution data output by microscale monitoring are acquired in real time; Based on the conductivity model of the conductor material and the aforementioned data, the predicted conductivity of the current forming region is calculated in real time. The conductivity model of the conductor material adopts a formula based on grain boundary scattering theory, and the specific expression is as follows: in To predict conductivity, The conductivity of a defect-free conductor material The grain boundary scattering coefficient was obtained through experimental calibration. Grain boundary density, The second-phase scattering coefficient was also obtained through experimental calibration. Given the volume fraction of the second phase, the calculation process for predicting conductivity is as follows: first, read the grain boundary density and the volume fraction of the second phase data, substitute them into the above formula to calculate the preliminary predicted value, and then correct it using a temperature correction coefficient, which is calculated based on the real-time temperature of the molding chamber. The predicted conductivity is compared with the target conductivity. If the deviation exceeds the set range, which is ±3% of the target conductivity, then in the hierarchical adjustment step, the field-aided parameter adjustment for optimizing the microstructure is triggered first. When the forming path enters the variable cross-section or thin-walled region of the component, the variable cross-section region is defined as the region where the change rate of adjacent cross-section dimensions is greater than 20%, and the thin-walled region is defined as the region with a thickness of less than 5 mm. The spatial sampling interval for macroscopic and microscopic monitoring is reduced from 0.5 mm to 0.2 mm, and the spatial sampling interval for microscopic monitoring is reduced from 10 μm to 5 μm. The monitoring data acquisition frequency is also increased from 50 Hz to 100 Hz. For the variable cross-section or thin-walled region, a dedicated feature extraction algorithm for structural abrupt changes is enabled in the fusion analysis step. This algorithm includes structural abrupt change identification and stress concentration factor calculation. Structural abrupt change identification uses a curvature abrupt change detection method; a structural abrupt change is determined when the rate of change of contour curvature is greater than 5% / mm. The stress concentration factor is calculated using the following formula: in The stress concentration factor is... The characteristic size of the mutation region. The radius of the fillet in the abrupt change region is used; in the analysis logic step, a dynamic threshold table preset for this type of region is called. The difference setting of this dynamic threshold table is based on the structural bearing capacity and conductivity requirements of the variable cross-section and thin-walled regions. The tolerance range is reduced by 30% compared to the ordinary region. For example, the macroscopic size deviation tolerance of the ordinary region is 0.2mm, while that of this type of region is reduced to 0.14mm.

[0028] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention. The scope of protection claimed by the appended claims and their equivalents is defined.

Claims

1. A method for near-net-shape manufacturing of core components of a conductor device across scales, characterized in that, Includes the following steps: S1. During the forming process, macro-scale and micro-scale monitoring are performed simultaneously, and the two types of data are correlated through coordinate calibration to obtain cross-scale correlated data; S2. Based on the cross-scale correlation data, perform fusion analysis with real-time process and environmental parameters, and output the process parameter adjustment direction according to the conductor-specific analysis logic. S3. Based on the adjustment direction, determine the type of deviation and adjust the macroscopic process parameters or field auxiliary parameters in a hierarchical manner. After linkage verification, a monitoring-analysis-adjustment closed loop is formed. S4. In the closed-loop cycle, an adaptive control step is embedded for the conductor's conductivity and complex structural characteristics to optimize the molding quality in real time.

2. The method for cross-scale near-net-shape manufacturing of core components of a conductor device according to claim 1, characterized in that, In S1, the macro-scale monitoring is specifically performed as follows: A laser scanning array and a vision sensor are used to collect surface geometric data of the forming area, and the three-dimensional point cloud coordinate sequence and two-dimensional texture image sequence of the deposition layer or the forming body are obtained in real time over the entire area. Outlier filtering and smoothing are performed on the three-dimensional point cloud coordinate sequence to extract the contour edge point set. Based on the two-dimensional texture image sequence, the contour boundary is corrected with the help of edge detection algorithm and fused to generate a high-precision surface contour topology mesh. Based on the preset component design model, the surface contour topology mesh is registered with the corresponding theoretical model in real time, the deviation between the actual shape and position dimensions of each monitoring point and the theoretical value is calculated, and macroscopic morphological deviation data and dynamic deformation gradient distribution map are output.

3. The method for cross-scale near-net-shape manufacturing of core components of a conductor device according to claim 2, characterized in that, In S1, the microscale monitoring is specifically performed as follows: An online electron backscatter diffraction probe and a Raman spectrometer are used to synchronously trigger acquisition in a preset key area of ​​the molded component, thereby acquiring the backscatter electron diffraction pattern sequence and Raman spectral signal sequence of the selected micro-area in real time. The backscattered electron diffraction pattern sequence is automatically calibrated and matched with the zone axis, the crystal orientation of each acquisition point is analyzed, an orientation mapping map is generated, and the grain boundary position and type data are extracted; the Raman spectral signal sequence is filtered and denoised and baseline subtracted, characteristic Raman peaks are identified by peak position fitting, and the phase composition and relative content are quantitatively analyzed. Based on the orientation mapping, the grain size distribution is statistically analyzed, and the texture intensity index and orientation difference angle distribution function are calculated. Simultaneously, based on the phase composition data, the type, morphology, and spatial distribution density of the second phase in the micro-region are identified. Combined with the peak position shift of the Raman spectrum, the local residual stress tensor is calculated. Integrating crystal orientation characteristics, second-phase distribution characteristics, and residual stress data, it outputs a set of microscale microstructure state parameters and real-time evolution curves.

4. The method for cross-scale near-net-shape manufacturing of core components of a conductor device according to claim 3, characterized in that, In S1, the cross-scale correlation data is obtained as follows: Using the base coordinate system of the molding equipment and the preset reference points on the surface of the component as a common spatial reference, the global coordinates of each vertex in the macroscopic monitoring topology grid are obtained, and the real-time robotic arm pose data of the probe center of the microscopic monitoring equipment is obtained simultaneously. The microscopic probe pose data is converted to the base coordinate system, and the three-dimensional spatial domain of the probe detection spot or electron beam spot acting on the surface of the component at the current moment is calculated. This domain is defined as the source space region for microscopic data acquisition. Based on the spatial nearest neighbor matching algorithm, the macro data point that is closest to the vertex set of the micro source spatial region is found in the macro monitoring topology grid, and the initial correspondence between the micro source region and the macro nearest neighbor point set is established. An affine transformation model is used to iteratively optimize the initial correspondence and solve for the optimal spatial coordinate transformation matrix, so as to accurately map the set of microscopic tissue structure state parameters from the probe coordinate system to the corresponding spatial position in the macroscopic base coordinate system. Based on the mapping results, each micro-analysis unit is bound to its corresponding macro-morphological deviation data and spatial coordinates, generating an integrated cross-scale associated data list indexed by spatial location. The micro-analysis unit refers to the basic data unit that characterizes the organizational structure state of a specific micro-region after being processed and analyzed in the micro-monitoring step.

5. A method for cross-scale near-net-shape manufacturing of core components of a conductor device according to claim 4, characterized in that, In S2, the fusion analysis specifically involves: Based on cross-scale correlation data, real-time process parameter sequences and environmental parameter sequences are obtained. The process parameter sequences include: energy input, feed rate and path planning coordinates. The environmental parameter sequences include: temperature distribution in the molding chamber and concentration of atmosphere components. Timestamp alignment and spatial coordinate unification are performed on cross-scale correlated data, process parameter sequences, and environmental parameter sequences to construct a multi-source data cube; Extract subsets of macroscopic morphological geometric features, microscopic organizational structure features, process state features, and environmental disturbance features from the multi-source data cube; The macroscopic morphological geometric feature subset is coupled and correlated with the process state feature subset to generate process and morphological response feature vectors; the microscopic structure feature subset is coupled and correlated with the environmental disturbance feature subset to generate environmental and structural evolution feature vectors. Based on the process and morphology response feature vectors and the environment and structural evolution feature vectors, a cross-scale multi-physics joint feature matrix is ​​constructed as the input data for the analysis logic.

6. A method for cross-scale near-net-shape manufacturing of core components of a conductor device according to claim 5, characterized in that, In step S2, the adjustment direction of the process parameters output based on the conductor-specific analysis logic is specifically as follows: Based on the aforementioned cross-scale multiphysics joint feature matrix, the conductor material-structure-performance association knowledge base and dynamic threshold table are invoked. The knowledge base stores the mapping relationship between the macroscopic morphology characteristics, microstructure characteristics and final conductivity and mechanical properties of different conductor materials under different process histories. Multidimensional feature decoupling is performed on the joint feature matrix to identify the macroscopic morphology-dominated deviation mode and the microstructure-dominated evolution mode in the current forming state, and the three-dimensional region where each mode occurs is located based on the spatial location index. Based on the aforementioned conductor material-structure-performance correlation knowledge base, the performance impact weight of the identified deviations and evolution patterns is evaluated, and their potential impact on the target conductivity and mechanical properties of the component is calculated. Based on the performance impact weight evaluation results and the tolerance range preset for the current material and structure in the dynamic threshold table, a multi-objective decision analysis is performed. The generated adjustment vector is input into the process parameter-field parameter coupling influence model for simulation verification. The chain effect of implementing the adjustment on non-target performance dimensions and other related regions is evaluated. The magnitude and timing of the adjustment vector are adjusted according to the verification results. The output includes a set of process parameter adjustment instructions, which includes the parameter adjustment category, direction, magnitude, area of ​​effect, and expected control target, and serves as the input for hierarchical dynamic feedback.

7. A method for cross-scale near-net-shape manufacturing of core components of a conductor device according to claim 6, characterized in that, The multi-objective decision analysis specifically refers to: If the macroscopic dimensional deviation exceeds the tolerance and has a high weight, a first type of adjustment vector is generated with the correction of geometric accuracy as the core, pointing to the macroscopic process parameters of energy input and feed rate; if the microstructure evolution is not conducive to the target performance and has a high weight, a second type of adjustment vector is generated with the optimization of microstructure as the core, pointing to the electromagnetic field and ultrasonic field auxiliary parameters. If the two are intertwined and have equal weights, a composite adjustment vector is generated, and the order of primary and secondary regulation is determined.

8. A method for cross-scale near-net-shape manufacturing of core components of a conductor device according to claim 7, characterized in that, In S3, the deviation type is determined by hierarchical classification as follows: Receive the set of process parameter adjustment instructions and parse the parameter adjustment categories and their effective areas; Based on the parameter adjustment category, it is determined whether the parameter to be adjusted belongs to a macroscopic process parameter or a field auxiliary parameter. If macroscopic process parameters need to be adjusted, the energy input and feed rate should be adjusted according to the adjustment direction and magnitude in the instruction set. If field auxiliary parameters need to be adjusted, adjust the electromagnetic field strength and frequency or the ultrasonic field power and frequency according to the adjustment direction and amplitude in the instruction set. If it is necessary to adjust both macroscopic process parameters and field auxiliary parameters simultaneously, the corresponding parameters should be adjusted sequentially according to the primary and secondary control order specified in the instruction set.

9. A method for cross-scale near-net-shape manufacturing of core components of a conductor device according to claim 8, characterized in that, In S3, the monitoring-analysis-adjustment closed-loop cycle formed after the linkage verification is specifically as follows: Before performing parameter adjustments, the macroscopic process parameters and field auxiliary parameters to be adjusted are input into the preset process coupling influence model. Run the process coupling effect model to predict the chain effect of the parameter adjustment on the current forming area and the adjacent unforming area; If the predicted chain reaction does not exceed the preset safety threshold, the parameter adjustment is confirmed to be executed; if the predicted chain reaction exceeds the safety threshold, the adjustment range is reduced or the adjustment timing is postponed, and the modified parameter adjustment instruction is generated and executed. After the parameter adjustment is completed, the next round of cross-scale collaborative monitoring of the adjusted area is triggered, and the newly collected cross-scale correlation data is re-input into the fusion analysis and analysis logic steps. Adjust the direction based on the newly output process parameters, repeat the hierarchical adjustment and linkage verification to form a closed loop until the entire component is formed.

10. A method for manufacturing near-net-shape core components of a conductor device across scales according to claim 9, characterized in that, In step S4, the adaptive control step for the conductor's conductivity and complex structural features is specifically as follows: In the closed-loop cycle, the grain boundary density and second phase distribution data output by microscale monitoring are acquired in real time; Based on the conductivity model of the conductor material and the data, the predicted conductivity of the current forming area is calculated in real time. The predicted conductivity is compared with the target conductivity. If the deviation exceeds the set range, the field-aided parameter adjustment for optimizing the microstructure is triggered first in the hierarchical adjustment step. When the forming path enters the variable cross-section or thin-walled region of the component, the spatial sampling interval between macroscopic and microscopic monitoring is reduced, and the frequency of monitoring data acquisition is increased; For the variable cross-section or thin-walled region, a dedicated feature extraction algorithm for the region with abrupt structural changes is enabled in the fusion analysis step, and a dynamic threshold table preset for this type of region is called in the analysis logic step.

Images