Geometric modeling method of cable assembly parameterized model

Abstract

The invention relates to the technical field of geometric modeling, in particular to a geometric modeling method of a cable assembly parameterized model. The method comprises the following steps: deploying a multi-source sensor array to collect cable space trajectory data, constructing a feature matrix, and generating a cable joint feature matrix; key parameter extraction is carried out on the cable joint feature matrix, and node feature vectors are constructed; performing adaptive graph construction by using the node feature vectors, and constructing a topological structure to obtain cable topological coding data; therefore, by introducing multi-source heterogeneous data fusion, curvature-torsion sensitivity analysis, adaptive topology construction and a time sequence deformation mapping mechanism based on LSTM, the problems of incomplete multi-scale deformation response identification and insufficient physical structure expression ability in the traditional cable geometric modeling process are solved; and the accurate modeling and dynamic regulation and control capabilities of the cable geometric model in a complex environment are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of geometric modeling, and in particular to a geometric modeling method for a parametric model of a cable assembly. Background Art

[0002] Existing geometric modeling methods for cable assemblies generally rely on regularized modeling templates or empirical construction methods, lacking the comprehensive response ability to aspects such as multi-scale deformation, complex spatial postures, and microscopic material heterogeneity of cables. On the one hand, traditional modeling means based on CAD or parametric curves usually construct cable geometries by means of static path definition, cross-sectional profile stretching, etc., and it is difficult to effectively capture physical characteristics such as curvature changes, local torsion, and non-linear flexure shown by cables after being stressed and deformed, resulting in large deviations in the modeling results during dynamic simulation or stress analysis. On the other hand, most existing digital modeling technologies focus on the macroscopic reproduction of the overall path of the cable, ignoring the potential impact of microscopic structural features on the outer surface of the cable (such as surface irregular protrusions, shrinkage of the coating layer, etc.) on the mating performance and conduction performance of the plug-in connectors, and it is difficult to support the extraction of local geometric response indicators from multi-source heterogeneous data such as point clouds, images, and sensors for fine modeling. In addition, in complex scenarios such as high-density wiring, three-dimensional bending laying, or loaded operating states, existing modeling methods usually cannot achieve the collaborative characterization between physical coupling parameters (such as stress-strain, curvature-torsion, etc.) and topological structures, restricting the extended application ability of the cable geometric model in tasks such as multi-field coupling simulation, deformation prediction, and intelligent optimization, and further exposing the significant deficiencies of the existing modeling process in terms of accuracy and parameter regulation ability. Summary of the Invention

[0003] Based on this, it is necessary to provide a geometric modeling method for a parametric model of a cable assembly to solve at least one of the above technical problems.

[0004] To achieve the above object, a geometric modeling method for a parametric model of a cable assembly, the method includes the following steps:

[0005] Step S1: Deploy a multi-source sensor array to collect cable spatial trajectory data, construct a feature matrix, and generate a cable joint feature matrix;

[0006] Step S2: Extract key parameters from the cable joint feature matrix and construct a node feature vector; use the node feature vector to construct an adaptive graph and construct a topological structure to obtain cable topological coding data;

[0007] Step S3: Use the cable joint feature matrix to output the symbol distance value, and initialize the neural network weight parameters with the cable topology coding data to obtain the initial parameters of the cable geometric model; Use the preset LSTM deformation predictor to perform deformation data fusion on the initial parameters of the cable geometric model, and perform isosurface extraction optimization to output the cable parameter geometric model;

[0008] Step S4: Perform multi-scale verification on the cable parameter geometric model to obtain cable model evaluation data; Use the cable model evaluation data to perform parameter fine-tuning on the cable parameter geometric model, and construct a cable component parameter model output report.

[0009] The beneficial effects of the present invention are as follows. By constructing a joint feature matrix of the cable space trajectory through a multi-source sensor array, it is possible to integrate data information of the cable in all dimensions, from multiple angles, and across scales, breaking through the limitations of traditional cable geometric modeling that only relies on a single sensing path or static measurement means, and significantly improving the spatial resolution and dynamic response ability of data acquisition. On this basis, key parameter extraction and node feature vector construction further map the high-dimensional original trajectory data into a structured and sparse expression form, enabling the subsequent graph structure construction to have good feature discrimination and adaptive capabilities, which helps to improve the stability and robustness of cable topology modeling. By initializing the neural network weights with topology coding data and introducing an LSTM deformation predictor to perform temporal deformation fusion on the initial parameters, the network's ability to capture the complex geometric features and dynamic deformation trends of the cable is effectively enhanced, and the fitting accuracy of the neural network for non-linear deformation is improved. In the isosurface extraction optimization stage, converting the predicted deformation information into a parameter model with geometric entity significance not only enhances the visualization expression ability of the model but also provides interpretability support for subsequent parameter fine-tuning. Finally, through the multi-scale evaluation and parameter fine-tuning mechanism, the model accuracy is verified at different scales, and the model deviation is dynamically corrected to ensure that the output cable component parameter model still has high fidelity and high stability under complex working conditions. The overall process is centered on data-driven, and through layer-by-layer abstraction and feature evolution, the modeling results have computability, iterability, and engineering adaptability while ensuring expression integrity, promoting the development of cable digital modeling towards higher precision and stronger generalization ability. Therefore, the present invention solves the problems of incomplete recognition of multi-scale deformation response and insufficient physical structure expression ability in the traditional cable geometric modeling process by introducing multi-source heterogeneous data fusion, curvature-torsion sensitivity analysis, adaptive topology construction, and LSTM-based temporal deformation mapping mechanism, and improves the accurate modeling and dynamic regulation ability of the cable geometric model in complex environments.

[0010] Preferably, step S1 includes the following steps:

[0011] Use a lidar array with a scanning frequency of 100Hz to perform three-dimensional scanning on the cable to obtain the original point cloud coordinate data;

[0012] Measure the motion trajectory of the cable through an IMU inertial measurement unit to obtain cable deformation data;

[0013] Arrange fiber Bragg grating sensors on the cable at intervals of 5 mm to collect cable strain value data;

[0014] Collect wavelength data through a signal collector and convert the wavelength data into cable wavelength data using an optical signal demodulator;

[0015] Perform cubic spline interpolation on the original point cloud coordinate data and cable deformation data to obtain lidar-deformation data; <T

[0016] Perform downsampling on the cable strain value data and cable wavelength data to obtain fiber optic sensor sampling data;

[0017] Align the spatial coordinate systems according to the lidar-deformation data and fiber optic sensor sampling data, and iterate the closest points to obtain a cable joint feature matrix.

[0018] The present invention realizes multi-modal, multi-dimensional, and high-precision data acquisition and fusion of the cable state by integrating a lidar, an IMU inertial measurement unit, and fiber Bragg grating sensors arranged at high density, and effectively constructs a cable joint feature expression system with spatio-temporal continuity and structural integrity. Specifically, the original three-dimensional point cloud data obtained by a lidar array with a scanning frequency of 100 Hz has high temporal resolution and spatial density, and can accurately reflect the external contour and geometric changes of the cable; the sequential motion trajectory provided by the IMU introduces the information dimension of cable dynamic deformation and enhances the temporal coupling characteristics of the data; at the same time, fiber Bragg grating sensors arranged along the cable at intervals of 5 mm, combined with the wavelength data collected at high frequency and converted by an optical signal demodulator, provide high-sensitivity local strain response information, enabling the perception of the internal stress distribution of the cable. At the data processing level, cubic spline interpolation reconstructs the spatial continuity of the point cloud data and deformation data, making the lidar-deformation data have stronger surface fitting ability and shape description ability; the downsampling of the data collected by the fiber optic sensor not only ensures the integrity of key structural information but also significantly reduces the amount of redundant data and improves the calculation efficiency. By aligning the spatial coordinate systems and using the iterative closest point algorithm to fuse the lidar-deformation data and fiber optic sensor data, not only is the mapping of multi-source data in the same geometric semantic framework realized, but also a cable joint feature matrix with both macroscopic morphology and microscopic strain characteristics is constructed, providing a highly consistent and structurally complete data information basis for subsequent cable structure modeling, state recognition, and dynamic evolution analysis, and significantly improving the physical consistency and prediction reliability of data-driven modeling methods.

[0019] Preferably, the UGC data desensitization described in step S2 includes:

[0020] Step S21: Perform curvature-torsion sensitivity analysis on the cable joint feature matrix to generate a curvature-torsion parameter index sub-matrix;

[0021] Step S22: Calculate the radial distance from the matrix node to the center line based on the curvature-torsion parameter index sub-matrix with the cable center line as the reference to obtain the cable spatial enhancement vector; perform equivalent stress conversion on the curvature-torsion parameter index sub-matrix through Hooke's law to obtain the cable physical enhancement vector;

[0022] Step S23: Use the cable spatial enhancement vector and the cable physical enhancement vector to construct an adaptive graph and build a topological structure to obtain cable topological coding data.

[0023] Through the curvature-torsion sensitivity analysis of the cable joint feature matrix and the construction of the curvature-torsion parameter index sub-matrix, the present invention effectively improves the expression accuracy of the spatial curve change information in the cable geometric features. By quantifying and extracting the geometric response degree of different nodes through this matrix, the double sensitivity analysis of the local deformation characteristics and the overall structure trend is realized, making up for the deficiency of the traditional model established based on a single parameter of position or strain. On this basis, the calculation of the radial distance based on the cable center line is further introduced, which not only enhances the structural expression of the cable in the spatial distribution, but also enables each feature node to have a fine geometric relationship relative to the global coordinate system through the construction of the spatial enhancement vector. This enhancement mechanism improves the expression ability of the node in the graph structure for the spatial position from the data level, making the subsequent graph construction more directional and coherent. At the same time, based on Hooke's law, the equivalent stress conversion of the curvature-torsion parameter index sub-matrix effectively introduces the mapping relationship of physical properties in the data expression. By constructing the physical enhancement vector, each node not only has geometric attributes but also carries stress response characteristics, realizing the deep coupling of spatial structure information and mechanical properties in the same data expression system. Finally, the adaptive graph structure is constructed by using the vector information that fuses spatial enhancement and physical enhancement, which not only improves the accuracy and convergence efficiency of the graph neural network in representing the cable topological structure, but also strengthens the physical consistency and geometric coherence of the graph structure through the topological coding data, laying a solid data foundation and theoretical support for the structural reasoning, deformation prediction and health assessment of complex cable systems.

[0024] Preferably, step S21 includes the following steps:

[0025] Step S211: Obtain the actual measurement error of the cable laser scanning; perform curvature and torsion screening on the cable joint feature matrix at a level higher than 1.5 times the actual measurement error of the cable laser scanning to obtain high-precision cable curvature-torsion data;

[0026] Step S212: Apply a 5% perturbation of curvature - torsion to the high - precision data of cable curvature - torsion, and calculate the sensitivity of perturbation parameters to obtain the preliminary sensitivity data of cable curvature - torsion;

[0027] Step S213: Conduct a coupled effect matrix analysis based on the preliminary sensitivity data of cable curvature - torsion, and evaluate the error contribution value to obtain the curvature - torsion parameter index sub - matrix.

[0028] The present invention introduces the measured error of cable laser scanning as a reference benchmark, establishes a high - precision screening mechanism with the actual observation error as a constraint condition, and can extract a data subset with curvature and torsion significantly higher than the error level from the original joint feature matrix, thereby effectively filtering out low - credibility data introduced by measurement noise or non - structural interference, and enhancing the robustness and reliability in the data processing stage. On this basis, by applying a ±5% perturbation to the high - precision curvature - torsion data, the small fluctuations of curvature and torsion in actual working conditions are systematically simulated, and the change trend of parameter response before and after perturbation is further analyzed, so as to obtain the preliminary sensitivity data of cable curvature - torsion parameters. This process not only improves the data's perception ability of local geometric changes, but also provides a necessary sensitivity basis for subsequent parameter tuning and model adaptation. Further, through the coupled effect matrix analysis of the preliminary sensitivity data, the non - linear interaction relationship between curvature and torsion and between them and other potential variables can be revealed. Then, through the quantitative evaluation of the error contribution value, the key parameter regions that have the greatest impact on the overall geometric accuracy of the cable can be effectively identified. The finally generated curvature - torsion parameter index sub - matrix has three characteristics: high precision, high stability, and high structural correlation. It can significantly compress the data dimension without losing key geometric deformation features, improve the data processing efficiency and model adaptability of subsequent tasks such as modeling, graph structure construction, and dynamic prediction, and lay a solid data foundation for constructing a cable digital model with physical consistency and fine expression ability.

[0029] Preferably, step S23 includes the following steps:

[0030] Step S231: Perform triangulation processing using the cable space enhancement vector and the cable physical enhancement vector, and output nodes according to the physical parameter similarity greater than 0.7 to obtain the cable hybrid adjacency list;

[0031] Step S232: Calculate the similarity of the cable hybrid adjacency list using cosine similarity to obtain the cable space - physical weighted combined data;

[0032] Step S233: Construct node weights for the cable space - physical weighted combined data according to the stress difference of boundary points, thereby constructing the cable component topological structure; perform graph data object conversion on the cable component topological structure and generate low - dimensional encoded data to obtain the cable topological encoded data.

[0033] By introducing the measured error of cable laser scanning as a reference benchmark, the present invention establishes a high-precision screening mechanism with actual observation error as a constraint condition, which can extract a data subset with curvature and torsion significantly higher than the error level from the original joint feature matrix, thereby effectively filtering out low-confidence data introduced by measurement noise or non-structural interference, and enhancing the robustness and reliability in the data processing stage. On this basis, by applying a ±5% perturbation to the high-precision curvature-torsion data, the small fluctuations of curvature and torsion in actual working conditions are systematically simulated, and the change trend of parameter response before and after perturbation is further analyzed, so as to obtain the preliminary sensitivity data of cable curvature-torsion parameters. This process not only improves the data's perception ability of local geometric changes, but also provides a necessary sensitivity basis for subsequent parameter tuning and model adaptation. Further, by performing a coupling effect matrix analysis on the preliminary sensitivity data, the non-linear interaction relationship between curvature and torsion and between them and other potential variables can be revealed, and then through the quantitative evaluation of the error contribution value, the key parameter area that has the greatest impact on the overall geometric accuracy of the cable can be effectively identified. The finally generated curvature-torsion parameter index sub-matrix has three characteristics: high precision, high stability and high structural correlation. Without losing key geometric deformation features, it can significantly compress the data dimension, improve the data processing efficiency and model adaptability of subsequent tasks such as modeling, graph structure construction and dynamic prediction, and lay a solid data foundation for constructing a cable digital model with physical consistency and fine expression ability.

[0034] Preferably, the personalized model construction described in step S21 includes:

[0035] Step S211: Obtain the measured error of cable laser scanning; screen the cable joint feature matrix with the measured error of cable laser scanning as the benchmark for curvature and torsion higher than 1.5 times to obtain high-precision cable curvature-torsion data;

[0036] Step S212: Apply a 5% curvature-torsion perturbation to the high-precision cable curvature-torsion data and calculate the perturbation parameter sensitivity to obtain the preliminary sensitivity data of cable curvature-torsion;

[0037] Step S213: Perform a coupling effect matrix analysis based on the preliminary sensitivity data of cable curvature-torsion and evaluate the error contribution value to obtain the curvature-torsion parameter index sub-matrix.

[0038] The present invention constructs a geometric feature screening mechanism based on data credibility by introducing the measured error of cable laser scanning as an evaluation criterion, significantly improving the effectiveness and accuracy of curvature and torsion information at the data level. First, as an objective physical observation result, the measured error can provide a quantitative reference for the upper limit of precision in data screening, and then screen the curvature and torsion values in the joint feature matrix that are more than 1.5 times the error, effectively eliminating the abnormal data with low confidence and affected by environmental disturbances, and only retaining the geometric features that have a significant response ability to the real structural changes of the cable, thus establishing a high-precision curvature-torsion feature data set. On this basis, a curvature and torsion perturbation of ±5% is applied to this high-precision data to simulate the small deformation effect of the cable caused by mechanical environment, installation stress or time evolution in the actual working condition, and a sensitivity analysis is carried out on the change trend of the parameter response before and after the perturbation to capture the dynamic response behavior of the geometric parameters at the data level. This process not only reveals the local stability of the geometric parameters but also provides a preliminary sensitivity basis for subsequent parameter optimization. Further, by constructing a coupling effect matrix for the preliminary sensitivity data, systematically depicting the mutual influence degree between curvature and torsion and the response diffusion relationship to other covariates, and combining the quantitative evaluation of the error contribution value, accurately identifying the key feature regions with the highest proportion in the systematic error, thus extracting a curvature-torsion parameter index sub-matrix with high sensitivity, high coupling degree and high representativeness. This matrix not only significantly compresses the original feature dimension but also strengthens the physical association characteristics contained in the data, providing a more recognizable and engineering interpretable data input basis for tasks such as structure generation, deformation mapping and health assessment in subsequent composition algorithms.

[0039] Preferably, step S23 includes the following steps:

[0040] Step S231: Perform triangulation processing using the cable space enhancement vector and the cable physical enhancement vector, and output nodes according to the physical parameter similarity greater than 0.7 to obtain a cable hybrid adjacency list;

[0041] Step S232: Calculate the similarity of the cable hybrid adjacency list using cosine similarity to obtain cable space-physical weighted combined data;

[0042] Step S233: Construct node weights for the cable space-physical weighted combined data according to the stress difference of the boundary points, thereby constructing a cable component topological structure; perform graph data object transformation on the cable component topological structure and generate low-dimensional encoded data to obtain cable topological encoded data.

[0043] The present invention strengthens the unified expression of the structural and physical responsiveness in the cable system at the data level by introducing the fusion processing of the cable space enhancement vector and the physical enhancement vector, and establishing a graph construction mechanism based on the dual characteristics of geometric position and mechanical properties. Specifically, first, the space and physical enhancement vectors are triangulated, and the effective connection relationships between nodes are screened by means of a physical parameter similarity threshold (greater than 0.7). The constructed cable hybrid adjacency list not only retains the structural geometric continuity, but also fully reflects the local consistency of the material properties, effectively avoiding the problems of topological information loss or incorrect clustering caused by isolated physical features or isolated spatial distributions. Subsequently, the cosine similarity calculation is introduced on the basis of the hybrid adjacency list to realize the weighted fusion processing of the space-physical dual characteristics, so that the similarity evaluation not only depends on the geometric adjacency relationship, but also introduces the decision-making weight of the physical properties in the node relationship judgment, thereby generating a more representative space-physical weighted combination data set. Based on this weighted combination data, by analyzing the stress differences of the boundary points and constructing node weights, the recognition ability and hierarchical expression ability of the graph structure in the edge region are further enhanced, and the sensitivity of the topological structure to the boundary response characteristics is improved. The cable component topological structure constructed on this basis is no longer just a static connection graph, but a multi-scale graph with stress-driven weight distribution, physical consistency between nodes and spatial continuity. Finally, by converting the graph structure object into an encodable data form and performing low-dimensional embedding encoding, not only the storage and calculation costs of the high-dimensional graph structure data are compressed, but also the topological semantic relationship and physical response characteristics are retained, providing lightweight and high-fidelity input features for subsequent deformation modeling, dynamic prediction and health assessment, etc., effectively improving the modeling efficiency and structural reasoning ability of the entire system.

[0044] Preferably, the use of the cable joint feature matrix to output the signed distance value and initialize the neural network weight parameters with the cable topology encoding data in step S3 includes:

[0045] Extract the cable surface point set from the curvature-torsion parameter index sub-matrix; use the cable surface point set to measure the shortest distance true value label of the sampling points to obtain the signed distance value;

[0046] Use the cable topology encoding data as the hidden layer weight matrix, and perform geometric representation initialization mapping with the signed distance value to obtain the initial parameters of the cable geometric model;

[0047] Use a preset LSTM deformation predictor to perform temporal modeling on the initial parameters of the cable geometric model, and fuse the deformation data with the output temporal affine transformation data to obtain a set of deformation parameters;

[0048] Perform isosurface extraction optimization on the set of deformation parameters and output the cable parameter geometric model.

[0049] The present invention constructs a point cloud subset based on real geometric boundary perception by extracting a set of cable surface points from the curvature-torsion parameter index sub-matrix, and uses this surface point set to perform the shortest distance true value label ranging on the sampling points to generate signed distance values, thereby providing a supervision quantity with physical meaning and structural accuracy for the entire geometric modeling process. The signed distance value not only reflects the Euclidean distance from the sampling point to the actual boundary of the cable, but also distinguishes whether the point is inside or outside the geometry through positive and negative signs, significantly enhancing the interpretability of the geometric representation. On this basis, the cable topology coding data is mapped to the hidden layer weight matrix of the neural network and embedded into the modeling process as a structural guiding signal, realizing the structural constraint coupling from topological semantics to spatial expression, and initializing the geometric representation with the signed distance value, effectively shortening the model training convergence path and improving the initial state accuracy. Further, an LSTM deformation predictor is introduced to perform temporal modeling on the initial parameters, enabling it to handle complex dynamic deformation sequences. By embedding the output temporal affine transformation data into the deformation data fusion process, each geometric node can dynamically respond to the small deformations caused by stress, temperature, or environmental disturbances while maintaining the original position relationship, thereby constructing a parameter response structure with strong temporal continuity and reasonable deformation trends. Finally, the isosurface extraction optimization is performed on the fused deformation parameter set, and the temporal fusion deformation signal is mapped to the visual geometric space to form a cable parameter geometric model with fine boundary layer expression, deformation behavior interpretation ability, and input feature consistency, providing a high-resolution and strong-semantic data expression carrier for subsequent health diagnosis, mechanical deduction, and simulation analysis. This process realizes a complete closed-loop from point set true value construction, topology-geometry coupling modeling, temporal-deformation prediction to optimized visual output in the data dimension.

[0050] Preferably, the isosurface extraction optimization is performed on the deformation parameter set, and the output cable parameter geometric model includes:

[0051] Gradient sampling judgment is performed according to the deformation parameter set, and isosurface extraction optimization is carried out to generate gradient sampling isosurface data, where the gradient sampling isosurface data includes high-gradient region data and low-gradient region data. The high-gradient region data is sampled at a sampling interval of 0.1 mm, and the low-gradient region data is sampled at a sampling interval of 1.0 mm;

[0052] The annular topological structure is extracted from the gradient sampling isosurface data, and the smooth step artifacts are eliminated, thus completing the construction of the cable parameter geometric model.

[0053] The present invention realizes the adaptive high-precision modeling of the cable deformation region by introducing a gradient-sensitivity-driven isosurface sampling mechanism. Specifically, gradient sampling judgment is performed based on the fused deformation parameter set, enabling the system to implement differential sampling strategies for different regions according to the deformation gradient intensity at each spatial node. In the high-gradient region, due to the drastic deformation changes, a high-resolution sampling interval of 0.1 mm is adopted to ensure the complete capture of geometric details; while in the low-gradient region, a sparse sampling of 1.0 mm is used to compress the data volume and computational overhead. Such a non-uniform sampling mechanism avoids the contradiction between information redundancy and key feature loss at the data level, realizes the dynamic matching of data density and deformation complexity, and greatly improves the structural reduction accuracy of the geometric model in highly variable regions. Subsequently, by extracting the annular topological structure from the above gradient-sampled isosurface data, not only is the closed-loop expression of the cable completed in the overall configuration, but also the stability of the annular structure is used to enhance the boundary continuity and global consistency of the geometric model in subsequent physical simulations and deformation predictions. More importantly, by actively eliminating the smooth step artifacts caused by multi-level sampling during the topological extraction process, the surface irregularities caused by gradient mutations and interpolation errors are effectively alleviated, thereby improving the smoothness and physical consistency of the isosurface in surface fitting and visual expression. The finally formed cable parameter geometric model has multiple advantages such as clear hierarchical structure, accurate local details, closed overall topology, and consistent physical response, constructing a modeling mechanism with both precision control and efficiency optimization at the data level, and providing a digital foundation with strong structural stability and high expression consistency for subsequent analysis.

[0054] Preferably, step S4 includes the following steps:

[0055] Step S41: Perform multi-scale verification on the cable parameter geometric model to obtain cable model evaluation data;

[0056] Step S42: Construct an elastic parameter list according to the cable model evaluation data; use the elastic parameter list to perform parameter fine-tuning on the cable parameter geometric model to obtain an optimized cable component parameter model;

[0057] Step S43: Construct a cable component parameter model output report based on the optimized cable component parameter model.

[0058] The present invention realizes the double-layer optimization of the geometric model of cable parameters from the macroscopic structural consistency to the microscopic physical response accuracy by introducing a multi-scale verification and elastic parameter-driven fine-tuning mechanism. Specifically, by implementing multi-scale verification on the geometric model of cable parameters, not only the stability and topological connectivity of its structural framework are evaluated at the global scale, but also its geometric accuracy, curvature continuity, and node stress response characteristics are detected at the local scale, thereby forming a set of cable model evaluation datasets at multiple resolutions. This dataset establishes a consistency baseline for the deformation response in the dimension of structural verification and also provides data support for subsequent parameter adjustment. Subsequently, referring to the cable model evaluation data, elastic parameters closely related to material behavior are extracted from the model response, including but not limited to the elastic modulus, Poisson's ratio, and the change trend of strain gradient, to construct an elastic parameter list with response adjustability and local sensitivity. This list participates in the iterative optimization of the cable geometric model as a fine-tuning constraint condition, enabling the model parameters to gradually converge to the true physical state. By applying microscopic elastic compensation to the high-error regions, not only the consistency between the model and the measured data is improved, but also the stability and generalization ability of the model under multi-condition simulations are enhanced. Finally, based on the optimized cable component parameter model, a model evaluation report including key parameter curves, error convergence trends, physical feature statistics, etc. is output, realizing a closed-loop feedback system from data extraction, model correction to result presentation, and thus building a cable digital twin basic unit with high fidelity, traceability, and application adaptability at the data level.

[0059] Preferably, the multi-scale verification of the geometric model of cable parameters includes the following steps:

[0060] Obtain the measurement data of the laser tracker and the scanning data of the surface topography of the electron microscope.

[0061] Set the macroscopic threshold CD to 0.15 mm; use the measurement data of the laser tracker to analyze the error between the model vertices and the measured points of the geometric model of cable parameters according to the macroscopic threshold CD to obtain the macroscopic error analysis data; perform dataset mapping based on the macroscopic error analysis data to obtain the cable macroscopic error distribution map.

[0062] Extract the normal vectors of the geometric model of cable parameters to obtain the surface normal vector map of the cable model; use the scanning data of the surface topography of the electron microscope to evaluate the microscopic structure similarity of the surface normal vector map of the cable model to obtain the cable microscopic similarity data.

[0063] The present invention obtains the measurement data of a laser tracker and the surface topography scanning data of an electron microscope, which respectively represent the spatial coordinate measurement of the cable structure at the macroscopic scale and the surface texture and normal distribution sampling at the microscopic scale, thus constituting a multi-source heterogeneous high-precision measured data basis. At the macroscopic level, by setting the threshold CD to 0.15 mm, the spatial point cloud data measured by the laser tracker is used as a reference, and the error vector analysis is performed on the vertex positions of the model to generate macroscopic error analysis data including the spatial offset trend, local distortion distribution, and error dispersion degree. Further, this data is mapped into the cable parameter geometric model to generate an error heat map with visual spatial distribution, facilitating the identification of weak areas in the global fitting of the model. At the microscopic level, a normal map is generated by extracting the normals of the model surface, and combined with the real surface structure data obtained by electron microscope topography scanning, a fine-grained similarity evaluation is performed to output the microscopic similarity data of the cable model in terms of surface roughness, microscopic profile continuity, and normal consistency. This similarity quantification information not only reflects the fitting quality of the model in terms of microscopic simulation accuracy but also provides support for subsequent local topography reconstruction and material identification. Overall, this step realizes the dual-precision evaluation of the cable geometric model from spatial positioning to surface topography by integrating multi-source measured data and constructing a multi-scale error analysis framework, providing a systematic data basis for model correction and credibility determination.

[0064] Preferably, step S42 includes the following steps:

[0065] Step S421: Extract high-sensitivity parameters according to the cable macroscopic error distribution map to obtain cable macroscopic sensitivity parameters; construct an elastic parameter list using the cable macroscopic sensitivity parameters;

[0066] Step S422: Fine-tune the cable microscopic similarity data with a small step size of 0.001 and generate fitted elastic parameters for the elastic parameter list to obtain optimized parameters for the geometric cable model;

[0067] Step S423: Use the optimized parameters of the geometric cable model to fine-tune the cable parameter geometric model and perform parameter feedback to obtain an optimized model of the cable component parameters.

[0068] The present invention extracts highly sensitive parameters from the macroscopic error distribution map of the cable, and uses the spatial distribution and error characteristics of the error data to accurately locate the key parameters that have the most significant impact on the model fitting. By extracting these macroscopic sensitive parameters, it is possible to clarify which physical properties (such as elastic modulus, stiffness, etc.) have a greater impact on the model performance, thereby establishing a list of elastic parameters. This list not only provides an important data basis for subsequent parameter adjustment, but also provides necessary constraint conditions for constructing a more physically realistic cable assembly. A fine-tuning method with a small step size (0.001) is adopted to finely adjust the microscopic similarity data of the cable, which helps to achieve a more accurate shape matching at the microscopic scale. The fitting data generated by fitting the list of elastic parameters makes the geometric cable model more consistent with the real conditions in terms of physical behavior and geometric shape, ensuring the accuracy and adaptability of the model. In addition, this fine-tuning process can also effectively avoid the local distortion phenomenon caused by overfitting during the global optimization process of the model. Finally, the parameters of the optimized geometric cable model are used for fine-tuning, and the parameter feedback is realized through the feedback mechanism, so that the geometric and physical properties of the cable assembly can be adjusted more efficiently and accurately, and finally an optimized parameter model of the cable assembly is generated. This model can not only more accurately reflect the performance of the cable in actual applications, but also improve its stability and applicability under different working conditions. From the data level, the whole process realizes the fine correction of the model from the large scale to the microscopic details through the joint optimization of macroscopic and microscopic errors, ensuring the high fidelity and actual adaptability of the cable geometric model at multiple scales BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 It is a schematic flow chart of the steps of a geometric modeling method for a parametric model of a cable assembly;

[0070] Figure 2 is Figure 1 a detailed implementation step flow chart of step S3 in

[0071] The realization, functional characteristics and advantages of the object of the present invention will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0072] The technical method of the present invention for the patent will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of them. All other embodiments obtained by those skilled in the art within the scope of the present invention without creative work belong to the scope of protection of the present invention.

[0073] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and thus repeated descriptions thereof will be omitted. Some of the block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor methods and / or microcontroller methods.

[0074] It should be understood that although terms such as "first" and "second" may be used herein to describe various units, these units should not be limited by these terms. These terms are only used to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0075] To achieve the above object, please refer to Figures 1 to 2 , a geometric modeling method for a parametric model of a cable assembly, the method comprising the following steps:

[0076] Step S1: Deploy a multi-source sensor array to collect cable space trajectory data, construct a feature matrix, and generate a cable joint feature matrix;

[0077] Step S2: Extract key parameters from the cable joint feature matrix and construct a node feature vector; use the node feature vector to construct an adaptive graph and construct a topological structure to obtain cable topological coding data;

[0078] Step S3: Use the cable joint feature matrix to output a symbolic distance value, and initialize the neural network weight parameters with the cable topological coding data to obtain the initial parameters of the cable geometric model; use a preset LSTM deformation predictor to perform deformation data fusion on the initial parameters of the cable geometric model, and perform isosurface extraction optimization to output the cable parameter geometric model;

[0079] Step S4: Perform multi-scale verification on the cable parameter geometric model to obtain cable model evaluation data; use the cable model evaluation data to perform parameter fine-tuning on the cable parameter geometric model, and construct a cable assembly parameter model output report.

[0080] In an embodiment of the present invention, referring to Figure 1 shown, it is a schematic flow chart of the steps of a geometric modeling method for a parametric model of a cable assembly of the present invention. In this example, the geometric modeling method for the parametric model of the cable assembly comprises the following steps:

[0081] Step S1: Deploy a multi-source sensor array to collect cable spatial trajectory data, construct a feature matrix, and generate a cable joint feature matrix;

[0082] In the embodiment of the present invention, the spatial trajectory data of the cable is collected by deploying a multi-source sensor array. The specific technical means include using different types of sensors to obtain information in multiple data dimensions. For example, a lidar sensor can provide high-precision spatial positioning data, an inertial measurement unit (IMU) can obtain the motion trajectory of the cable in real time, and a fiber optic sensor can collect the strain information of the cable. These data can comprehensively describe the spatial change characteristics of the cable through collaborative work. First, multiple sensors are simultaneously deployed in the monitoring area to record the spatial trajectory and morphological changes of the cable in real time, thereby generating a data stream containing parameters such as time, spatial coordinates, and motion state. Then, these data are processed through unified time synchronization and spatial coordinate alignment to ensure the consistency of data from different sensors under the same reference framework. For the raw data collected by each sensor, through feature extraction and data fusion techniques, the data collected by each sensor is transformed into a unified feature matrix. This feature matrix contains information such as the position, velocity, acceleration, and deformation of the cable in three-dimensional space. Further, in order to comprehensively consider the characteristics and contributions of different data sources, a weighted algorithm or other data fusion techniques are used to comprehensively process the output data of each sensor, and finally a cable joint feature matrix is generated. This matrix fuses multi-modal data from different sensors and can reflect the overall behavior and local change characteristics of the cable in space, serving as the basic data set for subsequent analysis and modeling.

[0083] Step S2: Extract key parameters from the cable joint feature matrix and construct a node feature vector; use the node feature vector to construct an adaptive graph and construct a topological structure to obtain cable topological coding data;

[0084] In the embodiments of the present invention, by extracting key parameters from the cable joint feature matrix, the aim is to identify the core features closely related to the cable state, morphology, and dynamic behavior from multi-dimensional sensor data. The specific technical means include using data processing algorithms to identify and extract parameters containing important physical features such as displacement, velocity, acceleration, strain, curvature, etc. These features reflect the movement trajectory and morphological changes of the cable in space. Through techniques such as principal component analysis (PCA) or independent component analysis (ICA), the dimension can be effectively reduced and the most representative features can be extracted. The extraction of these key parameters is achieved by analyzing each item of the joint feature matrix and screening out the indicators that contribute the most to the subsequent model construction. Next, the extracted key parameters are used to construct node feature vectors. The node feature vectors contain the feature information of each node (i.e., each cable spatial position point) in multiple physical dimensions, forming a multi-dimensional vector representation. Based on these node feature vectors, a graph construction algorithm is used for adaptive graph construction. The process of adaptive graph construction includes selecting an appropriate graph connection strategy according to the similarity or spatial proximity of node features. For example, using similarity metrics (such as cosine similarity, Euclidean distance, etc.), nodes with similar features are connected to form edges in the graph, thereby constructing a topological structure that can reflect the cable structure and behavior. In this process, deep learning methods such as graph convolutional network (GCN) can be used to further optimize the connection weights between nodes to ensure that the structure of the graph can accurately reflect the mutual relationship between various parts of the cable. Through these steps, the topological coding data of the cable is finally generated, which represents the relationship and structural features between each node of the cable in space, providing a structured data basis for subsequent model analysis, prediction, and optimization.

[0085] Step S3: Use the cable joint feature matrix to output the symbol distance value, and initialize the neural network weight parameters with the cable topological coding data to obtain the initial parameters of the cable geometric model; use the preset LSTM deformation predictor to perform deformation data fusion on the initial parameters of the cable geometric model, and perform isosurface extraction optimization to output the cable parameter geometric model;

[0086] In the embodiments of the present invention, symbol distance values are output through a cable joint feature matrix. This process involves further geometric analysis and symbolization of spatial trajectory data. The symbol distance value is a method for measuring the relative distance or similarity between points at different spatial positions in a cable, usually based on the geometric shape of the cable surface or path. The specific technical means can process the spatial coordinate information contained in the joint feature matrix, calculate the distance from each point to other points, and then output the symbol distance value. These symbol distance values serve as the basis for the cable geometric model and can reflect geometric features such as the curvature and stretching state of the cable. Use the cable topology coding data to initialize the weight parameters of the neural network. The technical means in this part involves using the topological structure constructed in graph theory (i.e., the connection relationship between cable nodes) as input to initialize the weights in the neural network. Through the processing of the topological coding, the neural network can receive prior information about the cable structure and physical characteristics, and thus be trained more effectively. The neural network can adjust these weights during the training process to capture more accurate cable behavior characteristics in subsequent deformation predictions. Subsequently, use a preset long short-term memory (LSTM) deformation predictor to perform a time series modeling on the initial parameters of the cable geometric model. The LSTM network is a deep learning architecture specifically designed for processing time series data, which can memorize and predict long-term dependencies in time series. In this step, the LSTM network fuses deformation data by analyzing the initial parameters of the cable geometric model (such as shape, position, speed, etc.) and their changes over time. This means that the LSTM network will, through training and prediction, predict the deformation state of the cable at a future moment based on the patterns of historical data. By optimizing the isosurface extraction of the deformation data, the parametric geometric model of the cable is output. The isosurface extraction technology is usually used to extract a set of points with specific eigenvalue (such as pressure, stress, etc.) from a three-dimensional dataset to form a three-dimensional surface. In this process, the optimization algorithm helps to remove noise and improve the accuracy of the data, ensuring that the finally generated cable geometric model can accurately reflect the actual deformation of the cable.

[0087] Step S4: Perform multi-scale verification on the cable parametric geometric model to obtain cable model evaluation data; use the cable model evaluation data to fine-tune the parameters of the cable parametric geometric model and construct a cable component parameter model output report.

[0088] In the embodiments of the present invention, multi-scale verification is performed on the geometric model of cable parameters. The technical means of this process mainly include systematically evaluating and verifying the cable model at different resolutions and scale levels. Specifically, the cable model exhibits different characteristics under different physical dimensions, stress conditions, and working environments, so verification needs to be carried out at multiple scales. The core technology of multi-scale verification is based on grids and data accuracies at different levels, and different verification criteria are adopted, such as the comparison of local and global geometric features, stress distribution, deformation conditions, etc., to comprehensively evaluate the accuracy and reliability of the cable model. Through multi-scale verification, local changes and global trends at different scales can be captured, thereby comprehensively evaluating the accuracy of the model. The evaluation data of the cable model obtained includes the performance of the model under different conditions, specifically including geometric feature errors, stress-strain distribution errors, deformation prediction errors, etc. This data can help evaluate the robustness of the model and identify existing biases and error sources. Based on these evaluation data, the next step is to perform parameter fine-tuning on the geometric model of cable parameters. The technical means of parameter fine-tuning include finely adjusting the parameters in the model (such as node positions, edge connection relationships, physical properties, etc.) based on the evaluation data. Usually, through optimization algorithms (such as the least squares method, gradient descent method, etc.), the model parameters are automatically adjusted according to the errors and biases in the evaluation data to make them more consistent with the actual measurement results. Using the fine-tuned cable geometric model parameters, a cable component parameter model is constructed and a report is output. The technical means of this process is to parametrically represent the fine-tuned geometric model to construct a representative cable component model. This model can accurately reflect the physical characteristics, geometric morphology, and behavior of the cable under specific working conditions. The output report will summarize all the fine-tuning results, including parameter changes, error corrections, optimized cable models, etc., and provide detailed technical basis for subsequent applications or engineering designs.

[0089] Preferably, step S1 includes the following steps:

[0090] Use a lidar array with a scanning frequency of 100 Hz to perform three-dimensional scanning on the cable to obtain the original point cloud coordinate data;

[0091] Measure the movement trajectory of the cable through an IMU inertial measurement unit to obtain cable deformation data;

[0092] Arrange fiber Bragg grating sensors on the cable at an interval of 5 mm to collect cable strain value data;

[0093] Collect wavelength data through a signal collector and convert the wavelength data into cable wavelength data using an optical signal demodulator;

[0094] Perform cubic spline interpolation on the original point cloud coordinate data and the cable deformation data to obtain lidar-deformation data;

[0095] Downsample the cable strain value data and the cable wavelength data to obtain the fiber optic sensor sampling data;

[0096] Align the spatial coordinate systems based on the lidar-deformation data and the fiber optic sensor sampling data, and iterate the closest points to obtain the cable joint feature matrix.

[0097] In the embodiments of the present invention, a lidar array with a scanning frequency of 100 Hz is used to perform three-dimensional scanning of the cable, and high-frequency and high-precision original point cloud coordinate data is obtained through the scanning function of the lidar. This data represents the accurate position and morphological information of the cable in space and is the basis for subsequent analysis. The point cloud data usually needs to be further processed to remove noise, fill in gaps, and perform fine-tuning to ensure the accuracy of subsequent analysis. Secondly, an IMU (Inertial Measurement Unit) is used to measure the movement trajectory of the cable to obtain cable deformation data. The IMU device is usually composed of an accelerometer and a gyroscope, which can provide the movement state of the cable at different times and its relative changes, thereby capturing the deformation characteristics of the cable. The acquisition of deformation data is crucial for describing the dynamic performance of the cable in the working environment. Fiber Bragg grating sensors are arranged on the cable at intervals of 5 mm to collect the strain value data of the cable. The fiber Bragg grating sensors measure the strain of the cable by the wavelength change of the reflected signal of the optical fiber, and these strain data provide an important basis for further analyzing the mechanical properties of the cable. At the same time, a signal collector collects the wavelength data of the cable and converts the wavelength data into the wavelength data of the cable through an optical signal demodulator. This conversion process can convert the detection results of the optical fiber sensor into actual physical quantities, facilitating subsequent data processing and analysis. Cubic spline interpolation is performed on the original point cloud coordinate data and the cable deformation data, aiming to smooth the original data and generate more refined lidar-deformation data. Cubic spline interpolation is a commonly used numerical interpolation method, which can effectively smooth the discrete data points and fill the gaps between the data without introducing too much error. Downsampling is performed on the strain value data and wavelength data of the cable. Downsampling reduces the computational burden by reducing the number of data points while trying to retain the key features of the data. This step helps to process large-scale data sets and provides a more concise and efficient data input for subsequent data analysis and model construction. Finally, spatial coordinate system alignment is performed based on the lidar-deformation data and the fiber sensor sampling data, and the Iterative Closest Point (ICP) algorithm is used for registration. This method realizes alignment by minimizing the distance between two sets of data points, thereby ensuring the consistency of all data sets in the same coordinate system. After alignment, a cable joint feature matrix is obtained, which further integrates the data collected by different sensors and provides a unified high-dimensional feature set for subsequent analysis, modeling, and optimization. The comprehensive application of these technical means combines the high-precision scanning of lidar, the dynamic monitoring of IMU, the strain measurement of fiber Bragg grating sensors, and signal processing methods, ensuring the comprehensive and accurate capture of the geometric shape and dynamic characteristics of the cable.

[0098] Preferably, step S2 includes the following steps:

[0099] Step S21: performing curvature-torsion sensitivity analysis on the cable joint characteristic matrix to generate a curvature-torsion parameter index submatrix;

[0100] Step S22: Calculate the radial distance from the matrix node to the centerline based on the curvature-torsion parameter index submatrix and the cable centerline to obtain the cable spatial enhancement vector; perform equivalent stress conversion on the curvature-torsion parameter index submatrix using Hooke's law to obtain the cable physical enhancement vector;

[0101] Step S23: constructing an adaptive graph using the cable spatial enhancement vector and the cable physical enhancement vector, and building a topology structure to obtain cable topology encoding data.

[0102] In an embodiment of the present invention, a curvature-torsion sensitivity analysis is performed on the cable joint characteristic matrix. The core of this process is to identify and quantify the cable's sensitivity to changes under specific curvature and torsion conditions. By analyzing the curvature and torsion changes at different nodes, it is possible to determine which parts of the cable are more susceptible to deformation under external forces, thereby generating a curvature-torsion parameter index submatrix. This matrix not only provides information on the cable's morphological characteristics but also provides basic data for subsequent calculations and optimization. The radial distances from the nodes to the cable centerline are calculated based on the curvature-torsion parameter index submatrix to generate a cable spatial enhancement vector. The calculation of the radial distances is based on the cable centerline, accurately capturing the positional changes of each node relative to the centerline. This spatial enhancement vector provides the spatial distribution characteristics of the cable for subsequent analysis. Furthermore, Hooke's law is applied to convert the curvature-torsion parameter index submatrix into equivalent stress. By calculating the mechanical properties, a cable physical enhancement vector is obtained. This physical enhancement vector, together with the spatial enhancement vector, provides the comprehensive mechanical and spatial characteristics of the cable, providing detailed physical data support for further model construction. An adaptive graph structure is constructed based on the spatial enhancement vectors and physical enhancement vectors obtained above, and cable topology encoding data is generated based on this topological structure. In this process, spatial and physical data are integrated into a graph model through graph theory methods. Each node represents a characteristic point of a cable, and the connections between nodes characterize the relationship between these characteristic points. This graph structure can more intuitively reflect the geometric shape and mechanical behavior of the cable and provide the necessary mathematical basis for subsequent analysis (such as mechanical optimization, deformation prediction, etc.). In summary, this series of steps, through sensitivity analysis, physical model conversion and graph structure construction, finely characterizes the spatial and physical characteristics of the cable from multiple dimensions, and ultimately generates highly structured topological encoding data.

[0103] Preferably, step S21 includes the following steps:

[0104] Step S211: Obtain the measured error of cable laser scanning; screen the cable joint feature matrix based on the measured error of cable laser scanning with a curvature and torsion higher than 1.5 times to obtain high-precision cable curvature-torsion data;

[0105] Step S212: Apply a 5% perturbation of curvature-torsion to the high-precision cable curvature-torsion data and calculate the sensitivity of the perturbation parameters to obtain the preliminary sensitivity data of cable curvature-torsion;

[0106] Step S213: Conduct a coupling effect matrix analysis based on the preliminary sensitivity data of cable curvature-torsion and evaluate the error contribution value to obtain the curvature-torsion parameter index sub-matrix.

[0107] In the embodiments of the present invention, through precise calculation and analysis, high-precision cable shape characteristic data is obtained, and sensitivity analysis and coupling effect evaluation are performed on it, so as to establish an accurate parameter index model. Its specific technical means can be clearly described through the following process. By obtaining the actual measurement error of cable laser scanning and screening the cable joint feature matrix based on this. Specifically, using the actual measurement error provided by the laser scanning data as the evaluation criterion, the curvature and torsion data in the cable feature matrix are screened, and the screening criterion is to select the values of curvature and torsion greater than 1.5 times. The key technical means of this step is the selection of the error benchmark and feature screening, which ensures the high precision and accuracy of the screened data by increasing the standard of the error threshold (such as 1.5 times the curvature and torsion). The screened data can be more effectively used for subsequent sensitivity analysis and coupling effect research, thus avoiding unnecessary complexity and inaccuracy caused by data with large errors. Then, perturbations are applied to the highly accurate data after screening. In this step, ±5% perturbations are applied to the cable curvature and torsion data, and preliminary sensitivity data is obtained through the sensitivity calculation of the perturbation parameters. The core of this technical means lies in calculating the influence of these perturbations on parameter sensitivity by simulating different amplitudes of perturbations (that is, controlling the fluctuations of curvature and torsion data). Specifically, the perturbation amplitude is set to ±5%, and by calculating the response of the system to the perturbation, it can be identified which data have a greater impact on the behavior of the cable when the parameters change, and then the sensitive data with important influences is screened out. Based on the preliminary sensitivity data, a coupling effect matrix analysis is performed, and the contribution values of each error source are evaluated. Through the coupling effect analysis, the mutual relationship between different parameters and their influence degree in the overall error are determined. This process can accurately analyze the contribution of each parameter to the final error by constructing a coupling matrix, providing a basis for subsequent error correction and optimization. This technical means can provide the necessary mathematical support for subsequent system optimization and correction, thus ensuring the reliability and high precision of cable curvature and torsion data in practical applications. Through these steps, the whole set of technical means gradually analyzes the shape and mechanical properties of the cable in depth from data collection, screening to sensitivity analysis and coupling effect analysis, and finally obtains an accurate and refined curvature-torsion parameter index sub-matrix, which can provide strong data support for subsequent work such as cable performance prediction, optimal design and fault diagnosis.

[0108] Of particular importance, the radial distance and equivalent stress conversion formulas are as follows:

[0109] According to the curvature-torsion parameter index sub-matrix, based on the cable center line, the radial distance from the matrix node to the center line is calculated to obtain the cable space enhancement vector, where the formula for the radial distance is:

[0110]

[0111] p i is the i-th node, C(s i ) is the i-th normal projection distance, and n is the number of nodes;

[0112] Among them, the equivalent stress conversion formula is:

[0113]

[0114] σ b is the bending stress; σ s is the shear stress; σ eq is the equivalent stress.

[0115] In the embodiment of the present invention, for the spatial distribution of discrete nodes of a cable, the method is based on the structural main axis of the cable center line. By introducing a projection transformation function C(s i ), each node p i is projected onto the center line coordinate system, and by minimizing the projection error the aggregation of the geometric positions of the nodes in the space towards the center line direction is realized. This step is essentially a vector space reconstruction and constrained optimization problem based on the Euclidean distance. Among them, n is the number of nodes, and C(s i ) represents the coordinates of the nearest projection point of the i-th node on the center line. The optimization goal is to minimize the sum of the distances between each node and its projection point, so as to form a spatially enhanced vector field data set that can be used for subsequent topology construction or stress analysis. The equivalent stress conversion formula is used to uniformly scalarize the physical stress data of nodes or local elements. Specifically, by combining the bending stress σ b and the shear stress σ s according to the weights, an equivalent stress expression is constructed. This formula is a simplified form based on the Von Mises criterion and is applicable to analyzing the comprehensive response level of each node under complex stress states. Through this conversion, the original stress tensor data can be uniformly converted into a scalar form of "equivalent response index", which is convenient for subsequent stress comparison, material response evaluation or fault prediction modeling at the node level.

[0116] Preferably, step S23 includes the following steps:

[0117] Step S231: Perform triangulation processing using the cable spatial enhanced vector and the cable physical enhanced vector, and output nodes according to the physical parameter similarity greater than 0.7 to obtain a cable hybrid adjacency list; [[ID=e46]]

[0118] Step S232: Calculate the similarity of the cable hybrid adjacency list using cosine similarity to obtain cable space-physical weighted combined data;

[0119] Step S233: The cable space-physical weighted combined data constructs node weights according to the stress differences of the boundary points, thereby constructing the topological structure of the cable assembly; performs graph data object conversion on the topological structure of the cable assembly, and generates low-dimensional encoded data to obtain cable topology encoded data.

[0120] In the embodiments of the present invention, through technical means such as triangulation, similarity calculation, node weight construction, and graph data processing, cable topology coding data is finally generated. The following details its specific technical means at the data level. It is mentioned that triangulation processing is carried out using the cable spatial enhancement vector and the cable physical enhancement vector, and node output is performed according to the physical parameter similarity being greater than 0.7, thereby obtaining the cable hybrid adjacency list. The key technologies in this step are triangulation technology and similarity calculation. Triangulation divides the spatial and physical characteristic data of the cable into triangular units with certain geometric relationships, facilitating subsequent data processing and structural analysis. The condition of "physical parameter similarity greater than 0.7" is obtained by calculating the similarity between the spatial and physical characteristics and screening out node pairs with higher similarity, and finally obtaining the cable hybrid adjacency list. The construction of the adjacency list is based on the similarity of physical parameters, ensuring the rationality and high precision of the topological structure. The cosine similarity is used to calculate the similarity of the cable hybrid adjacency list to obtain the cable spatial-physical weighted combined data. In this process, the cosine similarity is used as a measure of the similarity between the spatial and physical characteristics. Specifically, the cosine similarity evaluates its similarity by calculating the angle between the spatial enhancement vector and the physical enhancement vector. The obtained similarity data can provide a basis for subsequent weighted combination, ensuring the effective fusion of the spatial and physical characteristic data to form the weighted combined data. This combined data can reflect the comprehensive performance of the cable in terms of spatial and physical characteristics. The cable spatial-physical weighted combined data constructs node weights according to the stress differences of the boundary points, and then constructs the cable component topological structure. The stress difference, as the basis for constructing node weights, can accurately reflect the force conditions of each part of the cable under physical load, thereby reasonably allocating weights to the nodes. This process ensures the matching of the weight distribution between nodes and the actual physical load through stress difference data, further improving the authenticity and accuracy of the topological structure. Finally, the constructed cable component topological structure can effectively express the connection relationships and physical characteristics between the various nodes of the cable component. The cable component topological structure is transformed into a graph data object and low-dimensional coding data is generated to obtain the cable topology coding data. In this process, through graph data object transformation, the topological structure is transformed into a data form suitable for calculation and analysis, and further low-dimensional coding is performed to reduce the dimension of the data, making it more suitable for subsequent storage and calculation operations. Low-dimensional coding can effectively reduce the computational complexity and facilitate further processing and application of the data. Generally speaking, these technical means combine spatial and physical data, and use methods such as geometry, similarity calculation, stress analysis, and graph data processing to construct efficient and accurate cable topology coding data at the data level, providing a solid foundation for subsequent cable analysis, optimization, and application.

[0121] Preferably, the step S3 of using the cable combined feature matrix to output the signed distance value and initializing the neural network weight parameters with the cable topology coding data includes:

[0122] Extracting the cable surface point set from the curvature-torsion parameter index sub-matrix; using the cable surface point set to measure the shortest distance true value label of the sampling points to obtain the signed distance value;

[0123] Taking the cable topology coding data as the hidden layer weight matrix, and using the signed distance value for geometric characterization initialization mapping to obtain the initial parameters of the cable geometric model;

[0124] Using the preset LSTM deformation predictor to perform temporal modeling on the initial parameters of the cable geometric model, and using the output temporal affine transformation data for deformation data fusion to obtain the deformation parameter set;

[0125] Performing isosurface extraction and optimization on the deformation parameter set, and outputting the cable parameter geometric model.

[0126] In the embodiment of the present invention, the cable surface point set is extracted from the curvature-torsion parameter index sub-matrix. This point set represents the apparent structure of the cable in three-dimensional space and has a high curvature and torsion feature density, which is the key data for characterizing the cable structure features. Subsequently, based on this surface point set, the shortest Euclidean distance of the measured sampling points is calculated, and at the same time, symbols are added according to the internal and external relationships of the point set to generate the signed distance value (Signed Distance Function, SDF). This value is important label information used to construct the continuous shape boundary in geometric modeling. Next, the previously constructed cable topology coding data is used as the hidden layer weight matrix of the neural network to enhance the network's perception ability of the cable structure topology, and the signed distance value is used as the target label for geometric characterization initialization to guide the neural network to construct the initial parameter distribution of the cable geometric model. Then, the defined long short-term memory network (LSTM) is used as the deformation predictor to perform temporal modeling on the initial parameters of the cable geometric model. By inputting the time step and historical deformation information, the affine transformation estimation during the deformation process is realized. The affine transformation data output by the LSTM network reflects the deformation trend of the cable over time, and then is fused with the initial parameters to obtain the deformation parameter set with time continuity. Finally, a three-dimensional isosurface is constructed based on the deformation parameter set, and isosurface extraction algorithms such as Marching Cubes are used to realize the reconstruction and optimization of the space surface to restore the true geometric profile of the cable under multiple moments and multiple stress states. Finally, it is output as the cable parameter geometric model. This model not only retains the continuity of the geometric shape, but also integrates physical topology and time series information, providing a reliable data basis for subsequent improvement of modeling accuracy and state analysis.

[0127] Preferably, performing isosurface extraction optimization on the deformation parameter set and outputting the cable parameter geometric model includes:

[0128] Gradient sampling judgment is performed according to the deformation parameter set, and isosurface extraction optimization is performed to generate gradient sampling isosurface data, where the gradient sampling isosurface data includes high gradient area data and low gradient area data. The high gradient area data is sampled at a sampling interval of 0.1 mm, and the low gradient area data is sampled at a sampling interval of 1.0 mm.

[0129] The ring topology structure of the gradient sampling isosurface data is extracted, and the smooth step artifacts are eliminated to complete the construction of the cable parameter geometric model.

[0130] In an embodiment of the present invention, for the previously obtained deformation parameter set, its gradient field in three-dimensional space is calculated. The gradient field is based on the rate of change between points and reflects the severity of the local geometric morphological changes. Based on the gradient field, gradient sampling judgment is performed to extract areas with significant changes (i.e., high gradient areas) and areas with gentle morphology (i.e., low gradient areas), and the entire model is divided into sub-areas with different resampling requirements. In terms of specific sampling strategies, high gradient areas are sampled with a higher spatial resolution, and the sampling interval is set to 0.1mm to capture detailed information of areas with severe deformation; while for low gradient areas, lower resolution is used for sampling, and the sampling interval is set to 1.0mm, so as to take into account both data density and computational overhead. After forming new gradient sampling isosurface data, the isosurface data is further subjected to ring topology structure extraction, and the isosurface is subjected to surface fitting and structural closure using a triangular mesh reconstruction algorithm (such as Delaunay or Poisson reconstruction method) to ensure that it has a continuous and closed three-dimensional topology. In order to eliminate the smoothing step artifact problem caused by excessive curvature changes, a smoothing function based on gradient edge filtering is introduced to iteratively optimize the fitting surface. Usually, a Gaussian kernel or bilateral filter is combined to perform edge-preserving smoothing operations on boundary-sensitive areas. This process not only ensures the consistency of the topological structure, but also improves the physical consistency of the model in visual and geometric expression. The final output cable parameter geometry model takes into account both local details and overall contours under multi-scale gradient control, and has the ability to reconstruct with high precision and express real structures.

[0131] As an example of the present invention, refer to Figure 2 As shown, in this example, step S4 includes:

[0132] Step S41: performing multi-scale verification on the cable parameter geometric model to obtain cable model evaluation data;

[0133] Step S42: Construct an elastic parameter list based on the cable model evaluation data; use the elastic parameter list to finely tune the cable parameter geometric model to obtain an optimized cable assembly parameter model;

[0134] Step S43: Construct a cable assembly parameter model output report based on the optimized cable assembly parameter model.

[0135] In the embodiment of the present invention, in the process of finely verifying and correcting the model through multi-scale analysis and elastic parameter regulation mechanism, the core technical path focuses on extracting the parameter response relationship from the evaluation data and constructing a cable assembly parameter model. Perform multi-scale verification on the constructed cable parameter geometric model, that is, introduce a standard geometric reference template and measured structure data at different spatial scales (such as millimeter level, sub-millimeter level, and micron level), and through evaluation indexes such as distance error analysis, curvature distribution fitting error, and node-to-node tension simulation deviation, perform multi-level quantification on the geometric fitting degree and physical consistency of the model, and finally form cable model evaluation data including an error distribution matrix, a boundary deformation analysis atlas, and the identification results of stress concentration points. Based on the structural response characteristics and error aggregation areas reflected in the evaluation data, construct an elastic parameter list, which consists of equivalent Young's modulus, Poisson's ratio, local stiffness coefficient, and stress-strain residual vector at different positions, and is indexed by node level or grid area. Subsequently, use the elastic parameter list as an optimization vector to act on the cable parameter geometric model, and through methods such as introducing a stiffness adjustment coefficient, a deformation constraint matrix, and an elastic coupling coefficient, perform adaptive adjustment on each sub-structure in the model to achieve the coupling consistency correction between the cable geometry and mechanical response, and then obtain an optimized cable assembly parameter model. Based on the optimized cable assembly parameter model, systematically integrate geometric feature information, mechanical property indexes, topological connection relationships, and structural stability evaluation data, and construct a cable assembly parameter model output report, which uses a parameter index table, a geometric response atlas, and a model optimization path diagram as carriers to realize the full-chain data closed-loop expression from the original data to the engineering parameter model.

[0136] Preferably, the multi-scale verification of the cable parameter geometric model includes the following steps:

[0137] Generate comprehensive multi-granularity result rendering data by using the industrial personalized output results, where the comprehensive multi-granularity result rendering data includes macro-layer rendering data and meso-layer rendering data;

[0138] Obtain laser tracker measurement data and SEM surface topography scan data;

[0139] Set the macro threshold CD to 0.15 mm; use a laser tracker to measure data and perform error analysis on the vertex of the cable parameter geometric model to the measured point according to the macro threshold CD to obtain macro error analysis data; perform dataset mapping based on the macro error analysis data to obtain the cable macro error distribution map;

[0140] Extract the normal of the cable parameter geometric model to obtain the surface normal map of the cable model; use the scanning electron microscope surface topography scan data to evaluate the microstructure similarity of the cable model surface normal map to obtain the cable micro similarity data.

[0141] In the embodiment of the present invention, at the macro scale level, a high-precision laser tracker is used to locate several control points of the cable entity in the three-dimensional space, collect the three-dimensional coordinate sequence of the target points, construct the measured point cloud dataset, and set the global geometric matching threshold CD = 0.15 mm, which is used as the upper limit of the allowable model error; then use the point cloud error analysis algorithm to measure the spatial distance between the vertices of the cable parameter geometric model and the laser measured point cloud one by one, and then construct the vertex error matrix, and mark and cluster the vertices with errors greater than CD to generate macro error analysis data. On this basis, based on the mapping relationship between the error dataset and the model grid coordinate system, a one-to-many function from the error value to the spatial position is constructed to generate the cable macro error distribution map and realize the visual expression of the error space. At the micro scale level, the micro texture structure diagram of the cable surface is obtained by scanning electron microscope imaging of the cable surface, and the feature point distribution, surface roughness pattern and micro-undulation information are extracted by using image registration and structural similarity algorithms; at the same time, the surface normal of the cable parameter geometric model is extracted to form a continuous normal vector field, that is, the surface normal map of the cable model. Subsequently, indicators such as the cosine value of the angle and the correlation of the local direction field are used to evaluate the microstructure similarity between the electron microscope image features and the normal map to generate the cable micro similarity data, and quantify the consistency degree between the micro structure of the model surface and the real physical representation. The whole process integrates four key data of spatial point cloud error, vertex normal direction, surface image features and structural similarity from the data dimension, providing data support for subsequent model correction and multi-scale fusion modeling.

[0142] Preferably, step S42 includes the following steps:

[0143] Step S421: Extract the high-sensitive parameters according to the cable macro error distribution map to obtain the cable macro sensitive parameters; use the cable macro sensitive parameters to construct an elastic parameter list;

[0144] Step S422: Fine-tune the cable micro similarity data with a small step size of 0.001, and generate the optimized parameters of the geometric cable model by fitting the elastic parameters to the elastic parameter list;

[0145] Step S423: Fine-tune the cable parameter geometric model using the geometric cable model optimization parameters, and perform parameter feedback to obtain the cable assembly parameter optimization model.

[0146] In the embodiment of the present invention, according to the generated cable macroscopic error distribution map, the spatial gradient recognition algorithm and the error response mapping mechanism are used to extract the control variables corresponding to the model vertices in the high-error region, and combined with the node geometric attributes, material identification and historical perturbation responses, various error influencing factors are attributed and sorted, so as to obtain the cable macroscopic sensitive parameter set. This parameter set mainly reflects the high responsiveness of the model to geometric errors under specific regions or specific material properties. On this basis, an elastic parameter list is established, and the material mechanics characteristics and the empirical rules of structural response are introduced to convert the macroscopic sensitive parameters into a set of initial elastic coefficient data with clear physical meanings and adjustability. Subsequently, enter the data fitting and correction link at the microscopic level. First, set the step size to 0.001, and obtain the local change rate of the microscopic similarity response curve by performing a small-range gradient perturbation on the cable microscopic similarity data, and use this as the input of the sensitive region of the parameter fitting function. Using the interpolation fitting and error backpropagation mechanism, the initial elastic parameter list is approximated by a function to generate an elastic parameter update set for geometric optimization, that is, the geometric cable model optimization parameters. Finally, this optimized parameter set is fed back to the original cable parameter geometric model, and an elastic fine-tuning strategy for the key region is executed. The model node attributes are updated synchronously through the vertex reconstruction, normal correction and local surface reconstruction algorithms to achieve the overall optimization of the cable assembly parameters, and a parameter optimization model with error adaptability and physical consistency is formed. At the same time, the optimized parameters form a closed-loop feedback to support the next round of iteration or model evaluation, enhancing the data coherence and control stability of the parameter correction process.

[0147] Particularly importantly, step S43 includes:

[0148] Step S431: Perform an integrity assessment of the geometric model based on the cable assembly parameter optimization model to obtain cable geometric model assessment data;

[0149] Step S432: Perform data visualization processing on the cable geometric model assessment data, and upload the visualization result to the cloud platform for storage to obtain cable assembly visualization processing data;

[0150] Step S433: Generate a cable assembly parameter model output report based on the cable assembly visualization processing data.

[0151] In the embodiments of the present invention, for the geometric model of the cable assembly whose parameters have been finely tuned, by quantitatively inspecting geometric constraint indicators such as node coordinate consistency, boundary continuity, surface normal consistency, and topological closure in the model structure, a data set with multi-dimensional outputs including local geometric deviation values, global topological consistency indicators, and surface residual distribution is generated, constituting the evaluation data of the cable geometric model. This data is usually stored in matrix form and annotated with specific position coordinates and corresponding error values. Subsequently, based on this evaluation data, a three-dimensional visualization expression model is constructed, and key geometric defect areas are prominently marked and structural differences are highlighted by means of point cloud reconstruction, error heat mapping, and transparent body structure hierarchical rendering, and cross-platform compatible interactive model data is generated through graphics interfaces such as WebGL or OpenGL. This visualization result is then compressed through JSON or binary stream encoding and uploaded to the specified cloud platform storage interface, and is associated with a database indexing mechanism for traceable identification, forming the visualization processing data of the cable assembly. On this basis, the visualization processing data is associated with the parameter structure of the geometric model, and information such as evaluation summaries, structural parameters, deformation trend maps, and material identification codes is extracted through an automated template generator, and finally a parameter model output report for the cable assembly is constructed, which is essentially a composite report file integrating structural data, multi-dimensional visualization results, and parameter analysis content.

[0152] Therefore, from any perspective, the embodiments should be regarded as exemplary and non-restrictive. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the application document are intended to be encompassed within the present invention.

[0153] The above description is only a specific implementation manner of the present invention, enabling those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown herein, but will conform to the broadest scope consistent with the principles and novel features disclosed herein.

Claims

1. A geometric modeling method for a parametric model of a cable assembly, characterized in that, It includes the following steps: Step S1: Deploy a multi-source sensor array to collect cable spatial trajectory data, construct a feature matrix, and generate a cable joint feature matrix; Step S2: Extract key parameters from the cable joint feature matrix and construct a node feature vector; Use the node feature vector to construct an adaptive graph and build a topological structure to obtain cable topological coding data; Step S3: Use the cable joint feature matrix to output a symbol distance value, and initialize the neural network weight parameters with the cable topological coding data to obtain the initial parameters of the cable geometric model; Use a preset LSTM deformation predictor to perform deformation data fusion on the initial parameters of the cable geometric model, and perform isosurface extraction optimization to output the cable parameter geometric model; Step S4: Perform multi-scale verification on the cable parameter geometric model to obtain cable model evaluation data; Use the cable model evaluation data to fine-tune the parameters of the cable parameter geometric model and construct a cable component parameter model output report.

2. The geometric modeling method of the cable assembly parametric model according to claim 1, characterized in that Step S1 includes the following steps: Use a lidar array with a scanning frequency of 100Hz to perform three-dimensional scanning on the cable to obtain original point cloud coordinate data; Measure the movement trajectory of the cable through an IMU inertial measurement unit to obtain cable deformation data; Deploy fiber Bragg grating sensors on the cable at an interval of 5mm to collect cable strain value data; Collect wavelength data through a signal collector and convert the wavelength data into cable wavelength data using an optical signal demodulator; Perform cubic spline interpolation on the original point cloud coordinate data and the cable deformation data to obtain lidar-deformation data; Perform downsampling processing on the cable strain value data and the cable wavelength data to obtain fiber optic sensor sampling data; Align the spatial coordinate systems according to the lidar-deformation data and the fiber optic sensor sampling data, and iterate the closest points to obtain a cable joint feature matrix.

3. The geometric modeling method of the cable assembly parametric model according to claim 1, characterized in that Step S2 includes the following steps: Step S21: Perform curvature-torsion sensitivity analysis on the cable joint feature matrix to generate a curvature-torsion parameter index submatrix; Step S22: Calculate the radial distance from the matrix node to the center line based on the curvature-torsion parameter index submatrix with the cable center line as the reference to obtain a cable spatial enhancement vector; Perform equivalent stress conversion on the curvature-torsion parameter index submatrix through Hooke's law to obtain a cable physical enhancement vector; Step S23: Use the cable spatial enhancement vector and the cable physical enhancement vector to construct an adaptive graph and build a topological structure to obtain cable topological coding data.

4. The geometric modeling method of the cable assembly parametric model according to claim 3, characterized in that, Step S21 includes the following steps: Step S211: Obtain the actual measurement error of cable laser scanning; Screen the cable joint feature matrix for curvature and torsion higher than 1.5 times based on the actual measurement error of cable laser scanning to obtain high-precision cable curvature-torsion data; Step S212: Apply a 5% perturbation of curvature-torsion to the high-precision cable curvature-torsion data and calculate the perturbation parameter sensitivity to obtain preliminary cable curvature-torsion sensitivity data; Step S213: Perform coupled effect matrix analysis based on the preliminary cable curvature-torsion sensitivity data and evaluate the error contribution value to obtain a curvature-torsion parameter index submatrix.

5. The geometric modeling method of the cable assembly parametric model according to claim 3, characterized in that Step S23 includes the following steps: Step S231: Perform triangulation processing using the cable space enhancement vector and the cable physical enhancement vector, and output nodes according to the physical parameter similarity greater than 0.7 to obtain a cable hybrid adjacency list; Step S232: Calculate the similarity of the cable hybrid adjacency list using cosine similarity to obtain cable space-physical weighted combined data; Step S233: Construct node weights for the cable space-physical weighted combined data according to the stress differences of the boundary points, thereby constructing a cable component topological structure; perform graph data object transformation on the cable component topological structure and generate low-dimensional encoded data to obtain cable topological encoded data.

6. The geometric modeling method of the cable assembly parametric model according to claim 1, characterized in that The steps for outputting the symbolic distance value using the cable joint feature matrix in Step S3 and initializing the neural network weight parameters with the cable topological encoded data include: Extract the cable surface point set from the curvature-torsion parameter index sub-matrix; use the cable surface point set to measure the true distance label of the sampling points at the shortest distance to obtain the symbolic distance value; Use the cable topological encoded data as the hidden layer weight matrix, and perform geometric characterization initialization mapping with the symbolic distance value to obtain the initial parameters of the cable geometric model; Use a preset LSTM deformation predictor to perform temporal modeling on the initial parameters of the cable geometric model, and fuse the deformation data with the output temporal affine transformation data to obtain a set of deformation parameters; Perform isosurface extraction optimization on the set of deformation parameters and output a cable parameter geometric model.

7. The geometric modeling method of the cable assembly parametric model according to claim 5, characterized in that, Performing isosurface extraction optimization on the set of deformation parameters and outputting a cable parameter geometric model includes: Perform gradient sampling judgment according to the set of deformation parameters, and perform isosurface extraction optimization to generate gradient sampling isosurface data, where the gradient sampling isosurface data includes high-gradient area data and low-gradient area data. Sample the high-gradient area data at a sampling interval of 0.1 mm, and sample the low-gradient area data at a sampling interval of 1.0 mm; Extract the annular topological structure from the gradient sampling isosurface data and eliminate smooth step artifacts, thereby completing the construction of the cable parameter geometric model.

8. The geometric modeling method of the cable assembly parametric model according to claim 1, characterized in that Step S4 includes the following steps: Step S41: Perform multi-scale verification on the cable parameter geometric model to obtain cable model evaluation data; Step S42: Construct an elastic parameter list according to the cable model evaluation data; use the elastic parameter list to fine-tune the parameters of the cable parameter geometric model to obtain an optimized model of cable component parameters; Step S43: Construct a cable component parameter model output report based on the optimized model of cable component parameters.

9. The geometric modeling method of the cable assembly parametric model according to claim 8, characterized in that, Performing multi-scale verification on the cable parameter geometric model includes the following steps: Obtain laser tracker measurement data and SEM surface topography scan data; Set the macroscopic threshold CD to 0.15 mm; use the laser tracker measurement data to analyze the error between the model vertices and the measured points of the cable parameter geometric model according to the macroscopic threshold CD to obtain macroscopic error analysis data; perform dataset mapping according to the macroscopic error analysis data to obtain a cable macroscopic error distribution map; Extract the normal of the cable parameter geometric model to obtain the cable model surface normal map; use the SEM surface topography scan data to evaluate the microstructure similarity of the cable model surface normal map to obtain cable microscopic similarity data.

10. The geometric modeling method of the cable assembly parametric model according to claim 8, characterized in that Step S42 includes the following steps: Step S421: Extract high-sensitivity parameters according to the cable macroscopic error distribution map to obtain cable macroscopic sensitivity parameters; construct an elastic parameter list using the cable macroscopic sensitivity parameters; Step S422: Fine-tune the cable microscopic similarity data with a small step size of 0.001, and generate fitted elastic parameters for the elastic parameter list to obtain optimized geometric cable model parameters; Step S423: Use the optimized geometric cable model parameters to finely adjust the parameters of the cable parameter geometric model, and perform parameter feedback to obtain an optimized cable assembly parameter model.

Images