Additive Manufacturing Model Feature Identification via Signed Distance Fields
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for additive manufacturing digital model preprocessing struggle to efficiently identify and extract geometric and process features, particularly internal features, due to limitations in feature identification and extraction methods based on contour, topology, and visual shapes.
Innovation Solution
A field-based method that converts digital models into signed distance fields, introduces simulated physical fields, and performs frequency domain analysis filtering, using multi-precision convolution units to analyze feature distance fields and combine multiple field data for feature classification and extraction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If conventional contour-based feature identification methods are used on STL files, then the method is simple to implement, but the number of identifiable features is limited and internal features cannot be detected
Solution Approach 1:
The patent introduces a signed distance field as an intermediary representation between the STL surface model and feature identification. This distance field serves as a mediator that preserves both surface and internal volumetric information, enabling comprehensive feature detection while maintaining computational feasibility through field-based operations rather than complex geometric processing
Solution Approach 2:
The patent transforms the geometric representation from discrete surface contours to a continuous distance field parameterization. By changing the representation parameter from boundary-only coordinates to volumetric distance values, the method enables access to internal features while maintaining computational efficiency through field sampling and convolution operations
2Measurement precision
If 3D-CNN network structures are used for feature identification, then identification precision is improved, but the method still cannot identify specific geometric features such as thin walls
Solution Approach 1:
The patent segments the feature identification process into distinct analytical components: distance field computation, gradient calculation, curvature analysis, and specific feature detection rules. This segmentation allows each component to be optimized independently, ensuring both high identification precision and accurate detection of specific geometric features like thin walls through dedicated computational rules
Solution Approach 2:
The patent replaces the black-box 3D-CNN mechanical processing with a physics-based field analysis system. By substituting neural network inference with explicit distance field computations and geometric analysis, the method achieves both high precision identification and accurate geometric feature detection through interpretable mathematical operations
3Productivity
If 2D slice-based contour feature identification is used, then the method is computationally efficient, but the scope of application is limited and cannot identify internal model information
Solution Approach 1:
The patent transitions from 2D slice-based analysis to 3D distance field analysis by introducing a volumetric dimension. This dimensional extension allows the method to capture internal features and spatial relationships in three dimensions while maintaining computational efficiency through sampled field representations and localized convolution operations
4Device complexity
If surface information only is processed, then data processing is simple, but feature extraction is limited and manufacturing quality cannot be optimized
Solution Approach 1:
The patent creates a composite data structure by integrating surface geometry information with volumetric distance field information. This composite representation combines the simplicity of surface processing with the richness of internal feature data, enabling both simple implementation and high manufacturing quality through unified field-based feature extraction
Data Source
AI summary
The disclosure belongs to the technical field related to additive manufacturing model preprocessing, and discloses a field-based additive manufacturing digital model feature identification and extraction method and device, the method can convert the digital model represented by the facet into a signed distance field and introduce a simulated physical field of the forming/service simulation, the feature distance field after the frequency domain analysis filtering, and the geometric feature field obtained by the multi-precision convolution unit analysis according to the requirement of feature to be identified, then, multiple fields are combined to realize feature classification determination and labeling of features, and finally feature extraction is completed based on field data and isosurface and isoline reconstruction algorithms.