Graph Data Model for CAD Design Intent Recognition

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current systems lack the ability to machine-comprehend geometric designs created using Computer-Aided Design (CAD) tools, leading to inefficiencies in design reuse and automation of pre-processing tasks, as computer programs cannot understand the complex geometric designs generated by CAD tools.

Innovation Solution

A computer-implemented method that generates a machine-comprehensible graph data model by parsing features of CAD designs to identify attributes, computing vector representations of sketches, and mapping these attributes to a predefined ontology, creating a property graph that can be stored and used for similarity analysis and design intent recognition.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Extent of automation

If geometric designs are stored in traditional CAD formats, then the designs can be created and viewed by CAD tools, but computer programs cannot comprehend or analyze the design intent and geometric features

Engineering Contradiction:
Improvemachine comprehension of geometric designVSAvoiddesign intent understanding
Core Design Contradiction:
Extent of automationVSLoss of information

Solution Approach 1:

The patent introduces an intermediary system that translates CAD geometric designs into graph data models. This intermediary layer extracts geometric features, relationships, and design intent from CAD files and represents them as structured graph data with nodes and edges, enabling machine comprehension while preserving the original design information

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent transforms the representation parameters of geometric designs from traditional CAD formats (which are not machine-comprehensible for analysis) to graph data model parameters (nodes, edges, attributes, relationships). This parameter transformation enables automated analysis, similarity detection, and design intent understanding while maintaining fidelity to the original design

Inventive Principle:
Principle #35Parameter changes

2Productivity

If designers manually review past designs for reuse, then design intent can be understood, but the process is time-consuming and inefficient

Engineering Contradiction:
Improvedesign reuse efficiencyVSAvoidtime for design review
Core Design Contradiction:
ProductivityVSLoss of time

Solution Approach 1:

The patent replaces the manual mechanical process of design review with an automated computational system. The graph data model enables algorithms to automatically compare designs, detect similarities, and identify reusable components without human intervention, dramatically improving productivity while minimizing time loss

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

3Extent of automation

If geometric designs are converted to simple formats for machine processing, then automation is enabled, but the complex geometric details and design intent are lost

Engineering Contradiction:
Improveautomation of pre-processing tasksVSAvoidgeometric design accuracy
Core Design Contradiction:
Extent of automationVSManufacturing precision

Solution Approach 1:

The patent segments the geometric design into discrete graph elements (nodes representing geometric features, edges representing relationships). This segmentation enables automated processing of individual elements while preserving the overall geometric accuracy and design intent through the structured relationships in the graph model

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP4343603A1System and method for managing geometric designs
Publication Date: 2024.03.27 SIEMENS AG
  • EP4343603A1 patent drawingFigure 1
  • EP4343603A1 patent drawingFigure 2A
  • EP4343603A1 patent drawingFigure 2B~2C

AI summary

A method for managing a geometric design created using a CAD tool, comprising the steps of receiving, by a processor, a request for generating a machine-comprehensible graph data model corresponding to a geometric design created using a CAD tool, from a requesting entity. Further, a plurality of features associated with the geometric design are parsed to identify a set of first attributes indicative of a design intent. One or more of the features are associated with at least one sketch used for creation of the geometric design and at least one operation performed on the sketch. Further, the at least one sketch is analyzed to compute a set of second attributes indicative of a vector representation of the sketch. The machine-comprehensible graph data model generated based on the set of first attributes and the set of second attributes is provided to the requesting entity.