3D Shape Modification Using Coarse Structural Analysis

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing topology optimization methods for 3D object design require computationally complex structural analysis operations, limiting the number of iterations and reducing the design space explored, which can result in sub-optimal designs being selected for manufacturing.

Innovation Solution

A machine learning model is trained to modify shape portions using coarse structural analysis data, reducing computational complexity by generating designs based on lower resolution data, allowing for a more comprehensive exploration of the design space.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If topology optimization applications perform computationally complex structural analysis operations for each shape element during each iteration, then the design quality and convergence with design objectives improve, but the computational complexity and time consumption increase significantly

Engineering Contradiction:

Improvedesign qualityVSAvoidcomputational complexity

Core Design Contradiction:

Manufacturing precisionVSDevice complexity

Solution Approach 1:

The shape is divided into multiple shape elements that can be processed independently. The structural analysis is performed on individual shape elements or small groups of elements rather than the entire shape, reducing the computational burden per iteration while maintaining overall design quality through cumulative optimization effects.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent performs preliminary structural analysis operations before the main optimization iterations. By pre-computing certain structural properties and responses, the system reduces the computational complexity required during each subsequent iteration, allowing more iterations to be performed within the same time budget.

Inventive Principle:

Principle #10Preliminary action

2Productivity

If the total number of structural analysis operations is limited to reduce computational complexity, then the processing time decreases, but the design space explored is reduced and sub-optimal designs may be selected

Engineering Contradiction:

Improveprocessing speedVSAvoiddesign optimization quality

Core Design Contradiction:

ProductivityVSManufacturing precision

Solution Approach 1:

The patent performs structural analysis on only a subset of shape elements during each iteration rather than all elements. This partial action approach reduces computational complexity per iteration while still capturing the essential structural behavior needed for meaningful optimization progress, allowing more iterations within fixed time resources.

Inventive Principle:

Principle #16Partial or excessive action

Solution Approach 2:

The system dynamically adjusts parameters such as the number of shape elements analyzed per iteration, the resolution of structural analysis, and the frequency of full structural analyses. By changing these parameters adaptively, the system balances computational complexity and design quality to maximize the exploration of design space within available time.

Inventive Principle:

Principle #35Parameter changes

3Loss of time

If the number of topology optimization problems solved is reduced to save computational resources, then the overall processing time decreases, but the variety and quality of generated designs decrease

Engineering Contradiction:

Improvetotal design timeVSAvoiddesign space exploration

Core Design Contradiction:

Loss of timeVSAdaptability or versatility

Solution Approach 1:

The generative design process is segmented into multiple independent topology optimization problems that can be solved in parallel. Each problem focuses on a specific design objective or constraint combination, allowing the system to explore diverse design spaces efficiently without sequentially solving all problems, thus reducing total computation time while maintaining design variety.

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS12524978B2Techniques for training a machine learning model to modify portions of shapes when generating designs for three-dimensional objects

Publication Date: 2026.01.13 AUTODESK INC

AI summary

In various embodiments, a training application trains a machine learning model to modify portions of shapes when designing 3D objects. The training application converts first structural analysis data having a first resolution to first coarse structural analysis data having a second resolution that is lower than the first resolution. Subsequently, the training application generates one or more training sets based on a first shape, the first coarse structural analysis data, and a second shape that is derived from the first shape. Each training set is associated with a different portion of the first shape. The training application then performs one or more machine learning operations on the machine learning model using the training set(s) to generate a trained machine learning model. The trained machine learning model modifies at least a portion of a shape having the first resolution based on coarse structural analysis data having the second resolution.