Additive Manufacturing Structural Simulation with Spatial Properties
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for estimating properties of articles manufactured by additive manufacturing (AM) are inaccurate due to heterogeneity, anisotropy, defects, and geometric deviations, leading to article-to-article variability.
Innovation Solution
A method involving a computer-based approach to obtain spatially resolved in-process parameters, infer corresponding attributes, and estimate properties using machine learning algorithms, accounting for heterogeneity, anisotropy, and defects in AM articles.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional estimation methods assuming homogeneous and isotropic articles are used, then the estimation process is simple, but the accuracy of property estimation deteriorates due to heterogeneity, anisotropy, and defects in AM articles
Solution Approach 1:
The patent segments the AM article into multiple regions based on spatial coordinates, where each region is assigned specific material properties inferred from in-process parameters. This segmentation allows the model to account for heterogeneity and anisotropy by treating different regions differently, rather than assuming uniform properties throughout the entire article.
Solution Approach 2:
The patent applies local quality by inferring region-specific material properties (such as density, mechanical properties) based on local in-process parameters at specific spatial coordinates. Each region's properties are tailored to its manufacturing conditions, capturing the local variations caused by heterogeneity, anisotropy, and defects in the AM process.
2Measurement precision
If spatially resolved in-process parameters are obtained and used to infer region-specific attributes, then the accuracy of property estimation improves, but the complexity of data processing and modeling increases
Solution Approach 1:
The patent performs preliminary action by obtaining and storing spatially resolved in-process parameters during the AM manufacturing process itself. This data is captured and organized by spatial coordinates before the estimation phase, creating a ready-to-use dataset that maps process conditions to specific locations in the article, thereby simplifying subsequent property inference.
Solution Approach 2:
The patent introduces an intermediary computational model that links in-process parameters to material properties through spatial coordinates. This intermediary layer processes the raw spatial data and transforms it into region-specific attribute estimates, bridging the gap between process data and property prediction while managing the complexity of spatial data processing.
3Reliability
If region-specific material properties are assigned based on in-process parameters, then article-to-article variability is accounted for, but the time and computational resources required for estimation increase
Solution Approach 1:
The patent performs preliminary action by capturing and organizing in-process parameters with spatial information during the manufacturing process. This pre-processing creates a structured dataset that can be efficiently queried and processed during estimation, reducing the computational burden when actual property predictions are needed for individual articles.
Solution Approach 2:
The patent creates a virtual copy or digital twin of the physical AM article, including its geometry and region-specific properties derived from in-process parameters. This digital representation can be used for repeated estimations and simulations without requiring additional physical testing or complex real-time computations on the actual article.
Data Source
AI summary
A method of estimating a property of an article manufactured, at least in part, by additive manufacturing, AM, preferably directed energy deposition, DED, and/or powder bed fusion, PBF, is described. The method implemented, at least in part, by a computer comprising a processor and a memory. The method comprises: obtaining a set of in-process parameters of the AM of the article, wherein respective parameters of the set thereof have positional information associated therewith; inferring a set of attributes of the article corresponding to the set of in-process parameters, wherein respective attributes of the set thereof have the positional information associated therewith; and estimating the property of the article based, at least in part, on the inferred set of attributes.