Powder Bed Fusion Reference Sample for Faster Condition Search
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In powder bed fusion additive manufacturing, determining optimal manufacturing conditions is labor-intensive and time-consuming due to numerous control parameters, leading to varying process windows and increased internal defect rates, especially when the molded object area changes.
Innovation Solution
An additive manufacturing condition search apparatus and method using a processor to generate a predictive model based on a reference sample's molding results, optimizing conditions through demonstration experiments and updating the model to achieve target evaluation values, with a reference sample featuring multiple surfaces and regions for precise condition determination.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional trial-and-error method is used to determine manufacturing conditions, then comprehensive evaluation of multiple parameters is possible, but enormous amount of time and cost is required
Solution Approach 1:
The patent applies preliminary action by pre-manufacturing reference samples with different known internal defect conditions before actual production. These reference samples are created in advance and stored, allowing the system to compare actual production results against pre-prepared benchmarks without performing trial-and-error experiments during the optimization process.
Solution Approach 2:
The patent uses copying by creating virtual models and digital twins of the manufacturing process. The image recognition system captures actual production samples and compares them against stored reference sample images, effectively copying the evaluation process digitally rather than requiring physical trial-and-error experimentation.
2Manufacturing precision
If multiple control parameters are adjusted to find optimal conditions, then manufacturing quality can be improved, but the complexity of experimentation increases
Solution Approach 1:
The patent introduces an intermediary - the image recognition system - that mediates between the complex manufacturing parameters and the final quality assessment. Instead of directly analyzing multiple control parameters, the system uses image data as an intermediary to indirectly evaluate the effects of parameter combinations on internal defects.
Solution Approach 2:
The patent applies the principle of color changes by using image recognition to detect visual characteristics of molded objects. The system analyzes color, texture, and visual patterns in images of the molded objects to identify internal defects, transforming complex quality assessment into visual signal analysis.
3Reliability
If process window is established through conventional experimentation, then baseline manufacturing conditions are obtained, but the conditions may vary between different operators
Solution Approach 1:
The patent implements feedback by using image recognition to automatically compare actual production results against reference samples. The system provides quantitative feedback on deviations from optimal conditions, enabling consistent evaluation regardless of which operator performs the analysis. This automated feedback loop ensures reliability across different operators.
Solution Approach 2:
The patent replaces the mechanical system of manual experimentation and human judgment with an automated image recognition system. Instead of operators visually inspecting and evaluating samples, the system uses computer vision algorithms to objectively assess internal defects, eliminating operator-dependent variability.
4Adaptability or versatility
If conventional methods are used to search for molding conditions, then existing process windows can be utilized, but efficiency is significantly reduced when molded object area changes
Solution Approach 1:
The patent applies dynamics by making the reference sample selection adaptive rather than static. The system dynamically selects appropriate reference samples based on the actual molded object's characteristics, including its area and geometry. This dynamic adaptation allows the system to efficiently handle different object sizes without requiring complete re-experimentation.
Solution Approach 2:
The patent uses parameter changes by adjusting which reference samples are selected based on the molded object's parameters such as area and geometry. The system changes the evaluation parameters and reference sample set according to the specific object being manufactured, enabling efficient adaptation to different scales without losing productivity.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach significantly enhances the efficiency and accuracy of finding optimal manufacturing conditions, reducing the time and cost of establishing process windows and minimizing internal defects by using a predictive model to iteratively refine settings.
Implementation Method 1
the processor generates a predictive model indicative of relation between the conditions and the molding result
Data Source
Figure 1
Figure 2A~2B
Figure 3A~3B
AI summary
The present invention relates to an apparatus, method, and reference sample for searching for molding conditions of an additive manufacturing apparatus 100 based on a powder bed fusion method, and provides a search apparatus 200, a search method, and a reference sample that are remarkably more efficient and accurate than conventional counterparts. The reference sample for searching for the molding conditions is shaped so that it has a surface having aggregated punched holes formed by straight lines and curved lines involved in a fill region of a molding region (in-skin), a region for forming an overhang (down-skin), and a region for forming an outermost surface in the direction of molding height (up-skin). Slice data of the reference sample has at least two independent regions in an optional layer at a center in a lamination direction, and includes a small region and a large region. The small region has a width of 1 mm or less and is cut off from the outer edge of the reference sample. The large region is formed by a portion other than the small region.