Powder Bed Fusion 3D Modeling from NIR Melt and Powder Detection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for monitoring additive manufacturing processes have limited accuracy, particularly in detecting variations in weld pool width and differentiating between welded metal and agglomerated powder particles, leading to significant errors in thin-walled structures and components with complex geometries.

Innovation Solution

The techniques described involve obtaining near-infrared (NIR) images of each layer during the welding process, generating a multi-dimensional dataset, detecting melted regions and agglomerated powder, and applying a multi-layer predictive model to account for inter-layer melt effects and porosity defects, thereby reconstructing a 3D model with high geometric accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If existing optical methods with threshold-based image processing are used to monitor additive manufacturing, then the monitoring process is simple and fast, but the measurement precision deteriorates significantly for thin-walled structures and components with complex geometries

Engineering Contradiction:
Improvemonitoring speedVSAvoidgeometric measurement accuracy
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The patent transitions from 2D threshold-based image processing to 3D volumetric analysis by stacking and processing multiple layer images through the build volume. This dimensional transformation enables accurate measurement of thin-walled structures and complex geometries by analyzing the three-dimensional context of powder agglomeration and weld penetration across multiple layers, rather than treating each layer in isolation.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Solution Approach 2:

The patent changes the analysis parameters from simple binary thresholding to multi-parameter 3D analysis including weld pool width variations, powder agglomeration detection, and inter-layer penetration depth. By introducing these additional measurement parameters and using predictive models to account for process variations, the system achieves high measurement precision while maintaining monitoring efficiency.

Inventive Principle:
Principle #35Parameter changes

2Ease of manufacture

If threshold-based image processing is used to differentiate melted regions from un-melted powder, then the processing is straightforward, but the reliability deteriorates when detecting variations in weld pool width and agglomerated powder particles

Engineering Contradiction:
Improveprocessing simplicityVSAvoiddetection accuracy
Core Design Contradiction:
Ease of manufactureVSReliability

Solution Approach 1:

The patent introduces an intermediary predictive model that acts as a bridge between simple image capture and reliable defect detection. This model incorporates knowledge of the additive manufacturing process, including expected weld pool dimensions, powder agglomeration patterns, and inter-layer penetration characteristics. The intermediary model processes the raw images to distinguish between normal process variations and actual defects, thereby achieving high reliability while maintaining processing simplicity.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent performs preliminary analysis by stacking and pre-processing multiple layer images before final defect detection. This preliminary action includes aligning layers, detecting powder agglomeration patterns, and establishing baseline weld pool dimensions. By performing these preparatory steps in advance, the system reduces the complexity of the final detection step while improving overall reliability through cumulative information from multiple layers.

Inventive Principle:
Principle #10Preliminary action

3Measurement precision

If CT scanning is used to obtain 3D model data, then comprehensive internal structure information is obtained, but the cost and time delay increase significantly

Engineering Contradiction:
Improve3D model accuracyVSAvoidmeasurement time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent creates a digital copy or digital twin of the physical component by processing images captured during the additive manufacturing process itself. Instead of using separate CT scanning after manufacturing, the system constructs a 3D digital model that mirrors the actual printed geometry, including weld pool dimensions, powder agglomeration, and inter-layer penetration. This digital copy provides comprehensive internal structure information at no additional time cost or expense beyond the normal manufacturing monitoring.

Inventive Principle:
Principle #26Copying

Solution Approach 2:

The patent performs 3D modeling as a preliminary action during the manufacturing process itself, rather than as a separate post-processing step. By capturing and processing images layer-by-layer as the component is being printed, the system builds the 3D model concurrently with manufacturing. This preliminary action eliminates the need for separate CT scanning, achieving both high measurement precision and zero additional time delay.

Inventive Principle:
Principle #10Preliminary action

4Measurement precision

If input CAD model masking is used to correct for surface temperature variations, then noise reduction is achieved, but the ability to detect weld pool width variations exceeding commanded size deteriorates

Engineering Contradiction:
Improveimage qualityVSAvoiddefect detection capability
Core Design Contradiction:
Measurement precisionVSAdaptability or versatility

Solution Approach 1:

The patent segments the image analysis into distinct functional regions: expected weld areas identified from CAD, actual weld areas detected from images, and unaccounted variations representing defects. By segmenting the analysis in this way, the system can use CAD masking to reduce noise from temperature variations in expected areas, while simultaneously detecting deviations in weld pool width or unexpected powder agglomeration as separate detectable features rather than masking them away.

Inventive Principle:
Principle #1Segmentation

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

These techniques provide non-destructive testing data equivalent to CT scans without additional cost or time delay, maintaining accuracy across varying component sizes, complexities, and material densities, and accurately detecting porosity defects and agglomerated powder.

Implementation Method 1

An NIR or IR camera is used to capture an image of each layer

Methodology Applied
Scientific EffectNear-infrared radiation detection: Infrared Radiation

Implementation Method 2

hot melted (solid) metal has a substantially lower emissivity in the near-infrared (NIR) and infrared (IR) spectrum than the heated powder at a similar temperature

Methodology Applied
Scientific EffectEmissivity difference in infrared spectrum: Infrared Radiation

Data Source

PatentUS20250157022A1Unsupervised 3D modeling and geometric measurement of parts built with powder bed fusion-based additive manufacturing
Publication Date: 2025.05.15 BWXT ADVANCED TECHNOLOGIES LLC
  • US20250157022A1 patent drawing
  • US20250157022A1 patent drawing
  • US20250157022A1 patent drawing

AI summary

A method is provided for unsupervised 3D modeling and geometric measurement of additively manufactured parts. The method includes obtaining near-infrared (NIR) images for a welding process for a welded part. The welded part includes welded metal and agglomerated powder particles. The method also includes generating a multi-dimensional dataset based on the NIR images, including cropping the multi-dimensional dataset to a region of interest. The method also includes detecting melted regions in image layers of the multi-dimensional dataset to obtain an output volume. The method also includes detecting agglomerated powder in the output volume to obtain melt masks. Each melt mask indicates weld pixels for a respective image layer. The method also includes applying a multi-layer predictive model to account for multi-layer weld penetration, based on the melt masks, to obtain an output data mask that represents a 3D model and geometric measurements for the welded part.