Adaptive Additive Manufacturing Control for Real-Time Defect Correction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Additive manufacturing processes face challenges in rapidly optimizing and adjusting process control parameters in response to changes, leading to suboptimal quality and efficiency in produced parts.

Innovation Solution

The implementation of a machine learning-based system for real-time adaptive control of free form deposition processes, utilizing a training data set that includes simulation, characterization, and inspection data to classify defects and adjust process parameters in real-time, enabling rapid optimization and improvement of process yield and quality.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If traditional additive manufacturing control methods are used, then the process is simpler to implement, but the ability to rapidly optimize and adjust process control parameters in response to changes is insufficient

Engineering Contradiction:
Improveability to adjust process control parametersVSAvoidcomplexity of control system
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent implements a closed-loop feedback control system where sensors continuously monitor process parameters and defect conditions, feed this information to a machine learning model, which then adjusts process control parameters in real-time. This feedback mechanism enables rapid adaptation to changing process conditions while maintaining systematic control through the coordinated interaction of monitoring, analysis, and actuation components.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The machine learning model autonomously analyzes sensor data, identifies defects, and determines optimal process parameter adjustments without requiring continuous human intervention. The system self-corrects process deviations by automatically modifying control parameters based on real-time feedback, enabling the manufacturing process to self-optimize and adapt to changing conditions.

Inventive Principle:
Principle #25Self-service

2Manufacturing precision

If real-time adaptive control using machine learning is implemented, then process optimization and quality improvement are enhanced, but the system complexity increases

Engineering Contradiction:
Improvequality of produced partsVSAvoidcomplexity of control system
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

The machine learning model serves as an intermediary between sensor data and process control adjustments. It processes complex sensor inputs, identifies defect patterns, and translates these into appropriate control parameter modifications. This intermediary layer manages the complexity by encapsulating the analytical complexity within the ML model while presenting a simplified control interface for parameter adjustment.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent replaces traditional mechanical and empirical control methods with an intelligent software-based machine learning system. Instead of relying on fixed mechanical control mechanisms or human operator judgment, the system uses computational algorithms to analyze process data and determine optimal control actions, substituting physical control complexity with computational intelligence.

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

3Productivity

If traditional process control is used, then the system is easier to operate, but productivity and process yield are suboptimal

Engineering Contradiction:
Improveprocess yield and throughputVSAvoidease of controlling the process
Core Design Contradiction:
ProductivityVSEase of operation

Solution Approach 1:

The machine learning-based control system autonomously monitors process parameters, identifies optimization opportunities, and adjusts control settings without requiring continuous manual intervention. This self-service capability maintains ease of operation by automating complex decision-making processes while significantly improving productivity through rapid, data-driven optimization of manufacturing parameters and defect detection.

Inventive Principle:
Principle #25Self-service

4Speed

If rapid adjustment of process parameters is enabled, then responsiveness to process changes improves, but the risk of instability increases

Engineering Contradiction:
Improvespeed of parameter adjustmentVSAvoidstability of process parameters
Core Design Contradiction:
SpeedVSStability of the object's composition

Solution Approach 1:

The closed-loop feedback system continuously monitors process parameters and adjusts them based on real-time feedback from sensors and machine learning analysis. This continuous feedback mechanism enables rapid response to process changes while maintaining stability through systematic, data-driven adjustments that consider the current process state and historical performance patterns.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The control system dynamically adapts process parameters based on real-time process conditions and defect detection. Rather than using fixed static control settings, the system continuously adjusts parameters to optimize performance while maintaining stability through adaptive control strategies that respond to changing process dynamics and material behavior.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP3635640B1Real-time adaptive control of additive manufacturing processes using machine learning
Publication Date: 2023.07.12 RELATIVITY SPACE INC
  • EP3635640B1 patent drawingFigure 1
  • EP3635640B1 patent drawingFigure 2
  • EP3635640B1 patent drawingFigure 3A~3C

AI summary

Disclosed herein are machine learning-based methods and systems for automated object defect classification and adaptive, real-time control of additive manufacturing and/or welding processes.