Additive Manufacturing Build Classification for Fleet Behavior Variation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing additive manufacturing technologies face challenges in diagnosing aborted or failed builds and identifying performance issues in additive manufacturing devices, requiring significant human labor and time, and lacking effective methods for real-time monitoring and adjustment.
Innovation Solution
The implementation of a system that monitors and analyzes additive manufacturing processes using machine learning and statistical methods to identify non-standard builds, determine contributing machine parameters, and provide real-time feedback for adjustments, thereby enhancing build reliability and repeatability across a fleet of machines.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual diagnosis and analysis of additive manufacturing builds is performed by experts, then diagnostic accuracy and root cause identification can be achieved, but significant time and human labor are required
Solution Approach 1:
The system enables automated self-diagnosis of additive manufacturing builds through machine learning models that independently analyze build data, identify failures, and determine root causes without requiring expert intervention. The automated evaluator subsystem continuously monitors builds and performs diagnostic functions that would otherwise require manual expert analysis.
Solution Approach 2:
The patent replaces the mechanical system of manual expert diagnosis with an automated computational system using machine learning algorithms. The automated learner and evaluator subsystems process build data, classify failures, and identify root causes through software-based analysis rather than human expert review.
2Reliability
If manual monitoring and diagnosis of additive manufacturing devices is performed, then performance issues can be identified, but the process is difficult and time-consuming
Solution Approach 1:
The system implements continuous feedback loops where build data is automatically collected, analyzed, and used to generate insights that feed back into the manufacturing process. The evaluator subsystem provides real-time feedback on build quality and machine performance, enabling proactive adjustments to maintain reliability.
Solution Approach 2:
The automated monitoring system performs multiple functions including data collection, build classification, failure diagnosis, root cause identification, and predictive maintenance planning within a single integrated platform. This multi-functional approach consolidates what would otherwise require multiple separate systems and manual processes.
3Manufacturing precision
If real-time monitoring and adjustment of additive manufacturing processes is implemented, then build quality and repeatability can be improved, but system complexity increases
Solution Approach 1:
The system performs preliminary analysis and classification of build data during the manufacturing process itself, rather than waiting for post-processing. The automated evaluator subsystem continuously assesses builds in real-time, enabling early detection of issues and immediate corrective actions to maintain build quality.
Solution Approach 2:
The patent introduces an intermediary automated evaluation system that bridges the gap between raw build data and actionable insights. This intermediary layer processes and interprets complex manufacturing data, translating it into meaningful metrics and recommendations that guide process adjustments without requiring direct complex interactions between all system components.
4Productivity
If automated machine learning systems are used to classify and evaluate builds, then human intervention is reduced and productivity increases, but the extent of automation must be carefully managed
Solution Approach 1:
The automated system is divided into distinct functional modules: an automated learner subsystem that trains models on historical data, an automated evaluator subsystem that performs real-time classification, and various specialized analyzers for different types of build data. This segmentation allows each component to be optimized independently and facilitates gradual implementation and validation of automation capabilities.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Apparatus and associated methods to classify and adjust builds across additive manufacturing machine(s) are disclosed. An example apparatus includes learner circuitry (620) to: process first data from a set of first builds to learn behavior; classify each build as a standard or non-standard build; model the learned behavior to form a standard reference behavior and a non-standard reference behavior, the standard reference behavior including first features and the non-standard reference behavior including second features; and output the standard reference behavior and the non-standard reference behavior to classify additional builds. The apparatus includes evaluator circuitry (630) to: ingest second data for a second build; process the second data in comparison to the standard reference behavior and the non-standard reference behavior; classify the second build as a standard build or a non-standard build; and, when the second build is classified as a non-standard build, output a corrective action.