Etch Indication Classification for Reproducible Defect Assessment

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing methods for classifying etch indications on components, particularly in high-tolerance applications, lack reproducibility and uniformity, leading to inconsistent and subjective assessments.

Innovation Solution

A method involving image capture, criterion detection, database comparison, and classification using neural networks to objectively categorize etch indications into defect and criticality classes, incorporating manufacturing history for enhanced accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If manual visual inspection is used to identify and interpret etch indications, then flexibility and adaptability are maintained, but reproducibility and uniformity of classification results deteriorate

Engineering Contradiction:

Improveflexibility in inspectionVSAvoidreproducibility of classification

Core Design Contradiction:

Adaptability or versatilityVSReliability

Solution Approach 1:

The patent replaces manual visual inspection with an automated image processing system that uses computer algorithms to detect, characterize, and classify etch indications. This substitution of mechanical/manual operations with automated computational methods eliminates subjectivity and ensures reproducible classification results while maintaining the ability to adapt to different component types and defect characteristics through software configuration.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Ease of operation

If manual inspection and interpretation of etch indications is performed, then subjective assessment can be applied, but objectivity and reliability of defect diagnosis deteriorate

Engineering Contradiction:

Improvesimplicity of inspection processVSAvoidobjectivity of defect assessment

Core Design Contradiction:

Ease of operationVSMeasurement precision

Solution Approach 1:

The system replaces subjective manual interpretation with objective image processing algorithms that automatically analyze etch indication characteristics. The automated classification based on quantitative image features eliminates human subjectivity while maintaining operational simplicity through user-friendly software interfaces that present clear classification results.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Reliability

If comprehensive analysis of etch indications is performed to ensure quality specifications, then reliability of defect detection is improved, but time and complexity of the inspection process increase

Engineering Contradiction:

Improveaccuracy of defect diagnosisVSAvoidinspection time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The system performs preliminary automated characterization of etch indications by extracting multiple features (area, shape, texture, etc.) and comparing them against stored reference data sets representing different defect classes. This preliminary automated analysis quickly identifies likely defect types, enabling rapid initial classification while maintaining high reliability through comprehensive feature analysis.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system uses stored image data sets representing known defect patterns as references for comparison. By copying and comparing against these reference patterns, the system rapidly classifies new etch indications without requiring time-consuming manual analysis, thus reducing inspection time while maintaining diagnostic accuracy.

Inventive Principle:

Principle #26Copying

Data Source

PatentUS12518371B2Method for classifying an etch indication of a component

Publication Date: 2026.01.06 MTU AERO ENGINES GMBH

AI summary

A method (100) for classifying an etch indication (11) of a component (10), the method including the steps of: providing a captured image (13) of the at least one etch indication (11); detecting at least one criterion (15) of the etch indication (11) based on the captured image (13); comparing the determined criterion (15) to at least one criteria data set (16) of etch indications (11) that is stored in a database; and classifying the etch indication (11) into at least one predetermined defect class (D) based on the comparison.