Manufactured Part Assessment Using K-Fold Functional Tolerance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for assessing the quality of manufactured gas turbine engine components, such as integrally bladed rotors, rely solely on geometric tolerances, which often incorrectly identify parts as unacceptable when they are actually functionally tolerant.
Innovation Solution
A method combining physical testing and machine learning, using K-fold validation, to evaluate component quality by comparing generated data from multiple locations against training data, ensuring functional tolerance beyond nominal dimensions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual inspection of dimensions is used to assess component quality, then measurement precision is maintained, but productivity is reduced and false rejections increase
Solution Approach 1:
The patent replaces manual mechanical inspection with an automated system that uses optical scanning or measurement devices to capture dimensional data, followed by machine learning algorithms for assessment. This substitution maintains measurement precision while dramatically increasing productivity by eliminating manual inspection bottlenecks.
Solution Approach 2:
The system enables self-assessment of component quality through automated data capture and machine learning evaluation. The component's dimensional data is automatically collected and evaluated against functional tolerance criteria without requiring manual inspection, thereby improving both productivity and consistency.
2Manufacturing precision
If geometric tolerances are used to assess component quality, then manufacturing precision is enforced, but reliability is reduced due to false rejections of functionally tolerant components
Solution Approach 1:
The patent transitions from using fixed geometric tolerance parameters to functional tolerance parameters that are derived through machine learning. The system learns the actual functional impact of dimensional variations from training data, allowing components to be assessed based on their true functional tolerance rather than conservative geometric limits. This resolves the contradiction by maintaining manufacturing precision requirements while improving assessment reliability.
Solution Approach 2:
The system uses feedback from training data consisting of previously assessed components to continuously improve its evaluation criteria. By learning from historical data about which components actually performed functionally, the system refines its tolerance assessment to distinguish between geometric deviations that truly affect function and those that do not, thereby reducing false rejections while maintaining precision standards.
3Reliability
If K-fold validation with machine learning is implemented, then reliability of quality assessment is improved, but device complexity increases
Solution Approach 1:
The patent segments the quality assessment process into distinct modules: data capture, data processing, machine learning model application, and result evaluation. The K-fold validation is implemented as a structured approach within the machine learning branch, dividing training data into folds for systematic evaluation. This segmentation manages complexity by organizing the system into manageable, independent functional blocks.
Solution Approach 2:
The machine learning model serves multiple functions: it learns from training data, validates using K-fold cross-validation, assesses new components, and continuously improves through feedback. This multi-functionality justifies the increased device complexity by providing comprehensive quality assessment capabilities that significantly improve reliability beyond simple geometric tolerance checking.
Data Source
AI summary
A method of assessing the quality of a manufactured component includes the steps of manufacturing a component and generating data at each of a plurality of locations on the component. The generated data is passed to a machine learning branch wherein the generated data is compared to training data at each of the plurality of locations and across a plurality of folds using K-fold validation to determine whether the component is of a functionally tolerant dimension at each of the plurality of sections. A system is also disclosed.