Manufactured Part Failure Diagnosis via Data Correlation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Manufacturing inspection systems can flag parts as out-of-specification but fail to diagnose the underlying cause of deviations in physical properties, lacking the capability to identify parts with similar failure mechanisms.
Innovation Solution
A system and method that normalizes test data against historical means and standard deviations, correlates it with other parts to determine similarity, and displays correlation values to identify parts with the same failure mechanism, enabling diagnosis of manufacturing errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If inspection systems measure and store physical property values of manufactured parts, then measurement capability is improved, but the ability to diagnose failure causes remains insufficient
Solution Approach 1:
The system implements feedback by comparing measured physical property values against historical data and specifications, then providing diagnostic information back to operators about potential failure mechanisms. This closed-loop approach transforms raw measurement data into actionable diagnostic insights, resolving the contradiction between measurement capability and diagnostic information.
Solution Approach 2:
The patent introduces an intermediary diagnostic layer that sits between measurement systems and failure analysis. This intermediary component processes measurement data, compares it with historical patterns, and generates diagnostic information about failure mechanisms, thereby bridging the gap between physical measurements and cause diagnosis.
2Quantity of substance
If the system stores test data for multiple manufactured parts, then data availability is improved, but the complexity of analyzing similarities between parts increases
Solution Approach 1:
The system transforms raw test data into standardized parameters by normalizing physical property values against historical means and deviations. This parameter transformation enables efficient comparison across multiple parts without requiring complex analysis of raw data, resolving the contradiction between data quantity and analysis complexity.
Solution Approach 2:
The patent segments the analysis process into distinct stages: data collection, normalization against historical parameters, correlation calculation, and diagnostic interpretation. This segmentation breaks down the complex task of comparing multiple parts into manageable steps, reducing overall system complexity while handling large volumes of test data.
3Reliability
If the system correlates normalized test data to identify similar parts, then diagnostic accuracy is improved, but computational processing requirements increase
Solution Approach 1:
The system implements partial action by calculating correlations only for parts that meet specific criteria or are suspected of having similar failures. Rather than computing all possible pairwise correlations in the dataset, the system selectively processes relevant subsets, thereby maintaining diagnostic accuracy while reducing computational power requirements.
Data Source
AI summary
A system and method can identify manufactured parts. A user can select a particular manufactured part, which can be out-of-specification. The system can retrieve test data for the selected part and for other manufactured parts. The system can normalize the retrieved test data against historical means and historical standard deviations to form normalized test data. The system can correlate the normalized test data for the selected part against normalized test data for each of the other manufactured parts to form correlation values. The system can display the correlation values with identifiers corresponding to the manufactured parts. Each correlation value can represent a degree of similarity between the selected part and a respective manufactured part. The manufactured parts with the highest correlation values can have the same failure mechanism as the selected part, which can help diagnose why the selected part can be out-of-specification.