Assembly Error Detection Using Sensor Data Reconstruction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing end-of-line (EOL) testing methods for assemblies in manufacturing are inefficient and require significant development effort to detect new production defects, especially in complex products, as they do not effectively utilize correlations between different measurement data, leading to low sensitivity in identifying faulty assemblies.
Innovation Solution
A method involving dimensional reduction of sensor data from assemblies, using techniques like PCA or auto-encoders, to reconstruct data and calculate a reconstruction error, which is then compared against a threshold to detect production errors, leveraging correlations between measurements to enhance detection sensitivity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple post-processing steps are used to detect different error modes, then detection coverage is improved, but development effort and complexity increase
Solution Approach 1:
The patent combines multiple post-processing steps into a single unified post-processing step that uses dimensional reduction to analyze correlations across all measurement data simultaneously. This merging approach maintains comprehensive error detection capability while eliminating the need to develop and maintain multiple separate detection algorithms for different error modes.
Solution Approach 2:
The unified post-processing step serves multiple functions: it detects various types of production errors, analyzes correlations between measurements, and provides comprehensive quality assessment. This universal approach replaces multiple specialized detection steps, reducing development effort while maintaining broad detection coverage.
2Reliability
If multiple post-processing steps are performed for different error modes, then detection coverage is improved, but processing time increases
Solution Approach 1:
Multiple sequential post-processing steps are merged into a single parallel processing operation. The dimensional reduction technique enables all measurement dimensions to be analyzed simultaneously in one post-processing step, eliminating the sequential execution overhead and reducing total processing time while maintaining comprehensive error detection.
3Ease of operation
If individual measurement limits are used for error detection, then simplicity is maintained, but detection sensitivity is reduced
Solution Approach 1:
The patent transitions from analyzing individual measurement dimensions separately to analyzing the correlation structure across all dimensions simultaneously. By performing dimensional reduction and examining relationships between measurements, the system achieves higher detection sensitivity while maintaining operational simplicity through a unified analysis approach.
4Ease of manufacture
If correlations between measurement data are not utilized, then computational simplicity is maintained, but detection sensitivity is reduced
Solution Approach 1:
The patent introduces a new analytical dimension by performing dimensional reduction to reveal correlation structures in the measurement data. This approach leverages relationships between different measurements to improve detection sensitivity while the automated nature of dimensional reduction keeps computational complexity manageable.
Data Source
AI summary
A method for detecting a production error of an assembly in a manufacturing facility includes (i) providing sensor data having at least two dimensions, wherein a respective dimension of the sensor data comprises measurement data with respect to the assembly, (ii) performing a dimensional reduction of the sensor data, wherein at least one feature is extracted based on the at least two dimensions of the sensor data, (iii) reconstructing the dimension-reduced sensor data based on the at least one extracted feature to provide reconstructed sensor data, (iv) determining a reconstruction error based on a comparison of the sensor data with the reconstructed sensor data, and (v) detecting the production error of the assembly based on the determined reconstruction error. Also disclosed is a computer program, a device, and a storage medium for this purpose.