Trace Data CNN Modeling for Manufacturing Fault Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current technologies lack an efficient method to convert raw trace data from display panel manufacturing processes into a format suitable for machine learning models, limiting the ability to detect or predict faulty conditions in electronic devices.
Innovation Solution
The method involves receiving time-series sensor data from manufacturing processes, arranging it into a two-dimensional data array, and using a convolutional neural network model to identify patterns correlated with fault conditions, thereby providing a fault indicator.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If raw trace data is used directly as input to machine learning models, then the data collection process is simple, but the model cannot effectively detect or predict faulty conditions due to incompatible data format
Solution Approach 1:
The patent transforms one-dimensional time-series sensor data into two-dimensional data arrays by organizing sequential sensor readings into matrix structures with rows representing time steps and columns representing different sensor parameters. This dimensional transformation enables the application of convolutional neural networks (CNNs) which were originally designed for image data, thereby improving fault detection accuracy while providing a structured approach to data preprocessing
2Loss of information
If trace data is collected from multiple manufacturing processes, then the information completeness for fault detection improves, but the data processing complexity increases
Solution Approach 1:
The patent merges data from multiple manufacturing processes and sensor types into a unified two-dimensional array structure. By consolidating trace data from different sources into a standardized matrix format where each row represents a time step and each column represents a sensor parameter across different processes, the system achieves comprehensive information integration without proportionally increasing processing complexity
Solution Approach 2:
The patent creates a universal data preprocessing framework that can handle various types of sensor data from different manufacturing processes through a single standardized transformation pipeline. The same 2D array construction and CNN model architecture can process data from multiple process types, making the system multi-functional and reducing the need for process-specific processing logic
Data Source
AI summary
A method for detecting a fault includes: receiving a plurality of time-series sensor data obtained in one or more manufacturing processes of an electronic device; arranging the plurality of time-series sensor data in a two-dimensional (2D) data array; providing the 2D data array to a convolutional neural network model; identifying a pattern in the 2D data array that correlates to a fault condition using the convolutional neural network model; providing a fault indicator of the fault condition in the one or more manufacturing processes of the electronic device; and determining that the electronic device includes a fault based on the fault indicator. The 2D data array has a dimension of an input data to the convolutional neural network model.