Debug Method Identifying Excursive Machines via Data Continuity Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for identifying machine failures in manufacturing processes, such as those in semiconductor production, often rely on sampling and simulation under static conditions, which can lead to inaccurate representations of real-world environments, resulting in delayed detection of malfunctioning machines and increased production of defective products due to subjective judgments and additional costs from dummy material testing.
Innovation Solution
A process debug method and system that collects validity identification data from end products, clusters machines based on their functions, conducts continuity analysis, and ranks machines to identify excursive machines causing invalid products, eliminating the need for sampling programs and dummy material testing, thereby allowing for timely identification and prevention of failures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If sampling analysis is conducted under static conditions, then the manufacturing process can be tested, but the testing results may not represent the real manufacturing environment and excursive machines cannot be found in time
Solution Approach 1:
The patent transitions from static sampling analysis to dynamic real-time monitoring. The system continuously collects validity identification data from end products and dynamically updates the status of each machine, allowing immediate detection of excursive machines as failures occur rather than waiting for batch sampling results.
Solution Approach 2:
The patent implements a feedback mechanism where validity identification data from end products is continuously fed back to update machine status. The sorting module and continuity analysis module process this feedback to identify patterns and detect excursive machines in real-time, creating a closed-loop system that improves detection accuracy and speed.
2Measurement precision
If dummy material testing is conducted to simulate manufacturing processes, then machine failures can be identified, but additional costs are incurred and subjective judgments may lead to incorrect identification
Solution Approach 1:
Instead of using physical dummy materials to simulate products, the patent creates a data copy of the manufacturing process by collecting validity identification data from actual end products. This digital replication eliminates the need for physical dummy materials while maintaining accurate representation of the manufacturing process.
Solution Approach 2:
The system uses the end products themselves to provide diagnostic information about machine status. The validity identification data embedded in or associated with end products serves the dual purpose of product validation and machine health monitoring, eliminating the need for separate testing resources.
3Productivity
If small sample testing is used to represent the whole manufacturing environment, then testing costs are reduced, but the testing results may be inaccurate due to imprecise sampling programs or inappropriate tolerance levels
Solution Approach 1:
The patent implements continuous data collection from end products throughout the manufacturing process rather than intermittent sampling. This continuous monitoring ensures that all production data is captured and analyzed, providing comprehensive and accurate representation of the manufacturing environment while maintaining high productivity.
Data Source
AI summary
A process debug method used to identify at least one excursive machine in a manufacturing process comprising the following steps: First, a series of validity identification data is collected, and the serial validity identification data is associated with its pathway to obtain a plurality of validity identification data sequences in corresponding to the machines. Subsequently, a sorting process is conducted to cluster the validity identification data sequence into several groups, and the clustered groups are ranked into a first order. The validity identification data sequences are subjected a continuity analysis to determine the continuity of the defects occurring in a particular machine. And the continuities of the machines involved in a particular group are ranked into a second order. Accordingly, the excursive machines causing the defective end products in the manufacturing process can be identified by the way of joining the second orders according to the first order.