Alarm Frequency Visualization for Substrate Processing Failure Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Substrate processing apparatuses are complex, making it difficult to accurately specify and maintain failure occurring areas due to correlated and chained failures, resulting in numerous alarms and requiring skilled maintenance engineers.
Innovation Solution
A failure monitoring system that collects and analyzes alarms, displaying alarm issuing frequency on a two-dimensional space to visually identify patterns and specify failure areas, enabling precise maintenance by correlating alarm data with maintenance records.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If substrate processing apparatus uses complex configuration with multiple processing areas, then processing capability is improved, but failure detection and specification becomes difficult
Solution Approach 1:
The monitoring system segments the complex apparatus into individual monitoring units, each responsible for specific processing areas or components. Each unit collects and manages alarms from its designated area, allowing failures to be localized to specific segments rather than the entire complex system, thus resolving the difficulty of detecting and specifying failures in multi-area processing apparatus.
Solution Approach 2:
The system introduces a temporal dimension by monitoring alarm issuance over time periods and displaying patterns across multiple dimensions (spatial distribution of alarm sources, temporal frequency, and correlation patterns). This multi-dimensional analysis transforms the complex failure detection problem into visual patterns that reveal underlying failure modes and correlations, making it easier to specify failure areas in complex apparatus.
2Productivity
If substrate processing apparatus performs batch processing of multiple wafers, then productivity is improved, but failure occurrence increases due to multiple processing areas
Solution Approach 1:
The monitoring system establishes continuous feedback loops that collect alarm information from all processing areas during batch processing, analyze patterns in real-time, and provide feedback to operators about system status and potential failures. This feedback mechanism enables early detection of failures in any processing area, allowing corrective actions before failures propagate through the batch processing sequence, thus maintaining reliability while preserving productivity.
Solution Approach 2:
The system performs preliminary analysis of alarm patterns and correlations before failures become critical. By monitoring and analyzing alarm issuance patterns during batch processing, the system can predict potential failures and alert operators to take preliminary corrective actions, preventing minor issues from developing into major failures that would halt productivity.
3Device complexity
If traditional alarm monitoring is used without pattern analysis, then system simplicity is maintained, but maintenance efficiency decreases
Solution Approach 1:
The monitoring system creates a virtual copy or representation of the physical apparatus's alarm behavior through pattern analysis and visualization. Instead of directly managing complex physical diagnostics, the system copies alarm patterns into visual displays that represent failure modes and correlations, allowing maintenance personnel to efficiently identify and address issues through pattern recognition rather than complex diagnostic procedures.
Data Source
AI summary
Disclosed is a failure monitoring system for monitoring a failure of a substrate processing apparatus that performs a predetermined processing on a substrate to be processed, the failure monitoring system including: an alarm collecting unit configured to collects alarms issued from the substrate processing apparatus; and an analyzing unit configured to analyze the alarms collected by the alarm collecting unit and display, as an image, an alarm issuing frequency in each monitoring period on a two-dimensional space, of which one axis represents an alarm ID that specifies an alarm issuing area and another axis represents a predetermined monitor period.