Adaptive Sampling Control for Substrate Processing Time-Series Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data collection methods in substrate processing apparatuses fail to provide detailed data before an abnormality occurs due to inappropriate switching of sampling periods.
Innovation Solution
A data processing method that controls sampling periods to a normal period initially, and upon detection of abnormality, switches to a shorter abnormal period for more frequent data collection, using evaluation values, alarms, threshold values, and data variation as triggers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a fixed sampling period is used for data collection, then the data collection process is simple and stable, but detailed data cannot be obtained before an abnormality occurs
Solution Approach 1:
The sampling period is changed from a fixed value to a dynamic value that adjusts based on the operational state of the substrate processing apparatus. The control unit monitors operation status and automatically shortens the sampling period when abnormalities are detected, ensuring detailed data collection precisely when needed while maintaining simple fixed sampling during normal operation.
Solution Approach 2:
The sampling period parameter is modified based on the detected operational state. When the apparatus operates normally, a standard sampling period is used; when an abnormality is detected, the sampling period is shortened to capture more detailed temporal information about the abnormal event, thus improving measurement precision without permanently increasing system complexity.
2Loss of information
If the sampling period is shortened to capture detailed data, then detailed data before abnormality can be obtained, but data collection burden and processing load increase
Solution Approach 1:
The sampling period dynamically adapts to the operational state: a longer period during normal operation reduces data collection burden, while automatically shortening only when abnormalities are detected ensures detailed pre-abnormality data is captured without continuously increasing the data burden.
Solution Approach 2:
The system performs preliminary monitoring of operation status and proactively shortens the sampling period before the abnormality fully manifests, capturing detailed data in advance. This preliminary action prevents information loss while avoiding continuous high-frequency sampling that would increase energy burden.
3Measurement precision
If the sampling period is switched based on operation status, then detailed data can be obtained, but inappropriate switching timing reduces effectiveness
Solution Approach 1:
The control unit continuously monitors operation status and uses this feedback to determine the appropriate sampling period. This real-time feedback mechanism ensures the sampling period is switched at the correct timing based on actual apparatus state, capturing detailed data before abnormalities occur while avoiding premature or delayed switching.
Data Source
AI summary
A data processing method includes a sampling step of obtaining time series data based on a measurement result of a physical quantity in a substrate processing apparatus, an evaluation value calculation step of obtaining an evaluation value of the time series data by comparing the time series data with reference data, and a sampling period control step of controlling a sampling period used in the sampling step for each time series data. In the sampling period control step, all sampling periods are controlled to a normal period in an initial state, and when the evaluation value of the time series data is abnormal, the sampling period used when obtaining the time series data is controlled to an abnormal period shorter than the normal period.