Plasma Process Diagnostics Using Waveform Segmentation and Adaptive Thresholds
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional diagnostic technologies for plasma processing apparatuses face challenges in accurately determining maintenance needs due to false reports and oversights, caused by offset changes in sensor waveform data and differences in processing histories among apparatuses, leading to incorrect deterioration sign detection and noise susceptibility.
Innovation Solution
A diagnostic device that separates sensor waveform data into predefined components, calculates deterioration degrees for each component, and dynamically sets thresholds based on filtered data to distinguish between noise and deterioration signs, reducing false reports and improving accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional diagnostic technology is used to monitor plasma processing apparatus, then maintenance can be performed periodically, but false reports and oversights occur frequently due to offset changes in sensor waveform data and part deterioration signs being buried in unrelated changes
Solution Approach 1:
The patent segments sensor waveform data into multiple components (offset component, trend component, fluctuation component) to separate part deterioration signs from unrelated offset changes. By analyzing each component separately, the system can accurately detect deterioration without false reports caused by normal operational variations.
Solution Approach 2:
The patent introduces an intermediary processing layer that decomposes sensor waveform data into distinct components before analysis. This intermediary step acts as a mediator between raw sensor data and diagnosis results, filtering out noise and unrelated changes while preserving genuine deterioration signs.
2Ease of operation
If a common threshold is used for deterioration diagnosis across multiple plasma processing apparatus, then the diagnostic system is simple to operate, but false reports occur due to differences in processing histories and sensor waveform data among apparatus
Solution Approach 1:
The patent implements dynamic threshold setting that adapts to each plasma processing apparatus based on its specific processing history and operational characteristics. Instead of using a static common threshold, the system learns and adjusts thresholds individually for each apparatus, maintaining simplicity of operation while improving diagnostic accuracy through adaptation.
Solution Approach 2:
The patent changes the threshold parameter from a fixed common value to dynamic apparatus-specific values. By adjusting the threshold parameter based on individual apparatus characteristics and processing histories, the system eliminates false reports while maintaining ease of operation through automated adaptation.
3Productivity
If deterioration diagnosis is performed for each plasma process with short process intervals, then continuous monitoring is achieved, but noise susceptibility increases making it difficult to determine whether deterioration signs are real or caused by noise
Solution Approach 1:
The patent segments the sensor waveform into distinct components (offset, trend, fluctuation) to separate genuine deterioration signals from noise. By analyzing specific components rather than raw data, the system can perform frequent monitoring while maintaining high signal-to-noise ratio and reducing false positives from random fluctuations.
Solution Approach 2:
The patent performs preliminary decomposition of sensor waveform data into components before conducting deterioration diagnosis. This preliminary action prepares the data by separating signal from noise, enabling accurate diagnosis even with frequent monitoring at short process intervals.
Data Source
AI summary
A diagnostic device separates sensor waveform data obtained in each plasma process into components of individuals of a plurality of predefined sensor waveform change types, calculates a deterioration degree indicating a deterioration state of the part for each separated sensor waveform component based on sensor waveform components at a normal time and a diagnosis time or sensor waveform components at a deterioration time and the diagnosis time, diagnoses necessity for the maintenance of a part using the deterioration degree, executes a filtering process on time-series data of the deterioration degree calculated for each plasma process, and sets a threshold used for deterioration diagnosis for each plasma processing apparatus based on a distribution calculated using a plurality of deterioration degrees after the filter processing is performed during a learning interval from time of component maintenance to time after the process is executed a predetermined number of times.