Substrate Processing Data Modeling for Temperature Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies fail to effectively utilize data from substrate processing apparatuses for efficient temperature control and abnormality detection, leading to inefficiencies and potential quality issues in semiconductor manufacturing.
Innovation Solution
An information processing method utilizing dynamic mode decomposition to analyze time series data from substrate processing apparatuses, incorporating control input data, to predict future temperature conditions and detect abnormalities, enabling precise temperature control and real-time monitoring.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If traditional data analysis methods are used for substrate processing apparatus, then data utilization is insufficient, but implementing advanced methods like dynamic mode decomposition increases system complexity
Solution Approach 1:
The patent introduces dynamic mode decomposition as an intermediary computational method that bridges raw time series data and actionable insights. This mathematical technique decomposes complex temperature data into interpretable modes, enabling effective data utilization without requiring fundamental changes to the substrate processing apparatus itself.
Solution Approach 2:
The patent replaces traditional mechanical or heuristic data analysis approaches with a computational mathematical model. By substituting conventional analysis methods with dynamic mode decomposition algorithms, the system achieves superior data utilization while maintaining reasonable complexity through software-based solutions rather than hardware modifications.
2Manufacturing precision
If real-time temperature prediction is implemented using dynamic mode decomposition, then temperature control accuracy improves, but processing time and computational load increase
Solution Approach 1:
The patent performs preliminary decomposition of temperature time series data into dynamic modes during periods when computation can be accommodated. By pre-processing the data and establishing the modal structure in advance, the system reduces real-time computational requirements while maintaining high temperature control accuracy through the use of pre-computed decomposition modes.
Solution Approach 2:
The patent employs dynamic mode decomposition which adapts to changing temperature patterns in real-time. The method dynamically identifies and tracks evolving thermal modes, allowing the system to maintain high prediction accuracy even as processing conditions change, while the computational framework is designed to balance accuracy requirements with processing time constraints.
Data Source
AI summary
Provided are an information processing method, a computer program, and an information processing apparatus that can be expected to effectively utilize data obtained from a substrate processing apparatus. An information processing method according to the present embodiment is executed by an information processing apparatus and includes acquiring time series observed data in which a state of a substrate processing apparatus is observed, and calculating parameters of a model that predicts, based on first observation information regarding the observed data at a first point in time, second observation information regarding the observed data at a second point in time after the first point in time, by dynamic mode decomposition based on the acquired time series observed data.