RTP Chamber Control With Reduced Thermal Models From DMDc
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing model-based controllers for semiconductor processing tools, such as RTP tools, face challenges with complex detailed models that require excessive computing power and frequent modifications due to parameter variations and uncertainties, making them unsuitable for real-time control in multi-input-multi-output (MIMO) systems.
Innovation Solution
The use of dynamic mode decomposition with control (DMDc) to generate a reduced order model (ROM) from snapshots of temperature data in the processing tool, allowing for a simplified representation of the system that can be used for real-time control without the need for proprietary solvers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If detailed models are used for model based control, then control accuracy is improved, but computing power requirements increase excessively
Solution Approach 1:
The patent creates a reduced order model (ROM) that serves as a simplified copy of the detailed model. The ROM captures the essential dynamics of the system using fewer state variables and simpler equations, enabling real-time control calculations while maintaining sufficient accuracy. This copying approach allows the controller to use a tractable model representation that does not require excessive computing resources.
Solution Approach 2:
The patent transforms the detailed model into a reduced order model by changing the parameter representation. Instead of using the full set of parameters from the detailed model, the ROM uses a reduced set of parameters that capture the dominant system behavior. This parameter reduction enables real-time control while preserving the essential control characteristics of the original detailed model.
2Measurement precision
If detailed models are used for model based control, then control accuracy is improved, but model complexity increases
Solution Approach 1:
The patent creates a reduced order model (ROM) that serves as a simplified copy of the detailed model. The ROM captures the essential dynamics of the system using fewer state variables and simpler equations, enabling real-time control calculations while maintaining sufficient accuracy. This copying approach allows the controller to use a tractable model representation that does not require excessive computing resources.
3Measurement precision
If detailed models are used for model based control, then control accuracy is improved, but ease of operation deteriorates due to frequent modifications
Solution Approach 1:
The patent creates a reduced order model (ROM) that serves as a simplified copy of the detailed model. The ROM captures the essential dynamics of the system using fewer state variables and simpler equations, enabling real-time control calculations while maintaining sufficient accuracy. This copying approach allows the controller to use a tractable model representation that does not require excessive computing resources.
4Measurement precision
If reduced order models are generated from proprietary solvers, then control performance is improved, but accessibility deteriorates
Solution Approach 1:
The patent creates a reduced order model (ROM) that serves as a simplified copy of the detailed model. The ROM captures the essential dynamics of the system using fewer state variables and simpler equations, enabling real-time control calculations while maintaining sufficient accuracy. This copying approach allows the controller to use a tractable model representation that does not require excessive computing resources.
Data Source
AI summary
Embodiments disclosed herein include a method of developing a reduced order model (ROM) for a model based controller. In an embodiment, the method comprises obtaining a design of a plant, and building a detailed model of the thermal network of the plant from the design of the plant. In an embodiment, the method further comprises obtaining a training input recipe, and running the detailed model using the training input recipe. In an embodiment, the method further comprises generating a plurality of snapshots, wherein each snapshot includes the temperatures of a plurality of components in the detailed model, and utilizing a dynamic mode decomposition with control (DMDc) operation in order to extract the ROM from the plurality of snapshots.