Adaptive Virtual Metrology for Sparse Manufacturing Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing manufacturing processes face challenges in optimizing process control and monitoring due to limited data availability, data drifts, shifts, and sparsity, which affect the accuracy and scalability of virtual metrology (VM), image metrology (IM), and root cause analysis (RCA) systems.
Innovation Solution
The system employs a data and machine learning layer that includes a data collector, dataset generator, model management module, and inference module to optimize process control and monitoring. It utilizes an aggregated adaptive online model (AggAOM) that captures commonalities across equipment and hierarchical structures to improve prediction accuracy, especially in high-volume manufacturing environments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If virtual metrology (VM) is implemented using predictive modeling, then productivity is improved and quality is enhanced, but accuracy deteriorates due to data drifts, shifts, and sparsity
Solution Approach 1:
The system implements dynamic model adaptation by continuously updating predictive models with incoming process data. The model management module enables automated retraining and adjustment of VM models in response to changing process conditions, allowing the system to adapt to data drifts and shifts while maintaining prediction accuracy in high-volume manufacturing environments.
Solution Approach 2:
The system incorporates feedback mechanisms where prediction results and actual measurements are continuously compared. This feedback loop enables the system to identify accuracy degradation due to data drifts and shifts, triggering automated model updates or alerts for manual intervention, thereby maintaining measurement precision while sustaining high productivity.
2Measurement precision
If data collection is expanded to improve prediction accuracy, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The system employs a unified data collection framework that serves multiple functions: data acquisition for prediction, data storage for historical analysis, and data validation for quality control. This multi-functional approach consolidates what would otherwise require separate systems, improving measurement precision without proportionally increasing device complexity.
Solution Approach 2:
The system introduces intermediary components such as data preprocessing layers and feature engineering modules that bridge raw data collection and prediction algorithms. These intermediaries organize and structure incoming data, making it more suitable for prediction while preventing the complexity from propagating through the entire system.
3Device complexity
If sparse measurement data is used, then device complexity is reduced, but manufacturing precision deteriorates
Solution Approach 1:
The system performs preliminary data processing and feature extraction before feeding data into prediction models. By pre-processing sparse measurement data to extract relevant features and patterns, the system maintains manufacturing precision without requiring complex real-time processing, thus keeping device complexity manageable.
Solution Approach 2:
The system transforms sparse measurement data by changing parameters such as aggregation time windows, feature selection criteria, and prediction intervals. These parameter adjustments allow the system to work effectively with limited data while maintaining sufficient manufacturing precision for process control applications.
Data Source
AI summary
Described are systems and methods for optimizing process control or monitoring of manufacturing processes in manufacturing environments. Systems and methods can generate predictions or recommendations for process variables, target properties, or root causes of anomalies. Systems can include data and machine learning layers that can include: a data collector configured to receive data from the client application layer; a dataset generator configured to enable a user to create customized datasets from the data; a model management module configured to enable the user to build, train and/or update machine learning models; and an inference module configured to use the machine learning models for generating predictions or recommendations. Machine learning models can include aggregated adaptive online models (AggAOM) for generating predictions with scarce or sparse data.