Batch Process Monitoring Based on Quality-Related Time-Batch 2D Evolution Information.

Area of Science:

• Chemical Engineering
• Process Monitoring
• Data Analysis

Background:

• Batch processes require robust quality monitoring systems.
• Traditional methods often struggle with complex, time-varying process dynamics.
• Analyzing both temporal and batch evolution is crucial for accurate process understanding.

Purpose of the Study:

• To develop an online monitoring strategy for batch processes.
• To leverage two-dimensional time and batch evolution information for enhanced quality control.
• To address limitations of single-dimension analysis in traditional methods.

Main Methods:

• Utilizing Partial Least Squares (PLS) model regression coefficients to divide batches into phases and parts based on temporal evolution.
• Employing sliding windows on the score matrix (T) for mode division, considering slow time-varying characteristics.
• Integrating phase and part division with batch evolution analysis for a comprehensive model.

Main Results:

• The proposed strategy effectively combines time and batch two-dimensional evolution information.
• Demonstrated superior monitoring results compared to traditional single-dimension analysis methods.
• Successfully applied to injection molding processes, validating its effectiveness.

Conclusions:

• The developed online monitoring strategy offers improved accuracy and robustness for batch processes.
• Integrating multi-dimensional evolution data provides a more comprehensive understanding of process dynamics.
• This approach enhances quality-related monitoring by overcoming the limitations of analyzing data in isolation.