Multi-Delay Identification of Rare Earth Extraction Process Based on Improved Time-Correlation Analysis.

Area of Science:

• Chemical Engineering
• Process Control
• Computational Modeling

Background:

• Rare earth extraction processes exhibit complex time-delay characteristics, hindering accurate mathematical modeling and process optimization.
• Identifying these time delays is crucial for improving the efficiency and predictability of rare earth element separation.

Purpose of the Study:

• To propose and validate an improved multi-delay identification method for rare earth extraction processes.
• To enhance the accuracy and stability of mathematical models used in rare earth component content prediction.

Main Methods:

• Data preprocessing using grey relational analysis to construct time-delay sequences and correlation matrices.
• Defining a time-correlation analysis matrix and using the H∞ norm to quantify data sequence correlation.
• Transforming the multi-delay identification into an integer optimization problem solved by an improved discrete state transition algorithm.

Main Results:

• The proposed method significantly improves optimization accuracy, convergence speed, and stability compared to existing methods.
• A Neodymium (Nd) component content model built with a wavelet neural network showed substantially enhanced performance after applying the time-delay identification.
• Validation confirmed the effectiveness of the multi-delay identification technique in the context of rare earth extraction.

Conclusions:

• The developed multi-delay identification method offers a robust solution for addressing time-delay challenges in rare earth extraction.
• Accurate time-delay identification leads to demonstrably improved performance in component content modeling and prediction.
• This approach provides a valuable tool for optimizing rare earth processing operations.