Coal and Rock Hardness Identification Based on EEMD and Multi-Scale Permutation Entropy

Area of Science:

• Geotechnical Engineering
• Artificial Intelligence in Mining
• Signal Processing

Background:

• Coal and rock hardness detection faces challenges with real-time performance and accuracy.
• Existing methods struggle to provide reliable data for efficient mining operations.

Purpose of the Study:

• To develop an efficient and accurate approach for identifying coal and rock hardness.
• To enhance the real-time performance of hardness detection systems for mining machinery.

Main Methods:

• Ensemble Empirical Mode Decomposition (EEMD) to decompose cutting motor current signals into Intrinsic Mode Functions (IMFs).
• Signal reconstruction using energy density and correlation coefficients to filter noise.
• Multi-scale Permutation Entropy (MPE) analysis of the reconstructed signal.
• Training an Adaboost-improved Back Propagation (BP) neural network for hardness recognition.

Main Results:

• Signal reconstruction effectively filters noise interference, improving signal quality.
• The Adaboost-BP model demonstrated a 0.0633 decrease in relative root-mean-square error compared to the standard BP model.
• The developed model achieved higher prediction accuracy for coal and rock hardness.

Conclusions:

• The proposed method offers an efficient and accurate solution for real-time coal and rock hardness identification.
• The speed control strategy based on identified hardness ensures efficient roadheader cutting.
• This approach has the potential to optimize mining operations and improve equipment performance.