Identification of Shearer Cutting Patterns Using Vibration Signals Based on a Least Squares Support Vector Machine...

Area of Science:

• Mining Engineering
• Machine Learning
• Signal Processing

Background:

• Accurate identification of coal shearer cutting patterns is crucial for advancing automation and safety in mechanized mining.
• Least Squares Support Vector Machine (LSSVM) shows promise for classification but requires effective parameter optimization.
• Existing meta-heuristic algorithms for LSSVM parameter tuning can be complex and slow to converge globally.

Purpose of the Study:

• To develop and validate an improved fly optimization algorithm (IFOA) for optimizing LSSVM parameters.
• To apply the proposed IFOA-LSSVM model for accurate identification of coal shearer cutting patterns.
• To compare the performance of the IFOA-LSSVM model against other LSSVM variants and demonstrate its practical applicability.

Main Methods:

• Utilized an improved fly optimization algorithm (IFOA) to optimize the parameters of the Least Squares Support Vector Machine (LSSVM).
• Extracted special state features from vibration acceleration signals using Ensemble Empirical Mode Decomposition (EEMD) and kernel functions.
• Collected vibration data for five distinct shearer cutting patterns and applied the IFOA-LSSVM model for classification.

Main Results:

• The proposed IFOA-LSSVM model achieved high accuracy in identifying shearer cutting patterns.
• Comparative analysis showed that IFOA-LSSVM outperformed standard LSSVM, PSO-LSSVM, GA-LSSVM, and FOA-LSSVM.
• The model's effectiveness was validated through an industrial application example on a coal mining face.

Conclusions:

• The IFOA-LSSVM approach provides a feasible and efficient method for shearer cutting pattern identification.
• This technique significantly enhances the automation level and safety in fully mechanized coal mining operations.
• The developed system offers a practical solution for real-world coal mining applications.