Mine water inrush source discrimination model based on KPCA-ISSA-KELM

Area of Science:

• Geotechnical Engineering
• Artificial Intelligence in Mining
• Water Hazard Prevention

Background:

• Mine water inrush accidents pose significant risks to coal mine safety and production.
• Accurate identification of water inrush sources is crucial for effective prevention strategies.
• Existing models often lack the required accuracy and stability for real-world applications.

Purpose of the Study:

• To develop a highly accurate model for identifying mine water inrush sources.
• To improve the prevention of water inrush accidents in coal mines.
• To enhance the reliability and stability of water source identification models.

Main Methods:

• Utilized Kernel Principal Component Analysis (KPCA) for data dimensionality reduction.
• Developed an Improved Sparrow Search Algorithm (ISSA) incorporating Sine Chaotic Mapping, dynamic adaptive weights, and Cauchy Variation with Reverse Learning.
• Optimized Kernel Extreme Learning Machine (KELM) parameters using ISSA to create the KPCA-ISSA-KELM model for water source discrimination.

Main Results:

• The KPCA-ISSA-KELM model demonstrated superior accuracy in identifying mine water inrush sources compared to other models (KPCA-SSA-KELM, KPCA-KELM, ISSA-KELM, SSA-KELM, KELM).
• Accuracy improvements ranged from 4.17% to 25% over comparative models.
• In a practical application in Shanxi, the KPCA-ISSA-KELM model achieved the lowest misjudgment rate (4.76%).

Conclusions:

• The proposed KPCA-ISSA-KELM model offers a significant advancement in mine water inrush source identification.
• The model exhibits excellent universality, stability, and accuracy, making it suitable for practical mine safety applications.
• This research provides a robust method for preventing water inrush accidents, thereby ensuring safer coal mine production.