Research on the data validity of a coal mine solid backfill working face sensing system based on an improved...

Area of Science:

• Mining Engineering
• Data Science
• Internet of Things (IoT)

Background:

• Solid backfilling is crucial for coal mining safety and environmental compliance.
• Traditional methods suffer from limited data, isolated sensors, and hinder intelligent process development.
• Real-time monitoring and intelligent process development are currently limited.

Purpose of the Study:

• To propose a perception network framework for solid backfilling operations.
• To address challenges in data acquisition and validity within coal mine backfilling.
• To develop an intelligent system for enhanced monitoring and process control.

Main Methods:

• Analysis of critical perception objects in the backfilling process.
• Development of a coal mine backfilling Internet of Things (IoT) perception network and functional framework.
• Implementation of a transformer-based anomaly detection model for data validation.

Main Results:

• The proposed framework facilitates rapid concentration of key perception data into a unified data center.
• The transformer-based anomaly detection model achieved 90% accuracy in identifying data anomalies.
• The model demonstrated good generalization ability for monitoring increased perception objects.

Conclusions:

• The developed perception network framework enhances data collection for solid backfilling operations.
• The anomaly detection model effectively ensures the validity of perception data.
• The proposed system supports intelligent process development and real-time monitoring in coal mining backfilling.