Optimization of monitoring and early warning technology for mine water disasters using microservices and long short-term memory algorithm

Optimization of monitoring and early warning technology for mine water disasters using microservices and long short-term memory algorithm

연구 분야: Software Development

학회: The Journal of Supercomputing

This study seeks to improve mine safety management through the design and implementation of a comprehensive monitoring and warning system. The system architecture is organized into four distinct layers: data collection, early warning, data processing, and user interface, utilizing a microservice architecture to ensure modularity, scalability, and high availability. During the data preprocessing phase, sensor data undergoes various operations, including missing value imputation, outlier detection, normalization, and feature engineering. A hybrid model integrating Long Short-Term Memory, Gated Recurrent Unit, and Transformer architectures is developed, incorporating a dynamic real-time adjustment mechanism. The model is trained and validated using a dedicated training dataset. Experimental evaluations were conducted following data preprocessing and model training, with performance assessed on a separate test set. The results revealed an average response time ranging from 120 to 200 ms, with the system handling between 200 and 400 requests per second. The hybrid model required two hours for training and 1.3 s per prediction, achieving optimal predictive performance with a Mean Absolute Error of 0.35 and a Root Mean Squared Error of 0.55. The Early Warning System demonstrated strong effectiveness, achieving 92% accuracy, 87% recall, and an F1 score of 0.89 on the test set. This study introduces a high-efficiency, reliable mine water hazard monitoring and warning system, offering a novel solution to mine safety management. By leveraging sensor data and deep learning algorithms, the system enables timely detection of mine water risks, contributing to a safer operational environment for both miners and equipment.