Keeping coal miners safe is a top priority, and understanding the water richness in coal seam roofs is crucial for preventing dangerous water hazards. This study employed advanced statistical methods, including Spearman correlation and the GeoDetector technique, to identify the key factors influencing water richness in coal seam roofs. The researchers then developed water richness evaluation models using entropy weight, coefficient of variation, and random forest methods. By validating the models through pumping tests, workface water inflow tests, and three-dimensional electrical surveys, the team demonstrated the effectiveness of their approach in accurately predicting water richness and guiding mining operations. This research represents a significant step forward in improving safety and reducing the risk of water-related accidents in coal mines.
Identifying the Key Factors
The abundance of groundwater in coal seam roofs can have a significant impact on mining operations, with water inrush posing a serious threat to human lives and property. To better understand and predict the water richness in coal seam roofs, researchers employed statistical techniques to analyze various lithological and structural factors.
Spearman correlation analysis was used to identify the factors that have the strongest correlation with water inflow. Six key factors were identified: the ratio of sandstone to mudstone (Rsm), the lithologic influence index (Lid), the equivalent ratio of sandstone to mudstone (ERsm), the equivalent thickness of sandstone (Meh), the aquifer thickness (Ts), and the lithology coefficient of sandstone (P).
Exploring the Interactions Between Factors
While the Spearman correlation provided insights into the individual factors, the researchers recognized the need to also understand the interactions between these factors. They turned to the GeoDetector method, which can reveal the spatial heterogeneity and the underlying driving forces behind water richness.
The GeoDetector analysis identified three combination factors that had a particularly strong influence on water richness: Meh ERsm, Rsm DF, and DS D. These combinations capture the interplay between the water supply capacity of the aquifer, the degree of tectonic activity, and the interaction between lithology and structure.
Evaluating Water Richness through Different Models
The researchers then used the selected factors to develop water richness evaluation models based on the entropy weight method, coefficient of variation method, and random forest method. The models were validated through pumping tests, workface water inflow tests, and three-dimensional high-density electrical surveys.
The results showed that the random forest method, which can effectively handle high-dimensional data and capture the complex relationships between factors, outperformed the other two methods in accurately predicting water richness. The random forest model was particularly adept at identifying areas with high water richness, even in regions where the entropy weight and coefficient of variation methods struggled.
Improving Safety in Coal Mining
This study represents a significant advancement in the field of water richness evaluation for coal mines. By incorporating advanced statistical techniques, the researchers were able to identify the critical factors and their interactions that influence water richness. The validated models can now be used to guide mining operations and enhance safety, helping to prevent devastating water-related accidents and protect the lives of coal miners.
The success of this research highlights the importance of combining scientific knowledge with cutting-edge data analysis methods. By leveraging the power of techniques like Spearman correlation, GeoDetector, and random forest, researchers can gain a deeper understanding of complex environmental and geological systems, ultimately leading to safer and more efficient mining practices.
Author credit: This article is based on research by Guichao Gai, Mei Qiu, Weiqiang Zhang, Longqing Shi.
For More Related Articles Click Here
