Background Water disasters in mines are occasional accidents occurring due to the influence or triggered by mining-induced intense changes in geological bodies. The complexity, heterogeneity, and dynamics of geological bodies, coupled with the random and instantaneous nature of mining influence, lead to the uncertainty, instability, occasionality, and transience of water disasters in mines. Given that mine water monitoring and early warning serve as a prerequisite for advanced disaster prevention and control, their research and system construction hold great theoretical significance and practical value.

Progress and Prospects  Fighting against water disasters in mines is conducted throughout the development and utilization of coals. Starting from scratch, the monitoring and early warning of these disasters have undergone human experience-based identification, physical mechanism-guided information acquisition and identification, and intelligent monitoring and early warning driven by physics and data, with systematic transformation, industrial demonstration, and large-scale applications having been achieved presently. Significant advances have been made in relevant basic research, technological research and development, and system construction, enabling advanced early warning in some typical scenarios. These achievements underscore the belief that the prevention, control, and early warning of water disasters in mines can be achieved. However, the unclear physical mechanisms underlying the disasters lead to incomplete index systems, insufficient information acquisition, and inaccurate assessment and prediction. These issues tend to result in frequent occurrences of missed, false, and inaccurate early warnings. Consequently, the overall goal of advanced and precise early warnings has not been fulfilled for a vast majority of scenarios, facing serious challenges. Based on the review of the advances in research, this study proposed a system architecture for the monitoring and early warning of mine water, followed by theoretical discussions about four key technologies: the index system, information perception (monitoring), evaluation and prediction, and identification and early warning. Accordingly, this study summarized the connotations and interrelationships of these technologies, as well as relevant challenges. Furthermore, it pointed out that the monitoring and early warning should shift from methodologies based on physical mechanisms to the physics-data dual-driving mechanism. This overall development direction includes seven specific directions: constructing a comprehensive index system, optimizing the layout of the perception system, total-factor joint monitoring of multiple disasters, establishing geological-hydrological models + deep learning-based prediction models, setting early warning rules and factor thresholds, real-time and advance warning, and image monitoring and big data processing. These directions will lay the foundation for theoretical research, technological research and development, and system construction in this field.