Meikuang Anquan (Aug 2025)
Application of time series database technology in coal mine safety monitoring system
Abstract
In the informationization construction of coal mine safety production, the real-time collection, storage, and analysis of massive time-series data is a key technical bottleneck that restricts the efficiency improvement of safety monitoring systems. Traditional relational databases are unable to meet the high concurrency and high stability monitoring requirements of complex underground environments due to issues such as insufficient write throughput, low storage efficiency, and real-time query latency. Aiming at the time series characteristics, high-frequency writing, and long-term storage challenges of multi-source heterogeneous sensor data streams in coal mines, a coal mine safety monitoring data management solution based on time-series databases is proposed. By analyzing the temporal characteristics of underground environmental parameters (gas concentration, temperature and humidity, wind speed), equipment status (fan speed, power supply current, pressure), and safety devices (gas sensors, power-off devices), a specialized temporal data model for coal mine scenarios is constructed. A timeline partitioning storage mechanism is designed to separate equipment tag metadata from time series data, reducing redundancy by up to 40%. A dynamic time sharding storage strategy is proposed for 2 Hz high-frequency data streams, combined with an improved run length encoding compression algorithm (RLE-X), to achieve a stable write throughput of ≥ 1 000 data streams per second and a storage space compression rate of over 85%. At the level of query optimization, establish a hierarchical index structure based on timestamp range, supporting millisecond level real-time data retrieval (average response time ≤ 50 ms) and multi-dimensional historical data backtracking analysis (span query efficiency increased by 3 times). The system integrates real-time anomaly detection and trend prediction modules, dynamically identifying gas concentration mutation events through a sliding window mechanism, with a warning accuracy rate of 92.6%. Actual deployment has shown that this solution can support efficient management of daily 2 GB level data, with a 70% increase in historical data query efficiency compared to traditional solutions. It provides high reliability technical support for building a preventive coal mine safety monitoring system and effectively reduces the risk of underground safety accidents.
Keywords
