Method for intelligently identifying underground safety accidents based on fully connected neural network

Abstract

Coal mine production is generally a multi-step and multi-step process of underground mining. The geology and mining conditions are complex, and there are many uneasy factors. They are often affected by gas, water and fire, carbon monoxide, ventilation, temperature, and roofs. Therefore, only by putting coal mine safety first in the work can keep the safety of underground workers and the normal progress of coal mine production work be ensured. Nowadays, predicting the cause of the alarm is the primary task of coal production. Moreover, many coal mining industries still remain at the manual recording stage for the cause of the alarm. Most of them rely on staff to manually record underground and then enter the system for storage. This makes it impossible to obtain timely feedback processing and statistics on hidden dangers. Moreover, underground work also includes a lot of sensor alarms caused by normal work, such as blasting, calibration and maintenance of circuits and other reasons, so it also caused a waste of time. Not only that, in the history of coal mining, it has been found that most of the casualties were caused by the “three violations”. In short, it was man-made. In order to avoid these errors and reduce the phenomenon of “three violations”, it is necessary to strengthen the protection of the mining personnel. Compliance inspections and one's own safety awareness can fundamentally reduce the occurrence of casualties.