Coal Spontaneous Combustion Early Warning Methods Based on Slope Grey Relation Analysis

Abstract

As the depth of coal mining increases, concealed fires from residual-coal spontaneous combustion in goaf pose a significant threat to underground mining safety. Preferred index gases are used to predict temperature of coal spontaneous combustion (CSC), providing ideas for an early warning system for concealed fires. Here, a new mathematical method of slope grey relation analysis (SGRA) is established and proved to be reasonable, the index gases obtained from experiments are calculated and screened according to the relation degree, and the coal temperature is predicted according to the screened index gases concentration and prediction model. The conclusions are as follows: The coal oxidation process is divided into a slow oxidation stage and a rapid oxidation stage according to the speed of oxygen consumption and gases generation, and the rapid oxidation stage approximates an exponential growth, and the trend of gases ratio changes shows an exponential growth in localized stages. Compared with index gases screened by other types of grey relation analysis, the index gases screened by SGRA accurately reflect the coal temperature, and the magnitude of the relation degree reflects the prediction accuracy. Although the SGRA has computational errors, when the relation degree of the screened index gases is greater than 0.93 in the slow oxidation stage and greater than 0.95 in the rapid oxidation stage, the prediction results can satisfy engineering applications, and the method is considered reliable. Based on SGRA and CSC prediction model, combined with artificial neural network learning, an early warning system for CSC is proposed, which is expected to accurately forecast the temperature of CSC and guarantee the safety of mine production.