Research And Application Of Coal Mine Disaster Early Warning Based On Case-based Reasoning

The frequent occurrence of coal mine disasters seriously endangers people's lives and national public property.Early warning of coal mine disaster can provide rescue programs for coal mine disaster to reduce losses of individuals and countries.Aiming at the lack of research on early warning of various accidents in coal mine disaster by existing methods,this thesis proposes a coal mine disaster early warning method on case-based reasoning through analysis of early warning methods,combined with the characteristics and disaster-causing factors of coal mine disaster.Through the framework representation,the investigation report of coal mine disaster collected from the official website of the National Coal Mine Safety Supervision Bureau shall be expressed as the structured data required for case-based reasoning according to the accident category,and on this basis,a database of coal mine disaster cases shall be established;the attribute represented by the framework are divided into four categories,and corresponding attribute similarity calculation methods are given,due to the different impact of each attribute on the accident,the genetic algorithm is used to adjust and optimize the attribute weight value,combined with the current accident case category,and weight the similarity of each attribute according to the adjusted weight value,and use the nearest neighbor algorithm to calculate the global similarity between the current case and the historical case in the case database;select the case with high global similarity as the matching case of the current case and the early warning program of the matching case is extracted from the case database as the reference early warning program of the current case.If there is no historical case with high global similarity,the case needs to be modified by the experts of coal mine disasters or on ruled-based reasoning and saved.Based on the above theoretical methods,this thesis develops a coal mine disaster early warning system that can early warn coal mine disaster,designs the overall framework structure of the system,and realizes the functions of case representation,category retrieve and case learning.According to the classification of coal mine disaster,examples of application and analysis of flood and gas were implemented to realize the framework representation,attribute similarity calculation,attribute weight adjustment and global similarity calculation of flood and gas.It can verify the effectiveness of this method.Experimental results show that the early warning method in this thesis has good practicability.