Research On Risk Identification And Early Warning Of Abnormal Gas Emission In Coal Mines Based On Biological Immunity

The dominant position of coal in China’s energy structure will not change for a long period of time in the future.However,in recent years,due to the significant increase in coal production capacity,shallow coal resources have been gradually depleted,and coal mining has had to extend to the deep,with increasing gas content and pressure,resulting in frequent occurrence of abnormal gas surges during coal mining,increasing disaster risks,and posing a great threat to coal mine safety in production.Abnormal gas emission is a disaster phenomenon caused by multiple factors such as geological factors and mining factors,which is characterized by uncertainty,concealment,and strong risk.It is one of the sources of gas accidents such as gas explosion and gas suffocation.Moreover,due to the complex geological environment conditions in coal mines,it is generally difficult to identify the risk of abnormal gas emission through human perception.Therefore,it is very important to explore efficient,dynamic,and accurate risk identification methods and early warning technologies in order to maintain a safe and stable state in the constantly changing complex environment.There are many factors that can induce and correlate abnormal gas emission,with complex structures,and interact with each other.In terms of risk identification and early warning,only through automatic risk identification,rapid response,and multi-system collaborative processing can system data form a regular,identifiable,memorable,and responsive information service function,which poses new challenges to the risk identification and early warning of abnormal gas emission in coal mines.In view of the limitations of traditional risk identification and early warning methods,and in combination with the new demand for dynamic identification and early warning of abnormal gas emission risk,this paper draws on the relevant theories of biological immunity to carry out research on risk identification and early warning of abnormal gas emission based on biological immune mechanisms.The main research contents and conclusions are as follows:（1）Abnormal gas emission is a complex chaotic phenomenon influenced by geological factors,mining factors,and mine environmental factors.During mining operations,gas mainly comes from coal falling,coal wall release,coal desorption from goaf,and gas influx from upper and lower adjacent layers,presenting characteristics of varying magnitude and temporal variability.Using the PSR model in environmental assessment for reference,the abnormal gas emission is distinguished from three aspects:pressure state response.The emission status mainly includes three forms:fast rising and fast falling,fast rising and slow falling,and periodic fluctuation.（2）Based on the analysis of the biological immune system,this paper analyzes the principles of immune recognition,immune early warning,specific immune response,and the purpose and characteristics of gas abnormal emission risk identification and early warning.Using bionics theory,the similarity between the biological immune system and gas abnormal emission risk identification and early warning is analyzed from the aspects of action principle,action object,action environment,function realization,information processing process,and mode,explore the similar fusion mechanism between the two.（3）From the categories of natural factors,mining factors,and gas characteristic factors,based on the construction basis and principles of the risk identification indicator system,using the PSR environmental evaluation system model for reference,using AHP analytic hierarchy process to determine the weight coefficients of specific characteristic indicators,a gas abnormal emission risk identification indicator system was established.Combining the characteristics of abnormal gas emission risk in coal mines and the need for risk identification,improvements and optimizations have been made in data coding,risk identification pattern set generation,matching rules,and other aspects.A gas abnormal emission risk identification algorithm based on improved optimized negative selection algorithm has been proposed,and a risk immune identification model has been established.（4）Based on the background of a coal mine in Gansu Province,a risk identification model is validated.According to the needs of risk identification of abnormal gas emission,a gas emission level pattern set Z,Y₁,Y₂,and Y₃is generated according to four levels:normal,general,relatively abnormal,and severe abnormal.The threshold value is set to 0.002,and the number of samples is 10000.Candidate detection sets J_C1,J_C2,and J_C3that can cover the entire risk identification space are uniformly generated in each set,detect and match the characteristic index index of gas emission in various underground areas,and output risk identification signals.The results are as follows:the risk of abnormal gas emission in January and February in the East 1100 fully mechanized top coal caving face,the West 2109 fully mechanized top coal caving face,the East 1104 return air fully mechanized top coal caving face,and the East 2+900m rail crosscut rock caving face are all general anomalies,consistent with the actual situation on site.The variation factors of the East 1100 fully mechanized top coal caving face were both serious anomalies in January and February,while the geological factors were relatively abnormal in January and in February.The pressure indicators in the project layer were relatively abnormal.The environmental status of the rock excavation face at the East 2+900m rail crosscut in February was normal,while the others were normal.These were consistent with the actual situation on site,verifying the effectiveness of the model.The model was applied to identify the risk of abnormal gas emission in the underground production area in June,and the results were all normal.（5）According to the selection principle of gas abnormal emission risk early warning indicators,the gas abnormal emission risk feature vector is determined,and combined with the divided abnormal degree grades,a gas abnormal emission risk early warning feature indicator system is constructed.A gas abnormal emission risk early warning model based on biological immunity is established.Taking the East 1100 fully mechanized caving face of a coal mine in Gansu Province as the background,through collecting gas concentration monitoring data to form samples,using XGBoost algorithm to determine the weights of volatility,variation coefficient,and relative change speed as 0.3,0.44,and 0.26,forming a gas change index.Three prediction models,namely RBF neural network,IOA-RBF neural network,and CIOA-RBF neural network,are used to predict gas concentration,and a comparison shows that the CIOA-RBF neural network has the best prediction effect for gas concentration.Based on the prediction concentration and gas change index indicators,a gas abnormal emission risk early warning feature vector is formed.Based on the constructed risk early warning model,the risk degree of gas abnormal emission in the future t=1500～2000 min period is determined to be Level II.（6）Based on the design objectives of the early warning system,the architecture of the gas abnormal emission risk early warning system was designed based on immune mechanisms,and the main functional interfaces were determined.A biological immune based gas abnormal emission risk early warning system was developed.By calling three models in the system,namely RBF neural network,IOA-RBF neural network,and CIOA-RBF neural network,the gas concentration is predicted,and the predicted gas concentration and abnormal emission risk warning level are displayed in real time.