Font Size: a A A

Coal Mine Methane Monitoring Method Based On Improved D-S Evidence Fusion

Posted on:2019-07-22Degree:MasterType:Thesis
Country:ChinaCandidate:T LiFull Text:PDF
GTID:2371330572959788Subject:Control engineering
Abstract/Summary:PDF Full Text Request
Coal is an important primary energy source in China.Due to the poor working conditions in underground working face of coal mines,changes in gas concentration in coal mining operations have led to frequent occurrence of gas outbursts and gas explosions.In order to stabilize economic growth and ensure the safety of coal mining working environment,a complete coal mine methane monitoring system was introduced to monitor downhole gas conditions.How to accurately and dynamically perform risk warning for coal mine gas disasters based on the analysis and monitoring of monitoring information has always been a matter of concern.In view of the nonlinear,time-varying and multi-parameter strong coupling problems of coal mine gas monitoring systems,multi-sensor information fusion technology can effectively analyze the data information collected by various sensors in mines.The data collected by the coal mine downhole sensor in a poor working environment will be mutated due to external factors,can not truly reflect the underground safety conditions,using the support vector machine for denoising and classification processing,the application of nuclear independent principal component analysis algorithm to reduce the data sample The feature extraction of dimensionality,the preprocessed data input weighted average DS fusion model for decision fusion,and the evolutionary differential bat algorithm to optimize the weights assigned to the evidence,effectively solves the fusion problem of conflict evidence.Accelerated the convergence of evidence fusion.Then,according to the fusion results,the hazard rating of gas disasters can be forecasted to reduce the casualties and economic losses caused by the accident.The data samples of Xinzhi coal mine were selected for experimental verification and simulation analysis.The fusion results showed that the improved weighted average D-S evidence fusion improved the “one-vote veto” phenomenon of evidence fusion and increased the accuracy of the fusion result to the final decision.The improved method improves the quickness and accuracy of monitoring of coal mine gas,and realizes the purpose of warning information of gas disasters.
Keywords/Search Tags:gas monitoring, multi-sensor information fusion, nuclear independent principal component analysis, D-S evidence fusion, differential evolutionary bat algorithm
PDF Full Text Request
Related items