Font Size: a A A

The Application Of Gas Emission Prediction Based On HIMMAS-WNN Algorithm

Posted on:2013-10-25Degree:MasterType:Thesis
Country:ChinaCandidate:H C KangFull Text:PDF
GTID:2248330395469275Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
The face gas emission not only related to the mine ventilation design specificationbut also impacted to the coal mine production safety, so enhanced the design research ofthe gas emission prediction method is very important to strengthen mine safety. Due tothe relationship of the gas emission factors exist dynamics and fuzzy non-linear, presenta complex nonlinear dynamic change processes, leading to the accuracy of predictivemodels is difficult to meet actual field requirements based on the traditional method oflinear prediction.First of all, the paper makes a simple introduction about the gas emission researchsituation at home and abroad and then summarized some prediction method simply,proposed a combination algorithm which used the improved ant colony algorithm andwavelet neural network to create the gas emission projections the model. Paper analyzesthe basic theory of wavelet and the classification of network and the structurecharacteristics, in view of the neural network low convergence precision and easy to fallinto the local extremum shortcomings, proposed the ant colony algorithm to improve thesituation. And then analyzes the basic operating mechanism of ant colony algorithm andthe various parameters on the effect of convergence, according to basic ant colonyalgorithm still easy to converge to a local optimal problem, the paper proposessmoothing the ant optimization path pheromone trajectory and using the disturbancefactor and punish factor to solve local convergence of the algorithm and slowconvergence speed problem based on MMAS, its application in solving the TSP problemand showing a good solution quality. Then combines with the HI-MMAS and waveletneural network, uses the intelligent control concept to establish gas emission forecastingmodel, and simulated by the coal mine monitoring historical data, the experimentalresults show that the model’s prediction have a higher convergence accuracy and morerobustness better than the basic ant colony algorithm and the traditional BP algorithm,so as to provide a good idea to solve the problem of coal mine gas emission forecasts.
Keywords/Search Tags:The amount of mine gas gushing, Wavelet neural network, Antcolony algorithm, MMAS
PDF Full Text Request
Related items