Development And Application Of Coal Mine Gas Emission Volume Real-time Forecasting System

China is the world’s abundant coal resources of the country, but because of theparticularity and complexity of the coal geological conditions, combined coal miningtechnology and management techniques are lagging behind more developed countries,making China the world’s coal mine accident-prone one of the countries where gasdisaster is one culprit leading to mine accidents. Therefore, to understand the impact ofgas emission factors, and effective gas emission monitoring and forecasting is aneffective measure to guide the management of coal mine safety, as well as an importantmeans of preventing the gas disaster. So, the question for the development of a coalmine gas emission in real-time intelligent forecasting system.Gas Emission not just affect one or several factors, but by many factors, coupledbetween a variety of factors have complex non-linear relationship, but also by thedynamic interaction of various factors at different times impact, whereby we wanted tofind a method capable of having to consider these factors, as long as the sequence iscapable of real time monitoring according to a single variable can be predictedaccurately. For the time series of a single variable, Gas Emission Study gas problem isthe most direct and convenient information collected on chaotic phase spacereconstruction technique on doing it, it believes that a single variable Gas Emission timesequence contains a number of factors affecting its gush of information, as long as thereconstruction of the time series, we can recover the phase space characteristics of gasemission, but the method also avoids human subjectivity.First, in determining the amount of gas emission characteristics of chaotic timeseries, while the use of mutual information and Cao’s method to get the delay time τ andembedding dimension m, on Gas Emission time series phase space reconstruction. Andthen were using the weighted one-rank local law instantiated generalized neuralnetworks based authentication chaotic time series prediction method and reconstructiontechnique with PSO-LSSVM prediction method based on phase space, throughcomparative analysis of the final choice of the phase space reconstruction and PSO-LSSVM combining forecasting model developed in this paper as a core algorithmsoftware.Secondly, in the case of determining the amount of gas emission prediction model, according to the theory and techniques, the use of VC++6.0, Matlab2008a, SQL2005as a platform, developed with Gas Emission data collection, storage, user management,and more forecasting system to predict real-time and historical analysis to predict thefunction of latitude. Finally, a coal mine in Shanxi by the software application andpost-field data in the analysis, the software can be obtained to complete the real-timemonitoring of gas emission, prediction, warning and other tasks, and the predicted valueand the actual value of the five time dimension more than75%of the absolute error is0.02or less, with better reliability and stability. Therefore, the software development isconducive to mine managers more convenient and intuitive for underground gasemission in the security situation analysis, judgment and decision making.