Resesrch On Nonlinear Dynamic Prediction Of Coal And Gas Outburst In Heading Face

Coal and gas outburst is a kind of mine disaster phenomenon which is extremely complex,highly emergent and destructive;the effective method for preventing gas disaster is the most important research topic in the filed of mine safety all the time.Considering that the scientific nature of outburst can't be clarified completely so that the sensitive index and critical value of prediction outburst can't be confirmed form theories.This paper based on the thesis called the study on the dynamic identification and the based theory by National Nature Science Foundation of China researches nonlinear characteristics of gas concentration on tunneling face and proposes a nonlinear and dynamic prediction theory of coal and gas on multi-factor coupling tunneling face.Thesis baesd on the research of the methods of predicting gas outburst and the main factors of influencing coal and gas proposes a dynamic prediction model based on ASGSO-LS-SVM by coupling multi-factor.The in-depth research is proceed from the certainty andchaos characteristics of dynamic prediction dindex,fractal or multi fractal characteristics and the realization and verification of model system.Thesis analyzes the weight of influencial factors about coal and gas outburst by using analytic hierarchy process and the relationship of gas emission and other factors such as initial speed of gas emission,gas content,speed of gas emission and gas concentration by selecting 9 parameters as sample set and sorting by weight.Then this paper proposes 5 main factors which are the gas concentration,gas pressure,coal's solidity coefficient,structure type of coal and mining depth as the major index system of nonlinear dynamic prediction about coal and outburst danger on tunneling face.The nonlinear charactistics of time series and gas concentration on tunneling face are resraeched.A series of disposal steps such as the disposal method of abnormal data based on ?-ridge regression estimation,the disposal of data missing based on cubic exponential smoothing method and the methods of wavelet de-noising which the threshold function improved are proposed.The opinion that the gas concentration series on tunneling face have chaos charactistics is proposed and verified by three index which arecorrelation dimension,index calculation of second Renyi entropy and the max index of Lyapunov so that the chaos chararctistics are proved and provides basis for prediction theory of nonlinear system.The viewpoint of the fractal and multi-fractal characteristics of gas concentration sequence in heading face is put forward and Hurst index is calculated by using the method of re labeled range analysis to prove The fractal feature of the sequence.In the research of multi-fractal characteristics,a consequence is concluded that,the steeper the trend of the Dq-q is,the more the amount of internal information contained in the sequence is.There is a danger of outburst.And the fractal dimension is higher,it can explain that more and more tend to highlight the occurrence of.The results obtained are consistent with the Dq-q analysis.The analysis result makes it possible that the multi fractal characteristics of the gas concentration time series in the heading face become the important characteristic index of the occurrence of the gas concentration time series,and further to provide support for nonlinear dynamic prediction theory.In order to improve the learning ability and generalization ability of the least square support vector machine,the ASGSO-LS-SVM algorithm is proposed,in which the dynamic adaptive step size strategy of firefly is replaced by the fixed step size.The theory of nonlinear dynamic prediction in heading face of coal and gas outburst is established which is based on ASGSO-LS-SVM to real-time array parameters of mine gas concentration monitoring,combined with the gas pressure,coal's solidity coefficient,coal structure,multi factor coupling prediction model of mining depth,and the feasibility of the theory is verified by an example of gas outburst in Shanxi two mine.At the same time,the nonlinear dynamic prediction management system of coal and gas outburst is designed and developed.The application of nonlinear dynamic prediction of coal and gas outburst in heading face is promoted too.