Predicting the release of methane into the longwall workings in Polish coal mines using neural networks
Unfortunately, the effectiveness of the methods of forecasting methane release used so far is not satisfactory.For this reason, a method based on neural networks has been proposed.An artificial neural network is usually a computer-programme or a device that imitates the functioning of the nervous system.Adapting a neural network to solving problems consists in teaching it by applying a sequence of typical stimuli and the corresponding desirable responses, but not by defining an algorithm and recording it in the form of a programme.The aim of the present paper is to confirm the possibility of using artificial neural networks to predict absolute methane-bearing capacity in longwalls mined with roof caving.Input data for forecasting were derived from the methanometric and anemometric systems of the mine as well as from the general mining and geological data.The algorithm of the neural networks has achieved a high level of accuracy confirmed by large values of correlation coefficients, and therefore can be implemented to forecast the release of methane into a longwall mined with roof caving and ventilated with a retreating U ventilation system.
neural networks methane hazard release of methane
BOROWSKI M. SZLAZAK N.
AGH University of Science & Technology, Poland
国际会议
International Mining Conference 2014(2014国际采矿论坛)
安徽淮南
英文
521-530
2014-09-19（万方平台首次上网日期，不代表论文的发表时间）