|
Mine gas is one of the major factors that cause safe hidden trouble,it’s explosion lower limit of volume fraction in the air is 5.3%,upper limit of volume fraction in the air is16.0%.Once the accident happened,not only caused a huge loss of state property,but also threat to the miner’s life directly.At present,aiming at gas disaster prevention,we have mainly adopt means is real-time detection to the coal mine gas concentration,although this method is mature and popularity relatively nowadays,but it cannot do predict to the change of gas concentration in the future,do not enough to improve the alert consciousness of people,do not give people enough time to analysis and deal with problem,thus cause the accident occurred frequently.Aiming at this problem,with the present research situation of domestic and foreign forecast of gas,In this paper,based on research of all kinds of gas forecast method,design a forecast and warning of gas device that have accurate results,low power consumption,portable and intelligent.First,this paper through studying for grey forecasting model, markov model,artificial neural network model,support vector machines (SVM) model and prediction model based on granularity and quotient space,then do different model combination,through experimental simulation,finally find a optimal combination of the model,to apply the model to the whole devices,achieve the most ideal effect for gas concentration prediction.For the realization of the gas device,this paper uses the high-end C2000 series DSP chip:TMS320F2812,it come from TI company in American.Relying on DSP fast data processing ability,bigger storage space and high levels of integration,not only improve the signal processing speed greatly,but also improve the signal operation precision,this is an undoubted optimal choice for gas concentration accurate prediction in mine.
|