Prediction Of Dust Concentration In Haerwusu Open-pit Coal Mine Based On Recurrent Neural Network

As a large-scale in terms of coal storage and mining volume in China,Haerwusu open-pit coal mine plays an important role in the energy supply of Inner Mongolia and even the country.There is a large variety of coal dust,rock dust and other material diffusion with mining activities,which could pose a threat to personnel health and safety,and has also exacerbated air pollution in the northwest region.It has a big meaning to grasp the characteristics and regularity of the dust and make scientific predictions.Based on the above situation,this paper explores the characteristics and influence factors of dust concentration,establishes a prediction model to forecast concentration of PM2.5 in open-pit coal mines.Based on the dust concentration and meteorological data from September to November,provided by two monitoring points which set up in work area and Nonworking area.Exploring the monthly and daily variation of dust concentration in openpit coal mines;analysising the impact of mining and loading intensity and meteorological factors on dust concentration;and finding the spatial relationship of dust concentrations.Results show that overall output of ore and waste,wind speed,wind direction,temperature,temperature change rate,humidity,humidity change rate,and wind speed*overall are affect the concentration of dust significantly in Harwusu openpit coal mine.Establish a traditional regression prediction model,random forest prediction model and LSTM prediction model for prediction PM2.5 concentration.Selecting time factors,meteorological factors,and mining and loading intensity as input variables.Selecting RMSE and MAPE as evaluation indicators.The results show that the LSTM prediction model works well,with an accuracy rate of 92.97%.There are 39 figures,17 tables and 73 references in this thesis.