Dust is one of the most significant challenges in the mining industry, which adversely impacts the environment and human health. This research develops a hybrid approach based on dimensional analysis and regression analysis (H-DAMRA) methods to predict dust emission distance due to the bench-blasting in a limestone surface mining located in West Azerbaijan, northwest of Iran. Twelve effective variables including hole diameter, number of holes, delay timing, stemming, charge per delay, air humidity, air temperature, wind speed, wind direction, atmospheric pressure, powder factor, and blasted rock per hole were considered to model dust emission distance. According to the correlation analysis, powder factor, charge per delay, air humidity, and stemming have the highest correlation with the dust emission distance. In the dimensional analysis phase, the utmost influential factors in predicting the blast-induced dust emission distance were determined. As a result, several dimensionless outputs were obtained to be the dependent variables for regression analysis. Different multivariate regression models, including linear, nonlinear, logarithmic, and exponential regression, based on the dimensional analysis outputs, were developed and the performance of the models was evaluated using R2, RMSE, MAE, MRE, and VAF indices containing 100 in situ blasting rounds (100 data sample). The non-linear multivariate mixed regression model constructed on the dimensionless variables had the highest accuracy and the lowest error with a coefficient of determination of 93.57 percent. In addition, the results of the sensitivity analysis between input and output variables indicated that the dimensionless variables of “atmospheric pressure / (powder factor × stemming)” and “air humidity” with the values of 0.951 and 0.691 had the greatest and the lowest effect on dust emission distance, respectively.
کلید واژگان :Air pollution · Dust emission · Blasting, Dimensional analysis · Multivariate regression
ارزش ریالی : 600000 ریال
با پرداخت الکترونیک