Akademik Veri Yönetim Sistemi
JOURNAL OF THE SOUTHERN AFRICAN INSTITUTE OF MINING AND METALLURGY, cilt.126, sa.5, ss.261-278, 2026 (SCI-Expanded, Scopus)
This study aims to develop a model for particulate matter concentration during the loading process in open pit mining. The researchers conducted simultaneous measurements of particulate matter and meteorological parameters and collected samples to determine the moisture content of the loaded materials. The analysis used 7,895 measurement data points from gypsum and limestone quarries, employing two data analysis programs, SPSS® and Waikato Environment for Knowledge Analysis, to derive equations for predicting particulate matter release. In the modelling, particulate matter measurements were the dependent variables, while meteorological parameters, laboratory measurement results, and loader bucket capacities were the independent variables. The classical regression models did not adequately capture the dependent variable, thus, the researchers explored the decision tree approach for further modelling. The M5P algorithm was used to generate regression equations for the different data sets, and the findings showed that the models had a satisfactory degree of predictive capability. The number of fine particles released during loading is influenced by various weather factors. Temperature, humidity, wind speed, and station pressure can all affect the dispersal of these particles. Additionally, the moisture content of the loaded material and the capacity of the loading equipment contribute to this process. The M5P decision tree algorithm, which is rarely used in particulate matter concentration models, provides an innovative approach to developing local concentration estimates for non-coal surface mining.