Owusu, Kwaku Boateng2025-12-172025-12-172022https://hdl.handle.net/11541.2/344541 ethesis (xxviii, 311, [9] pages) :colour illustrations and charts.Includes bibliographical references (pages 281-297)The hostile internal environment of industrial autogenous/semi-autogenous (AG/SAG) mills makes online monitoring unreliable using direct measuring methods. To develop an efficient online monitoring system in AG/SAG mills, this thesis investigated the acoustic sensing of feed ore heterogeneity in a laboratory AG/SAG mill. The study was extended to a large-scale SAG mill, where acoustic frequencies and the Gaussian process regression (GPR) model were used to predict mill weight. It was shown that the acoustic response of the laboratory AG/SAG mill was sensitive for monitoring different feed hardness and size distributions. Also, it was established that the GPR model makes good mill weight prediction combined with relevant acoustic frequencies.engrinding mills;acoustic sensing;real-time monitoringAutogenous grinding.Milling machineryRegression analysis.Size reduction of materials.thesis