Application of statistics in Mineral Engineering

Abstract

Probability theory, statistics and the principles of model building and hypothesis testing have found application in mineral engineering in three main fields: (1) Regressional analysis to find out the relation between several independent variables with a dependent variable for control and optimization. (2) Statistical design of experiments for screening of variables and determination of their mutual interaction to increase the efficiency. (3) Statistical screening of particulate material. In plant and laboratory work, correlation of variables can be of great help in understanding, prediction and control. Regression curves are empirical curves derived from a set of data and are used to predict results from an observed set of conditions. Besides their usual application in control and prediction, they can be used for the proper evaluation of a changed process and also for obtaining the optimum operating conditions. Experimental programmes using statistical designs offer an efficient method to solve the problem of interaction of variables in 'one variable at a time approach' in studying systems. The programme involves use of factorial designs to study the effect of varying two or more factors simultaneously according to an ordered plan. The data is treated to yield information for improving the measured response. Statistical screening is separation of particulate feed into 'size intervals' on the basis of probabilistic behaviour of individual particles. In this paper, an attempt has been made to review to select a few contributions in each of the above particular fields. Some of the fundamentals necessary to understand these approaches have also been presented in a rudimentary form. (G.K. Bandyopadhyay and T.C. Rao, Department of Metallurgy, Indian Institute of Technology, Kanpur)