Reliability Of Calculation Of Mineral Process Efficiencies And Rate Parameters From Balanced Data

Data is frequently collected in mineral processing plants in order to evaluate operating performance or to model process units. Metal concentrations, circulating loads, products flow rates, classifiers efficiencies, flotation units recoveries, grinding mill efficiencies are typical parameters calculated from plant data. The validity of these calculations depends strongly upon the reliability of the data used for that purpose. The objective of this paper is to illustrate that flowrates, efficiencies and rate parameters of mineral processes are more reliable when calculated from balanced data rather than from raw data. Using variance-covariance matrices of balanced data produced by automatic mass balance packages (including flow rates, concentrations, percentage of solids and size analyses), variances of separator efficiencies (flotation or classification) and of mill grinding rates are derived. The variances are compared with those of variables calculated from raw data. The advantage of using balance data is illustrated with real data from flotation, grinding and classification units.