Engineering & Mining Journal

JAN 2013

Engineering and Mining Journal - Whether the market is copper, gold, nickel, iron ore, lead/zinc, PGM, diamonds or other commodities, E&MJ takes the lead in projecting trends, following development and reporting on the most efficient operating pr

Issue link: https://emj.epubxp.com/i/102812

Contents of this Issue

Navigation

Page 49 of 99

BAUXITE Figure 6—Trend graph of on-line moisture measurements. analysis. The sample errors had an average value µε = –0.449%w and a standard deviation σε = 0.995%w. Two of the valid samples (indicated by red points in Figure 5(a)) were out of the 2-standard-deviation region (|εi | > 2σε) and were regarded outliers. Excluding those outliers, the remaining 11 sample values are the valid set of values from which the measurement uncertainty should be determined. The measurement uncertainty or accuracy (σmeasurement) is given by the root mean squared error (RMSE) between the valid composite sam- Figure 7—Histogram of moisture for a 20-day period. 48 E&MJ; • JANUARY 2013 ple values from the analyzer and their corresponding values from laboratory. It resulted as σmeasurement = 0.577%w. Finally, the measurement uncertainty is a quadrature combination of the sampling uncertainty and the calibration uncertainty or analyzer accuracy (σanalyzer) as seen in Equations 5 and 5a. The moisture measurement accuracy required by the plant operations was ±0.5%w. Since the accuracy achieved by the analyzer had met this requirement, the calibration was regarded as suitable. Step 3: On-site Verification (Model Validation)—After on-site calibration is achieved, the model should be validated. This is done with a new set of material samples, collected in the same way as for the on-site calibration step. If the sample measurements from the analyzer are consistent with their corresponding sample moisture results from laboratory, then the model is considered valid. Otherwise, a new model calibration should be done. A set of 10 three-cut composite samples of wet ore were collected from conveyor belt stops, and the moisture measurements from the analyzer associated to those samples were recorded. The sampling uncertainty for all of the cut sample values of all the composite samples resulted as σsampling = 0.717%w. This uncertainty was greater than the one from the on-site calibration, suggesting differences in the sampling, storage or laboratory practices between the two sampling campaigns. None of the composite samples presented cut sample deviations greater than two standard deviations (no cut sample outliers). All the 10 valid sample results are shown in Figure 5(b). The sample errors εi between each moisture value from the analyzer and its corresponding result from laboratory analysis had an average value µε = –0.354%w and a standard deviation σε = 1.258%w. One of the valid samples (indicated by the red point in Figure 5(b)) was out of the 2-standard-deviation region (|εi| > 2σε) and was regarded an outlier. Excluding this outlier, the remaining 9 www.e-mj.com

Articles in this issue

Links on this page

Archives of this issue

view archives of Engineering & Mining Journal - JAN 2013