Mineral Processing Grade Recovery Optimization using Data unification and gross error detection
CIM MineSpace 2001
Osvaldo A. Bascur, Ales Soudek,
Mineral Processing Grade Recovery Optimization using Data unification and gross error detection.
Osvaldo A. Bascur, Ph.D. and A. Soudek
OSI Software, Inc.
email@example.com and firstname.lastname@example.org
Industrial complexes rely on measurements of flow rates, qualities, and concentrations for monitoring and optimizing their operation. Corporations use the data from the various sensors and measurements to make business decisions which affect the profitability of the operation. However, gross errors, due to a variety of reasons like faulty sensors, bad instrument calibration, or incorrect process models may corrupt the actionable information supplied to operations, quality managers, and maintenance personnel. The solution is to use the sophisticated new tools to model the process, sift through the available data to locate the gross errors, and use statistical data reconciliation to validate the sensors and instruments. In this way the data is unified to give the most accurate plant information for use by the monitoring and decision making processes.
Most mining and metallurgical plants use predictive maintenance policies based on statistical analysis and special techniques, like vibration analysis, oil and lube analysis for critical equipment. Process control strategies are installed in several metallurgical sites. All of these depend on reliable and accurate data. The data unification methods validate the individual measurements by using the concept of balances and statistical techniques where each instrument is no longer independent of the other measurements in the facility.
This paper refers to the data unification and analysis methods required for profitable plant operations. Examples in Mine/Concentrator integration and Grinding/Flotation Performance Monitoring will be presented. The key is to have grade recovery maps available to manage the conditioning necessary for optimal economical recovery of valuable metals. Continuous improvement and monitoring of the measurement system at the plant and analysis of the reconciled data provide the best information for the decision-making process. On-line gross error detection and elimination methods are required to ensure efficient evaluation of process performance and process optimization. An open discussion of the new methods for gross error detection and data unification will be presented.
Keywords: Grade/Recovery Maps, Concentrator Optimization, plant information, data reconciliation, gross error estimation, Kalman filtering.
Plant Wide Optimization, Ore Recovery Optimization, Grade Recovery Economic Optimization, Grinding/Flotation Analysis, Gross Error Detection, Data Unification