Concurrent reconciliation of chemical and mineral assays for mineral processing circuits

• Optimization-based estimation of mineral assays from chemical assays is discussed.
• Concurrent data reconciliation method for mineral and chemical assays is presented.
• Methodologies are demonstrated with laboratory data on iron-ore slimes enrichment.

The steady-state data reconciliation of mineral processing plants has traditionally depended on chemical assays of samples collected from various streams in the plant. In some cases, additional information in the form of size assays and size-wise chemical assays have been employed to improve the reconciliation problem. Reconciliation algorithms, in addition to making the measured data consistent with the basic conservation principles, also help in predicting the material flow rates in different streams. This information helps the process engineer to identify and troubleshoot problem areas, and further use it to optimize the plant's performance. If all the minerals present in the ore system are known (established through X-Ray Diffraction), the chemical formulae of the minerals provide the relationship between the chemical assays and mineral composition. In this communication, first we propose an approach to estimate the mineral composition from the measured chemical assays for a given sample or stream. It is then extended to a plant-wide concurrent data reconciliation methodology, involving both the chemical assays and mineral composition. This concurrent data reconciliation procedure provides estimates of mineral composition around the plant even when mineral composition measurements are not available. This additional information will help the plant engineers to track and improve their plant's behavior more intuitively. When mineral compositions are measured (for example, through QEMSCAN) they can be reconciled along with chemical assays. The methodology is demonstrated with data from batch and semi-batch laboratory studies on selective dispersion-flocculation of iron ore slimes.