Use of Functional Performance Models to Increase Plant Grinding Efficiency

"The Functional Performance Equation calculated from a plant grinding survey defines two distinct ball mill circuit efficiencies and how they relate to total production rate through the target grind size. The survey data are also used to calculate a complete ball mill size distribution model, represented by the energy specific grinding rates, not just through the target grind size, but through all the size classes. Cyclone separation performance is represented by individual size recoveries to underflow. When these two unit operation performances are entered into the circuit modelling program (Streamline™), along with the circuit feed rate and size distribution, all the circuit survey size distributions and mass flows are perfectly regenerated by the program. The many issues associated with the historical Epstein (breakage-selection) models are eliminated. Once the circuit program is populated, pump and cyclone performance changes can be evaluated using a step-by-step procedure in order to maximize Classification System Efficiency. Mill water use and media sizing opportunities are also diagnosed in order to maximize Mill Grinding Efficiency at the size of interest. An example is provided. This circuit analysis and modelling system, and related training, is provided through a web-based application.IntroductionSince its development (McIvor, 1988), the Functional Performance Equation for ball milling has facilitated new understanding of industrial circuit cause and effect relationships, leading to significant strides in operational performance. Success stories of this refreshing approach to plant circuit performance improvement are now too numerous to list. Finch and McIvor (circa 1986) also discovered that the Functional Performance “Mill Grinding Rate” at the size of interest was the same cumulative grinding rate that is calculated from mill feed and discharge size distributions using a first-order rate equation. By calculating these cumulative grinding rates for all screen sizes, a complete ball mill model is generated (as used by Finch and Ramirez, 1981). Others in industry, Hinde and Kalala (2009), for example, have noted the same. Lacking the complexity needed to maintain their interest, most researchers have ignored this model, broadly opting instead for Epstein’s (1948) characterization of grinding as a chemical reaction, combining “selection rate” and “breakage size distribution” functions. The much simpler ball mill model presented here has previously been incorporated into a proprietary circuit modeling program (McIvor, 2005) and used for plant improvement work (for example, see McIvor and Finch, 2007, and McIvor, 2014). Now, by combining this extremely business-friendly ball mill model with optimization criteria from Functional Performance Analysis, a circuit modeling system (including the needed supplementary steps to assure successful plant implementation of circuit design changes) is at the disposal of every plant metallurgist, grinding equipment/material provider, and circuit designer."