Simulation for Ball Mill Sizing: The Path to a Successful Plant Start-Up

"Robex Resources interrupted production at the Nampala Mine in September 2014 following throughput and gold recovery results which were below expectations during the first months of operation. After a thorough plant audit, an important brainstorming took place in order to identify an economical path to resume production. It was clearly identified that a significant investment was necessary at the concentrator prior to restarting the plant, namely the addition of a ball mill and a new elution system.The mineral processing simulation software USIM PAC was used in order to simulate the grinding circuit in design. The simulations’ outputs allowed the ball mill characteristics and the surrounding classification devices to be designed with more accuracy. Many ball mills available from second hand retailers were therefore studied in the simulated circuit. The chosen ball mill, which was in perfect condition, had the advantage of being available for rapid delivery on site.The paper first describes the ball mill sizing method in comparison with the conventional Allis Chalmers method. It then describes the phase 2 concentrator start-up (that took place in 2016) with the commissioning of the ball mill, the classification circuit, the gravity concentration, the four (4) new carbon-in-leach (CIL) tanks, and the Zadra elution system (The phase 1 start-up took place in 2014). It finally provides production data compared to expected simulation data, as commercial production was reached in January 2017.IntroductionDesigning a new concentrator or new equipment within an existing concentrator requires significant efforts to identify possible bottlenecks and their causes (i.e. limiting equipment or sectors) and accurate sizing of major equipment can have a tremendous impact on the economy of the whole project. In the presented case study, the ball mill was the critical equipment and oversizing it could have brought the Capex to a level where the project would not have gone through. Inversely, undersizing the mill could have resulted in lower production and therefore no economic viability for the project once it is in operation.During engineering studies, equipment sizing often relies on a static mass balance and one-by-one sizing of all pieces of equipment. The sizing is then left to the equipment manufacturer or in the case of mills, evaluated with empirical models mostly relying on the Bond equations (Allis Chalmers method) [Rowland and Kjos, 1978]. This methodology shows the disadvantage of not considering interaction between equipment in a mineral processing circuit, especially when recirculation occurs. Also, it has the disadvantage of not using the more advanced grinding models that include the use of the whole particle size distribution as well as phenomenological components by only using limited parameters in the ore particle size distribution (i.e. the P80). This can be explained by the lack of data that sometimes needs to be gathered in order to tune the more advanced models but also by the conservative mindset in mineral concentration engineering which does not always favor the use of process simulation. In many aspects of process engineering, Soutex has demonstrated its affinity to using simulation in order to size equipment with more accuracy, consider process variability, simulate complex systems, and study supply chain management (Berton et al., 2015)."