Surveys of two types of crushers operating at Argyle (Allis Chalmers (Kobe)) and Bougainville and Mt. Isa (Nordberg (Symons)) were undertaken in order to develop mathematical models of the crusher performance for use in simulation and design studies. The effects of crusher gap, feed coarseness, throughput, head angle, throw, finer length, finer wear and eccentric speed were investigated.
Both explicit and implicit models were developed. The implicit models included the 80% passing size of the product, P80, a dependent variable. The models developed confirmed the dominating effect of the crusher gap on performance. Throughput and feed coarseness were both significant but minor influences.
For the explicit model based on all the data from the three sites, an increase in the finer length was found to increase the value of the K1 classification parameter, indicating that finer particles in the crusher are more efficiently "classified" to the product in the Allis Chalmers (Kobe) crushers than in the Nordberg (Symons} crushers. The crusher throw was noted to be a significant variable in the K2 classification parameter regression for the implicit form of the model. As the throw increased, the size of the particles certain to be broken increased, resulting in a coarser product. The throws were 5 1 mm for the Allis Chalmers (Kobe) and 111 mm for the Nordberg (Symons} crushers.
The breakage parameter was found to be weakly dependent on the crusher gap, throughput and feed coarseness, but the regression equations were not significantly improved by the inclusion of these variables. The breakage parameter was therefore fitted as a constant in all the models.
Breakage distribution functions and energy-size-reduction characteristics of ores from the three sites were determined u sing a modified form of the single particle pendulum impact test (Narayanan, 1986). A method was developed for the prediction of crusher power requirements using this information.
The models were shown to be powerful and accurate tools in the simulation of the performance of crushing circuits when used in conjunction with power and throughput constraints. The implicit models were particularly suitable for the design of new plants and the optimisation of existing plants where the P80 size was specified. The explicit models had more general application.