Design of Flexible Plants

Geometallurgy brought the promise of adaptively processing the ore in the best possible way, depending on its local properties. Particle based modelling of processing allows to compute economically optimal processing parameters and configurations, given the flexibility of the plant, its capacity limits and the feed properties. In theory a flexible processing could increase the economic value substantially.

Real plants are designed for a constant feed and real operations use blending to homogenize the feed. This contribution analyses what features would make a plant design sufficiently flexible to exploit this theoretical superiority of adaptive processing. One can quantify the economic effect of the various options for flexibility, such as extra capacity in various steps, controlled or regulated machine settings including online measurements, intermediate storage and sorted stockpiling, variable residence time in feedback loops, rerouting, and parallel processing units.

We model the adaptive processing situation by a bivariate distribution of the properties of and knowlege about selective feed units (SFU) analogous to selective mining units (SMU), but larger. Using a forward simulator we can compute the optimal parameterisation of the plant as a function of the knowledge about SFU. Depending on the flexibility of the plant we get a different mean economic effect of picking the choice with the highest value. This value is computed as a difference of selling price of products and the processing and dumping costs. Averaging over all units we can compute a mean value. The difference of this mean to the optimal fixed processing setting a global ore blend provides the value of the given flexibility option. A more flexible plant also has a higher CAPEX. Comparing the discounted value of flexibility to this CAPEX one can select the optimal flexibility for the plant.