Thanks for the reference material 2guys.
A couple of factors come to mind. Obviously it would be best, or even ideal, if the info used for project estimates was true to the run of mill material to be used. If the tests were strictly hydro-metallurgical, as in not chemical, then it may be reasonable to scale results from say 10% to 2.5% by just dividing by 4 but only under the assumption that the target MN is in all cases closely the same, i.e. same grain size, same chemical composition of grains, etc.. The last of these, same chemical composition, but not just of the MN bearing grains but of the entire tested ore, will come into play when the process is chemical. For example, what if the 10% ore is markedly lower in iron than the 2.5% ore, such that for example the first is 3 to 1 MN to FE whereas the lower is 1 to 1. Now the chemistry of the whole process might be significantly changed. The acid consumption could be much higher, like double or more, than what is estimated by just scaling from the 10% based only on MN content. For further example, what if the higher grade ore is very low in other species, like titanium, vanadium, uranium, etc. but the depositional environment that layered the 2.5% ore happened to be high in these, and some of the species, like vanadium, has a large impact on the reducing power of the leaching solutions that is available to reduce the MN ? Etc. Catalytic effects due to presence of certain species is more pronounced for organic than inorganic wet chemistry, but does still exist, and also the dynamics of reactions, that is the rate at which they happen, is a function of everything present and the proportions present.
So, all said, scaling from non-representative ore to actual, and use of aged, presumably more oxygenated, ore instead of mine run fresh ore, are both introducing a number of unknowns and uncertainties. Those introduced factors might be negligible, they also might not. The problem as I see it is that at this stage the effort should focus of removing unknowns and uncertainties, and the efforts clearly did not do this, and in fact introduced some needlessly. Simply bad engineering imo.
All is JMO of course, based on limited specific but in accord with general principles.
