It is normally not a big problem but it depends on the size of the small scale mill and the operation mode. In dry grinding we have close relation between small and large scale. We have used small lab units with 400 and 600 mm diameter and scaled it up to 3 m mills with good results. In both case we used air classifiers with similar efficiency which is critical. In wet milling, I guess it will be very much depending on your flow conditions, water concentration, efficiency of your screens or hydrocyclone etc. SO it will be much more difficult to control it. But if you make enough tests with different conditions, it will work. It is also important to use correct calculations for larger mills.

There's a set of efficiency factors that can be applied to the Bond ball mill work index to predict grinding power requirements in bigger plants from this test procedure.

Designing Rod and Ball mills from RWI and BWI values and using the proper efficiency factors is a well-accepted method in industry and works well for most ores (power-based calculations). Special cares to be taken for mica, asbestos bearing ore for example, and for super fine product target (<38 microns).

Fully autogenous grinding will require pilot testing as it is not possible to extrapolate the behaviour of the large rocks from small rocks. HPGR might require testing in pilot unit as well.

Closed circuit milling at small scale (Bond mill sizes) provides opportunities to simulate mineral deportment issues, which you can't do in a basic batch mill except by doing locked cycle test, which can be a messy task for wet milling. A simple elutriator is probably the best substitute for a hydrocyclone for closed circuit laboratory milling.