Please see this book: 

Ashok Gupta, Denis S. Yan-Mineral Processing Design and Operations. An Introduction-Elsevier (2016)

Regards,

Jaime, you are a sucker for punishment if you are doing these manually.  Most people use software like MolyCop Tools or JK SimMet to fit datasets.  I haven't seen any worked examples of this that are suitable for you to train on, but perhaps get a copy of MolyCop Tools (it's free, contact your local MolyCop rep) and reverse-engineer the procedure in that 10-tab spreadsheet.

A major complication is these types of equations are iterative and need to approximate the imperfect classification that occurs in milling circuits.  They look superficially like linear mass balance calculations, but that deceptively simple model quickly breaks down and becomes nonlinear.

As Alex said, if you love hardship, see this PDF. Its conclusions are that an in depth review of the PBM used to simulate industrial wet grinding has found major flaws in the model showing-

1. There in no unique solution to the steady-state or dynamic PBM models. Several solutions were identified to yield the same product with the same feed matrix.
2. When steady-state mill matrix solutions were squared, significantly different product matrices were obtained.
3. Solution of dynamic PBM models using normalisation rather than mass balance constraints can lead to physically unrealistic solutions.
4. Predictions of time-vatying performance of mills using different solution of dynamic PBM models, leads to predictions which deviate at long grinding times especially for the smaller size fractions.
5. Because of the non-uniqueness of PBM inverse solutions and the non-linearity of the mill matrices and selection and breakage rate functions for most commercial grinding circuits, no set of general mill selection functions and breakage rates have appeared or are likely to appear in the literature.
6. A useful line of comminution research. would be to measure how the mill selection and breakage rates vary with the size distribution of the particles within the mill, as well as other mill conditions, such as slurry density, rheology, temperature etc.