The model structure proposed for the aggregation kernel works for low and high L / S ratios. Although some improvements can still be made in terms of granule prediction for a low L / S configuration, the use of a single core is a big step forward. The objective function that uses the moments of the distribution is a good improvement. Since it contains information on both fine and larger particles, it is possible to obtain a very good fit of the measured data, using information from the whole GSD.