The time-dependent increase of the electrical conductivity (dynamic percolation) of carbon black filled polymethylmethacrylate under quiescent conditions was studied theoretically and experimentally. Therefore, a three phase model based on two different aggregate sizes is proposed leading to manageable fit functions, which are able to describe the experimental data. In order to test the model, the time-dependent electrical conductivity of the composites was monitored prior and after a short shear step. It was found that the shear step leads to a significant change in the conductivity and influences the aggregation kinetics of the system which is in agreement with our physical model proposed. Moreover, it could be shown, that the model has a much higher accuracy in describing the experimental data compared to the existing aggregation models where only one aggregate size is considered.