Random Forest Model for Sediment Deposition Prediction in Sewer Pipes

Experimental studies have been carried out for determination of sediment incipient deposition in open channels. This study is aimed to generalize the incipient deposition models through incorporating a cross-section shape factor into the model. Random Forest (RF) technique is used for the modelling and results are compared with the conventional regression method. Comparison of developed models in this study with those existing in the literature indicates that cross-section specific models provide poor performances on other channels. RF model outperforms its alternatives which can be related to the utilizing shape factor as input of the models. A suggested model can be used as practical tool for sewer design in practice.

Keywords:

open channel , random forest , sediment transport , sewer system