Author(s): Joongcheol Paik; Youngil Kim
Linked Author(s): Joongcheol Paik
Keywords: Debris flow; Deposition; Culvert; Laboratory experiment; Logistic regression model
Abstract: The degradation of discharge capacity of culverts installed in mountain regions due to debris flow may results in damages of properties as well as serious casualty. Understanding and accurate reproducing the behaviour and the deposition features of debris flows at various conditions of geometrical configurations and sediment material property are prerequisite to develop advanced design criteria for drainage culvert to prevent such damages. We carried out a series of laboratory experiments of debris flows in a rectangular channel of constant width with an abrupt change of bottom slope to investigate the behaviour of debris flows in the rectangular culvert in terms of approaching and drainage culvert slopes, representative grain sizes and sediment volume concentrations. The experimental results show that the depth of debris flow substantially increases up to more than four times compared to that of water flow of same flow rate and, at the same flow rate and sediment volume concentration, the sediment deposition of find-material debris flow starts further downstream with approximately 3ohigher surface slope rather than that of the coarse-material debris flow. Based on the experimental measurements and a series of complementary numerical simulations, we develop a logistic regression model to seek optimal design criteria to prevent debris flow from depositing in the culvert. The estimated logistic regression parameters and the derived deposition criteria curve indicate that the deposition of debris flow in the culvert most strongly depend on the bottom slope of culvert while the effects of the approaching channel slope and the representative grain size on the sediment deposition are not significant rather than the others. The application of the proposed logistic regression model to the real field cases demonstrates that the present model can successfully well predict whether the debris flow material deposits or not in the culverts that already installed in mountain areas.
Year: 2012