Study on the Local Scour Characteristics of Sloping Drop Structure by Free Overfall Flows

• Main Authors: Hsi-Hsin Hsu, 許錫鑫
• Other Authors: Dong-Shin Shih
• Format: Others
• Language: zh-TW
• Published: 2016
• Online Access: http://ndltd.ncl.edu.tw/handle/34942495577285632214

博士 === 國立中興大學 === 土木工程學系所 === 105 === In recent years, hydraulic structures crossing rivers have been widely used in channels for the water resource management. Although the upstream channels can be stabilized, the gravel deposition in the hydraulic structures may increase the slope of the upstream channel bed . As a result, the large impact forces of a free-falling nape from the free overfall flow may weaken the hydraulic structures. In order to increase our understanding of the scouring phenomenon of free overfall for supercritical flows, a series of laboratory experiments with different upstream channel bed slopes S(=0~6 %), weir height H(=0.15~0.25 m)and unit width discharges q(=0.0035~0.0237 cms/m) acting on a uniform diameter of bed materiald_s (=0.50 mm,1.69 mm,3.35 mm) were performed. The experimental results indicate that the maximum scour depth (Y_se) is proportional to the drop number(D) but it is inversely proportional to the dimensionless diameter of bed material(G) and the bed slope of the upstream channel(S). The horizontal distance from toe of drop structure to maximum scour depth (L_se) and the maximum scour length (L_ie) are proportional to both D and S but they are inversely proportional to G. The maximum unit width scour hole volume(A_se) is proportional to the dimensionless impact force(F ̅_d) but it is inversely proportional to G and S. Due to the rapid progress of computer technology, an engineer now can use computers to design the hydraulic structures with remarkable computer power. The study combines the data collected in this study and the previous research results to develop a windows-based design model. It could show the upstream flow conditions, hydraulic characteristics, impact forces and scour hole volume on a chart based on the calculation of the model. Therefore, designing hydraulic structures could be much faster than in the past. It’s been approximately 50 years since Genetic Algorithm (GA) and Artificial Neural Networks (ANN) were adopted and developed. In recent years, the GA and the ANN has been widely used in the engineering and science fields. This paper uses GA and feed-forward back-propagation neuronal network (FFBPNN) to predict the scour hole behavior of free over-fall flow in the downstream channel. In this study, the extent of flow characteristics was discussed by using the channel tests. Firstly, we executed the laboratory experiment with clear water to obtain the total 144 sets of data. Then we normalized and trained these experimental data sets, by GA and FFBPNN, in order to get the training parameters of FFBPNN and the fitness function of GA. Additional 36 sets of laboratory experimental data were used to validate the usefulness of GA model, ANN model and the regression equation developed in this study. The simulation results showed that ANN provided the most accurate estimations for the maximum scour depth, the horizontal distance from the toe of the drop structure to the maximum scour depth, the maximum scour lengthand and the scour hole volume. And the results of GA was remarked second.