Predicting the Status of the Bridge Foundation According to Rainfall Events

• Main Authors: Yu-WenHong, 洪鈺雯
• Other Authors: Chong-Wei Feng
• Format: Others
• Language: zh-TW
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/86420014242961666363

碩士 === 國立成功大學 === 土木工程學系碩博士班 === 101 === 　　The damage of bridge foundations is one of the important reasons to collapse the bridge, especially during the rainy season. An accurate estimation of the maximum local depth of the scour depth around bridge piers is essential for reliable and cost effective design. However, the difficulties of measuring the scour depth, researchers have suggested a number of empirical formulas to estimate the scour depth. Estimating the maximum scour depth, around bridge piers is however not straightforward due to the complexity of scour phenomenon which potentially involves numerous variables such as the geometry of the pier (type, shape and skewness), flow depth, flow velocity, the composition of bed materials, and debris or ice jam. 　　This research chooses rainfall events as the input data, including the intensity of rainfall, days of rainy, and number of times, and the sets the bed degradation as output. Using the genetic algorithm (GA) to define the classification’s standard, so rainfall events can be classified. An artificial neural networks (ANN) model is then developed to map the input and output. Base on the observed pattern of the local scour at bridge piers and the simulation of rainfall by the Monte Carlo method, this research predicts the status of the bridge foundation.