Evaluation of strength properties of gravel formations based on field investigation and the discrete element method

• Main Authors: Min-Chieh Cheng, 鄭敏杰
• Other Authors: 張光宗
• Format: Others
• Language: zh-TW
• Published: 2011
• Online Access: http://ndltd.ncl.edu.tw/handle/11851955284715054981

碩士 === 中興大學 === 水土保持學系所 === 99 === Gravel layers are widely distributed in Taiwan. Lots of construction is located at gravel layers covered basins, mesas, or hills. Slope failure and debris flow are problems on gravel slopes. Understanding the mechanical behavior of gravel layers can help engineering construction and slope disaster prevention. In this study, geological and topographical characteristics of gravel formations were investigated to correlate geological factors and strength properties. Geological factors include the size distribution of gravel formations, percentage of gravel volume, matrix strength and gravel strength. We measured topographical features of slopes such as height and inclination from DEM (Digital Elevation Model). The relationships between the geological factors and strength of gravel formations are established. The software PFC 2D is used to estimate strength of gravel formations. We use optimization and statistical design to determine the micro-parameters, which relate to geological factors in the field. The use of empirical model and the PFC model to estimate the strength of gravel formations are validated. According to the field investigation, the factors that affect the strength of gravel layers are form greatest to lest, matrix strength (qM), gravel content (CG), particle size (D30), and gravel strength (qG). The major factor, matrix strength is selected to determine the micro-parameters f and k. The importance order of the micro-parameters of PFC is as follows: friction coefficient (f), bond strength (bond), contact with stiffness (k). The f and k are eventually functions of bond. The results of validation show that the error of the empirical model are -8.95%, -30.06% and -12.4% for Jiujiu Feng, Sanyi and Bagua Terrace. The error of the PFC model are -21.8%, 19.6 and 0.1%. Overall, the empirical model and the PFC model are demonstrated to satisfactorily estimate the strength of gravel formations.