Study of the Ultimate Bearing Capacity Prediction Model of Foundation Piles using Artifical Neural Network

• Main Authors: Rong-Chang Tsai, 蔡榮昌
• Other Authors: Sou-Sen Leu
• Format: Others
• Language: zh-TW
• Published: 2002
• Online Access: http://ndltd.ncl.edu.tw/handle/67839829309278746449

碩士 === 國立臺灣科技大學 === 營建工程系 === 90 === The bearing capacities of foundation piles are greatly influnced by the structurstatus, the material of the piles, and also the interaction behaviors between pile foundation and surrounding soil. The ultimate bearing capacities of foundation piles were traditionally formulated as a function of the variables of the foundation materials, the strength parameters of soil, and the construction methods. Yet the mechanical behaviors and the bearing capacities of piles are quite complicated and can not be easily predicted by theoretical or trachitional empirical approach. When foundation piles are designed, in most case, an empirical method or a semi-empirical method is usually used. Nevertheless, as stated alove, the results from these methods may be different from the actual situation .In order to resolve this problem, the correlation between the independent variables of soil characteristics and pile design, and the dependent variable of the maximal bearing capacities of the foundation piles was first analyzed, and then linear and nonlinear regression functions, as well as a neural network model which can precisely resolve nonlinear mechanical behaviors were established.