Research On Methods Of Vertical Bearing Characteristic Of Statically Pressured Pipe Pile In Liao-shen Area

The vertical bearing capacity of statically-pressured pipe pile is influenced by many factors, such as parameters of pile body, properties of soil layers around pile and construction conditions of site, which have mutual relationship with each other to a certain degree. For this pile-soil system with a higher complex stochastic uncertainty and fuzzy nonlinear, it is difficult to evaluate these influencing factors with traditional methods or ascertain the contribution degree of every parameter to bearing capacity in certain geological condition. Thus, this paper introduces some data analyzing methods to solve these types of problems, which is suitable for the solution of stochastic uncertainty, fuzzy nonlinear and multi-factors with fewer samples. Grey correlation theory and partial least squares (PLS) regression are used to do the correlation degree analysis on the influencing factors which affect the vertical bearing capacity in Liao-shen area. Empirical regression formula based on PLS method and optimized model of artificial neural network (ANN) based on principal component analysis are established to predict the vertical ultimate bearing capacity of single pile. Also, the thought of using cellular automata (CA) technique to simulate and predict the load-bearing deformation mechanism is suggested in this paper.In the process of evaluating influencing factors which affect the vertical bearing capacity of single pile in this paper, some engineering information of statically-pressured pipe pile is chosen to do the engineering case analysis in typical geological condition of Liao-shen area. Relative data from site investigation, construction of pile foundation and static load test is collected and arranged. Analysis of importance of variable in projection is applied in proceeding of PLS regression to determine the number of independent variables, principal component analysis is used to pre-treat the input parameters before the establishment of ANN model based on conjugate gradient algorithm, which are all to achieve the purpose of involving important information as much as possible and avoiding the interference from the less-important information. Thus, the prediction model of vertical ultimate bearing capacity of single pile with better accuracy has been got for statically-pressured pipe pile in Liao-shen area. Moreover, compared the results from the proposed methods in this paper with the testing data of field tests, it is concluded that the prediction value are similar to the experimental results. The methods proposed in this paper are feasible to analyze the vertical bearing characteristic of statically-pressured pipe pile in Liao-shen area. The results of estimation of ultimate bearing capacity meet the requirement of practical engineering. Besides, the basic thoughts and methods of application of CA technique discussed in this paper will make a contribution in theoretical analysis of bearing characteristic of single pile from a new way, and even provide references for other relative issues in geotechnical engineering.