论文部分内容阅读
基于支持向量机(SVM)和独立分量分析(ICA)建立了超长大直径钢管桩极限承载力的预测模型。先采用独立分量分析FastICA算法从实际工程的超长大直径钢管桩试桩的实测数据样本中抽取相互独立的分量,这些分量不仅去除了相关性,还保持统计独立,并服从非高斯分布,能更好地表现数据间的本质结构;然后,确定支持向量机作为分类器,以抽取的独立分量作为支持向量机模型的输入参数,建立超长大直径钢管桩的承载力预测模型ICASVM_Q;最后,采用某大桥的工程数据对预测模型进行测试。结果表明,ICASVM_Q的预测效果明显优于以原始数据作为支持向量机模型输入的SVM_Q模型的预测效果。可见,采用将独立分量分析与支持向量机相结合的方法建模预测超长大直径钢管桩的承载力是可行的,ICASVM_Q模型的预测结果可用于超长大直径钢管桩承载力的设计参考,具有一定的工程应用价值。这种方法还可以用于其他领域的智能预测研究中。
Based on Support Vector Machine (SVM) and Independent Component Analysis (ICA), the prediction model of ultimate bearing capacity of ultra-long-large diameter steel pipe pile is established. Firstly, the independent component analysis FastICA algorithm is used to extract mutually independent components from the actual measured data samples of the test pile of the extra-long large diameter steel pipe pile. These components not only eliminate the correlation but also keep the statistical independence and obey the non-Gaussian distribution, Then, the support vector machine (SVM) is selected as the classifier and the extracted independent component is used as the input parameter to the support vector machine model to establish the bearing capacity prediction model ICASVM_Q of the super long and large diameter steel pipe pile. Finally, the engineering model of a bridge is used to test the prediction model. The results show that the predictive effect of ICASVM_Q is better than that of SVM_Q model with raw data as support vector machine model input. It can be seen that it is feasible to use the method of combining independent component analysis and support vector machine to predict the bearing capacity of super-long-large-diameter steel pipe piles. The prediction results of ICASVM_Q model can be used to design the bearing capacity of super-long- Reference, has a certain value of engineering applications. This method can also be used in other areas of intelligent prediction research.