基于主元分析法的井架损伤定位研究

由于实测FRF频谱受到噪声污染和其它干扰因素的影响,数据变量间存在相关关系,直接利用这些数据进行井架的损伤识别很难实现。为此,提出了基于主元分析法和FRF的井架损伤识别方法。用实测FRF数据作为损伤识别的基本变量,用PCA方法和变量重构对FRF数据进行降维处理,用T2控制图分离异常数据,进而识别井架的损伤。井架测试试验结果显示,前5阶主元几乎包含了原始数据的全部信息特征,因此,用主元分析法对原始数据矩阵进行降维处理是可行的,尤其对噪声和非线性状态下井架的损伤识别有良好的适应性。

Study of damage location of derrick based on the principal component analysis

Since the measured FRF frequency spectrum is affected by noise pollution and other interference factors and there exists interrelationship among data variables,it is very difficult to realize the damage identification of derrick by directly using these data.Therefore,the derrick damage identification method based on principal component analysis and FRF is put forward.With the measured FRF data as the basic variable of damage identification,the FRF data is treated with dimension reduction by using PCA method and variable restructuring.The T2 control diagram is used to separate abnormal data for identifying the derrick damage.The derrick dynamic test result shows that the former 5-step principal components almost cover the whole information characteristics of original data.It is feasible to carry out dimension reduction of original data matrix with principal component analysis,which has good adaptability for derrick damage identification under the condition of noise and nonlinear state.