PSI - Issue 64

Marco Pirrò et al. / Procedia Structural Integrity 64 (2024) 669–676 Author name / Structural Integrity Procedia 00 (2019) 000–000

675

7

natural frequencies, which is purged of the EOVs present in the monitored features. Lastly, the approach based on T 2 -Hotelling control chart on PCA residuals requires the manual selection of (a) the number of the retained principal components for PCA and (b) the number of elements in the subgroups to build the T 2 -Hotelling control chart. This implies the need for a preliminary sensitivity analysis for the selection of the best parameters and add subjectivity to the method. Cointegration PCA+ T 2 -Hotelling control chart a)

b)

c)

Fig. 6. Comparison between T 2 -Hotelling control chart obtained from the residuals of frequencies (PCA) and Cointegration procedure, under varying training period lengths: a) 3 months; b) 6 months; c) 12 months. Upper and Lower Control Limits (UCL and LCL) are set at 95th percentile of training values. 5. Conclusions The paper presents an application of the Cointegration technique to eliminate the effects of varying environmental factors on natural frequencies in continuous monitoring. A cointegration residual is computed, purged from the common trends present in the natural frequencies and mainly due to EOVs. The 2-years continuous monitoring of Baixo Sabor dam is employed for structural assessment purposes: the case study demonstrates that cointegration technique is able to eliminate the influence of temperature and water level variations and, consequently, to identify the health condition for the dam. Compared to traditional methods based on the use of PCA + control-charts to minimize EOVs from natural frequencies, the cointegration technique presents some advantages, since: (a) it reduces the length of the training period to account for EOVs and (b) it relies on simpler operations, such as the construction of a stationary linear combination of the monitored features.