Numerical Model Updating of the Ucsc-Cross (Concepcion, Chile) Using Experimentally Extracted Data and By Applying Regional Sensitivity Methods

Abstract eng:
Calibration of numerical models has become indispensable to reduce error in the prediction of structural performance of complex systems. Several methods have been developed in the last two decades to calibrate a numerical model, but the implementation of these procedures in engineering common practice is still infrequent. Hence, the analysis-verificationcalibration-design cycle is usually incomplete. In this study, a 24-m high cross-shape truss structure was model using a finite elements approach. The finite element code was validated by comparing its results with other finite element models commercial package. The numerical model was calibrated using experimental data extracted from the actual structure located at the San Andrés Campus of the UCSC in Concepción, Chile. A grid of 9 accelerometers was installed in the actual structure to record its response due to ambient excitation. The modal frequencies were extracted from these acceleration records using two operational modal analysis methods (SSI and FDD), identifying 4 predominant frequencies. Model calibration was performed by applying a regional sensitivity method based on Monte-Carlo sampling. This method adjusts the modal response of the structural model to the real response measured on the actual structure. More than 2000 realizations were performed. The parameters chosen for calibrating the model were the Young´s modulus, truss-section thickness, material´s equivalent density and support flexibility. The simulation was carried out using the Matlab Toolbox MCAT. A significant improvement of the model prediction was achieved by the calibration that allows a more precise estimation of structural performance under extreme conditions (wind gusts, earthquakes and others).

• Export as: BibTeX | MARC | MARCXML | DC | EndNote | NLM | RefWorks
• Add to your basket