Intelligent automatic operational modal analysis

Abstract

Operational modal analysis methods have been proven especially useful to identify existing structures and infrastructures under serviceability conditions. However, the installation of sensing systems for monitoring continuously an ever larger number of existing constructions has motivated significant efforts towards the automation of the available methods.Within this framework, the present paper introduces a new paradigm for the automatic output-only modal identification of linear structures under ambient vibrations, namely the intelligent automatic operational modal analysis (i-AOMA). It exploits the covariance-based stochastic subspace (SSI-cov) algorithm for the output-only identification of the modal parameters and its workflow consists of two main phases. Initially, quasi-random samples of the control parameters for the SSI-cov algorithm are generated. Once the SSI-cov algorithm is performed for each sample, the corresponding stabilization diagrams are processed in order to prepare a database for training the intelligent core of the i-AOMA method. This is a machine learning technique (namely a random forest algorithm) that predicts which combination of the control parameters for the SSI-cov algorithm is able to provide good modal estimates. Afterward, new quasi-random samples of the control parameters for the SSI-cov algorithm are generated repeatedly until a statistical convergence criterion is achieved. If the generic sample is classified as feasible by the intelligent core of the i-AOMA method, then the SSI-cov algorithm is performed. Finally, stable modal results are distilled from the stabilization diagrams and relevant statistics are computed to evaluate the uncertainty level due to the variability of the control parameters.The proposed i-AOMA method has been applied to identify the modal features of the Al-Hamra Firduos Tower, an iconic 412.6 m tall building located in Kuwait City (Kuwait). The final results well agree with a previous experimental study, and it was also possible to identify two new vibration modes of the structure. The implemented open-source Python code is made freely available.