Identification of Modal Characteristics in Order to Monitor Bridges' Health in Time Domain Based on Complexity Pursuit Method

Nowadays, maintenance and health monitoring of structures and infrastructures of the country are the top priority of engineers. Therefore, acceleration sensors, which are installed on the bridge are used to extract the bridge's modal response, and even the latest studies have shown that sensors that have been attached to the car's chassis can record the bridge's response to vibration. Thus, it is important to use an accurate method, which is based on the output-only, to measure frequency, damping ratio, and mode shape. As a result, practical modal analysis is used for that purpose. In this paper, a new method of blind source separation, called complexity pursuit (CP), is investigated which is able to measure and extract dynamic characteristics in the time domain. Complexity pursuit is a well-known method, that is able to recover hidden sources (independent responses of every degree of freedom) and it can calculate the mixing matrix from measured mixed output data only. Some various conditions are considered in this paper to verify complexity pursuit (3 degrees of freedom mass-spring with various damping ratios, in white Gaussian noise conditions, in non-diagonal damping matrix, close modes, and for 12 degrees of freedom mass-spring system). Comparison of the CP results and modal analysis under all the conditions showed that the proposed method is able to detect, measure and extract the modal characteristics with an error of less than 3% and in most cases less than 1%.