Bridge Structural Identification Using Moving Vehicle Acceleration Measurements

Identification of dynamic characteristics of structures is a desired objective for existing infrastructure and has been accounted as a serious challenge for civil engineers. In this research, a structural identification method is proposed, which is capable of identifying dynamics of structures using sensor data inside vehicles passing over a bridge. The methodology utilizes a special type of identification algorithm facilitated by Expectation Maximization (STRIDEX) that is capable of identifying systems using mobile data networks. In this study, it is assumed that the mobile sensor measurements are the accelerations inside rigid vehicles and are primarily a mixtures of accelerations caused by the road roughness and bridge dynamic acceleration. With this regard, a stochastic State-Space model represents the equation of motion for a linear dynamic vehicle-bridge system consisting of an impure input. The observation vector is treated as a linear mixture of two sources that are not known. Therefore, the problem turns to a Blind Source Separation (BSS) procedure that is aiming to draw out the bridge vibrations from the mixture. An algorithm called Second Order Blind Identification (SOBI) has been utilized for source separation and validated using simulation. The entire algorithm, including both SOBI and STRIDEX acting together, could successfully identify natural frequencies and mode shapes of a numerical bridge model.