An efficient reduced order model for control of traffic-induced stresses in an adaptive steel girder bridge

The main objective of this dissertation was to determine the most appropriate methodology for creating a reduced order model (ROM) from a higher order model of an adaptive steel girder bridge subject to vehicular loading. A ROM is necessary to be able to design the control system for a large order system. The model for the vehicle was a multi-axle, 7-degree of freedom suspension mass system with damping. The bridge was modeled using finite elements. Actuators were considered to be a part of the adaptive structure and roadway roughness was included. MATLABRTM programming was used for modeling and simulation. Models and results were verified and validated.; Four categories of reduced order methodologies were evaluated in detail. These include Guyan's Reduction, Modal Analysis, Ritz Vector Reduction and Component Mode Synthesis. After initial evaluation, three ROM methodologies were selected on which to apply the control system. Lyapunov control law was selected due to the nonlinear dynamics of the adaptive bridge structure and of the semi-active hydraulic actuators. Each ROM was evaluated based on its ability to control displacements, velocities and flexural stress in the bridge and the degree of difficulty encountered in utilizing the methodology. The most efficient ROM was then selected. Simulations show that a control system, so designed and implemented, can be used to reduce fatigue stresses in the bridge and so extend its useful life.; The ROM created using Guyan's Reduction Methodology was the most efficient ROM in controlling the traffic-induced flexural stresses and in extending the remaining safe life of the bridge. This ROM predicted flexural stress range reductions of 12.3% and a 25.6 year extension of the remaining safe life of the bridge.