Using Instrumented Vehicles to Detect Damage in Bridges

This paper describes a ‘drive-by’ method of bridge inspection using an instrumented vehicle. Accelerometers on the vehicle are proposed as a means of detecting damage on the bridge in the time it takes for the vehicle to cross the bridge at full highway speed. For a perfectly smooth road profile, the method is shown to be feasible. Changes in bridge damping, which is an indicator of damage, are clearly visible in the acceleration signal of a quarter-car vehicle on a smooth road surface modelled using MatLab. When road profile is considered, the influence of changes in bridge damping on the vehicle acceleration signal is much less clear. However, when a half-car model is used on a road with a rough profile, it is again possible to detect changes in bridge damping, provided the vehicle has two identical axles.

Dynamic axle force and road profile identification using a moving vehicle

The axle forces applied by a vehicle through its wheels are a critical part of the interaction between vehicles, pavements and bridges. Therefore, the minimisation of these forces is important in order to promote long pavement life spans and ensure that bridge loads are small. Moreover, as the road surface roughness affects the vehicle dynamic forces, the monitoring of pavements for highways and bridges is an important task. This paper presents a novel algorithm to identify these dynamic interaction forces which involves direct instrumentation of a vehicle with accelerometers. The ability of this approach to predict the pavement roughness is also presented. Moving force identification theory is applied to a vehicle model in theoretical simulations in order to obtain the interaction forces and pavement roughness from the measured accelerations. The method is tested for a range of bridge spans in simulations and the influence of road roughness level on the accuracy of the results is investigated. Finally, the challenge for the real-world problem is addressed in a laboratory experiment.