3D vibration response of structures using TLS

3D terrestrial laser scanners are nowadays frequently used for documenting the geometry of infrastructural or industrial objects. They measure contactless and with high spatial density three- dimensional coordinates based on polar observations that are performed within a constant angular grid of horizontal and vertical angles at consecutive time steps. Current 3D terrestrial laser scanners achieve a maximal vertical rotation frequency of about 50 Hz. Therefore, despite measuring single points with a frequency of approx. 2 MHz, a single laser scan of a 3D scene containing an object of some meters in size usually takes at least several tens of seconds. The 3D scanning process and, consequently, the repetition rate of a 3D scene, are rather slow. This impedes capturing the vibration response of a measuring object by 3D terrestrial laser scanning. Enhancing the latter in this sense is however attractive due to the advantages of the measuring method under consideration, which consist in the contactless and areal capture of the geometry. Capturing the vibration response of infrastructures is important for assessing their serviceability and load-bearing capacity. Furthermore, the determination of resonance frequencies and natural modes of vibration allows drawing conclusions about the existence, localisation and assessment of damages. In view of the advanced age of many infrastructures, such a gain in knowledge is very relevant. This project aims at establishing a methodological procedure for the areal capturing of the vibration response of acquired structures, e.g. bridges. For this, a suitable approach is to be developed which combines the 3D-measurement data with a processing model. This processing model compensates the too low repetition rate of a 3D scene when using a terrestrial laser scanner and pushes the limits of the temporal resolution of the recorded geometric changes. Different processing models relating single measured points by temporally and spatially constraints to each other are deployed. The redundancy of the measurements generated in this way is used for estimating the parameters describing the vibration response of the structure. The deployed processing models combine different vibration models parameterizing the vibrations spatial and temporal behaviour - with different parameter estimation models, e.g. regression, Kalman filtering and artificial neural networks. Furthermore, they also include measurement models describing the 3D terrestrial laser scanners measurement uncertainty focused on short-term effects. The Research Division Engineering Geodesy at TU Wien and the Chair of Engineering Geodesy at TU Munich jointly conduct the project.