Browse/search for people

Publication - Professor Colin Taylor

    Interrogation of 'big data' from shaker table testing using virtual reality

    Citation

    Turton, L, Crewe, A, Kloukinas, P, Oddbjornsson, O, Dietz, M, Dihoru, L, Horseman, T, Voyagaki, E & Taylor, C, 2018, ‘Interrogation of 'big data' from shaker table testing using virtual reality’. in: Proceedings of the 16th European Conference on Earthquake Engineering. European Association for Earthquake Engineering (EAEE)

    Abstract

    This paper describes how VRML (Virtual Reality Modelling Language) techniques can be used to quickly visualise and interpret large multidimensional experimental data sets. The data sets were created by a series of experimental tests investigating the seismic response of the UK’s Advanced Gas Cooled Reactors (AGRs). At the University of Bristol a quarter scale physical model of an AGR core was developed to support, via experimental testing the existing numerical models that assess the seismic resilience of the AGRs. Outputs from the rig consist of highly detailed acceleration and displacement datasets that contain all the challenges of ‘big data’ processing. A particular challenge has been the interpretation of the response data from even a single shake, which comprises 6DOF time dependent displacement responses of many components. Traditionally this data would be viewed as a series of time histories, but the combination of hundreds of linked 6DOF motions, which all vary with time, makes interpretation of such plots practically impossible. To combat these problems a 3D VRML model including all the instrumented components has been developed. The test data is mapped into this VRML model to produce a 3D animation with the ability to pan, tilt, rotate and zoom in. This allows the user, in real time, to discover any behaviour that looks significant. The VRML model can also be used to look at responses from different tests and allow visual comparison of multiple data sets. This visualisation technique also has applicability for interpretation of other large experimental data sets.

    Full details in the University publications repository