Timothy Rogers

I am currently a PhD student, supervised by Dr. E.J. Cross, Dr. G. Manson, and Prof. K. Worden. I joined the DRG in September 2015 after completing an MEng. in Mechanical Engineering at the University of Sheffield. My masters project was titled “On the Behaviour of Structures with Many Nonlinear
Elements” supervised by Prof. K. Worden. 

Research interests:

My research interests revolve around using machine learning methods, specifically Bayesian methods and Gaussian Processes, to add value in an engineering context. Of particular interest is the role of these methods in Structural Health Monitoring (SHM) and nonlinear system identification.

Current research:

My research is currently focused on structural monitoring and nonlinear system identification on offshore platforms working with the Structural Monitoring Systems department at Ramboll. Our research falls into two main themes, the first being to understand the physical mechanisms driving nonlinear behaviour on the structures of interest; the second being the combination of engineering insight with state of the art machine learning techniques to better model physical processes.

A particular interest is in the problem of wave loading on offshore structures, where the current physical models are not sufficient to fully describe the complex nonlinear loading that the structures experience. Since understanding the loading experienced by a structure is key to understanding possible damage and the dynamic behaviour, it is hoped that a model of wave loading with better predictive accuracy will be able to facilitate better analysis of the dynamic behaviour of the structure.

Figure showing OSA wave loading prediction using a Gaussian Process NARX model. The results show very good predictions (normalised mean square error less than 3% on the mean prediction), more importantly the model is able to accurately quantify the uncertainty in the model and measurements in terms of confidence intervals. Even in regions where the mean prediction is not as accurate the confidence intervals display this model uncertainty.

Publications:

  • Rogers, T.J., Holmes G.R., Cross E.J., Worden K. (2017). On a Grey Box Modelling Framework for Nonlinear System Identification. (Presented at IMAC XXXV, publication pending)
  • Rogers, T. J., Manson, G., & Worden, K. (2016). On the Behaviour of Structures with Many Nonlinear Elements. In Rotating Machinery, Hybrid Test Methods, Vibro-Acoustics & Laser Vibrometry, Volume 8 (pp. 509-520). Springer International Publishing.