Civil and Environmental Engineering Department Seminar: Stochastic Mechanics and Wavelets Techniques for Response Analysis, Parameter Identification, and Uncertainty Propagation in Complex Systems

Seminar | March 3 | 11 a.m.-12 p.m. | Davis Hall, 542 Davis Hall

 Ketson Roberto Maximiano dos Santos

 Civil and Environmental Engineering (CEE)

Abstract:

In the first part of the talk, a novel stochastic dynamics technique based on a Hilbert transform stochastic averaging treatment is developed for determining the response and assessing the reliability of diverse nonlinear systems and structures subject to random excitations. A significant advantage of the technique relates to the fact that it can account for systems endowed with fractional derivative, which constitute a generalization of classical calculus and have been successfully employed in engineering mechanics and biomechanics for viscoelastic material modeling. Further, this technique is used in the development of a computationally efficient stochastic version of the Incremental Dynamic Analysis (IDA) framework. Next, wavelets and sparse representation techniques are exploited for developing an identification technique of nonlinear time-variant structures endowed with fractional derivative elements, even in cases of limited or incomplete measured data. Furthermore, wavelet and sparse representation techniques are employed in biomedical engineering applications showing their versatility in solving multidisciplinary problems.

In the second part of the talk, a novel method for dimension hyper-reduction of the solutions of numerical models is developed based on the application of diffusion maps on the Grassmann manifold such that the hyper-reduced solutions retain the essential geometric structure of the full solutions; a probability measure can be inferred for the hyper-reduced solutions, which will enable sampling on the low-dimensional manifold; and the geometric structure of the underlying manifold can be exploited for interpolation purposes to efficiently approximate the solution at new parameter values. Indicative applications are found in structural engineering and material modeling.

Bio:

Ketson R. M. dos Santos received his B.S. (2012) in Civil Engineering from the Federal University of Alagoas, Brazil, and his M.S. (2014) in Civil Engineering (Structural Engineering) from the University of São Paulo, Brazil. He also holds a M.Phil. (2018) and a Ph.D. (2019) in Civil Engineering & Engineering Mechanics from Columbia University, USA. His doctoral research primarily focused on the development of efficient techniques for uncertainty propagation and reliability analysis of dynamical systems excited by random loadings. Moreover, he developed signal processing and data-analysis techniques for nonlinear system parameter identification and for modeling of biological systems. Dr. dos Santos is currently appointed as a Postdoctoral Fellow in the Department of Civil and Systems Engineering at Johns Hopkins University, USA, where he is also co-developer of UQpy (https://sites.google.com/site/jhusurg/UQpy), a general-purpose Python toolbox for modeling uncertainty in the simulation of physical and mathematical systems. Further, he was recently appointed as an Early Career Editorial Board (ECEB) member for the ASCE-ASME Journal of Risk and Uncertainty in Engineering Systems, Part B: Mechanical Engineering.

 CA, pong@berkeley.edu, 5106645249