Uncertainty Quantification in Applied Mechanics

Organisers

• Alice Cicirello, TU Delft, Netherlands
• Edoardo Patelli, University of Strathclyde, UK

Abstract

Several effective Uncertainty Quantification (UQ) strategies have been developed to address problems in applied mechanics (e.g. structural dynamics, Structural Health Monitoring, System identification, Statistical model updating, and sensor placement) in the past 20 years, yet there is still a tremendous amount of challenges to be addressed. Broadly speaking, the most pressing challenges include:

(a)the characterization of uncertainties given mixed type of information (measurements, observations, reports, experts’ opinion, previous knowledge) and because of: (i) limited previous knowledge and/or data to learn from towards the future performance (e.g. few tests carried out on small samples); (ii) limited knowledge of the manufacturing variability resulting from the manufacturing process; (iii) uncertainty in operating and environmental conditions; and (iv) computational models not capable of capturing completely the physics of the problem at hand.

(b)The propagation of uncertainty to quantify: (i) the uncertainty in the performance metrics of interest; (ii) the uncertainty in the latent parameters of the computational model after measurements of observable quantities become available.

This minisymposium welcomes contributions showcasing novel advanced techniques and industrial applications (including civil, mechanical, aerospace, manufacturing, nuclear and other related engineering disciplines) addressing these challenges.

Keywords: Uncertainty Quantification; System Identification; Structural Health Monitoring; Statistical Model updating; Uncertainty Modelling; Non-probabilistic uncertainty