Nonlinear reduced-order model for vertical sloshing by employing neural networks

Abstract

The aim of this work is to provide a reduced-order model to describe the dissipative behavior of nonlinear vertical sloshing involving Rayleigh–Taylor instability by means of a feed forward neural network. A 1-degree-of-freedom system is taken into account as representative of fluid–structure interaction problem. Sloshing has been replaced by an equivalent mechanical model, namely a boxed-in bouncing ball with parameters suitably tuned with performed experiments. A large data set, consisting of a long simulation of the bouncing ball model with pseudo-periodic motion of the boundary condition spanning different values of oscillation amplitude and frequency, is used to train the neural network. The obtained neural network model has been included in a Simulink® environment for closed-loop fluid–structure interaction simulations showing promising performances for perspective integration in complex structural system.

Sloshing reduced-order model based on neural networks for aeroelastic analyses

“…However, they suffer in being able to characterise well the dissipative capacity of the fluid at different values of amplitude and frequency as can be seen by comparing the energy maps provided in Refs. [10,16].…”

“…On the other hand, this work exploits the approach recently introduced in Ref. [16] for the generation of neural-network based ROM for vertical sloshing, by employing low-fidelity data and finally improved and experimentally validated for a sloshing beam in Ref. [17].…”

Design and GVT of a dynamically scaled wing structure for fuel sloshing investigations

“…Such experimental work is used, beside providing an understanding of the fundamental physical interactions, to create equivalent mechanical models (EMM) and reduced order models (ROM) with predictive capabilities. For vertical sloshing, such models involve bouncing balls interacting with vertically oscillating tanks, such as the models presented in [1,7]. De Courcy et al [8] proposed a bouncing-ball model coupled with an aeroelastic wing model that showed dissipative effects at multiple gust lengths.…”

Neural-network-based reduced order modeling for nonlinear vertical sloshing with experimental validation

“…An equivalent mechanical model consisting of a bouncing ball capable to reproduce the impact mechanisms, was proposed by Refs. [21,22]). These models provide fast prediction of sloshing forces, but since they are obtained by fitting the sloshing forces -or its induced damping -of the free response of a single degree of freedom system, they do not generally provide consistent results when working at frequencies different from those used in their identification.…”