On the robustness of a model-based inverse force identification applied on a structure submerged in different media
- PDF / 2,349,861 Bytes
- 13 Pages / 595.276 x 790.866 pts Page_size
- 105 Downloads / 200 Views
(2020) 42:511
TECHNICAL PAPER
On the robustness of a model‑based inverse force identification applied on a structure submerged in different media Ricardo P. Álvarez‑Briceño1 · Frank Naets2,3 · Wim Desmet2,3 · Leopoldo P. R. de Oliveira4 Received: 25 September 2019 / Accepted: 19 August 2020 © The Brazilian Society of Mechanical Sciences and Engineering 2020
Abstract This paper addresses an experimental study on the implementation of an Augmented Kalman filter algorithm complemented by displacement dummy measurements, aiming at identifying a stochastic point force. For this purpose, the present experimental approach features a cantilevered structure instrumented with two pairs of accelerometers. A force sensor is used to measure the actual input force for benchmarking. The main objective of this study is to evaluate the ability of the algorithm to predict forces when the system is perturbed by different fluids surrounding the structure. Firstly, the structure is identified via experimental modal analysis, in two conditions, in air and underwater. Predicted and measured forces, for tests in each media, are compared showing good agreement. Additionally, the method is used to estimate forces applied when the structure is submerged in water while using a system model identified in air, in order to assess the algorithm robustness in scenarios that are either difficult or impossible to be tested. Although results accuracy in such cross-identification conditions depend on the closeness between the reference model and the actual boundary condition, in general, the frequency content of the predicted forces match with those predicted in the direct scheme, allowing qualitative data assessment in an otherwise unfeasible scenario. Keywords Cantilever beam · Force estimation · Kalman filter · Structural dynamics · Inverse problems
1 Introduction Typical problems in engineering are posed as forward problems, where the response is calculated from the system model and a known input. Inverse problems search for the input based on a system model (obtained analytically or based on experiments) and its responses. The latter is often more challenging, mainly due to: (1) often falling into Technical Editor: Thiago Ritto. * Leopoldo P. R. de Oliveira [email protected] 1
Mechanical Engineering Department, Escuela Politécnica Nacional, Ladrón de Guevara E11 ‑ 253, P.O.Box 17‑01‑2759, Quito, Ecuador
2
Department of Mechanical Engineering, KU Leuven, Celestijnenlaan 300B, 3001 Heverlee, Belgium
3
DMMS Lab, Flanders Make, 3001 Heverlee, Brazil
4
São Carlos School of Engineering, University of São Paulo, Av. Trabalhador Sancarlense 400, São Carlos, SP 13566‑590, Brazil
ill-conditioned issues, vulnerable to measurement noise [1] and modeling uncertainties, (2) force reconstruction being performed from a limited number of sensors and (3) inverse problems being susceptible to indeterminacy such as singularities or non-uniqueness of solution, thus requiring a higher level of analytical rigor [2]. Despite these difficulties, sol
Data Loading...
