Design of experiments for MR damper modelling

Jorge Lozoya-Santos, Ruben Morales-Menendez, Ricardo Ramirez-Mendoza

Research output: Contribution to conferencePaper

10 Citations (Scopus)

Abstract

An approach to find out the training inputs for identification of a Magneto-Rheological (MR) damper is proposed. Reduction of overuse of the damper, number of experiments and configurations of training inputs are main features of this approach. Experimental validation with a commercial MR damper was carried out.Main results show inputs configuration with modulated frequency at fixed amplitude displacement, and random amplitude step with fixed period generate key information. A feed-forward neural network was selected as model emulator. Modelling results showed an error-to-signal ratio lower than milli-thousands. © 2009 IEEE.
Original languageEnglish
Pages1915-1922
Number of pages8
DOIs
Publication statusPublished - 18 Nov 2009
Externally publishedYes
EventProceedings of the International Joint Conference on Neural Networks -
Duration: 18 Nov 2009 → …

Conference

ConferenceProceedings of the International Joint Conference on Neural Networks
Period18/11/09 → …

Fingerprint

Feedforward neural networks
Design of experiments
Experiments

Cite this

Lozoya-Santos, J., Morales-Menendez, R., & Ramirez-Mendoza, R. (2009). Design of experiments for MR damper modelling. 1915-1922. Paper presented at Proceedings of the International Joint Conference on Neural Networks, . https://doi.org/10.1109/IJCNN.2009.5179003
Lozoya-Santos, Jorge ; Morales-Menendez, Ruben ; Ramirez-Mendoza, Ricardo. / Design of experiments for MR damper modelling. Paper presented at Proceedings of the International Joint Conference on Neural Networks, .8 p.
@conference{699b268bdc764d1e9736cfc9bb4f5aaf,
title = "Design of experiments for MR damper modelling",
abstract = "An approach to find out the training inputs for identification of a Magneto-Rheological (MR) damper is proposed. Reduction of overuse of the damper, number of experiments and configurations of training inputs are main features of this approach. Experimental validation with a commercial MR damper was carried out.Main results show inputs configuration with modulated frequency at fixed amplitude displacement, and random amplitude step with fixed period generate key information. A feed-forward neural network was selected as model emulator. Modelling results showed an error-to-signal ratio lower than milli-thousands. {\circledC} 2009 IEEE.",
author = "Jorge Lozoya-Santos and Ruben Morales-Menendez and Ricardo Ramirez-Mendoza",
year = "2009",
month = "11",
day = "18",
doi = "10.1109/IJCNN.2009.5179003",
language = "English",
pages = "1915--1922",
note = "Proceedings of the International Joint Conference on Neural Networks ; Conference date: 18-11-2009",

}

Lozoya-Santos, J, Morales-Menendez, R & Ramirez-Mendoza, R 2009, 'Design of experiments for MR damper modelling' Paper presented at Proceedings of the International Joint Conference on Neural Networks, 18/11/09, pp. 1915-1922. https://doi.org/10.1109/IJCNN.2009.5179003

Design of experiments for MR damper modelling. / Lozoya-Santos, Jorge; Morales-Menendez, Ruben; Ramirez-Mendoza, Ricardo.

2009. 1915-1922 Paper presented at Proceedings of the International Joint Conference on Neural Networks, .

Research output: Contribution to conferencePaper

TY - CONF

T1 - Design of experiments for MR damper modelling

AU - Lozoya-Santos, Jorge

AU - Morales-Menendez, Ruben

AU - Ramirez-Mendoza, Ricardo

PY - 2009/11/18

Y1 - 2009/11/18

N2 - An approach to find out the training inputs for identification of a Magneto-Rheological (MR) damper is proposed. Reduction of overuse of the damper, number of experiments and configurations of training inputs are main features of this approach. Experimental validation with a commercial MR damper was carried out.Main results show inputs configuration with modulated frequency at fixed amplitude displacement, and random amplitude step with fixed period generate key information. A feed-forward neural network was selected as model emulator. Modelling results showed an error-to-signal ratio lower than milli-thousands. © 2009 IEEE.

AB - An approach to find out the training inputs for identification of a Magneto-Rheological (MR) damper is proposed. Reduction of overuse of the damper, number of experiments and configurations of training inputs are main features of this approach. Experimental validation with a commercial MR damper was carried out.Main results show inputs configuration with modulated frequency at fixed amplitude displacement, and random amplitude step with fixed period generate key information. A feed-forward neural network was selected as model emulator. Modelling results showed an error-to-signal ratio lower than milli-thousands. © 2009 IEEE.

U2 - 10.1109/IJCNN.2009.5179003

DO - 10.1109/IJCNN.2009.5179003

M3 - Paper

SP - 1915

EP - 1922

ER -

Lozoya-Santos J, Morales-Menendez R, Ramirez-Mendoza R. Design of experiments for MR damper modelling. 2009. Paper presented at Proceedings of the International Joint Conference on Neural Networks, . https://doi.org/10.1109/IJCNN.2009.5179003