DocumentCode :
2713566
Title :
Design of experiments for MR damper modelling
Author :
Lozoya-Santos, Jorge ; Morales-Menendez, Ruben ; Ramirez-Mendoza, Ricardo
Author_Institution :
Tecnol. de Monterrey, Monterrey, Mexico
fYear :
2009
fDate :
14-19 June 2009
Firstpage :
1915
Lastpage :
1922
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.
Keywords :
damping; feedforward neural nets; magnetorheology; nonlinear control systems; shock absorbers; vibration control; MR damper modelling; feedforward neural network; magnetorheology; Damping; Shock absorbers;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Neural Networks, 2009. IJCNN 2009. International Joint Conference on
Conference_Location :
Atlanta, GA
ISSN :
1098-7576
Print_ISBN :
978-1-4244-3548-7
Electronic_ISBN :
1098-7576
Type :
conf
DOI :
10.1109/IJCNN.2009.5179003
Filename :
5179003
Link To Document :
بازگشت