Modelo inverso de amortiguador magneto-reológico basado en redes neuronales

Gómez Fernández, Agustín

Publication:
Modelo inverso de amortiguador magneto-reológico basado en redes neuronales

dc.contributor.advisor	López Boada, María Jesús
dc.contributor.author	Gómez Fernández, Agustín
dc.contributor.departamento	UC3M. Departamento de Ingeniería Mecánica	es
dc.date.accessioned	2015-03-04T20:11:26Z
dc.date.available	2015-03-04T20:11:26Z
dc.date.issued	2013
dc.date.submitted	2013-07-09
dc.description.abstract	The suspension of a vehicle plays an important role in both the stability and the comfort of the occupants. There are three types of suspension, the suspension first passive and active suspension semi-active suspension. Passive suspension is the simplest of all, to record such a shock absorber oil and typically a helical spring. The semi-active suspensions and active have a higher complexity but also better results. Active suspension include controlling the stability of each wheel independently, and consumes large power since it must be equipped with a large number of sensors that can detect variables produced during the movement of the vehicle, and actuators, which act accordingly to the sensed variables. This complex system of sensors and actuators price considerably expensive suspension therefore be handled another option, such as semi-active suspension. One of the main research is to install magneto-rheological dampers in semi-active suspensions, such as amortigudores respond, varying viscous stress to an applied magnetic field. The sensors detect a number of variables such as force applied to the damper, its speed, and frequency shift, so must be a current inside the damper that generates a magnetic field in response to the variables mentioned. There are different models which identify a relationship between force and current. On one side are the direct models, which attempt to find the relationship between the frequency, amplitude and current strength. Among the direct models, parametric models are based on differential or algebraic equations, and non-parametric models. On the other hand, are the inverse models, which relate the strength, frequency and amplitude with the flow that must be passed by the MR damper. In this final degree work presents a magneto-rheological damper based on neural networks, because neural networks are an excellent resource for establishing a non-linear relationship between the strength, frequency and offset against the current. Proposed and studied various network topologies, in order to train these different patterns by training and validation. Once trained to obtain the optimal network, it will be able to generate current in the damper to exert force indicating the driver.	es
dc.description.degree	Ingeniería Mecánica	es
dc.format.mimetype	application/pdf
dc.identifier.uri	https://hdl.handle.net/10016/20170
dc.language.iso	spa	es
dc.rights	Atribución-NoComercial-SinDerivadas 3.0 España	*
dc.rights.accessRights	open access	es
dc.rights.uri	http://creativecommons.org/licenses/by-nc-nd/3.0/es/	*
dc.subject.eciencia	Ingeniería Mecánica	es
dc.subject.other	Amortiguadores	es
dc.subject.other	Redes neuronales	es
dc.subject.other	Sistemas de suspensión mecánica	es
dc.subject.other	Tecnología automovilística	es
dc.subject.other	Fluidos magneto-reológicos	es
dc.title	Modelo inverso de amortiguador magneto-reológico basado en redes neuronales	es
dc.type	bachelor thesis	*
dspace.entity.type	Publication