Neuro-Fuzzy Control for Semi-active Dampers - Fuzzy Modeling and Control: Theory and Applications - page 174

Moreover, the backpropagation gradient algorithmhas been applied in the learning
algorithm. Once the ANFIS system has been trained, it is included in a simulation
platform along with the rest of helicopter pieces. As the inputs are the frequencies,
it is required a FFT block between the vibration produced by the helicopter and
the ANFIS. These blocks are in charge of obtaining the spectral composition of the
signal on each axes applying the Fast Fourier Transform at intervals of 0.5 s. Hence,
the ANFIS analyzes the input frequencies and provides the best C value at each 0.5 s.
In this way, the damping value provided, is introduced into the semi-active control
system. Note that the actuator in this case consists in a semi-active damper placed
between the fuselage and the base of the vision system.
9.3 Semi-active Damper for Vibration Control
in Footbridges
Floor vibrations are very common problems since nowadays engineering structures
tend to be light, flexible and therefore, they become sensitive to these vibrations
which can be as small as order 1mm, still producing discomfort when walking on the
floor. Particularly, footbridges are subjected to external disturbances on daily basis.
According to several studies (Casado et al. 2008 ; Sanchez 2011 ) vibrations can be
generally triggered by pedestrians who can induce a footbridge response near to the
resonance. This contributes in detriment of pedestrian comfort and in some cases, it
could compromise its safety. Therefore, it is advisable to consider the incorporation
of a device in these types of structures such that undesirable vibrations are suppressed.
In recent years, Tuned Mass Dampers (TMDs) are being used for this purposes due
to their effectiveness in damping vibrations in both machines and structures. It is
a subsystem consisting of a mass, a spring, and a damper that is attached to the
vibrating system (expected to be isolated). In order to apply this method in our study,
it has been decided to use a semi-active TMD, instead of a conventional passive
TMD. The reason for considering semi-active methods (Yoshioka et al. 2002 )are
because their cost is lower and they provide reliable solution for the problem of
vibration isolation/suppression. They have similar architecture as a passive damping
methodology, but with a controlled damper, as it could be seen in Fig. 9.4 .
This provides better vibration reduction on the structures under consideration
since the damping properties can be modified at any time. In this case, a magneto-
rheological (MR) damper is adopted as isolation device in order to reduce the vibra-
tions suffered by the footbridge. These devices use controllable fluids composed by
micron-sized, magnetically polarizable particles dispersed in a fluid. Their properties
are changed when a magnetic field is applied, since particles are stood in chains form
modifying the fluid behavior.
There exist many mechanical models for predicting the response of a MR damper
(Dyke et al. 1996 ). In this work the used model consists of a Bouc-Wen model
in parallel with a damper (Yoshioka et al. 2002 ). These systems aim to reduce
the dynamic response of a structure that is symbolized by M 1 and a spring-damper
(K 1 -C 1 ). This is achieved with a body (M 2 ), a passive spring K 2 and an adaptive