Neuro-Fuzzy Control for Semi-active Dampers - Fuzzy Modeling and Control: Theory and Applications - page 170

vibration reduction are usually designed in order to generate an anti-vibration field, as
close as possible to the system vibration (measured by a sensor), but with an opposite
phase. In order to determine an anti-vibration field, a mathematical model of the sys-
tem should be included in the working environment. The most common techniques in
this sense are Higher Harmonic Control (HHC), Individual Blade Control (IBC) and
Active Control of Structural Response (ACSR) (Anusonti-Inthra 2002 ). As it can be
deduced, active elements are more reliable than passive elements. However, active
elements in general, require large external power systems. Semi-active elements have
become a widely used approach in order to take advantage from both passive and
active elements, mitigating their respective disadvantages. Semi-active based vibra-
tion systems need less power requirements since they only modify the structural
properties of the system. This also implies that control system stability is almost
guaranteed for the semi-active devices case (Dunbabin et al. 2004 ; Anusonti-Inthra
2002 ). Many researchers have proposed different semi-active strategies to deal with
the vibration suppression problem. As it can be seen below, many of the effort in this
field is related to the application of magneto-rheological fluid dampers (Guglielmino
et al. 2008 ; Paciello and Pietrosanto 2011 ). In fact, Kamath et al. ( 1998 ) carried
out an extensive comparative study of fluid-elastomeric and magneto-rheological
dampers, concluding that the fluid-elastomeric dampers show a mild degradation in
stiffness and damping under dual frequency excitation conditions while magneto-
rheological ones showed no modifications in their properties. Jansen and Dyke
( 2000 ) evaluated a set of semiactive algorithms (Lyanpunov controller, decentralized
bang-bang controller, modulated homogeneous friction algorithm, and a clipped
optimal controller) for magneto-rheological dampers.
Nitzsche ( 1996 ) considered a BO-105 type dynamically scaled model rotor in
order to introduce the concept of stiffness variation. Gandhi and Anusonti-Inthra
( 2003 , 2012 ) use a semiactive controllable stiffness device for tonal disturbance
rejection applications. The spring coefficient of this device can be modulated in
real-time through a frequency-domain design control algorithm. In their studies,
the authors point that the designed semi-active isolation methods could produce an
extra 30% vibration reduction if it is compared to a passive isolator. Titurus ( 2013 )
designed a systembased on semi-active hydraulic lag dampers for vibration control in
helicopters by using periodic flow restrictor modulation. Moreover, other researchers
have proposed new approaches in order to reduce helicopter vibrations (Zhao and
Were l ey 2004 ; Kothera et al. 2011 ). Indeed, Agarwal ( 2005 ) had previously proposed
a semi-active friction based lead-lag damper as a replacement for hydraulic and
elastomeric dampers, where damping is provided by optimized energy dissipation
due to frictional forces in semi-active joints. Amongst all of them, Fuzzy-based
approaches are emerging as promising techniques, Kasemi et al. ( 2012 ) propose a
Fuzzy-PID controller for semi-active vibration control using magneto-rheological
fluid dampers. In the following, an approach based on Neuro-Fuzzy techniques will
be presented. This Neuro-Fuzzy controller was designed for autonomous small-
scale helicopters. In this particular case, a Caliber 30 was used. It is a very stable
radio controlled model that is manufactured by the Japanese company Kyosho. The
main and tail rotors are both powered by an internal combustion engine that burns a