Transcript Slide 1

Civil Engineering Applications of
Vibration Control
(Structural Control)
Naresh K. Chandiramani, Associate Professor
Room 141, Dept. of Civil Engineering
Structural Control
• Vibration control of civil structures is more recent as
compared to machines & aerospace vehicles.
• Earthquakes and wind loads - main sources of
structural vibrations.
• Control vibrations by: changing rigidity, mass,
damping, shape, or applying passive or active
control forces.
• > 20 full scale active control appl. in Japan
• Passive base isolation used in USA.
• Retrofitting reqd. if new seismic activity detected
• High strength may result in high acceleration levels,
so increasing strength alone wont always work.
Structural control versus
Mechanical & Aerospace control
• Environmental disturbances (wind, earthquake
excitations) occur over wide range of frequency
and amplitudes, i.e., they are uncertain, whereas
mechanical loads are usually deterministic.
• Civil structures (without control) are stable and
may get destabilized with active control,
whereas aerospace structures require active
control for stabilization.
• Performance specifications for civil structures
are coarse (e.g., peak amplitude, time for motion
to settle down).
Mathematical model of structure.
Fig. 1: (a) Mathematical model, (b) Schematic of a building
•In a simplified model, the masses correspond to
slab masses and stiffnesses correspond to column
stiffnesses (i.e, the force required per unit lateral
displacement of column)
Passive control: Base isolation
Fig. 2: (a) Schematic of base isolated building, (b) Model, (c) Rubber bearing
Passive control: Base isolation
• Structure mounted on a suitably flexible base such that
the high frequency component of ground motion is
filtered out and the fundamental vibration period is
lengthened. This results in deformation in the isolation
system only, thus keeping the structure above almost
rigid. However, if the earthquake excitation contains a
major component of this fundamental period, there will be
large sidesway (albeit almost rigid) motions.
• San Fransisco city hall (retrofitted, 530 rubber bearings),
International terminal at SF airport (267 Friction
pendulum sliding bearings).
• Not suitable for tall slender buildings (subject to high wind
loads). For these auxiliary dampers (viscous,
viscoleastic) are deployed (eg. WTC).
Passive control: Tuned Mass
Damper (vibration absorber)
 
2
Fig. 3: (a) TMD schematic, (b) Response
Passive control: Tuned Mass
Damper (vibration absorber)
• TMD, usually having mass about 1% that of structure,
fitted to top of building. It is tuned to reduce vibration for
given frequency range.
• Absorber mass takes up vibratory energy, leaving the
main mass (building) almost static.
• Not very useful for earthquake excitations which occur
over wide frequency range.
• Main system properties (stiffness-k1, mass-m1) known,
absorber system properties (stiffness-k2, mass-m2) to
be designed such that absorber frequency equals
excitation frequency (w2=w).
• Examples: John Hancock Tower (Boston), Citicorp
Building (New York).
Passive control:
Untuned viscous absorber
Fig. 4: (a) Model of untuned viscous absorber, (b) Response
Types of passive control devices
• Metallic yield damper: relies on the principle that
the metallic device deforms plastically, thus
dissipating vibratory energy. Used in earthquake
applications.
• Friction devices: here friction between sliding
faces is used to dissipate energy. When used in
base isolation systems, the friction coefficient
has conflicting requirements. It should not be too
large otherwise shear forces from ground during
a strong earthquake will transmit to the structure.
Also it should not be too small or the entire
structure will move due to small/medium
wind/earthquake loads. These devices can also
be fitted between two storeys to damp their
relative motion. Used in earthquake applications.
Types of passive control devices
• Viscous/ Viscoelastic devices: Example is fluid in
a cylinder with piston having an orifice. These
can also be semi-active (eg., variable orifice,
variable viscosity). Used in earthquake and wind
applications.
• Tuned mass dampers: problems are size of the
mass to be used and its displacement relative to
the structure, in order that damping is effective.
• Liquid sloshing dampers, Impact dampers.
Classification of Control Methods
Active/Feedback control:
• External source of power drives actuators (i.e., provides
input voltage) .
• Voltages required are computed by controller using
certain algorithms with inputs from sensors.
• Sensors measure motion (strains, displ, vel, accl.)
• Actuators apply forces to structure, thereby adding or
dissipating energy.
• Examples of sensors are acceleromters, strain gauges.
• Examples of actuators are tendons, solenoids,
piezoelectric stacks, active mass dampers (AMD).
• Destabilization possible.
• External power may not be available during earthquake.
Classification of Control Methods
Passive control:
• No external power required.
• Passive control device (TMD, Base Isolator) imparts
forces that are developed directly as a result of motion of
structure (i.e., no actuator involved).
• Total energy (structure + passive device) cannot increase,
hence inherently stable.
• Relatively inexpensive.
• Reliable during earthquake
• Not as effective as active, hybrid, semi-active control.
Classification of Control Methods
Hybrid control:
• Uses active & passive devices.
• Advantages of both active and passive systems are
present and their limitations are reduced.
• Essentially an active control system
• Examples: viscous damping with AMD, base isolation with
actuators, TMD+AMD).
Classification of Control Methods
Semi-active control:
• Uses devices where input power requirements are orders
of magnitude less than fully active devices. In fact in some
cases battery power is sufficient.
• These devices usually don’t add energy to the system,
hence stability ensured.
• These devices can be viewed as controllable passive
devices (eg., Magneto-Rheological Fluid damper where
voltage input applied to change viscosity depending on
motion measured by sensors, variable orifice damper,
controllable friction devices, variable stiffness devices).
Active Control
• The goal is to design a control system to
keep stresses/strains/displ./accel. (called
outputs) at certain locations below
specified bounds (peak, rms) when
disturbances (wind, earthquake) below
specified bound are applied.
• Designer decides choice of outputs based
on comfort (e.g. accelerations) and safety
(e.g. stresses).
Active Control
Fig. 5: Schematic of an active control system
Active control
Fig. 6: Implementation of control
Active control
mx  cx  (k  k ) x  0
k  chosenand fixedin case of passivecontrol
u  (k ) x  appliedactuationforcefor passivecontrol
The above provides a simple comparison between active
and passive control. In passive control, the additional
stiffness is chosen and fixed, i.e., like a re-designed
structure. In passive control the actuator applies a force to
the original structure, the force being proportional to the
displacement measured by sensor (which is proportional to
the sensor’s output voltage) . In active control the main task
of design is determining the proportionality constant k
Active Control
Outline of Design Process
• Develop mathematical model of the structure and the
chosen sensors and actuators.
• Adopt a mathematical model for the disturbances (i.e.,
wind, earthquake load).
• Decide performance specifications (eg., peak
accleration, time to settle down after disturbance applied,
etc).
• Choose type of control algorithm (i.e., how to obtain the
‘proportionality constant’ and hence actuation
voltages/forces from sensor voltages). Then design
controller so that performance specs are met. Examples
of control algorithms are proportional, integral, derivative,
PI, PID, optimal control, robust control, etc.
Active control with TMD
Fig. 7: Schematic of AMD applied to building
Active control with TMD
Fig. 8: AMD on Kyobashi Seiwa building
Active control
• First full scale application of active control to a
building was done on Kyobashi Seiwa building
(Japan) in 1989 (Fig. 7,8). Two AMD’s were
used. Primary one weighs 4t and damps
transverse motion. Secondary one weighs 1t
and damps torsional motion.
• Can also use Magnetorheological fluid dampers
(semi-active), active tendons, etc. (Fig. 9, 10, 11)
Semi-Active control with MRD
Fig. 9: Control using MR dampers (a) two dampers (b) single damper
Actuators – MR Damper
Fig. 10: Magnetorheological damper
Active control with Tendons
Fig. 11: Active tendons used in control
Active Control with Tendons
u[t ]  actuationmot ionof tendon
x[t ]  slab mot ionmeasuredby sensor,i.e., what we want tocontrol
x0 [t ]  ground mot ion
m  mass of slab
k c  stiffnessof tendon
m 02  stiffnessof columns
  dampingratio