Semiactive Management of Structures Subjected to High Frequency

Semi-active Management of Structures Subjected to High Frequency Ground Excitation C. M. Ewing, R. P. Dhakal, J. G. Chase and J. B. Mander 19 th ACMSM, Christchurch, New Zealand, 2006

The Scene • Structures can be highly vulnerable to a variety of environmental loads • These days, man-made events can also have significant impact on the life, serviceability and safety of structures, and must be accounted for in new designs – i. e. blast loads • However, what do you do about already existing and potentially vulnerable structures? – In particular, how do you manage to protect the structure without overloading shear or other demands? – Particularly true for relatively older structures • Semi-active methods offer the adaptability to reduce response energy without increasing demands on the structure, but add complexity • Passive methods offer simplicity and ease of design, but are not adaptable or as effective.

Characteristics of BIGM Typical Seismic excitation Horizontal 50 m Large amplitude (~100 g) May cause sudden collapse. Short duration (<0. 05 sec) Impulsive nature. Post-BIGM response is also important.

Characteristics of BIGM Typical Seismic excitation Horizontal 50 m High frequency (~200 Hz) May excite high frequency vibration Modes during major shock duration.

Impulse Shock Spectra • If t 1/T < critical (0. 4 -0. 5), - The maximum response of a linear structure depends on t 1/T.

Impulse-Response Relationship T = 1 sec • If t 1/T < critical (0. 4 -0. 5), – The maximum response factor is proportional to the total energy applied, regardless of the impulse shape.

A Simple Structure & Damage 450 k. N live • Loads are impulsive • Excite higher order modes • Plastic first peak response is not unusual • Plastic deformation on return or second peak response may also occur • After initial pulse the response is transient free response from a large initial value • Main forms of damage: – – Residual deformation Low cycle fatigue Blast load based on pressure wave and face area 630 k. N live 1000 kg/story E = 27 GPa

General Dynamic Response Fundamental global mode Higher order global mode Frequency increases Acceleration increases Displacement decreases Fundamental local modes

More Detailed Model Basic Elements: • Multiple elements per column to capture higher order responses [Lu et al, 2001] • Mass discretised over all elements in column • Blast load discretised to each storey based on pressure wave and face area • Simple frame used to characterise basic solutions available for something more complex than a SDOF analysis • Non-linear finite elements (elastic-plastic with 3% post yield stiffness) • Fundamental Period = 1 sec • Main structure model captures all fundamental dynamics required for this scenario P

Typical Load • • • Short duration impulse (< T 1/5) Any shape will give the same result, as the basic input is an applied momentum Provides an initial displacement Pblast = 350 k. Pa pressure wave Triangular shaped pulse of duration Dt = 0. 05 seconds or 5% of fundamental structural period

Typical Uncontrolled Response • • A first large peak that is plastic Second and third peaks may also have permanent deformation Free vibration response after initial pulse (not linear) Residual deformation Permanent deflection may be larger or even negative depending on size of the load

Possible Solutions • Passive = Tendons – – • Restrict first peak motion = initial damage Add slightly to base shear demand on foundation Match overturning moment diagram [Pekcan et al, 2000] Tendon yields by design during initial peak Tendon in shape of moment diagram Semi-Active = Resetable devices using 2 -4 control law – Do not increase base shear – Reduce free vibration response = subsequent damage • Therefore, in combination these devices are designed to reduce different occurrences of damage in the response • However, can devices hooked to story’s manage damage for this case characterized by higher column mode response? • Paper also considers device on 2 nd story and from ground to 2 nd story Resetable device 1 st floor

Becoming A Proven Technology End Cap Seal Cylinder Piston More later in conference from Mulligan et al, Rodgers et al and Anaya et al on resetable devices and semi-active applications/experiments

Semi-Active Customised Hysteresis Resist all velocity Viscous Damper Resist all motion Reset at peaks 1 -4 Resetable Resist motion away from 0 From 0 Peak Resist motion toward 0 From Peak 0 1 1 -3 Resetable 3 4 2 -4 Resetable 2 Only the 2 - 4 control law does not increase base-shear

The Very Basic Ideas a) Cylinder b) Valves Piston Cylinder Piston Independent two chamber design allows broader range of control laws

Specific Results Device on first floor and tendon versus uncontrolled First peak and free vibration reduced ~40 -50% 1 st story response Displacement • • • Time

Device Stiffness is Critical • • • Results normalised to uncontrolled response Device stiffness in terms of column stiffness k 50 -100% of column stiffness = good result in free vibration per Response Energy 2 -norm 1 st Peak 2 nd Peak [Rodgers et al, 2006]

Conclusions • Blast can be completely represented by the applied momentum rather than shape, pressure or other typically unknown values • Simple robust system shows potential in this proof of concept study on an emerging problem of importance for structural designers • Complexity added is minimal • Results show that significant improvements that could be critical to safety and survivability can be obtained • Minimal extra demand on foundations makes it particularly suitable for retrofit of existing (relatively older) structures