Paul Bonnet – Oxford University – paul.bonnet@eng.ox.ac.ukPaul Bonnet – Oxford University – paul.bonnet@eng.ox.ac.uk

Developing Real TimeDeveloping Real TimeSubstructureSubstructure

Dynamic TestingDynamic Testing

(RTS testing)(RTS testing)

The problem:The problem:analyse the behaviour of a structure in an earthquakeanalyse the behaviour of a structure in an earthquake

Let’s do aLet’s do acomputer model ofcomputer model of

the structurethe structure

Let’s build aLet’s build areplica of the structure replica of the structure

on a shaking tableon a shaking table

Let’s conductLet’s conducta pseudo-dynamic test a pseudo-dynamic test

on the structureon the structure

All these methods All these methods have their inherent have their inherent

drawbacksdrawbacks

Modelling cracks, fatigue,Modelling cracks, fatigue,

concrete, strain rate effects, etcconcrete, strain rate effects, etc..Scale effects and/orScale effects and/orload capacity & costload capacity & cost

No experimental velocity and No experimental velocity and inertial effectsinertial effects

Let’s conductLet’s conducta real-time substructure a real-time substructure

test test

1 numerical substructure1 numerical substructure

++

1 physical substructure1 physical substructure

++

1 emulation interface1 emulation interface

The idea behind RTS testing:The idea behind RTS testing:divide structure into a part solved numerically and a part divide structure into a part solved numerically and a part

tested physically.tested physically.

separateseparate

butbut

complementarycomplementary

ensuring correct ensuring correct boundary conditions boundary conditions on each substructureon each substructure

affect each affect each other at all other at all

timestimes

For a RTS experiment to work, the 2 substructures and the For a RTS experiment to work, the 2 substructures and the interface emulation need to be solved / tested / conducted interface emulation need to be solved / tested / conducted

SIMULTANEOUSLYSIMULTANEOUSLY & in & in REAL-TIMEREAL-TIME..

have dynamic have dynamic effectseffects

Base isolators

M Top motion: xt

Ground motion: xg

Building structure

Passive mass-damper

P C

xtM

F2

F1

Actuator

Emulated system = Physical substructure + Interface + Numerical substructure

Base motion: xb

xb

xb

F2

xtF1

Example:Example:behaviour of a dynamically isolated 10 storey building behaviour of a dynamically isolated 10 storey building

during an earthquake?during an earthquake?

Benefits of RTS testing:Benefits of RTS testing:

Structural parts of interest => physically tested Structural parts of interest => physically tested with the rest offering an infinite repeatability.with the rest offering an infinite repeatability.

Complete structure is emulated with much lower Complete structure is emulated with much lower cost & capacity than shaking table test.cost & capacity than shaking table test.

Dynamic effects are exact because reproduced Dynamic effects are exact because reproduced in real-time.in real-time.

=> Practical & financial benefits while retaining => Practical & financial benefits while retaining good scientific validity.good scientific validity.

Results of preliminary tests:Results of preliminary tests:

Results of preliminary tests:Results of preliminary tests:

Challenges / current work:Challenges / current work:

Quick numerical substructure computation.Quick numerical substructure computation.=> Using discrete time integration scheme (several schemes to => Using discrete time integration scheme (several schemes to compare).compare).

““Instantaneous” actuation is not possible (presence of an “actuator Instantaneous” actuation is not possible (presence of an “actuator delay” delay” → instability→ instability).).=> Need to know how much delay, minimise it, know how & why it => Need to know how much delay, minimise it, know how & why it varies.varies.=> Need to compensate for actuation delay (several methods to => Need to compensate for actuation delay (several methods to compare).compare).

Method needs to be applied & proven for non-linear numerical & Method needs to be applied & proven for non-linear numerical & physical substructures of earthquake engineering size & properties.physical substructures of earthquake engineering size & properties.=> Conduct a validation test on a 2 storey building column.=> Conduct a validation test on a 2 storey building column.

If you’d like to know more…If you’d like to know more…

……come and visitcome and visitus in the Oxfordus in the OxfordUniversity StructuralUniversity StructuralDynamics LaboratoryDynamics Laboratory

paul.bonnet@eng.ox.ac.ukpaul.bonnet@eng.ox.ac.uk Tel: 01865 273 112 Tel: 01865 273 112