Pierre Gélat National Physical Laboratory 3 April 2003 Developments in Acoustic Emission at the UK’s National Physical Laboratory.

Pierre GélatPierre Gélat

National Physical LaboratoryNational Physical Laboratory

3 April 20033 April 2003

Developments in Acoustic Developments in Acoustic Emission at the UK’s National Emission at the UK’s National Physical LaboratoryPhysical Laboratory

The Finite Element MethodThe Finite Element Method

The FE method is a numerical method to solve arbitrary PDEs The FE method is a numerical method to solve arbitrary PDEs

The method consists of approximating the structure in small The method consists of approximating the structure in small domain portions called finite elementsdomain portions called finite elements

Each element has a set of material properties associated with it Each element has a set of material properties associated with it (Young’s modulus, material density, Poisson’s ratio, etc.)(Young’s modulus, material density, Poisson’s ratio, etc.)

A set of output quantities (e.g. displacement) can be obtained A set of output quantities (e.g. displacement) can be obtained for given forcing/boundary conditionsfor given forcing/boundary conditions

http://www.npl.co.uk/npl/acousticshttp://www.npl.co.uk/npl/acoustics

Axisymetric and Three Dimensional Axisymetric and Three Dimensional Transducer ModellingTransducer Modelling

Use of PAFEC 8.6 Vibroacoustics software (soon to be upgraded to 8.8) Use of PAFEC 8.6 Vibroacoustics software (soon to be upgraded to 8.8) used for modelling the behaviour of ultrasonic and audio-range transducer, used for modelling the behaviour of ultrasonic and audio-range transducer, with fluid loading or in vacuum, for both continuous and transient excitationwith fluid loading or in vacuum, for both continuous and transient excitation

Allows the coupling of vibrating structures to the modelling of finite and Allows the coupling of vibrating structures to the modelling of finite and infinite regions of fluid (boundary elements, wave envelope elements)infinite regions of fluid (boundary elements, wave envelope elements)

Piezoelectric elements can be defined Piezoelectric elements can be defined

PAFEC’s mathematicians are working with NPL both to extend the PAFEC’s mathematicians are working with NPL both to extend the functionality of the software and to provide NPL with bespoke software to functionality of the software and to provide NPL with bespoke software to tackle specific modelling problemstackle specific modelling problems

Quantities routinely obtained from PAFEC include: acoustic pressure, Quantities routinely obtained from PAFEC include: acoustic pressure, structural displacement and electrical impedance structural displacement and electrical impedance

Sensitivity analysis and optimisation capabilities Sensitivity analysis and optimisation capabilities


Sensor Modelling Using the Finite Element MethodSensor Modelling Using the Finite Element Method

Understanding the dynamics of piezoelectric sensorUnderstanding the dynamics of piezoelectric sensor

Investigating the effect of variations in design on overall Investigating the effect of variations in design on overall sensitivitysensitivity

Designing novel sensor configurationsDesigning novel sensor configurations

Sensitivity analysis and optimisation (sensitivity Sensitivity analysis and optimisation (sensitivity vsvs. . bandwidth)bandwidth)


AE sensor modellingAE sensor modelling

Sensor is assumed to be axisymmetricSensor is assumed to be axisymmetric

Piezoelectric material is PZT5-APiezoelectric material is PZT5-A

Apply unit voltage across PZT5-A between 0.1 MHz and 1MHzApply unit voltage across PZT5-A between 0.1 MHz and 1MHz

Obtain electrical impedanceObtain electrical impedance


Electrical Impedance of AE sensorElectrical Impedance of AE sensor

Modelling of AE reference facilityModelling of AE reference facility

Aim is to develop a methodology to model a simple AE system from the Aim is to develop a methodology to model a simple AE system from the electrical excitation of the reference transducer to the electrical output of the electrical excitation of the reference transducer to the electrical output of the sensorsensor

Requires the combination of more than one techniqueRequires the combination of more than one technique

- one to model the transducer and sensorone to model the transducer and sensor

- the other to model the stress wave propagation in the mediumthe other to model the stress wave propagation in the medium

NPL’s Finite Element (FE) method used toNPL’s Finite Element (FE) method used to

- predict displacement output of transducer for a given electrical inputpredict displacement output of transducer for a given electrical input

- predict electrical response of sensor for a given displacement at its facepredict electrical response of sensor for a given displacement at its face

Fraunhofer’s Finite Difference/Integral (EFIT) method used toFraunhofer’s Finite Difference/Integral (EFIT) method used to

- predict displacement as a function of time at a given point in an elastic predict displacement as a function of time at a given point in an elastic mediummedium


Experimental arrangementExperimental arrangement

• The reflective The reflective coating is coating is Aluminium or Aluminium or ChromiumChromium

• The displacement The displacement equivalent noise equivalent noise floor of the floor of the interferometer is interferometer is around 3.5 pmaround 3.5 pmRMSRMS

1)1) Out-of-plane displacement history of surface measured using Out-of-plane displacement history of surface measured using interferometerinterferometer

2)2) Repeated with sensor coupled in place of interferometerRepeated with sensor coupled in place of interferometer

Modelling of AE reference facilityModelling of AE reference facilityComparison of NPL’s FE and Fraunhofer’s EFIT methodComparison of NPL’s FE and Fraunhofer’s EFIT method

Comparison performed using a simple Comparison performed using a simple problem for validationproblem for validation

- point displacement excitationpoint displacement excitation

- small cylindrical glass block allows small cylindrical glass block allows treatment as an axisymmetric problemtreatment as an axisymmetric problem

- displacements compared at chosen displacements compared at chosen pointspoints

Peter D. Theobald:

Peter D. Theobald:

. .

Force Force f = A sin (2f = A sin (2ft) . B sin (ft) . B sin (ft)ft) excited excited at at r = 0r = 0, , z = 0z = 0


Reference source - Reference source - Finite Element modelling of conical transducer

Reference source - Reference source - Finite Element modelling of conical transducer

Modelling of AE reference facilityModelling of AE reference facilityComparison results at 200 kHzComparison results at 200 kHz



Modelling of AE reference facility - Modelling of AE reference facility - ConclusionsConclusions

Validation is complete and shows good comparison between Validation is complete and shows good comparison between techniquestechniques

Work is now progressing on the modelling of a complete system Work is now progressing on the modelling of a complete system comprising of source transducer, glass cylinder and sensorcomprising of source transducer, glass cylinder and sensor

If successful this should produce a complete transfer function for the If successful this should produce a complete transfer function for the AE system for each wave mode received at the sensorAE system for each wave mode received at the sensor

Next stage of work is to model the NPL test facilityNext stage of work is to model the NPL test facility

- would provide more information for the calibration of sensorswould provide more information for the calibration of sensors

- could be combined with a calibrated reference source to provide system could be combined with a calibrated reference source to provide system calibrationcalibration

Pierre Gélat National Physical Laboratory 3 April 2003 Developments in Acoustic Emission at the UK’s National Physical Laboratory.

Documents

sensor modelling

modelling of finite

given displacement

electrical response

finite element method

finite elements

electrical impedance

electrical output