InFocusIssue 02 | 2012

Optical Measurement Solutions

Simulation and Modal Analysis in Product Development

Identification and Simulation of a Milling Robots Flexibility Behavior

Page 4

Design of a Novel Longi t udinal-torsional Ultrasonic Transducer

Page 6

Vibration Measurement During Milling Operations

Page 10

Laser Vibrometers Meas ure Where no Other

Instruments Can Reach (Interview)

Page 15

Dr. Hans-Lothar Pasch

Dear Reader,Industrial product development relies heavily on computer-based, highly

efficient design and simulation tools. But how close will a model approach

reality with regard to the functionality of the product or component? This

can only be revealed by testing, where fast and precise tools are necessary to

provide the experimental data needed for the assessment and optimization

of the model.

Regarding vibration and acoustic testing, our customers have trusted

Polytecs Scanning Vibrometers for 20 years. The PSV-500 is the newest,

and the 5th generation of our renowned full-field vibration sensors, again

setting new standards in performance, flexibility and user friendliness.

We invite you to celebrate this milestone with us. Your steady input of

suggestions and demands has contributed a lot to both our development of

new products and our ability to remain the clear market leader, an achieve-

ment that began 25 years ago with the first fiber-optic vibrometer.

In this issue wed like to present you many interesting applications from

our customers, such as simulation studies on industrial robots, ultrasonic

drilling tools and machining equipment. You will also find an interview

about using industrial vibrometers on the production line and learn more

about our latest products.

Have fun reading!

Editorial

Eric Winkler

Eric WinklerOptical Measurement Systems

Dr. Hans-Lothar PaschManaging DirectorPolytec GmbH

Polytec News Page 3

Special Feature:Simulation and Modal Analysis

Identification and Simulation of a Milling Robots Flexibility BehaviorPage 4

Design of a Novel Longitudinal-torsional Ultrasonic TransducerPage 6

Thermomechanical Characterization of Materials under Extreme Conditions Page 8

Vibration Measurement During Milling OperationsPage 10

R&D and Production Applications

Vibration Imaging on Facial Surfaces During PhonationPage 12

Noise, Vibration and Harshness (NVH) Analysis at MAN in NurembergPage 14

Laser Vibrometers Measure Where no Other Instruments Can ReachInterview with Olaf Strama, MEDAVPage 15

Events & MediaPage 16

2

Polytec News

Polytec Expanding into New Premises

Polytecs new building was finished on schedule in September covering 8,000 square meters of ground. Polytecs manu-facturing team was happy to move into the new premises together with their pro-duction machinery. They were joined by

their Optical Measurement Systems sales and application colleagues and a couple of administrative departments. Work is ongoing at the specially designed large acoustic testing hall that will provide robot-automated vibration measurements.

Despite the upheaval, the transition ran so smoothly that all of our office and lab employees were able to continue support-ing their customers seamlessly.

Polytecs PSV-500 Scanning Vibrometer will continue to meet our customers cur-rent and future need to bridge the gap between computer generated simu lation models and the real dynamic and acous -tic properties of their products. Thus we have reason to celebrate and we do it with a special edition of our InFocus Cus-tomer Magazine where we intro duce

the innovations of the new system and say thank you to all customers for providing generous feed back and con-tinuous support. If you didnt receive the magazine, please down load here or order your printed copy: www.polytec.com/InFocus

The Next Scanning Vibrometer GenerationTechnical Excellence in a Compact Design

The RSV-150 Remote Sensing Vibro meter, designed to measure vibration and dis-placement from remote dis tances, offers expand ed applications when configured with the new RSV-E-150-M Controller. The new unit extends the bandwidth to 2 MHz and the maximum velocity to 24 m/s, which is almost the limit of the current HSV High Speed Vibrometers (30 m/s), but with a much wider meas -urement range and higher optical sensi-tivity. The development was driven by applications in ultrasonic NDE, e.g. test -ing of railroad tracks or leak detection, as well as applications with the high speeds and ac -celerations experi e nc -ed in falling tower and pyro shock experi ments. www.polytec.com/rsv

New RSV ControllerRemote Vibration Sensing Now at High Vibration Velocities and Frequencies

3

Robots Under ControlIdentification and Simulation of a Milling Robots Flexibility Behavior

mine the static flexibility behavior of the gearbox, the bearings and structural components with both high accuracy and low measuring complexity. This meth od, which involves the precise measurement of displacements at in -di vidual points on the robot structure, employs a 3-D Scanning Vibrometer.

Measurement Setup

The measurements were carried out on a KR 240 R2500 prime industrial robot made by KUKA Roboter GmbH. A mil -ling spindle is mounted on the robots flange, on the housing of which a force in the range of several kilonewtons can be applied via a double-action pneumatic cylinder (fig. 1, left). The loading cycle involves both tensile and compressive loading of the robot. To improve the sig-nal quality of the optical measurement system, all measurement points have reflective film attached to them. In addi-tion, a mirror is positioned behind the robot, so that points on the reverse side can also be measured.

The 3-D Scanning Vibrometer, supplied by Polytec, comprises three scanning heads that are attached to a tripod. The laser beams from all scanning heads are targeted at a point and the vibration velocity determined in those three direc-

The tilt resistance of both the gear box and the bearings can also be calculated using this. The processing of these para-meters in a computation-efficient manner makes it possible to determine the deflec-tion of the robot in real time.

Currently, machine tools are used almost exclusively for machining work. While milling robots do indeed have many ad -vantages, because of the high flexibility of the robot structure such robots are only suitable for machining jobs with low accuracy requirements and low cutting forces. To increase the accuracy achiev-able during machining using industrial robots, the Institute for Machine Tools and Industrial Management at the Tech-nical University of Munich is using a model-based control system for com -pensation of the force-induced static displacements.

The challenges in designing the simu-lation model are both the need for an absolute real-time capability and the requirement to precisely determine the robot flexibility parameters. In this respect, the tilt resistance of the gear-box and bearings must be considered.

No method has existed until now for determining the stiffness parameters of a robot that makes it possible to deter-

tions. From knowing the positions of the scanning mirrors and sensor heads in space, the raw data can be trigonometri-cally transformed into the three-dimen-sional movement of that point. In this way, the velocity vector is determined for each measurement point during a load cycle of the robot. The displacement of all measuring points is determined by a numerical integration of the velocity data.

Calculation of the Stiffness Para-meters from the Measurement Results

Fig. 2 shows the resulting displacement images for a measurement in which the robot was only loaded in the z-axis direction.

Tilting can be seen about the gear shaft of the second and third linkage can be seen. In addition, a rotation of the first linkage about the y-axis is apparent which, because of the relationship of the levers, has a marked effect on the dis-place ment of the tool-center point (TCP). This ex am ple highlights the necessity to also consider the bearing stiffness of the individual axles when modeling the flexibil i ty behavior of the robot. The stiffness para meters of the linkages are determin ed from the rota tion of the

Simulation

This article introduces a method to identify the stiffness

parameters of an industrial robot through measurements

made using a 3-D Scanning Vibrometer.

4

Vibrometer, its possible to predict in real time and therefore compensate via robot control for the displacement of the tool center point due to the applied force. In the next step, an algorithm for post-processing of the measurement data will be developed to further improve the accuracy of determining the linkage rotation angle. This should also allow the flexibility of the structural components of the robot to be measured and to inte-grate this into the flexibility model.

respectively adjoining struc tural compo-nents using a calcula tion algorithm that is not de scribed in more detail here.

Further Use of the Stiffness Parameters

The determined parameters are used in an analytical stiffness model of the robot in order to calculate its flexibility behav -ior in real-time. The model is coupled to the robot control so that the actual axis position of the robot can also be consid-ered. Depending on the calculated dis-placements, offset signals are transmitted to the control so that force-induced dis-placements of the robot, for example caused by milling, can be compensated in a control-based manner.

Summary and Outlook

The work presented here has shown that, with the aid of the PSV-400-3D Scanning

Authors Contact Oliver Rsch, Prof. Dr.-Ing. Michael F. Zholiver.roesch@iwb.tum.de

Institute for Machine Tools and Industrial Management, Technical University of Munich (TUM), Munich, Germany

www.iwb.tum.de/en/Institute.html

Fig. 2: Deflection of the robot during tensile and compressive loading.

0 20 40 60 s 100100

50

0

%

100

Time

Forc

e

Loading cycleIndustrial robot KR 240 R2500 prime

Milling spindleHSD ES501

Force measure-ment cell

Pneumaticcylinder

Mirror Reflectivefilm

Compressive phase

Tensile phase

Fig. 1: Measurement setup (left) and applied force curve (right).

New Look and Feel Newly Designed PSV 9.0 and VibSoft 5.0 Software

Aside from functionality supporting the updated PSV-500 hardware, the brand new Version 9.0 Scanning Vibrometer Software has powerful new features for all users, including those participating in one of our software maintenance plans such as the special university program. The newly designed user interface pro-vides more flexibility and is customizable to meet individual needs. The long-term user will notice that all of the familiar func-tions necessary to perform a measurement with ease have been retained. Shortcuts improve access to standard features for the power user. There are many beneficial changes in core functionality. The align-ment process for example is dramatically improved with the High Contrast Laser Display: a machine vision technology to avoid glare in the camera image from the laser reflecting off shiny surfaces. The SignalProcessor post processing feature

can now be used to instantly compare stored and live data, which is beneficial especially when setting up excitation sig-nals for experimental modal analysis tests. New analyzer functionality is imple ment - ed in the presentation mode, and the ears will benefit from the PSV-S-Audio option, listen ing to live and stored data files to, for example, assess the tuning of the struc-ture. Version 9.0 will be the last release to support Windows XP, but near ly all PSV models can be upgraded to the latest operating system. The new Version 5.0 VibSoft data acquisition and analysis sys-tem for single point systems will of course also benefit from the latest features.

www.polytec.com/software

5

Drill Tools for Earth and SpaceDesign of a Novel Longitudinal-torsional Ultrasonic Transducer

augmenting the oscillatory displace ment amplitude provided by an ultrasonic trans ducer. The device is necessary to effi -cient ly transfer the acoustic ener gy from the ultrasonic transducer into the mediumbe ing treated. The ultrasonic horn is com-monly a solid metal rod with a round transverse cross-section and a variable-shape longitudinal cross-section. The length of the device must be such that there is mechanical resonance at the de -sired ultrasonic frequency of op era tion one or multiple half wave lengths of ultrasound in the horn material.

Efforts have been made to excite longi -tudinal-torsional responses in devices using two different techniques; by coup-ling the longitudinal and torsional modes,

Tool Modeling

Fig. 1: The interaction of modes as the length of an ultrasonic step horn is incrementally altered.

Measurement of complex combinations of different vibration modes

operating together at ultrasonic frequencies can be carried out using

3-D laser vibrometry.

These measurements are being used to optimize the performance of longitu di - nal-torsional (L-T) ultrasonic horns at the University of Glasgow. L-T ultrasonic vibration has many applications includ -ing surgical devices, industrial welding and ultrasonic motors. Researchers at Glasgow are even developing ultra sonic drill tools for planetary explora tion. Due

to the low gravity, traditional drilling will be difficult and the next generation Mars landers will require low-reaction devices for drilling into the surface.

Development of Ultrasonic Properties

An ultrasonic horn, also known as a sono-trode, is a metal bar commonly used for

6

Experimental Modal Analysis

In order to evaluate the behavior of the transducer, the operating modal res -ponse is determined by experimental modal analysis (EMA). The effectiveness of the transducer is characterized by its torsionality, which is the ratio of torsional to longitudinal vibration amplitude. The 3-D laser vibrometer was used to meas -ure the response because it allows data to be obtained without affecting the nat-ural frequency, mode shape, or damp-ing, regardless of whether the device is ex cited in air (unloaded) or under a load representing a real-world application.

Using the 3-D laser vibrometer, responses in three orthogonal directions are ac -quir ed at a grid of surface points on the transducer and the modal frequencies and animated mode are extracted using MEscope software. The results enable us to assess the vibrational characteristics of

Authors Contact Hassan Al-Budairi, Dr. Patrick Harkness, Prof. Margaret Lucash.al-budairi.1@research.gla.ac.uk; Margaret.Lucas@glasgow.ac.uk

University of Glasgow, School of Engineering, UK

www.gla.ac.uk/schools/engineering

Fig. 3: Modelled and measured mode shapes of the transducer represented in Simulia Abaqus and Vibrant Technologies MEscope.

and by causing the degeneration of a longitu dinal mode into a longitudinal-torsional response by incorporating heli-cal flutes and slits. Coupling the longitu-dinal and torsional modes of vibration has been found to be difficult because, as the geo metry of a typical ultrasonic horn is modified incrementally, the two modal frequencies have been found to approach each other and then move apart but with no crossover point where the modes are fully coupled.

For example, using a Polytec 3-D laser vibrometer system, the interaction of the first longitudinal mode (L1), the second torsional mode (T2), and a bending mode (B) can be observed (fig. 1) in terms of longitudinal (L) and tangential (T) ampli-tude in a simple titanium half-wavelength step-horn as the length of the base sec-tion of the horn is gradually decreased.

As it is very difficult to achieve effec tive coupling between these modes, horns exploiting this approach tend to be char acterized by low responsiveness or, alternatively, need to incorporate two differently poled piezoceramic stacks in the transducer to excite the two modes. Therefore, the mode degeneration meth -od is considered to be more promising and we have developed a transducer (fig. 2) to take advantage of this tech-nique. This transducer can deliver a lon-gitudinal-torsional output when excited by the longitudinal vibration mode of a single piezoceramic stack.

the operating mode and the frequency spacing between the desired mode and surrounding unwanted modes of vibra-tion. The vibrometry measurements can also be used to validate finite-element (FE) models of the transducer (fig. 3).

Conclusion

The results show that the model can be used reliably to design novel transducer shapes and evaluate the longitudinal-torsional response characteristics to maxi-mize performance of devices.

Fig. 2: The new L-T transducer.

7

Material Science

A new dynamic method for thermomechanical characterization of candidate

target materials (tungsten, tantalum and molybdenum) has been developed

as part of the program advancing high power targets for the UK Neutrino

Factory.

The materials used in the target systems of next generation high power particle accelerators are frequently exposed to a combination of high stresses, high strain rates and very high temperatures. To estimate the lifetime of target and target system components in this environment,

the candidate materials have to be tested under extreme conditions.

Experimental Setup

A thin wire made from the candidate material was heated and stressed by a fast high current pulse (see fig. 1) gen -

e rated by a power supply for the ISIS syn-chrotron (Rutherford Appleton Laborato-ry) kicker magnets. The wires were sup-ported in a vacuum chamber to avoid oxidation and heated to 2650 C by ad -justing the pulse rep etition rate. To allow the current to generate a sufficient ther-mal stress, the wire had to be thin (less than 1 mm diameter). A single-point Laser Doppler Vibrometer (LDV) compris-ing the OFV-534 optical sensor head and OFV-5000 vibrometer control l er meas-ured the longitudinal (by sensing at the

Metals at the LimitsThermomechanical Characterization of Materials under Extreme Conditions using a Laser Doppler Vibrometer

8

Fig. 1: The experimental setup. Coaxial cables (1) carrying the current pulse from the power supply are combined into a sin -gle cable (2) which is connected to the test wire. The wire is supported in a vacuum chamber (3) and its oscillations are meas-ured by the Laser Doppler Vibrometer (4).

Fig. 3: The measured and calculated radial velocity of a 0.5 mm diameter tungsten wire at different temperatures.

Fig. 4: Youngs modulus of tungsten compared between experiments

Fig. 5: The yield strength versus peak temperature for tantalum, tungsten and molybdenum wires. The characteristic strain rate values are also shown.

wire tip) as well as radial velocities and displacements of the wires. Three differ-ent LDV decoder units (VD-02, VD-05 and DD-300) cover the complete range of amplitudes and frequencies of the wire vibrations. An optical pyrometer measured the wire temperature at the same point that was measured by the LDV.

Results and Conclusions

Fig. 3 shows the radial velocity of the tungsten wire measured by the Laser Doppler Vibrometer at different tempe -ratures. The current pulse began at t = 0 and the wire started to contract and ex -pand. After about 1 s it reached a new equilibrium position and started oscillat -ing around it. The time scale in fig. 3 begins with t < 0 in order to illustrate a relatively low background noise level of the LDV. The correlation between ex -periment and finite element modeling (LS-DYNA) calculations is very good. The radial oscillation frequency obtained was used afterwards to extract the Youngs modulus of the wire material as a func -tion of temperature. In fig. 4 the Youngs modulus results from the new measure-ments are compared with earlier results.

To determine the yield strength, the current pulse amplitude in the wire was increased in steps. This was continued until the wire started bending or kinking. The radial surface velocity measured by

the LDV (see fig. 3) was used to extract the strain rate during the test. The LDV optical sensor head includes a fast cam -era for monitoring strain in the wire. In addition, it was noticed that the LDV radial velocity signal became very noisy when signs of plastic deformation first appeared. This change in LDV signal quality indicated that the wire was near its yield point. Fig. 5 shows the stress that is needed to reach the yield point of molybdenum, tantalum and tungsten wires.

To conclude, the candidate materials for the Neutrino Factory target were tested under conditions expected at this accelerator. A new dynamic method for thermomechanical characterization of materials under extreme conditions was therefore developed that can help test the consistency of different constitutive models.

Authors ContactDr. Goran koro, Dr. J. R. J. Bennett,Dr. T. R. Edgecockgoran.skoro@stfc.ac.uk

ISIS Facility, STFC, Rutherford Appleton Laboratory, Harwell, Oxford, UK

www.isis.stfc.ac.uk

AcknowledgementsThis work was supported by Science and Technology Facilities Council (UK).

Fig. 2: Photograph of a 0.75 mm diameter tungsten wire during current pulse tests (a) and an illustration of wire clamping (b).

9

structures may experience regenerative lateral vibrations for some cutting con -ditions. The inherent nature of variable dynamic conditions of these milling and machining processes are likely to be the culprits by which the finished parts ex -hibit poor surface finish and lower pro-ductivity of those manufacturing process-es. Stability lobes diagram methods are common techniques that use dynamic information to define stability regions in which it is possible to find the appropri -ate and desired combinations of machin-ing parameters. With this technique, the experimental Operational Frequency Response Functions (OFRF) and regular FRF are required to feed the Enhanced, Multistage Homotopy Perturbation Method (EMHPM).

In milling, structural modes of the machine tool-workpiece system are initially excited by cutting forces. Surface finish profiling due to oscillatory excitation left by a sec-tion of the tooth is subsequently removed by the incoming and advancing cutting tooth surface which causes increased struc-tural vibrations. This self-excited cutting phenomenon can become unstable, and chatter vibrations grow until the tool jumps out of the cut, ruins the expected surface tolerances and can even break under excessive cutting forces.

High speed machining is a widely used process in the aeronautical industry to machine low stiffness structures with thin walls and floors where high tolerances are required. Machining of these types of

Comparison Between Laser Vibro-metry and Existing Measuring Methods

It is well-known that accelerometer mass load on heavy structures has neg li gible in -fluence on dynamic measurements. How-ever, those effects are not negligible when the workpiece mass is small. Since the accuracy of the OFRF directly affects the stability lobes diagram, it is important to study the ac celerometer mass loading effects over the stability diagrams to pre-dict accu rately the dynamic behavior when milling thin walled parts.

In order to study accelerometer mass-loading effects on thin walled structures, we performed several impact tests at different locations over aluminum 7075 thin-walls (1 x 35 x 50 mm) and col l - ect ed the corresponding FRFs by using a 0.6 gram accelerometer. We repeat -ed the measurements using a Polytec CLV-2534-2 Compact Laser Vibrometer that allows dynamic measurements with -

Process Optimization

Chatter vibrations continue to be the major factor limiting the increase in

material removal rates of machine tooling. Machine tool chatter vibrations

occur due to a self-excitation mechanism in the generation of chip thick-

ness during milling operations.

Improving Machining TechniquesVibration Measurement During Milling Operations

MMaacchhhiiinnniiinngg TT

10

Authors ContactProf. Dr. Alex Elas-Ziga, Daniel OlveraInstituto Tecnolgico y de Estudios Superi-ores de Monterrey (ITESM), Monterrey, Mexico.

www.topmak.com.mx

Mario Pineda, Polytec Inc.

Acknowledgements: This work was funded by Fondo Mixto Conacyt #145045 and by Tecnolgico de Monterrey through the research chairs in Nanotechnology and Intelligent Machines.

Parts of this article are based on the paper Experimental Stability Analysis of Thin Wall Structures using Laser Doppler Vibro-meter Devices, Proc. 9th Int. Conf. on High Speed Machining, San Sebastian, Spain, March 7 8, 2012.

out adding mass to the workpiece. The dotted lines shown in fig. 1 represent the frequency response functions ac -quired using the accelerometer and Laser Doppler Vibrometer (LDV) under the same conditions.

Fig. 1 illustrates significant differences between the FRF functions obtained by using the accelerometer and the laser Doppler device. The laser vibrometer without the accelerometer attached captures two fundamental vibrational modes with peak values at 1105 Hz and 1722 Hz. However, the measurements performed with the accelerometer ex -hibit the same dynamic modes but with peak values at 580 Hz and 1366Hz. These differences in the recorded FRF spectra were noticed during experimental tests because of the sound pressure level pro-duced by the thin wall workpiece. The 525 Hz shift of the FRF value of the first peak mode is attributed to the effect of the accelerometer mass.

In order to verify our experimental obser-vations, we performed measurements by using the LDV with the accelerometer attached to the workpiece. The results, shown by the dashed line in fig. 1, are the same as those collected with the ac -celerometer. This experimental test con-firmed the accelerometer mass-loading effects over FRF values, which not only causes a shift of the frequency value of about 48%, but also changes consider-ably the modal damping of the system.In addition, we expect changes in the calculat ed stiffness. As expected, the ef - fect of the accelero meter mass in creases as 1 mm thickness of wall material is re moved from the workpiece during the machining pro cesses.

In order to demonstrate the effects of the accelerometer mass-loading on the dynamics of the cutting processes, we compute the stability lobes by using the EMHPM for both accelerometer and vibrometer measurements. The stability lobes plotted in fig. 2 are generated for a inch diameter end mill with 2 teeth, a helix angle of 20, and a radial depth of cut of 0.8 mm.

As we can see from fig. 2, the stable depth of cut values on the computed

stability lobes are strongly influenced by the ac celerometer mass load. Accelero-meter measurements produce a shift in stability lobes not only on spindle speed direction, but also axial depth direction in comparison with vibrometer measure-ments. For this reason, unstable predict -ed cutting conditions are experimentally explored by means of the LDV. The beam was aimed on the top center of the thin wall.

For a 26,500 rpm spindle speed, the maxi-mum vibration amplitudes were recorded up to 0.3 m/s (fig. 3), with a confirmed stable behavior through the frequency spectrum which corresponded to the tool passing-frequency and its harmonics. On the other hand, an accelerometer-predict-ed unstable boundary region was explored by LDV OFRF responses (fig. 4). In this case, the velocity amplitude at 30,000 rpm reached 0.6 m/s. The frequency domain exhibits a chatter frequency.

Conclusion

Fig. 5 shows a comparison between the modal parameter values obtained by using both the FRF of the accelerometer and the laser vibrometer recorded data.It can be clearly seen that an accelero-meter of 0.6 grams attached to the thin wall workpiece has a significant effect on FRF measurements.

Fig. 2: Stability diagrams for the FRF using both devices.

Fig. 3: Stable cutting at 26,500 rpm.

Fig. 4: Unstable cutting at 30,000 rpm.

Fig. 1: Laser measurements versus accelerometer measurement.

Fig. 5: First dominant frequencies measured with both laser sensor and an accelerometer.

11

Head-to-HeadVibration Imaging on Facial Surfaces During Phonation

Bio Acoustics

Getting Set

The vibration patterns of head and neck surfaces, and their contribution to the overall sound, has been inadequately studied until now. A male (22 years of age) with no speech disorder participat -ed in these laser vibrometry measure-ments, intended to image vibration pat-terns that are generated by speaking.

The vibration velocity was obtained with a scanning laser Doppler vibrometer sys-tem (Polytec PSV-400-M4). The laser

Sound from speech is radiated not

only from the mouth and nostril

openings but also from the surfaces

of the head and neck.

Japanese scientists are there fore able to use vibrometers to study speech by meas-uring vibrations caused by phonation, taking advantage of their ability to ana-lyze vibration for a particu lar frequency band of interest.

Doppler vibrometer is an optical trans-ducer that senses the frequency shift of light reflected from a vibrating sur -face caused by the Doppler effect. It can determine the vibration velocity and dis-placement at a certain point. The scan -ning vibrometer can also scan and probe multiple points of a vibrating surface automatically.

Fig. 1 shows the experimental setup. The scanning head of the vibrometer was mounted on a tripod, perpendicular to

12

the floor. The participant was positioned to lie directly beneath the scanning head.

The participant was asked to articulate utterances repeatedly while keeping his head immobile during the measurements. Speech sounds were recorded through a microphone.

The vibration patterns of the facial surface were measured from the frontal direction, which is perpendicular to the forehead, and from an oblique direction, which is nearly perpendicular to the left cheek and the left side of the nose.

In the experiment, scanning points on the facial surface were first determined using system control software. During the meas-urement, the vibrometer scann ed each point and determined the vibration velo-city. It took approximately one second to probe each point. The vibra tion velocity and speech sounds were measured up to 5 kHz.

Results

The two upper images in fig. 2 show the vibration patterns of the frontal facial surface during sustained phonations.

There were significant differences bet-ween the vibration velocity patterns for the vowel (left) and nasal (right) conso-nant. For the vowel, the facial surface around the mouth opening vibrated the most compared to the other regions. In contrast, for the nasal consonant, the facial sur faces of the nose and its vicinity vibrated strongly owing to resonances in the nasal sinuses. The forehead sur -face also vibrat ed to some extent, pos-sibly indicating that the frontal sinuses reso nated during the production of the nasal consonant.

The two lower images in fig. 2 show the vibration velocity patterns of the left facial surface for the phonemes. The vibration of the side of the nose was observed to be stronger for both phonemes than that indicated in the upper images of fig. 2. This means that the direction of the laser light is a significant factor in this measure-ment method. The result also revealed that, for the participant in the present study, the nose surface vibrated even when the vowel, not only just the nasal consonant, was articulated.

How Can It Be Used?

The proposed method enables us to evalu-ate the speech of patients with cleft pal-ate or velopharyngeal insufficiency. The vibration pattern may be helpful as visual feedback of a speaking exercise for such patients. The vibration pattern may be easier to relate to their somesthesis than spectra of their speech sounds. This could also be useful for singing exercises.

In conclusion, the proposed method allows fast, non-contact and multi-point measurements of the vibration velocity of skin surfaces. The results will expand our knowledge of speech pro duction. The next step that needs to be taken is to investigate the relationship between the vibration velocity pattern of skin surfaces and the formants and antiformants of speech sounds.

Author ContactProf. Tatsuya KitamuraFaculty of Intelligence and Informatics, Konan University, Kobe, Japanhttp://basil.is.konan-u.ac.jp/index_e.html

This article is based on the publication Measurement of vibration velocity pattern of facial surface during phonation using scanning vibrometer which can be found at www.jstage.jst.go.jp/article/ast/33/2/33_2_126/_pdf.

AcknowledgmentsThis study was supported by JSPS KAKENHI (21300071). The author wishes to thank Dr. Kazuhito Ito (National Institute of Advanced Industrial Science and Techno-logy), Mr. Francois Bouteille, Mr. Ryo Ishiyama, Ms. Shoko Wakatsuki (Polytec Japan), and Professor Ken-ichi Sakakibara (Health Sciences University of Hokkaido) for their generous assistance and valuable advice.

Fig. 1: Experimental setup.

Fig. 2: Vibration velocity patterns of frontal and left facial surface during articulation of vowel (left) and nasal (right) consonant. The unit is m/s [dB] and 0 dB is equal to 1 m/s.

13

of the individual engine components from the signals, with the following aims:

Identification of engines with conspicu-ous vibration and noise behavior.

Identification of component and assem-bly faults based on the supplied signals

Identification of the fault for effective support of the rework.

MAN exclusively uses Polytec laser vibro-meters as vibration sensors, alongside microphones for airborne noise measure-ment.

To evaluate the sensor signals, ANOVIS offers comprehensive signal processing and classification methods ranging from simple frequency analysis and rotation-angle synchronized methods (order analy-sis) to automatic limit value adaptation. The signal components that arise are re lated to engine kinematics and, based on their characteristics, are allocated to moving parts of the engine (example in fig. 2).

The ANOVIS test system can, through its modularity, be flexibly adjusted to match a wide variety of different engine

MAN Truck & Bus AG in Nuremberg produces a wide range of modern com-mercial vehicle engines. Each individ u al engine is subject to comprehensive tests before it leaves the factory, including tests for undesirable noise and vibrations (noise, vibration and harshness, NVH). This is where the ANOVIS system is used, supplied by MEDAV. These vi bration and noise analyses are effective in helping to safeguard the high quality standards of the Nuremberg engine plant.

ANOVIS Detects Component and Assembly Faults

The NVH analysis carried out using MEDAVs ANOVIS system constitutes an important component of the so-called engine cold test in which engines are filled with oil and then driven by an elec-tric motor. In comparison with other methods, this approach is economical, ecological and simultaneously provides in-depth testing. The ANOVIS test system (fig. 1) records noise and vibrations at various points on the engine and deter-mines characteristic values for assess ment

Sounds GoodNoise, Vibration and Harshness (NVH) Analysis at MAN in Nuremberg

variants. This also refers to the selection of the sensor type. The IVS-200 and IVS-400 laser vibrometers that are used also permit vibration measurements on difficult-to-access measurement points or loca tions that are in some way non-cooperative. Thus, for example, it is in some cases necessary to measure the black-painted surfaces of a high-pressure pump. Process-reliable signal measure-ment with 20 kHz bandwidth is also no problem here using the IVS industrial vibrometer and intelligent speckle elimi -nation integrated into ANOVIS, The few manual steps required as part of the maintenance, for example, to check the position and focus, are taught in a train -ing course to enable reliable operation during production. In this way, ANOVIS makes it possible to detect a large num-ber of potential types of defects, which are otherwise undetectable using other testing methods. For example:

Damage and manufacturing deviationin the camshaft

Noise in the valve-train assembly, e.g. caused by too much clearance

High pressure pump and other auxiliary component faults

Toothing faults, damage and geometri-cal faults on cogs, incorrect tooth-flank backlash

Fig. 2: Sideband energy measure based on order analysis for assessment of toothing vibrations.

voIadocsbpmpvnmm

Fig. 1: ANOVIS system variants for flexible vibration and noise analysis in production.

Engine Cold Test

14

Authors ContactDr. Michael Weidner, Olaf Stramainfo@medav.de

MEDAV GmbH, Uttenreuth, Germanywww.medav.de/?&L=2

Turbocharger faults, for example, dam age and geometrical faults

Missing connecting rod bearing shells and for eign bodies in the combustion chamber

Systematic errors are practically non-exis-tent in modern engine production. The particular advantage of using ANOVIS for

vibration and noise measurements is its ability to identify individually occurring, random errors. Additionally, the vast ex -perience of the MEDAV division, Indus -trial & Automotive Solutions (IAS), in engine testing allows the identification of the cause of the faults. Also, practical tools are provided to quality engineers for analysis and statistical evaluations,

with the help of which, production can be continuously further optimized.

Laser Vibrometers Measure Where no Other Instruments Can ReachInterview with Dipl.-Ing. Olaf Strama, director of Industrial & Automotive Solutions (IAS) at MEDAV

Mr. Strama, you lead the IAS division of MEDAV. What are your application areas?

The Industrial & Automotive Solutions Division specializes in vibration and noise analyses, including notably the so-called NVH test. It is used by our industrial custo-mers for such tasks as the end-of-line test-ing of engines, gearboxes and steering components.

How long have you been using laser vibrometers and what first caused you to do so?

We have been using laser vibrometers for vibration measurement since the 90s. We started with a Polytec CLV and then we purchased a batch of IVS-200 indus-trial vibration sensors. In the meantime, we have used its successor, the IVS-400, especially for measurement points that are difficult to access or where a high measurement bandwidth is required for automatic production.

The breakthrough came about 10 years ago. Since then, our ANOVIS testing sys-tems have been equipped with an intelli-gent speckle elimination feature. With its introduction, the number of false meas-urements in engine production lines producing more than 1000 engines per day could be reduced to 1 2 cases per month. Measuring with a bandwidth of

20 kHz on machined metal surfaces is reliable.

How do you view the use of laser vibrometers up until now and what do you consider to be the essential advantages of optical measurements when compared with the alternatives?

As already mentioned, accessibility to the optimal measurement points is one primary criterion, as is high flexi bili ty for use with different types of test pieces. Also, the wide bandwidth of the laser vibrometer achievable in automat ed testing is another argument, as some error types are only identifiable at high frequencies. Laser vibrometers also score high because they lack mechanical parts that wear out and do not require fre -quent calibration.

How reliable is the laser vibrometer under harsh industrial conditions?

Many IVS-200 units have been used daily in engine testing for over 10 years. With many customers, maintenance is only implemented once it becomes necessary as indicated by the signal quality, which is statistically determined by ANOVIS.

What is the reaction of your cus -tomers and what proportion of orders have laser vibrometers as the vibration sensor?

From our point of view, users react positively if the sensor fulfills its task and no problems occur during daily use. The evaluations carried out by our customer support department regularly indicate that this is the case with laser vibrome-ters. The IVS industrial vibration sensors have contributed significantly to the attractiveness of the solution we offer in a considerable number of installations.

How do you see the potential of laser vibrometers for cold and hot testing applications in the automotive industry?

As a provider of vibration and noise meas-urement systems we offer our customers complete testing solutions for their pro-duction right from the start. These solu-tions consist of optimized sensor technol-ogy for the required application, match ing data acquisition hardware and the needed analysis and evaluation of the defined task. The engines to be tested are becoming more complex and, consequently, the optimal measurement points are more difficult to access. Hence, and because of the frequency range re quired to carry out the tasks, the choice is increasing ly likely to be a Polytec industrial vibration sensor.

Thank you very much for the interview, Mr. Strama!

15

EventsTe

chni

cal s

pec

ifica

tions

are

sub

ject

to

chan

ge w

ithou

t no

tice.

OM

_InF

ocus

_201

2_02

_600

0_E

Advancing Measurements by Light

ImprintPolytec InFocus Optical Measurement SolutionsIssue 2/2012 ISSN 1864-9203 Copyright Polytec GmbH, 2012Polytec GmbH Polytec-Platz 1-7 D-76337 Waldbronn, Germany

CEO/Publisher: Dr. Hans-Lothar Pasch Editorial Staff: Dr. Arno Maurer, David E. Oliver,

Philipp HassingerProduction: Regelmann KommunikationImages courtesy of the authors unless otherwise specified.

New VideoBridging Design and Real World

Did you know how precise, fast and convenient vibration measurements can be? Please view our brand new video to learn how the PSV-500 Scanning Vibrometer is easily applied to complex vibro-acoustic problems: www.polytec.com/psv.

Polytec25 Years of Expertise with Laser Vibrometers

The worlds first fiber-optic vibrometer was introduced by Polytec in 1987. Being capable of performing single point and differential vibra-tion measurements up to 1 MHz, it soon became obvious that this technology was extremely well suit ed for studying the vibrational be -havior of tiny hard disk components in the disk drive industry. Only one year later, the vibro meter was honored with the Photonics Circle of Excellence Award the beginning of a success story that eventually led to Polytec becoming the accepted world leader in laser vibrometry.

Trade Shows and ConferencesDate Event Location

Nov 29, 2012 Non-Contact Topography Measurement Workshop Loughborough University, UK

Dec 04 06, 2012 Seed Expo 2012 Hyatt Regency Chicago, USA

Dec 10 11, 2012 Netherlands MicroNano Conference 12 De ReeHorst, Ede, The Netherlands

Jan 20 24, 2013 MEMS 2013 Taipei, Taiwan

Mar 18 20, 2013 Automotive Testing Expo Korea Seoul, Korea

May 08 10, 2013 CMMNO Condition Monitoring Conference Ferrara, Italy

For the most up-to-date information please refer to our web site, www.polytec.com.

Stay Well Informed

Polytec News-letter and RSS Feed Subscribe to Polytecs com-plimentary e-mail newsletter or to our RSS feed to get the latest information on all our products, programs and benefits the minute they are released!www.polytec.com/news

Polytec Web AcademyFor more information visit: http://polytec.webex.com.

Date / Time Event

Nov 22, 201210:00 11:00 CET

Optimizing the performance of ultrasonic tools and transducers

Nov 27, 201210:00 11:00 CET

Simplify your rotational vibration analysis: characterizing the dynamics of rotating structures by using LDV

Dec 04, 201210:00 11:00 CET

Go ahead with optical, non-contact length and speed measurement: LSVs for optimized reliability and minimized operational costs

Dec 13, 201210:00 11:00 CET

Accelerate your design process: optimizing NVH/Modal Testing and FE correlations with 3-D vibration mapping system

www.polytec.com

Polytec GmbH (Germany)Polytec-Platz 1-776337 Waldbronn Tel. + 49 7243 604-0info@polytec.de

Polytec France S.A.S.Btiment Orion 1er tage39, rue Louveau92320 ChtillonTel. +33 1 496569-00info@polytec.fr

Polytec Ltd.(Great Britain)Lambda House, Batford Mill Harpenden, Herts A L5 5BZTel. + 44 1582 711670info@polytec-ltd.co.uk

Polytec Japan Arena Tower, 13th floor3-1-9, Shinyokohama Kohoku-ku, Yokohama-shiKanagawa, 222-0033Tel. +81 45 478-6980info@polytec.co.jp

Polytec, Inc. (USA) North American Headquarters16400 Bake ParkwaySuites 150 & 200Irvine, CA 92618Tel. +1 949 943-3033info@polytec.com

Central Office1046 Baker RoadDexter, MI 48130Tel. +1 734 253-9428

East Coast Office25 South Street, Suite AHopkinton, MA 01748Tel. +1 508 417-1040

Polytec South-East Asia Pte. Ltd.Blk 4010 Ang Mo Kio Ave 10#06-06 Techplace 1Singapore 569626Tel. +65 9680 8933www.polytec-sea.cominfo@polytec-sea.com