Tutorial: 3D Scanning Vibrometry for. IMAC XXVII D. E. Oliver, Polytec, Inc.

Transcription

1 Tutorial: 3D Scanning Vibrometry for Structural Dynamics Measurements IMAC XXVII D. E. Oliver, Polytec, Inc.

2 Content Principles of Laser Doppler Vibrometry Scanning Laser Doppler Vibrometry (SLDV) Limitations of 1-D SLDV Introducing 3-D SLDV Data Import and Export Post-processing Some Tips and Tricks for Optimal Measurements Example Measurement Procedure Analysis of Results

3 What is Laser Doppler Vibrometry? PDV-100 Portable Digital Vibrometer Laser Doppler l Vibrometry is a non-contact, "point and shoot" technology that directly measures the vibration of a test object using the Doppler effect. Analogy: Acoustic Doppler Effect Sound emitted from stationary car has frequency enc f = c/λ c: velocity of the sound wave λ: emitted wavelength f: emitted frequency λ For car moving at velocity V, the observer hears the frequency f D = c/(λ -V/f). v c Emitted frequency f Observed frequency f D

4 The Optical Doppler Effect Vibrometer = source of frequency f Target = receiver of frequency f Target moving away from source with velocity v et unit vector c v o f = ( ) f c

5 The Optical Doppler Effect Target reflects light back Target = source of frequency f Vibrometer = receiver of frequency f er unit vector c f = f c v s

6 The Optical Doppler Effect f c c v c v o o = f = f c v c c v 1 s s 2 v c f Shift in frequency: u 0 =u s v << c Δf = (f f) = 2v v λ (v parallel to laser beam) Velocity determined from the observed frequency shift Calibration-free sensor head Vibration causes a frequency modulated measurement signal

7 Why do we Use an Interferometer? t A. Electronics are not fast enough to measure the frequency of light ( nm) directly The Doppler shift cannot be measured directly Solution: Interferometry allows us to see RELATIVE changes in phase / frequency of light at frequencies that can be detected B. The Doppler frequency has no sign Direction of movement unkown Solution: A heterodyne interferometer with a fixed frequency offset using a Bragg cell allows us to distinguish between positive and negative Doppler shifts.

8 The Heterodyne Interferometer t Measurement Beam f 0 f 0 ± f D Reflected Beam Bragg cell He-Ne Laser <1mw (633nm) x(t) () v(t) f MHz Photo-detector Frequency Modulated signal 40 MHz ± f D Δ fd = 2V/λ

9 Signal Demodulation Photo-detector t t system FM Doppler signal Controller Voltage ~ Velocity AM electrical signal FFT Spectrum Vlt Voltage ~ Displacement

10 Scanning Vibrometry Sensor Head Test Object Test Object Scan Electronics Video LDV Sensor Scanning Mirrors LDV Controller Data Management System Up to 250,000 points scanned Easy-to-use software for data acquisition, display & manipulation Animated data visualization Efficient interfaces for modal analysis or FEM validation Geometry file imported or measured Stitched data

11 Working Principle i Scan signal Ref signal Measure V and ϕ at many points relative to a reference signal Amplitude and relative phase at each scan point Spatial vibration pattern of measurement surface

12 Operating Deflection Shapes

13 Scan Modes Frequency FFT Zoom FFT Frequency Time Domain FastScan Time Time

14 PSVSoft

15 PSVSoft

16 Some Applications

17 Typical l1-d DScanning Vibrometer PSV-400-H4 4 (or 8) channel data acquisition with 80 khz bandwidth 4 reference channels 40 x 40 scan angle 6 velocity ranges up to 10 m/s Resolution to µm/s/ Hz Autofocus MIMO capable. 3 independent generator channels; Principal Component Analysis

18 Directionality A vibrometer only measures the velocity component in the direction of the laser beam Target V x Vibrometer λ Wave front e Θ v Δ f = 2v.e λ = 2v λ cos θ θ= % error θ= 8 1% error θ= % error

19 Limitations it ti of a 1-D DVibrometer Laser vibrometers measure vibration in laser direction only Careful data interpretation due to vibration directions, surface shape and varying scan angles Limitations when validating FE models Laser Vibrometer Θ

20 An Introduction to 3-D Scanning Laser Vibrometry

21 i i i i 3 D object coordinate system defined by means Principles of 3-D Scanning Laser Vibrometry 3-D object coordinate system defined by means of a minimum of 3 known reference points Laser beam unit vectors for all scan heads are y z determined using the transformation matrix: v l l l l l l v z y x x z = v v l l l l l l v v z y x z y x z y x 3D alignment

22 Principles i of 3-D Scanning Laser Vibrometry 3 synchronized laser scanners with central control Geometry imported or measured Lasers intersect at all scan points on surface Simultaneous measurement of 3 vibration components y z x

23 Options for Obtaining i Geometry File Data import (via UFF): a) Other test software packages b) Finite Element Model Geometry File c) Extracted from other methods such as CAT scan When there is no geometry file: a) Alignment of all 3 laser beams manually for every scan point b) Measure geometry for every scan point

24 Geometry Scan Unit Time-of-flight Geometry Unit detector laser source beam splitter x-deflection mirror Scan Head Interfero- meter actuated mirror y-deflection mirror Geometry data combined with video image

25 PSV-400-3D-M Junction Box Hi. Res. Camera Data Management System Laser Controllers Scan heads Motorized tripod Tabletop tripod 3-D data acquisition system and scanning heads 4-channel data acquisition system (8 channels 80kHz PSV-400-3D - MIMO possible) ~DC - 1 MHz vibration frequency range 10 m/s max. velocity

26 Solution Measurement Process Segment Geometry Measure 3D Vibration

27 Solution Measurement Process Input Segment Measure Stitch Deflection Export for Geometry Geometry Vibration Segments Shape Processing Analysis Experimental Modal Analysis Experimental Modal Analysis FE Correlation & Model updating

28 Sample Data Deflection shape at 82 Hz showing individual measurement points

29 Geometry and dvibration Data Export Data export (for example via UFF): Multi-channel FFT data for each measurement point Universal Data Block: Header Units 2411/2 - Geometry Nodes 55 - Frequency Bands 58 - Data Scan_Broadband_2kHz.svd d d PSV Version Oct 07 09:20:18 None None PolyUFFExport 1.60 Compatible to VIBRANT 10 Oct 07 16:31: METRIC_ABS_(SI) E E E E e e e+00

30 Signal Processing with SigPro Integrated functions Recalculates data and presents results in PSVSoft Spreadsheet functionality: Time and FFT data copied to cells Mathematical Functions: +,-,*,/ scalar or vectorial Cos, sin, exp, log, Re, Im FFT Inverse FFT Digital filtering Re-sampling Extract Statistics Filter transmission Integration, differentiation...

31 Polytec File Access Object Model Collection of objects written according to Component Object Model (COM) of Microsoft Allows access to all data stored in PSVSoft files from any (scripting or programming) language that supports COM; e.g. Visual Basic, Delphi, C++, Matlab, LabVIEW,...) Automatically installed with PSVSoft, but can also be installed independently (no hardlock required) Data organized in hierarchy the same as menu structure of PSVSoft Domain Channel Signal Display Type

32 Example Code Installed on local hard drive with Polytec File Access Examples for DotNET, Excel, Macros, Matlab, VC Excel Polyfile.xls (access and display data stored in PSV/VibSoft data file Matlab: - GetPointData.m (reads.svd and.pvd data into Matlab) - SetPointData.m (writes processed data back into.svd /.pvd file) - Get0dBReference.m (is used with GetPointData.m and SetPointData.m) - GetXYZCoordinates.m (read in geometry data) - GetIndexOfPoint.m (returns original index) - GetBandData.m

33 Advanced d Signal Processing with Matlab Read Polytec data into Matlab using Polytec File Access and perform advanced d signal processing Save processed signal back into PSVSoft for easy display and animation Example Matlab scripts are available PSVSoft data Data processed in Matlab Waterfall display Spectrogram

34 Macro Programming with Visual Basic Engine Automation of recurrent tasks Every user action in PSVSoft can be automated with integrated Visual Basic Editor Object-oriented programming language using Polytec File Access Easy navigation through h the object model with dot (.) operator Various example macros included with software

35 Example Macro: SteppedFastScan.bas Macro for automatically performing a set of measurements (FastScans) while stepping up the excitation ti frequency by a set increment Result: frequency response functions across defined frequency range for all the measurement points.

36 Remote Operation Using LabVIEW Remotely perform measurements and operation of PSVSoft - With PSV 8.6, PSVSoft not necessary any more, but hardlock required Open ActiveX within LabVIEW allows to access objects through Polytec File Access Example LabVIEW vi: remotely set generator parameters

37 Optimizing S/N and Reducing Effects of Speckle Noise Specular Mirror 1 mm/s/v, 20 khz Noise level: 2-5 nm/s/ Hz

38 Optimizing S/N Diffuse White plate 1 mm/s/v, 20 khz Noise level: µm/s/ Hz Diffuse Black plate 1 mm/s/v, /V 20 khz Noise level: µm/s/ Hz

39 Tips and Tricks What is Speckle? Laser beam Bright speckle Dark spec kle D σ 1.22λL/D S 8λL² /D² L Speckle patterns are produced when a coherent light is focused onto a rough surface

40 Tips and Tricks What is Speckle? If there is a lateral movement*, different speckles arrive at the vibrometer lens (and therefore detector) High probability that a dark speckle passes over the detector at some time. Optical phase shift Short periods of low optical signal = speckle dropouts! * lateral movement > focused beam diameter => speckle pattern changes noticeably

41 Strategies t Against Dropouts Collect more light Optics Treat surface to backscatter more light Set optimal working distance and focus Reduce electronic noise level l in decoder Electronics Turn on tracking filter to bridge low signal periods Apply adaptive filter for periodic signals Search for bright speckle, (weighted) averaging: signal enhancement, Post process data to eliminate Data processing dropouts Track moving surfaces

42 Tips and Tricks - Optimize Decoding Relation between HF-Bandwidth and dropoutprobability (example: p VD-07 decoder) Range (mm/s/v) BW (khz) Dropout Probability* % % % % Necessary bandwidth: 2 x (Δf Doppler + f Vib ) * Depends on setup Example: f Vib, max = 20 khz, v max = 10 mm/s -> Δf Doppler, max = 32 khz Necessary BW = 104 khz

43 The Training Structure A Typical Measurement Procedure 12 diameter globe on a foam base Shaker and amplifier Impedance sensor with force and acceleration output Stinger set Wooden d baseboardb

44 Initial lset-up Set up globe + shaker + stinger on table Position console Connect cables

45 Cabling 1

46 Cabling 2

47 Setting up Globe and Shaker System

48 Setting up Globe and Shaker System Don t set shaker amplifier on same table!

49 Positioning i Scan Heads Set heads at a visibility maxima distance D=99mm +/-n*204mm (in this case 711mm or 28 ) Separation i ~ equilateral l triangle with 500mm (20 ) sides Surface treatment is recommended for optimal 3-D measurements

50 Sequence of Events Turn on controllers Align heads Boot up computer Set preferences

51 Video Camera Setup Autofocus camera image Maximize image size covering area of interest

52 2-D DAlignment Perform 2-D coordinate alignment using and

53 3-D DAlignment t1 Perform 3-D coordinate alignment using at least 3 coordinate points with known x, y, z coordinates and 4 alignment points Calculate alignment quality and realign manually if necessary

54 3-D DAlignment t2 Select alignment mode : Origin, i Axis, Plane Define 3 coordinate points eg origin, point on x axis and point on x,y,yplane Define 1 additional alignment point Now geometry of any point can be measured with the Now geometry of any point can be measured with the geometry scanner and x,y,z calculated

55 3-D DAlignment t3 When setting up on another side of globe don t delete alignment points only replace coordinate points.

56 Geometry Import To define a scan grid using imported geometry click on (define scan points) then (points) then (geometry import)

57 Scan Area Preparation Delete any points that you don t want to scan. Assign Focus Fast i f l t ll i t f t d t Assign Focus Fast assigns focus values to all points from geometry data Overlap one row of measurements for each scan of a different side.

58 A/D Parameter Settings By viewing sample live data in analyzer window Set A/D parameters such as: Acquisition mode: FFT, zoom FFT, time domain Averaging Window Signal Enhancement and Speckle Tracking Remeasure Channels Frequency span FFT lines Reference channels settings AC or DC coupling Digital filters Trigger Time = (# lines/bandwidth) idth) x (# averages) x (# scan points)

59 A/D Parameter Settings Set the vibrometer: Range in mm/s/v Tracking filter Analog low and high pass filters Set the signal generator: Signal type Check the ACTIVE box

60 A/D Parameter Settings Adjust all of the above to achieve strong coherence

61 Stitch Data Click File. New..Combined File Select data sets to be stitched

62 Dfi Define Frequency Bands

63 CAE workflow RoboVib