Download presentation

Presentation is loading. Please wait.

Published byWilson Avery Modified over 2 years ago

1
System Identification: a Cornerstone of Structural Design in the Aerospace and Automotive Industries Herman Van der Auweraer SCORES Workshop Leuven,

2
SCORES 2004 Leuven 12/10/04 2 Overview mObjective: To discuss the vital importance of System Identification in the Mechanical Design Engineering Process mTo identify the specific challenges for this kind of problems and to illustrate the research needs mIllustrate with typical products: cars, aircraft, satellites, …. where adequate mechanical product behaviour is vital

3
SCORES 2004 Leuven 12/10/04 3 Overview mIntroduction: the role of Structural Dynamics in Mechanical Design Engineering mApproach and methodology for Structural Dynamics Analysis: Experimental Modal Analysis mModal Parameter Identification methods mApplications of modal analysis mRecent evolutions and challenges for the future mConclusions

4
SCORES 2004 Leuven 12/10/04 4 Introduction Mechanical Design Engineering mMarket Demand: Delivering products with the required mechanical characteristics: Excel in m Operational quality (performance specifications…) m Reliability (load tolerance, fatigue, life-time…) m Safety (vehicle crash, aircraft flutter….) m Comfort (noise, vibration, harshness) m Environmental impact (emissions, waste, noise, recycling…) mProcess process challenges: Excel in m Time-to-Market: reduce design cycle m Reduce design costs m Product customization

5
SCORES 2004 Leuven 12/10/04 5 Introduction Economic Impact: Some Figures Typical vehicle development programs require investment budgets of B$Typical vehicle development programs require investment budgets of B$ New Mercedes C-class (Automotive Engineering Intl., Aug. 2000) :New Mercedes C-class (Automotive Engineering Intl., Aug. 2000) : 600 M$ development + 700M$ production facilities600 M$ development + 700M$ production facilities Developed in less than 4 yearsDeveloped in less than 4 years New Mini: 200M£ development costs (+ as much in marketing...)New Mini: 200M£ development costs (+ as much in marketing...) Chrysler minivan (“The Critical Path” by Brock Yates) :Chrysler minivan (“The Critical Path” by Brock Yates) : 2 B$ development budget, of which 250 M$ R&D2 B$ development budget, of which 250 M$ R&D 36 different body styles, 2 wheelbases, 4 engines36 different body styles, 2 wheelbases, 4 engines

6
SCORES 2004 Leuven 12/10/04 6 Introduction Time Pressure Increases Recall Risks Warranty costs may explode the overall budget 2000 warranty cost (Mercedes-Benz) : 1.5 b$2000 warranty cost (Mercedes-Benz) : 1.5 b$ Warranty cost exceeds R&D costWarranty cost exceeds R&D cost Warranty cost x 3 in 2 years...Warranty cost x 3 in 2 years...

7
SCORES 2004 Leuven 12/10/04 7 Introduction Mechanical Design Engineering mEarly Design Optimization is Essential mProduct design has to go beyond the “Form and Fit” mFocus on “Functional Performance Engineering” mFor mechanical performances: structural analysis mStatic: strength, load analysis mKinematic: mechanisms, motion mDynamic: vibrations, fatigue, noise mBasic approach: is through the use of structural models mA priori (Finite Element) and experimental (Modal) mAnalyze the effect of dynamic loads mUnderstand the intrinsic structural dynamics behaviour mDerive optimal design modifications

8
SCORES 2004 Leuven 12/10/04 8 System Transfer SourceReceiver Accessories Environmental Sources Total Vehicle System Road Input Wheel & Tire Unbalance Steering Wheel Shake Seat Vibration Rearview mirror vibration Noise at Driver’s & Passenger’s Ears TACTILE VISUAL ACOUSTIC Engine Introduction: A Systems Approach A Source-Transmitter-Receiver Model X =

9
SCORES 2004 Leuven 12/10/04 9 Overview mIntroduction: the role of Structural Dynamics in Mechanical Design Engineering mApproach and methodology for Structural Dynamics Analysis: Experimental Modal Analysis mModal Parameter Identification methods mApplications of modal analysis mRecent evolutions and challenges for the future mConclusions

10
SCORES 2004 Leuven 12/10/04 10 Experimental Modal Analysis Principles mStructural dynamics modelling: relating force inputs to displacement/acceleration outputs mMultiple D.o.F. System: mContinuous structures approximated by discrete number of degrees of freedom -> Finite Element Matrix Formulation mMajority of methods and applications: Linear and Time- Invariant models assumed

11
SCORES 2004 Leuven 12/10/04 11 Experimental Modal Analysis Principles mModal Analysis: Related to Eigenvalue Analysis mTime domain equation mLaplace domain equation mEigenvalue analysis -> system poles and Eigenvectors m System pole -> Resonance frequency and damping value m Eigenvector -> Mode shape m Transformation vectors to “Modal Space”

12
SCORES 2004 Leuven 12/10/04 12 Experimental Modal Analysis Principles mModal Shape: Eigenvector in the physical space: physical interpretation (Example “Skytruck”)

13
SCORES 2004 Leuven 12/10/04 13 Modal Analysis Principle; Decomposition in Eigenmodes mModal Analysis: The modal superposition = a1a1a1a1 a2a2a2a2 a3a3a3a3 a4a4a4a4 x x x x + + … …

14
SCORES 2004 Leuven 12/10/04 14 Experimental Modal Analysis Principles mModal Analysis: An input/output relation mTransfer Function Formulation: mModel reduction (Finite number of modes):

15
SCORES 2004 Leuven 12/10/04 15 Experimental Modal Analysis Principles mExperimental Analysis: using input/output measurements mNon-parametric estimates (FRF, IR) -> Data reduction mBlack box models (ARX, state-space) mModal models mStandard experimental modal analysis approach: Fitting the Transfer Function model by Frequency Response Function measurements H u(t) U(ω) y(t) Y(ω) Input System Output

16
SCORES 2004 Leuven 12/10/04 16 Experimental Modal Analysis Test Procedure ExcitationExcitation Shakers (Random, Sine) or Hammer (Impulsive)Shakers (Random, Sine) or Hammer (Impulsive) Load cell for force meas.Load cell for force meas. ResponseResponse AccelerometersAccelerometers Laser (LDV)Laser (LDV) Cross-spectra averaging to estimate FRFsCross-spectra averaging to estimate FRFs Measurement systemMeasurement system FFT analyzer (2-4 channel)FFT analyzer (2-4 channel) PC & data-acquisition front-end ( channels)PC & data-acquisition front-end ( channels) “patching” -> non- simultaneous data“patching” -> non- simultaneous data

17
SCORES 2004 Leuven 12/10/04 17 Experimental Modal Analysis: Aircraft Test Setup ExampleInputsResponses 1 row or column suffices to determine modal parameters1 row or column suffices to determine modal parameters ReciprocityReciprocity

18
SCORES 2004 Leuven 12/10/04 18 Experimental Modal Analysis A Typical Experiment Vehicle Body Test F : 2 inputsF : 2 inputs Indicated by arrowsIndicated by arrows X : 240 outputsX : 240 outputs All nodes in pictureAll nodes in picture H has 480 elements H FXInputSystemOutput Vertical force Horizontal force X = H * F

19
SCORES 2004 Leuven 12/10/04 19 Experimental Modal Analysis Typical FRFsIndustrial Gear box Vehicle Subframe

20
SCORES 2004 Leuven 12/10/04 20 Experimental Modal Analysis Typical FRFs Engine block driving point FRF Engine block FRF

21
SCORES 2004 Leuven 12/10/04 21 Experimental Modal Analysis Ambient Excitation Tests mMany applications do not allow input/output tests m No possibility to apply input m Typical product loading difficult to realise (non-linear effects) m Large ambient excitation levels present mSpecific approach: m Use output-only data (responses) m Assume white noise excitation mReduce output data to covariances or cross-powers

22
SCORES 2004 Leuven 12/10/04 22 Experimental Modal Analysis The Analysis Process mModal Analysis: identification of modal model parameters from the FRF (or Covariances) mSpecific problems: mLarge number of inputs/outputs, long records (noisy data) mNon-simultaneous I/O measurements mHigh system orders, order truncation, modal overlap mLow system damping ( %), Large dynamic range mSpecific approach: mSimultaneous (“global”) analysis of all reduced (FRF) data mOrder problem: Repeated analysis for increasing orders -> The stabilisation diagram -> The stabilisation diagram

23
SCORES 2004 Leuven 12/10/04 23 Experimental Modal Analysis Principles mExperimental Modal Analysis: using FRF measurements in a reduced set of structural locations

24
SCORES 2004 Leuven 12/10/04 24 Overview mIntroduction: the role of structural dynamics in Mechanical Design Engineering mApproach and methodology for structural dynamics analysis: experimental modal analysis mModal Parameter Identification methods mUsually taking into account the physical model mUse of raw time data exceptional -> reduced FRF models mTime and frequency domain approaches mIndustrial and societal applications of modal analysis mRecent evolutions and challenges for the future mConclusions

25
SCORES 2004 Leuven 12/10/04 25 Modal Model Parameter Identification Main Methods Frequency domain methods: rational polynomial FRF model Nonlinear in the unknowns Common denominator methods Partial fraction expansion methods Linearized methods State space formulations (“Eigensystem Realization”)

26
SCORES 2004 Leuven 12/10/04 26 Modal Model Parameter Identification Main Methods Linear frequency domain methodLinear frequency domain method Weighted or notWeighted or not LS, TLSLS, TLS Maximum Likelihood: takes data variance into account -> Non- linear error formulation -> iterative; Error bounds!!Maximum Likelihood: takes data variance into account -> Non- linear error formulation -> iterative; Error bounds!! Continuous or discrete frequency domainContinuous or discrete frequency domain Preferred approach: “PolyMAX”, Least Squares Discrete Frequency Domain LS/TLS, originating from VUB.Preferred approach: “PolyMAX”, Least Squares Discrete Frequency Domain LS/TLS, originating from VUB.

27
SCORES 2004 Leuven 12/10/04 27 Modal Model Parameter Identification Main Methods Time domain: Complex damped exponential approach (UC)Time domain: Complex damped exponential approach (UC) Impulse responses or correlations are solutions of the “characteristic equation”Impulse responses or correlations are solutions of the “characteristic equation” Poles: found as eigenvalues of [W i ] companion matrixPoles: found as eigenvalues of [W i ] companion matrix Modeshapes: Least-squares fit of FRF matrixModeshapes: Least-squares fit of FRF matrix

28
SCORES 2004 Leuven 12/10/04 28 Time domain: Discrete time state space model -> Subspace method In particular used with output-only data: stochastic subspace Estimate [A] and [C] from output-only data (KUL…) covariances (INRIA): Modal Model Parameter Identification Main Methods

29
SCORES 2004 Leuven 12/10/04 29 Modal Model Parameter Identification Main Methods mStabilisation diagram: discrimination of physical poles versus mathematical/spurious poles -> heuristic approach

30
SCORES 2004 Leuven 12/10/04 30 Overview mIntroduction: the role of structural dynamics in Mechanical Design Engineering mApproach and methodology for structural dynamics analysis: experimental modal analysis mModal Parameter Identification methods mApplications of modal analysis mRecent evolutions and challenges for the future mConclusions

31
SCORES 2004 Leuven 12/10/04 31 EMA Example: Aircraft Modal Analysis Component DevelopmentComponent Development Engine, landing gear, ….Engine, landing gear, …. Aircraft Ground Vibration TestsAircraft Ground Vibration Tests Low frequency: 0 … 20… 40 HzLow frequency: 0 … 20… 40 Hz > 50 orders, > 250 DOF> 50 orders, > 250 DOF Model Validation & updatingModel Validation & updating Flutter predictionFlutter prediction

32
SCORES 2004 Leuven 12/10/04 32 EMA Example: Aircraft Modal Analysis (Dash 8)

33
SCORES 2004 Leuven 12/10/04 33 EMA Example: Aircraft Modal Analysis for Aeroelasticity (Flutter) Frequency (Hz ) Damping (%) Airspeed (kts)

34
SCORES 2004 Leuven 12/10/04 34 EMA Example: Aircraft FE Model Correlation and UpdatingFEM GVT FEM GVT FEMEigenfrequencycorrelation Mode shape Correlation (MAC) Courtesy H. Schaak, Airbus France + 5% - 5%

35
SCORES 2004 Leuven 12/10/04 35 EMA Example: Business Jet, Wing-Vane In-Flight Excitation In-flight excitation, 2 wing-tip vanesIn-flight excitation, 2 wing-tip vanes 9 responses9 responses 2 min sine sweep2 min sine sweep Higher order harmonicsHigher order harmonics Very noisy dataVery noisy data PolyMAX

36
SCORES 2004 Leuven 12/10/04 36 In-Operation Modal Analysis Example: PZL-Sokol Helicopter Testing Flight tests in different conditions (speed, climbing, hover…)Flight tests in different conditions (speed, climbing, hover…) 3 flights needed, 90 points3 flights needed, 90 points Correlation lab. / flight resultsCorrelation lab. / flight results No problem with rotor frequenciesNo problem with rotor frequencies MR-I ODS 6.4 Hz mode

37
SCORES 2004 Leuven 12/10/04 37 EMA Example: Car Body and Suspension Tests Suspension EMA for a rolling-noise problem : Booming noise at 80HzSuspension EMA for a rolling-noise problem : Booming noise at 80Hz Main contribution from rear suspension mountsMain contribution from rear suspension mounts Body EMA for basic bending and torsion analysis (vehicle stiffness)

38
SCORES 2004 Leuven 12/10/04 38 EMA Example: Civil Structures Dynamics Øresund Bridge Input-output testing Output-only testing

39
SCORES 2004 Leuven 12/10/04 39 Example: Civil Structures - The Vasco da Gama Bridge In-operation Modal Analysis Covariance Driven Stochastic Subspace

40
SCORES 2004 Leuven 12/10/04 40 Overview mIntroduction: the role of structural dynamics in Mechanical Design Engineering mApproach and methodology for structural dynamics analysis: experimental modal analysis mModal Parameter Identification methods mApplications of modal analysis mRecent evolutions and challenges for the future mConclusions

41
SCORES 2004 Leuven 12/10/04 41 Industrial Model Analysis: What are the issues and challenges? Optimizing the Test processOptimizing the Test process Large structures (> 1000 points, in operating vehicles…)Large structures (> 1000 points, in operating vehicles…) –Novel transducers (MEMS, TEDS…) –Optical measurements Complex structures, novel materials, high and distributed damping (uneven energy distribution)Complex structures, novel materials, high and distributed damping (uneven energy distribution) –Multiple excitation (MIMO Tests) –Use of a priori information for experiment design –Nonlinearity checks, non-linear model detection and identification –Excitation Design: Get maximal information in minimal time

42
SCORES 2004 Leuven 12/10/04 42 Industrial Model Analysis: What are the issues and challenges? Optimizing the Analysis processOptimizing the Analysis process High model orders, numerical stabilityHigh model orders, numerical stability Discrimination between physical and “mathematical” polesDiscrimination between physical and “mathematical” poles Automated modal analysisAutomated modal analysis Test and analysis duration and complexityTest and analysis duration and complexity Test-right-first-timeTest-right-first-time Support user interaction with “smart results”Support user interaction with “smart results” Automating as much as possible the whole processAutomating as much as possible the whole process Quantifying data and result uncertaintyQuantifying data and result uncertainty -> bring intelligence in the test and analysis process

43
SCORES 2004 Leuven 12/10/04 43 1 1 x2x2 x1x1 hid1 hid2 2 2 x2x2 Automatic Assessment and Classification of FRF Quality and Plausibility Innovation and Challenges: Data Quality Assessment Coherence analysis (225 spectral lines X 540 DOFs)

44
SCORES 2004 Leuven 12/10/04 44 Uncertainty and Reliability: A Research Context Methods to assess uncertainty and variability of CAE models:Methods to assess uncertainty and variability of CAE models: Input distribution -> response distributionInput distribution -> response distribution Fuzzy-FE, transformation method, Monte-Carlo…Fuzzy-FE, transformation method, Monte-Carlo… Robust design and reliability considerationsRobust design and reliability considerations What about test data confidence limits?What about test data confidence limits? IN OUT Uncertainty in front craddle Young’s modulus ( GPa)Young’s modulus ( GPa) mass density ( kg/m 3 )mass density ( kg/m 3 ) shell thickness ( mm)shell thickness ( mm)

45
SCORES 2004 Leuven 12/10/04 45 Mimic the human operator (rules, implicit -> NN)?Mimic the human operator (rules, implicit -> NN)? Iterative methods (MLE)Iterative methods (MLE) Fundamental issue: discriminate mathematical and physical polesFundamental issue: discriminate mathematical and physical poles Indicators (damping value, p-z cancellation or correlation…) Indicators (damping value, p-z cancellation or correlation…) Fast stabilizing estimation methods Fast stabilizing estimation methods Clustering techniques Clustering techniques Innovation and Challenges: Automating Modal Parameter EstimationPolyMAX

46
SCORES 2004 Leuven 12/10/04 46 Industrial Model Analysis: What are the issues and challenges? Novel applicationsNovel applications Combined Ambient – I/O testingCombined Ambient – I/O testing Nonlinear system detection and identificationNonlinear system detection and identification Build system-level models combining EMA and FE modelsBuild system-level models combining EMA and FE models Vibro-acoustic modal analysis: include cavity modelsVibro-acoustic modal analysis: include cavity models Mechatronic and controlMechatronic and control End-of-line controlEnd-of-line control Model-based monitoring Model-based monitoring …..….. Healthy structure Damaged structure 2 nd mode shape

47
SCORES 2004 Leuven 12/10/04 47 HSS Engine Mounts Bushings Subframe&Crossmember BodyVibro-acoustics Engine&Brackets Hybrid System Synthesis Innovative Applications: Building Hybrid System Models

48
SCORES 2004 Leuven 12/10/04 48 Acoustic resonances, coupled structural-acoustical behaviour can be modelled by vibro-acoustic modal modelsAcoustic resonances, coupled structural-acoustical behaviour can be modelled by vibro-acoustic modal models Innovative Applications: Vibro-Acoustic Modal Analysis Excitation by shakers and loudspeakers -> Balancing of test data needed (p/f, x/f, p/Q, x/Q)Excitation by shakers and loudspeakers -> Balancing of test data needed (p/f, x/f, p/Q, x/Q) Non-symmetrical modal modelNon-symmetrical modal model Through structural acoustic couplingThrough structural acoustic coupling Different right and left eigenvectorsDifferent right and left eigenvectors

49
SCORES 2004 Leuven 12/10/04 49 f = 32.9 Hz = 8.5% Vibro-Acoustic Modal Analysis Example: Aircraft Interior Noise f = 78.3 Hz = 7.0% ATR42 F100

50
SCORES 2004 Leuven 12/10/04 50 Summary and Outlook Early product optimization is essential to meet market demandsEarly product optimization is essential to meet market demands Mechanical Design Analysis and Optimization heavily rely on Structural ModelsMechanical Design Analysis and Optimization heavily rely on Structural Models Experimental Modal Analysis is the key approach, it is a de-facto standard in many industriesExperimental Modal Analysis is the key approach, it is a de-facto standard in many industries While EMA is in essence a system identification problem, particular test and analysis issues arise due to model size and complexityWhile EMA is in essence a system identification problem, particular test and analysis issues arise due to model size and complexity Important challenges are related to supporting the industrial demands (test time and accuracy) and novel applicationsImportant challenges are related to supporting the industrial demands (test time and accuracy) and novel applications Research efforts should also pay attention to “state-of-the-use” breakthroughsResearch efforts should also pay attention to “state-of-the-use” breakthroughs

Similar presentations

© 2017 SlidePlayer.com Inc.

All rights reserved.

Ads by Google