Structural validation of offhighway vehicles using service data






















































- Slides: 54
Structural validation of off-highway vehicles using service data Colin J Dodds and Associates Website: www. doddsandassociates. co. uk Contact: cjdodds@doddsandassociates. co. uk Phone: +44 788 055 4590 20 th May 2015 Dodds and Associates 1
Test Engineer’s Handbook Structural testing of complete vehicles, Available at www. doddsandassociates. co. uk Password = walterscott aggregates and components in the laboratory 20 th May 2015 Dodds and Associates 2
Summary Limited number of prototype vehicles and components available for structural validation. The service loading is uncertain and often abusive The weld quality is variable! Validation program Predefined reliability Limited field measurements on 3 -4 prototypes Classical life estimation methodology Introducing the “Gassner” life estimation methodology Results compared with warranty data 20 th May 2015 Dodds and Associates 3
Test vehicle & methodologies • Vehicle – off-road BHL • Specimen Back-hoe • 2 methodologies – Classical (Wöhler) BS 7608 – Gassner prediction based on maximum stress 20 th May 2015 Dodds and Associates 4
The test specimen 20 th May 2015 Dodds and Associates 5
Duty cycle events 20 th May 2015 Dodds and Associates 6
Modelling – boom Trenching event Courtesy ARAI, Pune Combination of Multi-Body Simulation & FEA Max bucket force First approach at design validation 20 th May 2015 Dodds and Associates 7
Poor correlation CAE vs. measurement 20 th May 2015 Dodds and Associates 8
Static load cases to check the force-stress relationship Courtesy ARAI, Pune 31 different load cases covered! 20 th May 2015 Dodds and Associates 9
Let’s get some service data Classical (Wöhler) BS 7608 methodology published in SAE 20 th May 2015 Dodds and Associates 10
Life estimation Classical approach N 20 th May 2015 Dodds and Associates n 11
Gauge location and terminology are very important NOMINAL STRESS LOCAL STRESS 200 Mpa 100 Mpa CAE Measure All gauges should be placed such that the centre of the gauge is 0. 5 t from weld toe and the gauge length should be <0. 2 t where t is the plate thickness. If this is not possible then, the leading edge of the gauge should be 0. 3 t from the weld toe. 20 th May 2015 Dodds and Associates 12
Data acquisition • • • 7 -8 GB of data collected 74 channels of strain data (88 total) Approx. 12 hours of data on multiple events Trenching Rock chipping Rock unsettling Crane application Tree uprooting User specific events Time to collect data 14 days 20 th May 2015 Dodds and Associates Trenching MD 13
Data acquisition Tree uprooting 20 th May 2015 Duck walking Dodds and Associates 14
Events the designer never thought of Did we forget something? Truck loading 20 th May 2015 BH hitting ground Dodds and Associates 15
Yet another event! The Times, 19 th May 2015 20 th May 2015 Dodds and Associates 16
The problem with modelling! If the end conditions do not match the results will not Model trenching Trenching HD 20 th May 2015 Dodds and Associates 17
Weld challenges 20 th May 2015 Dodds and Associates 18
Weld classification 20 th May 2015 Dodds and Associates 19
Weld classifications in specimen D-class 20 th May 2015 F 2 -class Dodds and Associates 20
BS 7608 Weld SN curves N 50 SN curves 20 th May 2015 Dodds and Associates 21
Typical scatter in weld life data Recommended Hot-Spot Stress Design S-N Curves for Fatigue Assessment of FPSOs, S J Maddox 20 th May 2015 Dodds and Associates 22
Measured stress & load spectrum: Trenching 20 th May 2015 Dodds and Associates 23
Measured stress & load spectrum: Craning 20 th May 2015 Dodds and Associates 24
Measured stress & load spectrum: Rock chipping 20 th May 2015 Dodds and Associates 25
Load spectra – all events 20 th May 2015 Dodds and Associates 26
Create Load spectrum for duty cycle Example: Sum of 3 load spectra Duty Load spectrum A, B, C Concept 20 th May 2015 Dodds and Associates 27
Estimate percentile max stress in worst case event Stress time series Max value in Δt Max value PDF Scale load spectrum to maximum value Max value CDF 20 th May 2015 Dodds and Associates 28
Reliability concept Equate the weakest component to the highest load 20 th May 2015 Dodds and Associates 29
Methodology overview Palmgren-Miner 20 th May 2015 Dodds and Associates 30
Present results Numbers represent strain gauge locations Scatter in SN curve Uncertain service loads Weld quality and residual stress! 20 th May 2015 Dodds and Associates 31
A better approach The designer can apply it! 20 th May 2015 Dodds and Associates 32
Life estimation: Gassner approach Load spectra & corresponding SN curves 20 th May 2015 Dodds and Associates 33
Design life concepts Wöhler & Gassner life curves 1908 -88 Gassner established the topic of "Betriebsfestigkeit" (operational fatigue strength) which can be described as follows: dimensioning (sizing) of a component for finite, but sufficient fatigue life under variable loads 20 th May 2015 Dodds and Associates V. Grubisic 34
Wöhler & Gassner SN curves 20 th May 2015 Dodds and Associates 35
Simplicity Gassner life estimation 20 th May 2015 Dodds and Associates 36
Gassner life estimation concept 20 th May 2015 Dodds and Associates 37
Concept, continued Life is estimated based on an assumed Gaussian load spectrum (which has a low probability of occurrence) with the maximum range equal to the maximum range in the duty cycle. The method has been proven to give a safe working life estimate. The red dot is the design life (20, 000 hours) at the required reliability based on this Gaussian load spectrum. Depending on the real load spectrum of the duty cycle, A, B or C, life is estimated at points 1, 2 or 3. and is independent of the shape of the duty cycle. Point 1 is to the left of the 10% failure probability for the weld and hence will have a very low probability of failure if at all. Point 2 is the same as the design life even though the actual N 10 life will be longer as the duty cycle is less severe. Point 3 is to the right where the probability of failure will be > 10%. In this case the duty cycle lasts longer than the design duty cycle (20, 000 hours) 20 th May 2015 Dodds and Associates 38
Gassner step 1 Create the filter Convert to Stress Use trenching data from Gauge B 14 To create frequency filter for the Gaussian data 20 th May 2015 Dodds and Associates 39
Filters from other gauges The frequency spectra of all gauges on the component have the same shape 20 th May 2015 Dodds and Associates 40
Gassner step 2 Create a Gaussian signal • • Gaussian white noise through the filter Compare – Signals – PSDs – Load spectra 20 th May 2015 Dodds and Associates 41
Compare the Gaussian and service stress signals 20 th May 2015 Dodds and Associates 42
Compare the filter and response PSDs 20 th May 2015 Dodds and Associates 43
Compare the Gaussian & service load spectra Max range 350 MPa Number of cycles = 1. 25 E 4 to cut-off Time = 22 min Gauge B 14 20 th May 2015 Dodds and Associates 44
Compare level crossings Red = Gaussian Blue = service 20 th May 2015 Dodds and Associates 45
Other service load spectra 20 th May 2015 Dodds and Associates 46
Load spectra from a road vehicle chassis Stress cycles/km - 550 -600 km 20 th May 2015 Dodds and Associates 47
Gassner step 3 Generate the Gaussian SN curves
Test specimens under random loading Compared with sinusoidal tests the random test will take 50 ~ 100 times longer per specimen Not an impossible task – Random loading tests on motorcycle frame welds 20 th May 2015 Dodds and Associates 49
Compare the Gassner life curve with estimated lives from field data using Wöhler curve Gassner line 20 th May 2015 Life estimates from sampled field data Dodds and Associates 50
Easy to understand & use 20 th May 2015 Dodds and Associates 51
Does it work? Proc. IMech. E J Rail & Rapid Transit 226(5) 518 -529 2012 20 th May 2015 Dodds and Associates 52
Design stress for good welds for a 99. 9% Reliability in 20, 0000 hrs Note: 14, 000 was life of BH in a 20, 000 total life Max allowable STRESS RANGE = 108 MPa 20 th May 2015 Dodds and Associates 53
Gassner as a design tool 20 th May 2015 Dodds and Associates 54