James A. Harter Senior Consultant for LexTech, Inc. Effect of Grain Orientation on Round Robin Life Predictions

Transcription

1 James A. Harter Senior Consultant for LexTech, Inc. Effect of Grain Orientation on Round Robin Life Predictions

2 2017 AFGROW Workshop Round Robin Test Effort Test Specimens Specimen Testing Crack Measurements/Data Reduction Crack Growth Rate Data Round Robin Instructions

3 Round Robin Participants

4 Purpose Determine the ability to predict the evolution of crack shape and total life of a corner crack at a hole using the same loading spectrum, material data, and a given Initial Flaw Size (IFS),

5 Centered Hole Test Specimen Geometry

6 Test Specimen Replicate Data PC-CX-7075 Max Spectrum Stress 12 ksi Serial Feature Description SwRI Measure SAFE Measure Pass? SAFE Tool Used -1 1 Width Yes SAFE Caliper Thickness Yes SAFE Mike 59/ Hole Diameter Yes SAFE Inside Mike Length " Mito Caliper -1 7 Notch Width Nikon Microscope (W/ DRO) -1 8 Notch Length Nikon Microscope (W/ DRO) -1 Initial crack (bore) Nikon MM-60 (APES) -1 Initial crack (surface) Nikon MM-60 (APES) -1 Alternate IC (bore) Nikon MM-60 (APES) -1 Alternate IC (surface) Nikon MM-60 (APES) Max Spectrum Stress 12 ksi Serial Feature Description SwRI Measure SAFE Measure Pass? SAFE Tool Used -2 1 Width Yes SAFE Caliper Thickness Yes SAFE Mike 59/ Hole Diameter Yes SAFE Inside Mike Length " Mito Caliper -2 7 Notch Width Nikon Microscope (W/ DRO) -2 8 Notch Length Nikon Microscope (W/ DRO) -2 Initial crack (bore) Nikon MM-60 (APES) -2 Initial crack (surface) Nikon MM-60 (APES) -2 Alternate IC (bore) Nikon MM-60 (APES) -2 Alternate IC (surface) Nikon MM-60 (APES) Max Spectrum Stress 12 ksi Serial Feature Description SwRI Measure SAFE Measure Pass? SAFE Tool Used -3 1 Width Yes SAFE Caliper Thickness Yes SAFE Mike 59/ Hole Diameter Yes SAFE Inside Mike Length " Mito Caliper -3 7 Notch Width Nikon Microscope (W/ DRO) -3 8 Notch Length Nikon Microscope (W/ DRO) -3 Initial crack (bore) Nikon MM-60 (APES) -3 Initial crack (surface) Nikon MM-60 (APES) -3 Alternate IC (bore) Nikon MM-60 (APES) -3 Alternate IC (surface) Nikon MM-60 (APES)

7 Test Spectrum

8 Marking load sequence used to capture crack length and aspect ratio data

9 Crack Growth Rate Data

10 Participants MacAllan - A MacAllan - B MacAllan - C Dalwhinne FinLaggan Black Label Dewars Jura Cathead K-Solution - Center K-Solution - Offset Two Point Advanced Two Point Advanced Two Point Advanced Advance Classic FEA Advanced w/b Correction FEA Advanced Unknown Unknown Load Interaction None Hsu Hsu, b R b R None None None Unknown Unknown a/c constant No No No No No No No No No

11 Life Prediction Results

12 Crack Length, a-tip (in) Crack Length, c-tip (in) 0.30 PC-CX a-tip 2.0 PC-CX c-tip Test MacAllan A MacAllan B MacAllan C Finlaggan Black Label Test MacAllan A MacAllan B MacAllan C Dalwhinnie Finlaggan Black Label 0.05 Dewars A Dewars B 0.4 Dewars A Dewars B Jura 0.2 Jura ,000 20,000 30,000 40,000 50,000 Cycles Cathead ,000 40,000 60,000 80, , ,000 Cycles Cathead

13 Crack Length, a-tip (in) Crack Length, c-tip (in) 0.30 PC-CX a-tip 2.0 PC-CX c-tip Test 1.2 Test 0.15 MacAllan A MacAllan B 1.0 MacAllan A MacAllan B MacAllan C Finlaggan Black Label Dewars B Jura Cathead MacAllan C Dalwhinnie Finlaggan Black Label Dewars B Jura Cathead ,000 20,000 30,000 40,000 50,000 Cycles ,000 40,000 60,000 80, , ,000 Cycles

14 Crack Length, a-tip (in) Crack Length, c-tip (in) 0.30 PC-CX a-tip 2.0 PC-CX c-tip 1.8 Test MacAllan A MacAllan B MacAllan C Test 1.2 Finlaggan MacAllan A MacAllan B MacAllan C Finlaggan Black Label Dewars B Jura Cathead Black Label 0.05 Dewars B 0.4 Jura Cathead ,000 20,000 30,000 40,000 50,000 Cycles ,000 40,000 60,000 80, ,000 Cycles

15 Crack Shape Results

16 PC-CX7075-1

17 PC-CX7075-2

18 PC-CX7075-3

19 Life predictions were generally very good, but the crack shape predictions did not follow the trends in the test data Issues Most of the crack shape predictions converged at a/c ~ 1.5, while the tests converged at ~ 1.4 The crack shape prediction curves crossed the test data points All predictions were made using a single crack growth rate model for each growth direction

20 Single vs. Dual Crack Growth Rate Curve Fits Single Growth Rate Curve Fit: A single crack growth rate curve fit is used to predict growth in both directions (a and c-directions) Dual Growth Rate Curve Fit: Two crack growth rate curve fits are used to predict crack growth. One curve fit model is used for the a- direction, and a different model is used for the c- direction

21 Revisiting the Crack Growth Rate Data Using Tom Mills digitized crack length data and two methods for calculating DK Fawaz/Andersson Newman/Raju

22 Fawaz/Andersson Example: Dual Crack Growth Rate From Specimen Striation Measurements

23 1.E-04 Fawaz/Andersson 1.E-04 Newman/Raju 1.E-05 C Length (Specimen 1) A Length (Specimen 1) Through Crack (Specimen 1) C Length (Specimen 2) A Length (Specimen 2) Through Crack (Specimen 2) C Length (Specimen 3) 1.E-05 C Length (Specimen 1) A Length (Specimen 1) Through Crack (Specimen 1) C Length (Specimen 2) A Length (Specimen 2) Through Crack (Specimen 2) C Length (Specimen 3) A Length (Specimen 3) A Length (Specimen 3) Through Crack (Specimen 3) Through Crack (Specimen 3) C Length Fit (Dual Rate) C Length Fit (Dual Rate) A Length Fit (Dual Rate) A Length Fit (Dual Rate) Round Robin Fit Round Robin Fit 1.E E

24 1.000E-03 Specimen 1 Fawaz/Andersson Fits 1.000E-04 C Length A Length Through Crack 1.000E-05 A Length Fit C Length Fit 1.000E

25 1.0E-04 Specimen 2 Fawaz/Andersson Fits C Length 1.0E-05 A Length Through Crack C Length Fit A Length Fit 1.0E

26 1.000E-04 Specimen 3 Fawaz/Andersson Fits C Length 1.000E-05 A Length Through Crack C Length Fit A Length Fit 1.000E

27 AFGROW Dual Rate Curve Demonstration

28 Postdictions

29 Crack Length, a-tip (in) Crack Length, c-tip (in) 0.30 PC-CX a-tip 2.0 PC-CX c-tip 0.25 Test MacAllan A Test MacAllan A MacAllan B MacAllan C Finlaggan Black Label Dewars A Dewars B Jura MacAllan B MacAllan C Dalwhinnie Finlaggan Black Label Dewars A Dewars B Jura 0.05 Cathead Round Robin Fit Dual Rate Fit Cathead Round Robin Fit Dual Rate Fit ,000 20,000 30,000 40,000 50,000 Cycles ,000 40,000 60,000 80, , ,000 Cycles

30 Crack Length, a-tip (in) Crack Length, c-tip (in) 0.30 PC-CX a-tip 2.0 PC-CX c-tip 0.25 Test MacAllan A Test MacAllan A MacAllan B MacAllan C Finlaggan Black Label Dewars B Jura MacAllan B MacAllan C Dalwhinnie Finlaggan Black Label Dewars B Jura 0.05 Cathead Round Robin Fit Dual Rate Fit Cathead Round Robin Fit Dual Rate Fit ,000 20,000 30,000 40,000 50,000 Cycles ,000 40,000 60,000 80, , ,000 Cycles

31 Crack Length, a-tip (in) Crack Length, c-tip (in) 0.30 PC-CX a-tip 2.0 PC-CX c-tip 0.25 Test MacAllan A Test MacAllan A 0.20 MacAllan B MacAllan C Finlaggan MacAllan B MacAllan C Finlaggan 0.15 Black Label 1.0 Black Label Dewars B 0.8 Dewars B 0.10 Jura Cathead 0.6 Jura Cathead 0.05 Round Robin Fit Dual Rate Fit Round Robin Fit Dual Rate Fit ,000 20,000 30,000 40,000 50,000 Cycles ,000 40,000 60,000 80, ,000 Cycles

32 Crack Shape Postdictions

33 Aspect Ratio (a/c) Specimen Series1 Round Robin Fit Dual Rate Fit Front Face Length (c), inch

34 Aspect Ratio (a/c) Specimen Series1 Round Robin Fit Dual Rate Fit Front Face Length (c), inch

35 Aspect Ratio (a/c) Specimen Series1 Round Robin Fit Dual Rate Fit Front Face Length (c), inch

36 The Crack Shape Predictions Look Much Better Using Dual Crack Growth Rate Data, But the Life Predictions Are Nearly the Same Question? What happens if you assume a constant crack shape? (e.g. a/c = 1)

37 Crack Length, a-tip (in) Crack Length, c-tip (in) PC-CX a-tip PC-CX c-tip Test MacAllan A Test MacAllan A MacAllan B 0.20 MacAllan B MacAllan C 1.4 MacAllan C Dalwhinnie 0.15 Finlaggan Black Label Dewars A Finlaggan Black Label Dewars A 0.10 Dewars B Jura Cathead Dewars B Jura Cathead 0.05 Round Robin Fit Dual Rate Fit 0.4 Round Robin Fit Dual Rate Fit ,000 20,000 30,000 40,000 50,000 Cycles Constant Crack Shape ,000 40,000 60,000 80, , ,000 Cycles Constant Crack Shape

38 Life predictions and postdictions are in good agreement with the test results. Crack shape trends shown in this effort could not be predicted using a single crack growth rate curve. Dual crack growth rate models were required to more accurately match the crack shape data. Summary Additional work is needed to assess the value of using dual (or multiple) crack growth rate curves. More crack growth rate data are needed to model crack growth in multiple directions. An interpolation method will be needed to model crack growth rate for multiple points along a crack front. Crack shape changes can significantly impact life prediction capability for many practical problems (high stress gradients, cold-worked holes, etc.).

39 Final Point Do we expect crack growth rate data to be the same for different grain orientations? Probably not It may be time to consider a new approach.

40 Questions/Comments?