PERFORMANCE OF P/MCOMPONENTS DURINGDYNAMIC LOADING

Worcester Polytechnic InstituteOctober 22-23, 2003

Diana Lados & Diran Apelian

Morris Boorky Powder Metallurgy Research Center

OUTLINE

ß Background (and examples of fatiguestudies from the literature)

ß Objectives

ß Our Approach … Experimental Plan

ß Critical experimental details

BACKGROUNDGeneral perspectives …

Two designconcepts

Defect intolerant Defect tolerant

BACKGROUNDDefect intolerant …

s-N curve e-N curve CSS curve (Basquin eq.): (Coffin-Manson eq.): (cyclic stress-strain eq.):

( ) 'np

' K eD⋅=sD

Ferrous

Nonferrous

Stre

ss, s

HCF (N > 105) LCF (N < 105)

Stra

in r

ange

, De p

( )Ba NA ⋅=s

( )b'f N2

2⋅s=

sD ( )c'f

p N22

⋅e=eD

( )Dp NC ⋅=eD

Stre

ss, s

Strain, e

Monotonic

Cyclic

BACKGROUND

LCF – high plastic deformation/low loading

cycles

HCF – quasielastic behavior/very high

loading cycles

LCFHCF

(r low) 10 < Nftr < 1000 (r high) PM steels (Ni-Mo)

Defect intolerant … contd.

BACKGROUNDDefect tolerant …

Fatigue crack growth curve (LEFM)

(Paris eq.):

(Forman eq.):

( )mKCdN

daD=

( )( ) KKR1

KC

dN

da

FT

m

D-⋅-

D=

DKFT

DKth

I II III

logDK

log(

da/d

N)

P/M iron

BACKGROUNDHigh cycle fatigue (HCF) … P/M iron

Region A:- nucleation of microcracks;

Region B:-appearance of slip bands on the specimen

surface;

Region C:- characteristic S-N curves where final

failure is caused by macrocracking;

BACKGROUNDHigh cycle fatigue (HCF) … P/M iron … contd.

P/M iron

Region I:- mostly closed porosity;- cracking in the specimen interior;- transgranular path between isolated pores;

Region II:- transition from closed to open porosity;- cracks nucleate @ specimen surface at

isolated pores and pore clusters;- some broken sintering necks;

Region III:- pores connected to each other (open)- biphasic material: matrix phase + pore phase;- simultaneous cracks @ specimen surface- broken surface is smooth in both fatigue and

ductile fast fracture regions (brokensintering necks).

BACKGROUNDHigh cycle fatigue (HCF) … P/M iron … contd.

Wateratomized

Reducedsponge

Tension-compression

Planebending

Life (samples from reduced sponge powder) >Life (samples from water atomized powder)

• Axial testing – volume properties

• Bending – surface properties

• Fatigue limit (bending) < Fatiguelimit (axial loading)

BACKGROUNDHigh cycle fatigue (HCF) … P/M steels

Fe-1.75Ni-1.5Cu-0.5Mo-0.6C (TM&S) Fe-2Cu-0.8C (P&F)

Fatigue life increases with increasing density andpore shape factors BUT density alone can not

describe fatigue behavior of such PM materials pore/matrix interactions

BACKGROUNDHigh cycle fatigue (HCF) … P/M steels … contd.

Fe-1.75Ni-1.5Cu-0.5Mo-0.6C Fe-1.75Ni-1.5Cu-0.85Mo-0.6C

• no significant difference between binder-treated and diffusion alloyed

• If Mo Fatigue life

BACKGROUNDLow cycle fatigue (LCF) … P/M steels

Fe-1.75Ni-0.5Mo

With decreasing densitythe differences betweenthe strain life curvesbecome smaller

Increasing porosityreduces microstructuralinfluence on fatigue life

BACKGROUNDCyclic stress-strain … P/M iron

I. Pure elastic response

II. Microcracks opening (plastic strain-softening)

III. Pronounced opening of microcracksoverrides matrix hardening

IV. Growth of macro-cracks / final failure

Changes in hysteresis loop indicatehardening/softening of the material

BACKGROUNDCyclic stress-strain … P/M iron … contd.

P/M iron: cyclic softeningK’• r

• n’ unaffected by density

• high density materials K’~K’fully-dense

BACKGROUNDCyclic stress-strain … P/M steels

Fe-1.75Ni-0.5Mo: low strain - softening & high strain - hardening

HomogeneousP

InhomogeneousP+F+M

Fe-2Cu-2.5Ni: work hardening

Fe-1.5Cu-0.6C: softening

Fe-0.8P (F&P): cyclic hardening

BACKGROUNDFatigue crack growth studies, da/dN vs. DK …

ß Near threshold PM/C&Wsimilar behavior;

ß Higher DK, PM inferior to C&W,cracks grow one order ofmagnitude faster;

ß Pseudo fracture toughness DKc

much lower in PM, 20-50 MPa m1/2

(compared to 80-130 MPa m1/2

for quenched and tempered steels).

BACKGROUNDFatigue crack growth studies, da/dN vs. DK … contd.

Fe-1.75Ni-0.5Mo-0.5C(homogeneous - Divorced P)

Higher density

Enhanced resistance tofatigue crack growth

Uniform shifts

Density/porositydominates FCGR over the

microstructure of the matrix

ß Many investigations on various P/M materials, but little knowledge on fatigue mechanisms and fracture;

ß Pore/matrix interactions and how the presence of pores influences/changes the behavior of the matrix are not understood;

ß Characteristic microstructural features as well as inhomogeneities need to be individually understood and further correlated to the pore structure (deconstruct/reconstruct);

ß Fatigue life data corroborated with a fundamental understanding of the alloys behavior predictive abilities;

ß There are no systematic studies to provide “knowledge basedrecipes” to optimize material characteristics and processing

parameters for enhanced fatigue and fatigue crack growth response.

SUMMARY OF THELITERATURE REVIEW

ß Study the effects of density/porosity on the fatigue initiation and propagation in P/M components;

ß Investigate the porosity/microstructure interactions;

ß Understand the effects of different microstructural phases on dynamic properties – mechanisms;

ß Create guidelines for fatigue design corroborated with the fundamental understanding of the alloys behavior;

ß Optimize the material characteristics and processing parameters for enhanced fatigue response.

OBJECTIVES

EXPERIMENTAL APPROACHMaterials selection …

Pre-alloyed

(QMP ATOMET 4601Ni-Mo pre-alloyed powder)

Admixed

(QMP ATOMET 4001Mo pre-alloyed powderadmixed with Ni)

Molding grades particles (50-75 mm)

0.6~0.10.15-0.200.50-0.551.8~0.003[%]

Graphiteadditions

OMnMoNiCChemicalcomposition

EXPERIMENTAL APPROACHPhases …

Phase I (a):Phase I (a): Mechanistic understanding of the effects of pore amount/type on fatigue behavior;

ß Find the relationship density-open/closed porosity ratios for our composition-processes;

ß Pore/Microstructure (matrix) interactions;

Phase I (b):Phase I (b): Microstructure effects on fatigue response;

ß Microstructure 1 vs. Microstructure 2;

Phase II:Phase II: Is fatigue resistance a state function ???

ß Effects of pore size/shape on fatigue.

Open pores

EXPERIMENTAL APPROACHPhase I … Density –closed/open porosity relationship

ß Produce samples of our composition in both pre-alloyedand admixed conditions;

ß Adjust compaction (conventional press, warm compaction,powder forging, etc.) to get the full range of densities:

Set 1 Set 2 Set 3

100%Closed

100%OpenPorosity

7.75or

highestpossible

7.0<6.5Density[g/cm3]

Closed pores

EXPERIMENTAL APPROACHPhase I … Density levels selection

Micro-structure

Low level ofclosed porosity

30% openporosity

&70% closed

porosity

70% openporosity

&30% closed

porosity

Poreamount/

type

Set 37.75+

Set 27.2-7.25

Set 16.8-6.9

Density[g/cm3]

EXPERIMENTAL APPROACH

ß Compaction:

‘ low densities (Set 1): normal compaction;

´ intermediate densities (Set 2): controlled temperature compaction (warm compaction 145°F );

” high densities (Set 3): powder forging.

ß Sintering:

· temperature:T=2050°F ;

6 time: t=30 min;

ÿ T and t invariant for phase I.

Phase I … Compaction +Sintering

EXPERIMENTAL APPROACH

ß Post sintering heat treatment:

ÿ austenitize @ 1600-1700°F for 30 min (similar austenitic grains)

ÿ quench to 2 microstructures (for both pre-alloyed and admixed):

ÿ temper @ 350-450°F for 30 min-1 hr (similar matrix micro-hardness)

Phase I … Heat treatment

Martensite + ~10% R.A.

Martensite + Bainite

Martensite + Pearlite + ~10%R.A.

Martensite + Bainite + Pearlite

EXPERIMENTAL APPROACH

ß Two microstructural considerations are relevant:

Phase I(a): pores vs. matrix

- How porosity interacts with the matrix and when microstructurebecomes cause of failure

- Two microstructures (M / M+X) will be analyzed at three porositylevels and the pore-to-matrix transition will be investigated forall the 12 cases (6 for pre-alloyed and 6 for admixed)

Phase I(b): matrix 1 vs. matrix 2

- How different microstructures influence fatigue behavior

- Two microstructures will be studied and their effects on fatigueinitiation and propagation will be assessed for both pre-alloyedand admixed

Phase I … Effects of pores and microstructures

EXPERIMENTAL APPROACHPhase I … Two microstructural considerations

Low Highdensity density

Pore Pore/Matrix Matrixcontrol control control

A.

??B.Cooling Fatiguerate 2 behavior 2

Cooling Fatiguerate 1 behavior 1

Microstructure 1

Microstructure 2

EXPERIMENTAL APPROACHFatigue testing … Specimens and equipment

Dog-bone specimensfor pull-pull/push-pull

CT specimensfor FCGR

[Courtesy of Westmoreland]-

+

smin

smax

smean

sa

EXPERIMENTAL APPROACHFatigue testing …

-The experiments will beconducted by the WPI team incollaboration with FTA;

-1 sample for each of the 12conditions (prealloyed+admixed, 3density levels, 2 microstructures)

-The tests will be done at anoutside testing facility in parallelwith the fatigue crack growthwork;

-3 failed samples at 4 life levelsfor each of the 12 conditions):

* 103-104

* 104-105

* 105-106

* 106-107

2. Fatigue crack growthtests (E647)

1. Pull-pull / Pull-push tests(E466)

One density level is selected (~7.2 g/cm3, same as Set 2 in Phase I) and 4 waysof achieving are investigated in parallel (we choose the most attractive alloyfrom fatigue point of view [of the 12 combinations] and concentrate ourattention on how pore size/shape will influence its behavior):

1. Compaction - coarser powder, 100-105 mm;

2. Normal compaction to 7.0 g/cm3, followed by a differenttemperature/time sinter;

3. Double press/Double sinter

4. Surface densification (7.0 g/cm3 - core and 7.2 g/cm3 -outer shell)

EXPERIMENTAL APPROACHPhase II … Is Fatigue Limit a State Function ?

Pore size/shape effects on the fatigue behavior

EXPERIMENTAL APPROACHPhase II … Various pore sizes/shapes

Poremorphology(size/shape)

4.Surfacedensified(core 7.0g/cm3 andouter layer7.2 g/cm3)

3.DP/DS

2.Press to

7.0 g/cm3

anddifferentsinter to7.2 g/cm3

1.Coarserparticles

100-105 mm

Molding gradesparticles50-75 mm

Case study

7.0 / 7.2ß Fatigue crack growth work will be conducted for one selected microstructure and one density level

EXPERIMENTAL APPROACHOther experimental considerations …

ß Inclusion level is low shed light on pore and microstructure effects;

ß Sintering atmosphere: synthetic dissociated ammonia;

ß Lubricating additives: Acrawax C, AncorMaxD;

ß Low residual stress levels are critical to understand the true behavior of the materials and have a fair comparisonbasis:

- stress relief is done during tempering - additional stress relieving may be needed after machining.

IT IS ALL IN THE DETAILS …

ß Prepare samples of both pre-alloyed and admixed alloys for the 2 microstructures with the whole range of densities from 6.5 to 7.86g/cm3;

ß Determine closed and open porosity levels for each alloy and microstructure;

ß Develop relationships between density and closed to open porosity ratios for all the 2 microstructures in both pre-alloyed and admixed conditions;

ß Select three densities corresponding to: closed porosity only, two controlled mixtures of closed and open porosity (70%C+30%O and 30%C+70%O) for each case;

ß Perform static tensile tests to get YS, Young’s modulus, UTS for all the 12 cases in Phase I and 4 cases in Phase II (10 samples per case);

IT IS ALL IN THE DETAILS … contd.

ß Analyze microstructural results and microhardness for different heat treatments on both the homogeneous and non-homogeneous materials for each of the microstructures: -Different quenching conditions

ß Do a study on the austenitic grain size and assess the possibility ofconstant austenitic grain size for all the cases;

ß Check the residual stress level and decide if an additional stressrelieving is needed after the post-sintering heat treatment; Adjustthe stress relieving procedure to eliminate residual stress;

ß Prepare samples for the life study (200 dog-bone samples) and thefatigue crack growth work (16 compact tension specimens);

ß Machine all the samples;

ß Start fatigue work.

AFGROW

ß AFGROW was developed by the Analytical Structural Mechanicsbranch, Air Vehicle Directorate, U.S. Wright-Peterson Air ForceResearch Laboratory

ß Input data required: material properties, DKth, Paris coefficients,C, m, geometry and dimensions of the component, initial flaw

characteristics, maximum applied load, stress ratio, choice of constant/variable amplitude; retardation/closure corrections, residualstress adjustments for known residual stresses, etc

ß Output data: life predictions (cycles to failure) and failure modesfor various applications

http://afgrow.wpafb.af.mil/downloads/afgrow/pdownload.php

AFGROW … CONTD.

D UA MA UE ME UA-T4 MA-T4

0

0.02

0.04

0.06

0.08

0.1

0.01

0.03

0.05

0.07

0.09

Max

imu

m f

law

si z

e ( i

n)

s=20 ksiN=1 000 000 cycles

Case study I:s = 20 ksi

N = 1 000 000 cycles

a = ?W=0.5”

T=

4”

AFGROW … CONTD.

Case study III:s = 20 ksi

N = 10 000 cycles

a = ?

D UA MA UE ME UA-T4 MA-T4

0

0.1

0.2

0.3

0.4

0.05

0.15

0.25

0.35

Max

imu

m f

law

si z

e ( i

n)

s=20 ksiN=10 000 cycles

AFGROW … CONTD.

1 10

DK (ksi÷in)

10DK (MPa÷m)

10-9

10-8

10-7

10-6

10-5

10-4

10-3

10-2

10-1

da/d

N (

in/c

yc)

10-7

10-6

10-5

10-4

10-3

10-2

10-1

100

da/d

N (

mm

/cy c

)

R=0.1 Alloy 13%Si-M

13%Si-UM

AFGROW … CONTD.

D UA MA UE ME UA-T4 MA-T4

0

4000000

8000000

12000000

16000000

2000000

6000000

10000000

14000000

Nu

mb

er o

f cy

cles

s=17.4 ksis=18 ksis=21 ksi

Initial flaw size a=c=0.05 in

D UA MA UE ME UA-T4 MA-T4

0

4000000

8000000

12000000

16000000

2000000

6000000

10000000

14000000

Nu

mb

er o

f cy

c le s

s=17.4 ksis=18 ksis=21 ksi

Initial flaw size a=c=0.07 in