residual stress measurements of cast aluminum engine

Copyright ©JCPDS - International Centre for Diffraction Data 2005, Advances in X-ray Analysis, Volume 48.
RESIDUAL STRESS MEASUREMENTS OF CAST ALUMINUM ENGINE
BLOCKS USING DIFFRACTION
D.J. Wiesnera, T.R. Watkinsb, T.M. Elyc, S. Spoonerb, C.R. Hubbardb, J.C. Williamsd
a
Dept. of Materials Science and Engineering, University of Tennessee, Knoxville, TN 37916
b
Metals and Ceramics Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831
c
Rigaku/MSC, Inc., The Woodlands, TX 77381
d
Dept. of Materials Science and Engineering, Ohio State University, Columbus, OH 43210
ABSTRACT
Residual stresses and strains in as-cast and annealed aluminum engine blocks were mapped in
the main web area above the bearing split using diffraction. Automated coordinate-mapping of
the surface and bulk strains was accomplished with X-ray and neutron diffraction, respectively.
LabVIEW-based software was developed for improved efficiency, flexibility and quality of data
analysis. In the hoop direction, an overall reduction in compressive surface stress level was
found with increasing heat-treatment temperature. The radial surface stresses shifted similarly
except that the final strains were slightly tensile. Residual strain mapping in the bulk was
performed in two orthogonal directions with neutron diffraction. In the Y and Z (axial)
directions, the bulk tensile strains were reduced from tensile to slightly tensile/compressive with
increasing heat-treatment temperature. The opposite trend was observed in the surface.
INTRODUCTION
An engine casting plant identified the development of residual stresses in their cast aluminum
engine blocks to be due to differential cooling during processing. Concern for the presence of
these stresses derives from the distortion of critical features during machining, which increases
the difficulty of maintaining dimensional tolerances, particularly for main web area. In an effort
to remove these residual stresses, the engine blocks are subjected to a T5 stress relief treatment
prior to machining. The stress relief treatment has added complexity to the process, and the
stress relief oven has become a major production volume-limiting step. Alternate stress relief
cycles, capable of allowing greater throughput in the stress relief oven, would be beneficial in
this regard. To understand the problem better, the residual strains/stresses in a series of heat
treated engine blocks were mapped using X-ray and neutron diffraction.
EXPERIMENTAL
Four engine blocks were supplied by an automotive manufacturer for study. The engine blocks
are die cast from SAE 383.0 (UNS A03830) alloy aluminum. This alloy is a silicon-copper alloy
with compositional limits (wt %) [1]: 2.0 to 3.0 Cu, 0.1 Mg max, 0.50 Mn max, 9.5 to 11.5 Si,
1.3 Fe max, 0.30 Ni max, 3.0 Zn max, 0.15 Sn max, 0.50 others (total) max, bal Al. In Figure 1,
a typical microstructure is displayed showing four distinct phases. One engine block was left in
the as-cast condition, and three were heat treated (Alfe Heat Treating, Inc.) at 230oC, 248oC, and
270oC for 4 hours.
136
This document was presented at the Denver X-ray
Conference (DXC) on Applications of X-ray Analysis.
Sponsored by the International Centre for Diffraction Data (ICDD).
This document is provided by ICDD in cooperation with
the authors and presenters of the DXC for the express
purpose of educating the scientific community.
All copyrights for the document are retained by ICDD.
Usage is restricted for the purposes of education and
scientific research.
DXC Website
– www.dxcicdd.com
ICDD Website
- www.icdd.com
Copyright ©JCPDS - International Centre for Diffraction Data 2005, Advances in X-ray Analysis, Volume 48.
Figure 1. Microstructure of the aluminum in the engine block.
The X-ray system consists of a TEC Model 1600 X-ray stress analyzer mounted overhead to a
gantry which forms the frame for a 15 x 8.2 x 9.2ft (L x W x H) shielded enclosure. This setup
can accommodate small samples (100 lb. capacity) on small x-y-φ stages mounted to the bench
top, large samples (250 lb. capacity, see Figure 2(a)) on large X-Y-φ stages mounted to the floor,
or huge samples (+250 lb.) can be placed directly on the floor. Either sample stage can be
computer controlled via LabVIEW for automated mapping of strains. In an effort to apply a more
robust and efficient data analysis, LabVIEW software was developed to read the ASCII data files
and fit the entire peak profile, allowing for improved accuracy of the residual stress calculation.
(a)
(b)
Figure 2. (a) Enclosure, X-Y-φ sample stage and X-ray stress analyzer mounted to the gantry; (b) Area of interest
for stress/strain mapping: Main web area above the bearing split. Note coordinate axes with the Z or axial direction
normal to the surface.
X-ray stress mapping was done on the main web area above the bearing split shown in Figure
2(b). The sample stages were leveled and the rotation (φ) axis was centered under the collimator.
Gross sample positioning was achieved by marking the position of the engine block on the
137
Copyright ©JCPDS - International Centre for Diffraction Data 2005, Advances in X-ray Analysis, Volume 48.
sample stage. Fine sample positioning was achieved through X and Y translation of the engine
block to points on the 0, 90, and 180 degree lines.
Table I lists the details of the experimental conditions for the X-ray measurements. Briefly, a
single axis (ψ) goniometer [2] was employed for the stress measurements using the "Ωgoniometer geometry" [3(A)]. The (311) reflection from the Aluminum phase was utilized for
the strain measurements. During scanning, the ψ axis was oscillated ±2° to improve particle
statistics. Specimen alignment was accomplished using a contact probe, which was accurate to
±0.25 mm. Prior to data collection, goniometer alignment was ensured by examining an iron
powder pellet standard. The maximum observed peak shift for the (211) reflection of Fe (156
°2Θ) was less than 0.06 °2Θ for Ω tilting as described in Table I.
Table I. - Experimental conditions of the X-ray measurements made on the TEC large specimen
stress analyzer.
Parameter
Equipment
Power
Radiation
Source to specimen distance
Specimen to detector distance
Tilt axis and angles1:
Scans:
Mapping:
Direction:
Condition
TEC Model 1600 X-ray stress analyzer
Position sensitive detector (PSD), 14°2Θ range
52.5 W; 35 kV, 1.5 mA
Cr Kα1, λ = 2.28970 Å
220 mm
220 mm
Ψ = 0, ±9.6, ±13.6, ±16.8, ±19.5, ±21.9, ±24.1,
±26, ±28.1, ±30° (equal steps of sin2ψ)
0.06 °2Θ/step; 60 sec/scan; 135-149o2Θ
33 points; radial distribution
hoop and radial
The X-ray data was analyzed with LabVIEW software based on the SaraTECTM WindowsTM
software [2], which originated from James [4]. The OEM data analysis software, SaraTECTM,
assumes a biaxial stress state and fits a parabola to the top 20% of each peak to determine peak
position. This method overemphasizes the top portion of the peak. Consequently minor
perturbations of intensity can unduly influence the resulting peak position [5]. The LabVIEWbased software [6] addresses this by fitting the background and then fitting a Pearson VII
function to the entire peak profile, accounting for the kα1-α2 doublet. In addition to being able to
handle kβ data, this software allows flexibility in background definition and data handling (i.e.,
rejection of data with poor profile fits and manual insertion of d0). The stresses were calculated
using the “sin2ψ” technique [3(B)], assuming a biaxial stress state. The fully expanded equation
relating strain to stress,
εφψ = (dφψ-d0)/d0 = (1+ν)/E • {σ11cos2φ + σ12sin2φ + σ22sin2φ - σ33} • sin2ψ + (1+ν)/E • σ33
- ν/E • (σ11 + σ22 + σ33) + (1+ν)/E • {σ13cosφ + σ23sinφ} • sin2ψ ,
(1)
1
All angles were used to measure the hoop stress. For the radial stress, Ψ values of +16.8 - 30 ° were omitted due to
clearance issues.
138
Copyright ©JCPDS - International Centre for Diffraction Data 2005, Advances in X-ray Analysis, Volume 48.
may be reduced to
εφψ = (1+ν)/E • σφ • sin2ψ - ν(σ11+σ22)/E
(2)
Alternatively,
dφψ = (1+ν)/E • σφ • sin2ψ • d0 - ν(σ11+σ22)/E • d0 + d0
(3)
assuming σ13=σ23=σ33=φ=0. ε, d, ν, E and σ are the strain, interplanar spacing, Poisson’s ratio,
Young’s modulus and stress, respectively. The variables and subscripts φ, ψ and 0 refer to the
azimuthal angle, tilt angle and strain-free, respectively. The y-intercept of the plot of dφψ vs.
sin2ψ was taken as the strain free interplanar spacing, d0. While this selection of d0 is often not
ideal, the effect on the calculated stress is <0.1% for the biaxial stress state [3(B)]. Here, E and ν
were taken as 70 GPa and 0.33, respectively.
Neutron scattering experiments were carried out on the HB-2 spectrometer at the High Flux
Isotope Reactor (HFIR) of the Oak Ridge National Laboratory. The incident neutron beam
wavelength was determined by the orientation of the (331) planes of a double-focusing silicon
monochromator. The wavelength determination, 2θ zero error and ∆2θ/channel calibration of
the position-sensitive detector was accomplished using a nickel powder standard. In this study,
the (311) aluminum peak was located at around 83 °2θ with the neutron wavelength of 1.61 Å.
The diffracting volume was defined by a parallelepiped with 3 mm long sides and angles
between the sides of 90, 90 and 83 °, for a volume of ~26.8 mm3. The sample was mounted on
an automated X-Y-Z translation table, and the coordinates of the sampling positions were
referenced with respect to the surfaces of the sample. The locations of the external surfaces of
the sample were verified with an intensity-position scan. Here, the sample is translated relative
to the fixed gauge volume. When the scan displays a non-linear intensity decrease, the gauge
volume has begun to “exit” the sample. Macro-residual strain was determined from the shift in
the interplanar spacings of diffracting grains relative to that of a strain-free reference material (a
free 0.25 inch cube of the parent material). This strain is the average from the diffracting grains
within the sampling volume. Stress mapping was done on the main web volume above the
bearing split shown in Figure 2(b).
RESULTS
Figure 3 shows the surface residual stresses in the hoop and radial directions. The applied color
scheme was equally scaled for visual comparison of the different heat treatments. The surface
hoop stresses show compressive residual stresses decreasing towards zero stress with increasing
heat treatment. The reduced compressive residual stress can also be thought of as the addition of
tensile residual stress. The stress profile is fairly non-uniform but becomes more uniform with
increasing heat treatment. The radial stresses also show a trend of increasing of tension to 248°C
then decrease towards zero after 4 hours at 270°C. No explanation can be offered for this
behavior, particularly in light of the well-behaved hoop data from the same sample. As before,
this stress profile becomes more uniform with increasing heat treatment.
139
Copyright ©JCPDS - International Centre for Diffraction Data 2005, Advances in X-ray Analysis, Volume 48.
Figure 3. Stress profiles from the surface. The XY grid is the same for each profile with X and Y in mm. The dots
mark the measurement locations.
Two strain components from the bulk of the as-cast and 230 °C heat treated engine blocks were
measured with neutron diffraction (see Figure 4). The height of each colored cylinder is
proportional to the strain magnitude. Most of the strains are tensile with a maximum strain of
+0.09%. Without the measurement of the third orthogonal strain in the X-direction, it was not
possible to accurately calculate the residual stresses. In general, the strains are largest near the
cylindrical bearing interface and show a left-to-right asymmetry, which is consistent with the fact
that the web is connected to a thicker section at one end. This asymmetry is also clearly present
in the radial X-ray data (see Figure 3) but of opposite sign. The tensile strains in the Y-direction
are reduced with heat treatment, which is opposite to corresponding X-ray data. In the axial or Zdirection, the as-cast bulk strains are tensile and become compressive with 230°C heat treatment.
For comparison purposes, the maximum strain crudely corresponds to +63 MPa, assuming the
stresses and strains in orthogonal directions are zero.
140
Copyright ©JCPDS - International Centre for Diffraction Data 2005, Advances in X-ray Analysis, Volume 48.
Figure 4. The as-cast and heat-treated (230°C) strain profiles from the bulk. The arrows show the direction of the
strain components. The Z direction corresponds to the axial direction while the Y-direction corresponds to a
mixture of radial and hoop strains. Tensile strains are in red; compressive strains are blue. The maximum strain
above was 0.0009.
SUMMARY
Residual stresses and strains in as-cast and annealed aluminum engine blocks were mapped using
X-ray and neutron diffraction. Automated coordinate-mapping of the radial and hoop strains was
accomplished with manual X-Y-Z gantry supporting a large sample X-ray stress analyzer, which
is suspended above a computer controlled X-Y-φ sample stage. LabVIEW-based software was
developed for improved efficiency, flexibility and quality of data analysis. Neutron strain
mapping was performed in two orthogonal directions.
Overall, the surface stresses went from compressive to near zero or slightly tensile with
increasing heat treatment temperature. In concert, the bulk strains went from tensile to slightly
tensile or compressive with increasing heat treatment temperature. Given the complex geometry
of the engine block, complete stress relief may not be realistic given asymmetric constraints of
the engine block structure.
141
Copyright ©JCPDS - International Centre for Diffraction Data 2005, Advances in X-ray Analysis, Volume 48.
ACKNOWLEDGEMENTS
Research sponsored by the Department of Energy, Office of Science Education through the SULI
internship program and by the Assistant Secretary for Energy Efficiency and Renewable Energy,
Office of FreedomCAR and Vehicle Technologies, as part of the High Temperature Materials
Laboratory User Program, Oak Ridge National Laboratory, managed by UT Battelle, LLC, for
the U.S. Department of Energy under contract number DE AC05-00OR22725.
REFERENCES
[1]
“Properties of Cast Aluminum Alloys.” ASM Handbook – Volume 2, Properties and
Selection: Nonferrous Alloys and Special Purpose Materials, ASM International:
Materials Park, OH 1990.
[2]
TEC Model 1600 X-ray stress analysis system with SaraTECTM WindowsTM software
v.1.64, Technology for Energy Corporation, 10737 Lexington Dive, Knoxville, TN
37932, www.tecstress.com.
[3]
I. C. Noyan and J. B. Cohen, Residual Stress: Measurement by Diffraction and
Interpretation, Springer-Verlag: New York, 1987, A, p. 101 and B, pp. 122-3.
[4]
M. R. James, An Examination of Experimental Techniques in X-ray Residual Stress
Analysis, Ph.D. Thesis, Northwestern University: Evanston, IL 1977.
[5]
T. M. Ely, “Investigation of the Repeatability of Stress Values Reported by TEC
Software,” Unpublished work, 2001.
[6]
D. J. Wiesner, Residual Stress Analysis 1.01, Labview based software, ORNL, January
2004.
142