Influence of T6 Heat Treatment on Distortion-Relevant Cutting Forces and Surface Residual Stresses in Orthogonal Cutting of Aluminium Casting Alloy AlSi7Mg0.3
Paper details
Section titled “Paper details”| Field | Value |
|---|---|
| Journal | Journal of Manufacturing and Materials Processing |
| Published | 2026-10-20 |
| Authors | Tim Reeber, Martin Sicking, Vincent Altmann, Hans‐Christian Möhring, Dirk Biermann |
| DOI | 10.3390/jmmp10100445 |
T6 heat treatment raised cutting forces by 25% to 43% and feed forces by 7% to 44% across 15 orthogonal-cutting parameter combinations. The effect exceeded replicate scatter, which remained below 5.5%. Cutting-direction surface residual stresses stayed tensile for both conditions, with parameter means of 45 MPa to 147 MPa. The findings provide cutting and residual-stress inputs for process-chain models of thin-walled castings.
T6 raises cutting and feed forces
Section titled “T6 raises cutting and feed forces”Cutting force increased monotonically with feed for both conditions at every tested cutting velocity. The T6 condition produced higher forces in all 15 combinations. Welch t-tests gave p < 0.015 for cutting force in every combination and p < 0.05 for feed force in 14 of 15 combinations.
At 900 m/min and a 0.15 mm/rev feed, as-cast cutting force averaged 69.62 N/mm. At 1200 m/min and the same feed, it averaged 77.52 N/mm. For as-cast material, raising speed from 900 to 1200 m/min increased cutting force by 4% to 11%. The increase was significant at four of five feeds but did not appear for T6.
Experimental method and material conditions
Section titled “Experimental method and material conditions”AlSi7Mg0.3, also designated EN AC-42100, was tested as-cast and after T6 treatment. T6 used solution annealing at 545 °C for 130 min followed by water quenching. The sequence then included a 15 min room-temperature hold and artificial ageing at 180 °C for 360 min.
Machining used 4.5 mm thick blanks and a PCD insert with a 0.01 mm edge radius, 0° rake face, and 7° flank face. Cutting speeds were 600, 900, and 1200 m/min, with feeds from 0.05 to 0.15 mm/rev.
A CNC machining centre emulated turning to maintain orthogonal kinematics and constant surface speed. Kistler forces were sampled at 80 kS/s. Each result averaged the central 60% of three to five engagements after discarding the incomplete first engagement.
Each combination had at least three workpieces drawn from 12, 22, and 32 mm casting walls, yielding 154 valid force measurements overall.
XRD used the sin² ψ method at three surface positions. The Al {311} reflection was evaluated at 2θ = 139.3° with five tilts from -45° to +45°. Cr-Kα radiation sampled about 10 µm deep. Measurements with fitting uncertainty above 200 MPa were excluded, removing 16 of 924 results.
Dynamic strength explains the force increase
Section titled “Dynamic strength explains the force increase”Split-Hopkinson pressure bar tests covered approximately 1500 to 6000 s⁻¹. Three striker pressures and five temperatures from 20 °C to 500 °C were tested three times per condition. The T6 condition showed higher flow stress throughout the tested strain, strain-rate, and temperature range. At room temperature and low strain, its flow stress was about 40% to 50% higher than as-cast material.
Table 1. Johnson-Cook parameters fitted to the SHPB flow curves.
| Parameter | Unit | As-cast | T6 |
|---|---|---|---|
| Initial yield stress A | MPa | 224 | 319.95 |
| Strain hardening coefficient B | MPa | 60 | 60 |
| Strain hardening exponent n | - | 0.1800 | 0.1666 |
| Strain rate sensitivity C | - | 0.0010 | 0.0274 |
| Thermal softening exponent m | - | 1.1987 | 0.9573 |
A and B reached the ±20% fitting bounds for T6; A, B, n, and C did so for as-cast material. The as-cast value C = 0.0010 should be treated as a fit lower bound, not a material property. Maximum measured-to-fitted deviations reached about 160 MPa for as-cast and 390 MPa for T6. The stronger T6 flow stress was consistent with greater force dissipation in the primary shear zone.
Friction and chip formation
Section titled “Friction and chip formation”The apparent friction coefficient for as-cast material stayed between 0.27 and 0.31 across all cutting parameters. It fell slightly from about 0.30 at 600 m/min to about 0.28 at 1200 m/min. For T6, apparent friction fell from 0.29 to 0.30 at 0.05 mm/rev and 0.24 to 0.25 at 0.15 mm/rev.
All chips were continuous, tightly curled spirals, so chip formation appeared unchanged by T6. Contact length increased linearly with feed, with no clear speed effect for T6. At low feed, as-cast contact length tended to decrease with speed. Pixel counting gave about ±36 µm uncertainty, equal to 9% to 28%.
Residual stresses in the machined surface
Section titled “Residual stresses in the machined surface”Cutting-direction surface residual stress was tensile for every parameter combination in both conditions. Parameter means ranged from 45 MPa to 147 MPa. Per-combination standard deviations were 30 MPa to 63 MPa for as-cast and 54 MPa to 111 MPa for T6.
Scatter among three positions on one workpiece averaged 47 MPa, which the authors linked to coarse cast microstructure. Cutting speed had no significant effect on either condition, with p > 0.18.
For T6, stress decreased about 40 MPa per 0.1 mm increase in feed, with p = 0.03. The overall cutting-direction means were nearly equal at 109 MPa for as-cast and 111 MPa for T6. T6 produced lower perpendicular tensile stress, averaging 41 MPa against 80 MPa for as-cast material, with p < 0.001.
Depth profiles were measured on one specimen per condition. The as-cast surface was about 93 MPa at 600 m/min and 0.125 mm/rev. The T6 surface was about 70 MPa at 900 m/min and 0.1 mm/rev. T6 peaked near 150 MPa at 5 µm, then became compressive near 10 µm. At 50 µm, T6 stress was about -300 MPa.
As-cast stress fell from 60 MPa at 5 µm to near zero around 15 µm. As-cast compression reached about -190 MPa at 40 µm, then recovered to about -90 MPa at 50 µm. Neither profile reached a stable plateau by 50 µm, suggesting that the affected zone extended beyond the measured depth.
Coupled Eulerian-Lagrangian model performance
Section titled “Coupled Eulerian-Lagrangian model performance”The CEL model used an Eulerian workpiece and rigid tool, with 102,080 Eulerian elements at 0.0125 mm. Johnson-Cook described the workpiece. A local friction coefficient of 0.09 was calibrated against as-cast cutting forces. Heat generation converted 90% of plastic work into heat, and both bodies started at 20 °C. Mass scaling increased density by 100 and reduced specific heat by the same factor.
Table 2. Mean absolute percentage errors between simulation and experiment.
| Quantity | As-cast MAPE | T6 MAPE | Error range across conditions |
|---|---|---|---|
| Cutting force Fc | 9.9% | 15.3% | 1.7% to 33.8% |
| Feed force Ff | 28.4% | 58.2% | 0.2% to 108.1% |
| Surface residual stress σ∥ | 93.0% | 264.1% | 12.4% to 596.9% |
The model reproduced higher T6 forces and monotonic growth with feed, with cutting-force MAPE of 9.9% for as-cast and 15.3% for T6. Feed force was overestimated in both conditions, with 58.2% MAPE for T6.
Mass scaling added an estimated 2.7, 6.0, and 10.7 N/mm at 600, 900, and 1200 m/min, versus below 0.1 N/mm at physical density. This helps explain the simulated speed dependence and reduced chip curl.
In the representative T6 profile, the model predicted about -190 MPa at the surface, while XRD measured about +70 MPa. The simulated T6 profile stayed compressive through 50 µm. Absolute surface-stress errors averaged 102 MPa for as-cast and 271 MPa for T6. MAPE is less meaningful near zero or across a sign change.
Limits for distortion simulation
Section titled “Limits for distortion simulation”The model’s main errors were constant friction, mass scaling, and a Johnson-Cook fit limited to 6000 s⁻¹. The fit covered strain rates below those expected in the primary shear zone. No mesh sensitivity study tested the 0.0125 mm element size. The plane-strain model omitted out-of-plane effects and did not solve perpendicular stress independently.
Future work should use pressure- and temperature-dependent friction, lower mass scaling, higher-rate data, mesh convergence, and matched-condition depth-profile validation. For a first approximation, the authors propose a tensile cutting-direction layer of about 100 MPa, largely independent of cutting parameters.
The current CEL model can estimate cutting forces but cannot quantitatively predict T6 residual stresses. Process-chain simulations should treat the measured tensile layer as a calibrated input, not as CEL output.
View Original Abstract
Thin-walled aluminium castings, such as stator housings for electric motors, are susceptible to distortion during and after machining due to the superposition of residual stress states introduced by casting, heat treatment and cutting. This paper investigates the influence of T6 heat treatment on cutting forces and surface residual stresses in orthogonal cutting of the aluminium casting alloy AlSi7Mg0.3 and assesses the capability of a Coupled Eulerian-Lagrangian (CEL) model to predict them. Orthogonal cutting tests were conducted with a sharp polycrystalline diamond (PCD) tool at cutting speeds of vc = 600, 900 and 1200 m/min and feeds of f = 0.05 to 0.15 mm, with at least three replicates per parameter combination (154 force measurements in total). Johnson-Cook parameters for both material conditions were derived from Split-Hopkinson pressure bar tests at temperatures of up to 500 °C. The T6 heat treatment increases the cutting force by 25-43% and the feed force by 7-44%, consistent with the higher flow stress of the T6 condition, while the scatter between replicates is below 5.5%. Surface residual stresses measured by X-ray diffraction are tensile in the cutting direction for both conditions (mean values of 45-147 MPa), and show no consistent dependence on cutting speed and at most a weak decrease with feed for the T6 condition; perpendicular to the cutting direction, the T6 condition exhibits significantly lower stresses than the as-cast condition. The CEL model reproduces the cutting force with mean deviations of approximately 10-15%, but overestimates the feed force and fails to predict the tensile surface residual stresses of the T6 condition. The friction formulation, the mass scaling and the constitutive description are identified as the main sources of error that must be addressed before the model can be used for distortion prediction in the process chain.
Advanced Thermal Simulation Service Diamond Heat Spreader Performance Analysis
View Specs →
Diamond Heat Spreader With Copper Coated Metalized Surface
View Specs →