Undergraduate research · Thermal simulation

WAAM thermal modeling

Studying how interlayer dwell time changes the thermal history of an HSLA steel wall.

I developed and calibrated an ANSYS thermal model of a wire arc additive manufacturing build, then used it to compare cooling behavior across several dwell times. The model was based on an experimental ER120S-1 wall and checked against thermocouple data collected during the build.

Undergraduate research with Dr. Brodan Richter, supported by the University of St. Thomas Undergraduate Research Opportunities Program.

41

layers modeled

53 °C

calibration RMSE

30 to 150 s

dwell times compared

3

thermocouple locations

Experimental basis

The reference build

The physical build was a 70-layer ER120S-1 steel wall deposited on an ASTM A36 baseplate. Thermocouples at layers 20, 40, and 60 recorded the temperature history used to check the simulation. The experimental work was conducted by Zac Hong, Owen Lundstrom, and Aiden Woods.

The wall, wire, and thermocouple locations used for the reference build.

Experimental WAAM setup

Side view of the WAAM setup with wire, steel wall, and thermocouples labeled
Top view of the approximately 295 mm long wall.

Completed wall geometry

Top view of the completed steel wall with a 295 millimeter dimension

Build conditions

Wire
ER120S-1, 1.14 mm
Wall
70 layers, about 295 × 178 × 10 mm
Travel speed
348 mm/min
Wire feed
8.64 m/min
Arc settings
237 A, 18.6 V
Shielding gas
75% Ar / 25% CO₂ at 21 L/min

Simulation

Thermal model

I rebuilt the measured wall and process conditions in ANSYS Additive DED. New material was activated layer by layer, which allowed the model to follow heat input, cooling, and reheating as the wall grew.

Temperature-dependent ER120S-1 properties were generated in Thermo-Calc 2025b using the TCFE14 steels database and used as material inputs in ANSYS.

Software
ANSYS Additive DED 2025 R2
Build modeled
First 41 of 70 layers
Applied power
4.41 kW
Absorptivity
0.80
Layer height
2.54 mm effective
Ambient
22 °C
The first 41 layers were modeled using the measured geometry and process conditions.

ANSYS thermal model

ANSYS simulation showing the temperature distribution through the WAAM wall and baseplate

Comparison with measurements

Model calibration

I compared the simulated response with the thermocouple at layer 20 over the 2380 to 3220 second window. An effective dwell time of 67.5 seconds captured the programmed 60-second dwell plus the machine motion and process delays between layers.

Measured and simulated temperatures at the layer 20 thermocouple.

Thermocouple calibration

Measured and simulated thermocouple temperatures plotted over time

Calibration error: 53 °C RMSE

This was 5.9% of the measured 898 °C temperature range. The calibrated model was then used for the dwell-time comparison.

Results

Dwell-time comparison

I tracked cumulative t8/5, the total cooling time from 800 to 500 °C across repeated thermal cycles, for dwell times from 30 to 150 seconds. Shorter dwell times produced more heat accumulation, especially through the middle of the build. Increasing the dwell reduced cumulative t8/5 time, with the largest change occurring at the shorter dwell settings.

The middle layers accumulated the most t8/5 time before the value fell near the top of the build.

Cooling history by layer

Cumulative t8/5 time by layer for 30, 60, and 120 second dwell times
Longer dwell times reduced the mean cumulative t8/5 time, with smaller changes at the longest settings.

Mean cooling time by dwell

Mean cumulative t8/5 time plotted against interlayer dwell time

Early layers

The baseplate drew heat away from the first part of the wall.

Middle layers

Repeated deposition caused the greatest heat accumulation.

Final layers

Fewer remaining reheating passes reduced cumulative t8/5 time.

These thermal histories can help guide later experimental work on phase transformation and hardness. Confirming those material changes would require physical characterization of the deposited wall.