UNIVERSITY OF MICHIGAN / RESOURCEFUL MANUFACTURING & DESIGN LAB

Through-Die Cooling for Aluminum Extrusion

I modeled and bench-tested an insulated cooling channel to limit heat loss from a hot extrusion die while preserving a path for internal cooling.

My roleAbaqus thermal analysis, experimental setup, cooling trials, and comparison of insulation strategies.

Abaqus
Thermal analysis
Test fixtures
Manufacturing research

Physical die-cooling test rig with steel die, coolant hoses, and temperature measurement equipment. — Physical bench setup for the cooling-channel study

450 °CInitial die temperature in the thermal model

2Insulation materials compared

3Coolant mass-flow trials

01 / ENGINEERING PROBLEM

Cool the extrusion without overcooling the die

Internal quenching can reduce temperature differences across an aluminum profile, but coolant passing through the die can also remove too much heat from the steel. I investigated insulation between the channel tube and die to control that unwanted heat transfer. A heated steel test block provided a practical way to examine the thermal behavior.

02 / METHODS & ANALYSIS

Thermal model and bench comparison

Project figures and hardware
Click images to enlarge

Separate the coolant from the hot die

Steel test die with hose and adapter connections for the internal channel.

I compared an uninsulated channel with alumina and calcium aluminum silicate insulation. The model represented the steel, insulating layer, and coolant tube separately so the effect of the insulation could be assessed against the same initial conditions.

Set the thermal conditions and check them experimentally

Original Abaqus die model used for the thermal study.

Thermal model450 °C initial die; 25 °C surroundings
Heat-transfer boundaries25 W/(m²·K) air; 20,000 W/(m²·K) coolant
Bench checkThermocouple cooling traces and timed coolant collection

I used transient FEA to compare cooling histories, then checked cooling trends on the physical rig. Timed collection gave flow rates of 0.356–0.361 kg/s. Heating conditions, adapter losses, and thermocouple contact limited a direct model-to-test comparison.

03 / RESULTS & OUTCOME

Insulation reduced predicted die cooling

THERMAL FEA + BENCH TESTING

Predicted die temperature after 3,600 s

Same initial temperature: 450 °C

No insulation

30.1 °C

Alumina

182.2 °C

Calcium aluminum silicate

176.4 °C

Higher retained temperature means less unwanted cooling of the die.

0.356–0.361 kg/sMeasured coolant flow · three trials

2Insulated cases retained more die heat in the model

Both insulation materials slowed predicted die cooling relative to the bare channel. Bench tests checked flow and temperature trends, but did not establish production extrusion quality or a quantified model-error bound. The next comparison should control heat input and sensor contact more closely.