Digital Twin of a Thermal System Using MATLAB and ANSYS CFD

A student-laptop digital twin of a forced-air aluminium heat sink (6 fins in a 20 mm x 200 mm duct) that cools a 4-18 W power module. A CFD design of experiments over 10 inlet velocities (0.25-2.0 m/s) gives the thermal resistance (4.72 -> 2.02 K/W) and pressure drop (0.73 -> 37.9 Pa); a 6-mode POD reduced-order model and the correlation R_th = -0.289 + 2.997 U^-0.369 turn it into a millisecond model; an extended Kalman filter with a fouling state estimates the hidden junction temperature with 0.18 K RMSE over 24 hours, detects heat-sink fouling, isolates a biased thermocouple and drives the fan predictively (86.8 % less fan energy than a fixed fan). A role-based web dashboard shows all of it. Honest scope: MATLAB and ANSYS are NOT installed on the build machine - the MATLAB scripts and ANSYS SpaceClaim / Fluent / Workbench journals were written but never run. All numbers come from the Python implementation, in which a 2D laminar finite-volume conjugate-heat-transfer solver stands in for the 3D Fluent k-omega SST model, and the plant is simulated with injected faults.

Category: Mechanical Engineering / Thermal Engineering / Computational Fluid Dynamics / Electronics Cooling / Digital Twins / Reduced-Order Modelling / MATLAB and ANSYS (scripts) with Python

Mechanical Engineering Project M03 from ProTechSC Academy, with graphical abstract, video demonstration and project guidance.