The aim of the MUSCAT project, co-funded by the PR FESR 2021-2027 Regione Friuli Venezia Giulia program, is the realization of a Proof of Concept (PoC) for the optimization of microchannel heat exchangers (MCHX) with latex structure. The main result is the development of an innovative methodology for multiscale optimization in conjugate heat transfer (CHT) problems. Such problems are notoriously complex to address with traditional numerical techniques because of the high computational cost required to simulate the fluid dynamic and heat transfer phenomena occurring at the different spatial and temporal scales of the exchanger. The PoC was concerned with the optimization of a heat exchanger for aircraft applications. Specifically, the test application concerns the optimal design of a cooling system for power electronics in avionics systems realizable by Additive Manufacturing (AM).
The approach developed in MUSCAT is based on an innovative paradigm that combines traditional Computational Fluid Dynamics (CFD) tools with Machine Learning (ML) tools to accelerate CHT simulations within the optimization cycle. The main innovation lies in the fact that the heat transfer problem is reformulated using a homogenization technique, i.e., the micro-channel array is modelled as a non-homogeneous, nonlinear porous material whose (non-uniform) properties are described at the macro-scale only through a distribution of parameters related to the geometry of the micro-channels. The distribution of these parameters determines the shape and topology of the latex and is determined during optimization. The CHT problem is solved by classical fluid dynamics techniques only at the scale of the heat exchanger (macro-scale), while a ML model is used to simulate the flow instantaneously at the micro-channel scale. The use of a hierarchical approach allows the problem to be solved with a speed-up of several orders of magnitude compared to classical computational techniques, in which the CFD solver is used to solve the flow at all spatial and temporal scales.
Eligible Expense: 161.740,00€
Grant: 69.001,23€ (of which 40% UE 27.600,49€)
