Aerodynamic shape optimization of an electric aircraft motor surface heat exchanger with conjugate heat transfer constraint

•An adjoint-based shape optimization method minimized the drag of a surface heat exchanger for an array of heat transfer constraint values.•Optimal surface heat exchanger geometry changed from smooth to finned as the heat load constraint increased.•The designs optimized using the conjugate heat transfer model were smaller than those optimized with the convection-only model.•The fins designed by the optimizer had a forward tilt to reduce the aft adverse-pressure gradient and consequently separation. Electrified aircraft benefit from the versatile ways electric motors can be integrated with an airframe. However, thermal management is needed to move waste heat out of the motors because the heat is not expelled with the exhaust as in a conventional engine. Plate-fin, fin, and surface heat exchangers are incorporated as air-side heat exchangers for electrified aircraft thermal management systems. Typically, analytic tools are used to design heat exchangers within these categories. However, analytic tools lack the fidelity required for detailed shaping and assessment of general heat exchanger configurations. Tools based on first principles, such as finite element analysis or computational fluid dynamics, can verify heat exchanger performance but are too costly to use in a manual design loop. Shape optimization can be used with first-principles-based models to design heat exchangers without limiting the geometry to those previously well studied. In this work, we apply this methodology to design a heat sink for the high-lift motor of an electric technology demonstrator, the X-57 Maxwell. We use a gradient-based optimizer to modify the thickness distribution of the heat sink to find designs that minimize drag while meeting the heat load constraint. To model the heat transfer from the motor, we use both convection-only and conjugate heat transfer models and compare the resulting differences in the optimized shapes. We found that the convection-only model under-predicted heat rejection and thus led to larger than necessary heat sinks when used in optimization. To study the effect of the heat load on the design, we compare the heat sinks designed for the baseline motor and heat sinks designed for less efficient motors. Our study results show how the heat exchanger’s geometry changes from uniformly thick to designs with fins as the heat load increases. Furthermore, we found that the variation in drag across designs is driven by differences in the pressure drag due to f