Additive manufacturing-aware topology optimization of cooling channels using a reduced-order cold plate model for electric machine water jackets

Integrated liquid cooling jackets are widely used in the thermal management of high-power-density motors, but traditional designs rely on parametric channel layouts, failing to fully utilize the geometric degrees of freedom offered by additive manufacturing (AM). This paper proposes an AM-aware topology optimization framework based on a reduced-order cold plate model. A two-dimensional case study systematically investigates the effects of manufacturability-oriented regularization on channel topology and thermo-hydraulic performance. Results show that stronger regularization substantially reduces interface complexity and fluid area fraction, while generally increasing hydraulic resistance and degrading thermal performance. A sharp-interface reanalysis further demonstrates good agreement in the global thermal response, while revealing the sensitivity of the local peak temperature and hydraulic conductance to the density-based representation.