Loss-Optimized Design of a Triple Active Bridge DC-DC Converter for an Electric Vehicle Application

This paper proposes a design optimization methodology for an isolated triple active bridge (TAB) DC-DC converter for an electric vehicle (EV) application. The optimization algorithm computes the optimal values of the high-frequency transformer leakage inductances to minimize converter losses and enhance efficiency across a broad range of operating conditions. The TAB is modeled using the generalized harmonic approximation (GHA) technique, which reduces computational complexity compared to time-domain analysis while maintaining high modeling accuracy across all converter operating zones. The TAB control variables are generated using particle swarm optimization (PSO) for the discrete operating points defined within the design search space. The TAB model, with the selected optimization strategy, is implemented in the PLECS Blockset software for an 800 V EV application with a maximum on-board charger (OBC) power of 9.6 kW and an auxiliary power module (APM) rated at 3.6 kW. The findings demonstrate that efficacy of the proposed design methodology effectively optimizes the leakage inductances of the three-port transformer, extends the zero-voltage switching (ZVS) range, and results in approximately a 23% reduction in converter losses under full-load conditions.