Accelerated Lifetime Model-Based Design Optimization Strategy With Efficiency, Reliability, and Cost Trade-Off for High-Power Modular AFE Rectifiers

This paper presents a design optimization for grid-connected modular Active Front-End (AFE) rectifiers with an evaluation of efficiency, lifetime, cost, volume, and weight. This tool optimizes the rectifier switching frequency, component sizing and selection, and the number of parallel converters by designing and evaluating every possible configuration of the AFE rectifier system defined by the end user. The design of the LCL filter, magnetic design of inductors, selection of the inductor core and winding, and selection of SiC switches and MLC capacitors are discussed. Rapid Low-Fidelity (Lo-Fi) electro-thermal and lifetime models that are fast enough to be used in an optimization process have been developed. The rapid Lo-Fi models estimate component losses, temperatures, and lifetime for given load profiles and converter configurations. This Lo-Fi approach accelerates the processing time to generate the thermal data for the converter mission profile and allows us to skip rainflow-counting algorithm to assess the accumulated thermal damage to the SiC switches at the design and development stage. This in turn allows engineers to design converters with a longer predicted lifetime. Moreover, optimization of the grid-side three-phase LCL filter is performed considering efficiency, cost, and volume trade-off between the grid-side and converter-side inductors to achieve up to 50% decrease in losses, and 23% decrease in cost. Moreover, a 21-22% decrease in the system losses, 23-27% decrease in system cost, and tenfold improvement of the system lifetime can be achieved by optimizing the converter switching frequency and number of parallel modules. A 15 kW hardware prototype consisting of three 5 kW AFE rectifier modules is built and used to validate the efficiency from the fast Lo-Fi models. The validation of the detailed loss model and junction temperature swing is performed against a High-Fidelity (Hi-Fi) simulation in MATLAB Simulink environment.