The use of 3D thermal simulations of magnetic components for power converters is limited to validation of final designs due to time consumption. A two-step homogenization is proposed to simplify them, while keeping good accuracy. The required equations and some guidelines are described in this paper to apply it to solid round wire, litz wire or foil wound magnetic components. Considerations to extend its use to transformers, including the use of interleaving techniques and insulating tape are explained. This method can reduce the running time of 3D simulations (seven times faster than usual simulations) while the convergence is assured even for complex structures or cases with litz wire. Finally, the proposed method is experimentally validated with DC and AC excitation.
24 October 2019 in IEEE Transactions on Power Electronics
This paper analyzes the possibility to employ Artificial Neural Networks (ANN) for thermal modeling of magnetic components, with special emphasis on Medium Frequency Transformers (MFTs). The objective is to obtain a MFT thermal model that can be easily used in multivariable optimizations and automated design of power electronics converters, when it is necessary to rapidly explore the complete design space, optimizing the system from the point of view of the efficiency and power density, without losing the accuracy. This can be accomplished using low computational cost ANNs that are trained using the results obtained from thermal Finite Element Method (FEM) simulations. Nevertheless, depending on the approximations that are often applied to speed up the simulations (without modifying the mesh) and, consequently, to reduce the dataset generation time, the precision of the simulations and ANN based models vary significantly, especially in the case when the forced air cooling is used. For example, we demonstrate that different approaches can accelerate the dataset generation (460 simulations) by up to 40 hours, but would introduce an error of up to 40 °C. Therefore, a trade-off of the different alternatives to generate the ANN dataset is analyzed in this paper. The proposed design and optimization approach have been verified using the experimental platform, a Medium Frequency Transformer (MFT), designed and optimized for a 7.5 kW Dual Active Bridge (DAB), designed by brute force optimization using the ANN trained with FEM simulations. The experimental results obtained in different operating points show errors lower than 2.6 °C for the estimated average surface temperature. Additionally, the FEM obtained results are compared with the results obtained by state-of-the-art simplified analytical models and simplified FEM simulations, clearly showing the necessity of highly accurate FEM simulations to generate the training dataset. Finally, we discuss the drawback of the proposed ANN approach in the cases when the ANN based model should be scaled, trying to design a MFT with characteristics outside the range that was used for ANN training. We show that one possible solution for this issue may be ANN re-training, allowing to reduce the error from 50 °C to 5 °C by using only 40 new data points.
New magnetic composite based on ferromagnetic nano-particles embedded in a polymeric material can potentially offer significant electromagnetic advantages over the polymer used to protect the ferrite in Wireless Power Transfer systems. Since the nano-particles and the strands of the litz wire are very small, brute force simulation by Finite Element Method is not possible. In this work, a method to overcome these limitations is proposed based on the use of homogenization techniques that will allow replacing both the Litz and the magnetic compound material by homogenous equivalent materials that will behave, in a macroscopic sense, as the actual material.
3 March 2022 in IEEE Transactions on Industry Applications
A multi-level boost topology based on a flying capacitor is compared in this work with the hybrid, multi-level, partial power processing topology already presented in the literature for use as a dc/dc stage in 1500-V, two-stage, grid-connected PV systems. Different variations of the resonant stage of the hybrid topology are analyzed and compared in terms of losses, volume and complexity. Full multivariable optimization is conducted for the multi-level boost topology with different number of levels and for all the analyzed variations of the resonant stage of the hybrid topology. Appropriately rated new classes of 650-V and 900-V WBG devices are included in the optimization algorithm. Results of the comparison show that losses of both topologies are in the same range, while hybrid, partial power processing topology provides lower occupied volume. All the analysis related to the resonant stage of the hybrid topology is confirmed by detailed and comprehensive measurements of two prototypes of 4kW of nominal and 7kW of maximal power with reported euro-efficiency of 99.79% and 12.1 kW/kg of specific power. Additionally, all the loss models applied in the optimization algorithms presented in this paper are confirmed by detailed measurements, thermal calibration and finite element thermal simulations of the wide bandgap devices and inductor designs employed in this work.
The dissipation of the generated power losses of on-board magnetic components is a major concern regarding the thermal management of the power systems. In this study, two planar transformers designed for common space power converters are chosen as representative samples to analyze the effect of different elements on heat transfer (vias, connectors to a baseplate, gap filler pads, glue, integration in the main PCB). 3D thermal simulations are used to quantify it. It has been found out that the contribution of vias to heat transfer can reach a 8.9%, and the use of metallic connectors, window filling or gap filler pads can reduce the maximum temperature rise between 18.4% and 57.4%. Additionally, a simplified 2D model is proposed as a fast solution to characterize the thermal behavior of planar transformers and determine if a design is valid or if it requires any modification. Finally, some guidelines to address the thermal management of these components are explained.