High-Performance Solid-State Transformer Optimisation Using VITROPERM Nanocrystalline Cores: A Hybrid RSM and Fuzzy Logic Approach for Enhanced Electrical Grid Efficiency

Purpose: This paper aims to enhance the design of Solid-State Transformers (SSTs) operating at medium and high frequencies (MF/HF) by optimizing key performance attributes such as efficiency, power density, and thermal management through the use of advanced nanocrystalline materials. Design/Methodology/Approach: An interactive MATLAB-based magnetic characterization tool was utilized to evaluate transformer core losses and efficiencies with various core materials. A hybrid optimization methodology integrating L9 Taguchi design and Response Surface Methodology (RSM) was employed to design a 2 kVA, 20 kHz, 600/60 V shell-type SST. Design parameters such as duty cycle, window utilization factor, insulation material, and core series were optimized across three levels to achieve minimal losses and maximum efficiency. Research Limitations/ Implications: While the study demonstrates promising results, the research is limited to a specific transformer rating and operating frequency, indicating a need for further exploration across varying ratings and frequencies. Practical Implications: The findings contribute to the development of efficient and compact SSTs, benefiting renewable energy systems, electric vehicle charging, grid-connected converters, and compact power supplies by ensuring reduced thermal stress and enhanced durability. Originality/Value: This research highlights the efficacy of integrating advanced nanocrystalline materials and hybrid optimization methodologies in SST design, offering substantial improvements in energy efficiency and operational reliability, thus contributing valuable insights to transformer design and application fields. Major Findings: The optimized Solid-State Transformer (SST) design achieved an impressive efficiency of 99.5832% with core and winding losses of 6.2391 W and 2.2457 W, respectively. The temperature rise was minimized to 11.4541°C, confirming the effectiveness of the hybrid optimization approach. Validation through fuzzy inference systems demonstrated robust alignment with both theoretical and experimental results.