Kuo, Chi WeiChi WeiKuoKang, Chia YuChia YuKangYING-YIN HUANGCHIH-KUNG LEEYAO-JOE YANG2023-04-072023-04-072007-08-231420063421https://scholars.lib.ntu.edu.tw/handle/123456789/630029In this work, we present the optimization works of two disk-type piezoelectric transformers (PTs), the single-output PT and the dual-output PT, by using the elite Genetic Algorithm (GA) and a finite-element solver. The design goal is to maximize the efficiency under the constraint that the voltage gain is greater than 50 for igniting CCFL. The design parameters are the dimensions of the output-electrode sections, the electrode areas, and the devices structure. With different electrical loading impedances, the voltage gain and the efficiency were computed using the finite element solver, NTUPZE. The results were also verified with measured data. The optimization process is parallelized by the MPI library and a PC cluster for improving the computation efficiency. The characteristics of the optimal designs with different loads are also calculated.Finite element method (FEM) | Genetic algorithm (GA) | Multi-object optimization | Parallel computations | Piezoelectric transformerconference paper2-s2.0-34547988682https://api.elsevier.com/content/abstract/scopus_id/34547988682