| dc.description.abstract | Abstract
In this dissertation, the authors apply the numerical methods such as the least square method (LSM), the discrete Fourier transform (DFT), dyadic wavelet transform and the wavelet transform (WT) for applications in optical device and in fault detection for transmission line of power system.
The thickness-dependent renormalization of strain effects on self-organized InAs quantum dots has been systematically investigated. By means of observing the photoluminescence spectra, the related information of energy band of device microstructure can be obtained. The activation energy and the carrier dynamics of the quantum dots with various thickness are interesting phenomena for investing strain effects of InAs quantum dots grown on GaAs. We can utilize the plot of energy-integrated intensity to characterize their thermal quench behavior. With the LSM and the WT method to coefficients estimating, the activation energy of the quantum dots can be fitted. The results are also compared.
Two different approaches for transmission lines protection are proposed. First, an adaptive phasor measurement unit (PMU) based protection scheme is proposed. In this algorithm based on two-ends method which the authors proposed, real-time voltage and current measurements are obtained firstly and discrete data can be obtained by sampling. With the aid of the LSM, the phasors can be solved. By using DFT of different window size, the phasors are easily estimated. Based on the amplitude and phase information, a fault detection index is then derived. Second, a dyadic wavelet transform based fault detection scheme is also proposed. The coefficient of discrete approximation of the dyadic wavelet transform is used to be an index for transmission lines fault detection, fault classification, and fault direction discrimination. The results reveal the dyadic wavelet transform based approach is a simple and effective one.
The authors utilize the WT method to estimate the aberration coefficients for a simulated wave-front. The spherical aberration coefficient and defocus are both estimated, and the numerical results are compared with those obtained by the LSM. With noise added, the results reveal the excellency of the WT method. The wave-fronts are also reconstructed by two methods simultaneously. The signal-to-noise ratio is also used as a performance index for evaluation between two methods with Gaussian white noise added. As the variance of Gaussian white noise increases, the superiority of the WT method is obvious. | en_US |