Date of Award
2006
Degree Type
Thesis
Degree Name
Master of Engineering Science
Program
Electrical and Computer Engineering
Supervisor
Dr. Lyndon J. Brown
Abstract
This thesis presents two improved approaches to extend a periodic disturbance cancellation algorithm to achieve different desired goals in active noise control and resistance spot welding (RSW). The original algorithm is adopted from the internal model based frequency estimation and noise cancellation principle in the error feedback control field. One improvement of the algorithm is made and applied for an acoustic duct system to increase the stable frequency band. A design method of adaptively tuning the two control gains with the estimated frequency is presented. Simulations show that the improved algorithm not only broadly increases the working frequency band of the system, but also decreases the worst case error with minimum loss of speed. The other extension of the algorithm is interpreting its results as an instantaneous Fourier series. This is then used for online estimation of the dynamic resistance in the secondary circuit of a resistance spot welding machine by instantaneous Fourier decomposition. The Fourier decomposition system is developed behaving as a bandpass filter with notch filters. Simulations on synthesized data show good agreements between the estimates of the resistance and tip voltage magnitude and their respective known values. Application on experimental data gives very reasonable results.
Recommended Citation
Sun, Yujuan, "Instantaneous Fourier Series Estimation" (2006). Digitized Theses. 4958.
https://ir.lib.uwo.ca/digitizedtheses/4958