Optimizing Surplus Harmonics Distribution in PWM

• Published in: Intelligent Information Technology pp. 366 - 375
• Main Authors: Hu, Shiyan, Huang, Han
• Format: Book Chapter Conference Proceeding
• Language: English
• Published: Berlin, Heidelberg Springer Berlin Heidelberg 01.01.2004; Springer
• Series: Lecture Notes in Computer Science
• Subjects: Applied sciences; Artificial intelligence; Computer science; control theory; systems; Constrained nonlinear optimization; Exact sciences and technology; Learning and adaptive systems; Quantum-inspired Evolutionary Algorithm; Surplus harmonics; Waveform; Harmonic; Pulse width; Probabilistic approach; Evolutionary algorithm; Initial value problem; Local optimum; Pulse modulation; Constrained optimization; Frequency band; Genetic algorithm; Filter; Problem solving; Frequency spectrum; Pulse duration modulation; Artificial intelligence; Newton method; Mathematical programming
• ISBN: 9783540241263; 3540241264
• ISSN: 0302-9743; 1611-3349
• DOI: 10.1007/978-3-540-30561-3_38

The goal of optimal pulse-width modulation (PWM) is to select the switching instances in such a way that a waveform with a particular characteristic is obtained and a certain criterion is minimized. The conventional method to solve the optimal PWM problem would usually lead to large content of surplus harmonics immediately following the eliminated frequency band, which may increase the filter loss and reduce the efficiency and performance of the whole controller. Meanwhile, it may increase the probability of resonance between line impedance and filter components. To overcome the shortcomings of conventional PWM methods, in this paper, we propose an algorithm for pushing the first crest of the surplus harmonics backward, ameliorating the amplitude frequency spectrum distribution of the output waveform, and thus reducing the impact of surplus harmonics. The problem is first formulated as a constrained optimization problem and then a Quantum-inspired Evolutionary Algorithm (QEA) algorithm is applied to solve it. Other than Newton-like methods, the enhanced QEA does not need good initial values for solving the optimal PWM problem and is not stuck in local optimum. The simulation results indicate that the algorithm is robust and scalable for a variety of application requirements.