Selective harmonic elimination in cascaded H-bridge multilevel inverter using hybrid APSO algorithm / Mudasir Ahmed

Mudasir , Ahmed (2019) Selective harmonic elimination in cascaded H-bridge multilevel inverter using hybrid APSO algorithm / Mudasir Ahmed. PhD thesis, University of Malaya.

	PDF (The Candidate's Agreement) Restricted to Repository staff only Download (173Kb)
	PDF (Thesis PhD) Restricted to Repository staff only until 31 December 2021. Download (2165Kb)

Abstract

Observing the present scarcity of fossil fuel and emissions of greenhouse gases, electricity generated from Renewable Energy (RE) sources turns out to be the best alternative for generating power. In the RE system, the multilevel inverter (MLI) is normally used to condition the DC power into AC to meet the requirements of the load and transmission system. Furthermore, among the developed topologies of MLIs, cascaded H-bridge (CHB) is preferable because of its simplicity and use of less number of components. The MLIs offers myriad benefits, however, the presence of harmonics (particularly low order) in the output voltage affects the efficiency and performance of inverter, causes switching losses and decreases the lifetime of the system. In the last three decades, significant research has been done to develop an efficient control technique for eliminating unwanted harmonics. The preliminary review of existing control techniques revealed that the Bio-inspired intelligent algorithms (BIAs) based selective harmonic elimination pulse width modulation (SHEPWM) are more proficient to eliminate the loworder harmonics. Regardless of many benefits of BIAs, the issues like local minima, less precision, slow convergence rate, and computational complexity are the core optimization problems which required some serious consideration to acquire the maximum benefit from the SHEPWM technique. This study presents two efficient hybrid SHEPWM algorithms, one for low and second for high-level inverters to calculate the optimized firing angles. Both algorithms are designed to regulate the fundamental at the required level with the additional advantage of eliminating the undesired lower-order harmonics for modulation index (MI) which has solutions. The asynchronous particle swarm optimization (APSO)-newton raphson (NR) is proposed for low-level inverters, and APSO-genetic algorithm (GA) is proposed for high-level inverters. In both algorithms, global searching is performed by APSO algorithm. However, NR and GA respectively are used in each iteration for local refinement. In this study, APSO-NR is experimentally validated on the three-phase 7-level inverter to eliminate the 5th and 7th harmonics. The simulation results show that the proposed algorithm calculates highly precise firing angles in less number of iterations with the very high capability of tackling the local optima then Bee Algorithm (BA), differential evolution (DE), GA and PSO. For the 48% of MI range, APSO-NR minimized the objective function (OF) value lower than (10-25). The APSOGA is experimental validated on a reduced number of components based asymmetrical single-phase 7 and 49-level inverter. The 3rd & 5th harmonics in the 7-level inverter and 3rd to 47th harmonics in the 49-level inverter are chosen to eliminate from respective MLIs. Simulation results show that, at the high-level inverter, the proposed algorithm can easily find the feasible solutions, however, GA, PSO, bee algorithm (BA), and differential evolution (DE) face the difficulty due to less exploration capability. The proposed algorithm calculated precise switching angles particularly for high-level inverters with high capability to tackle local minima than BA, DE, GA and PSO. Also, the proposed algorithm is less computational complex than GA, BA and DE. For the 7-level inverter, of value of 19% of MI range is lower than 10-2. For the 49-level inverter, the value is below 1 for the 23% of MI range.

Actions (For repository staff only : Login required)