Optimal single-tuned passive filter design for harmonics mitigation using mixed integer distributed ant-colony optimization

This thesis was submitted for the award of Doctor of Philosophy and was awarded by Brunel University LondonPower system harmonics has existed since the distorted current and voltage were observed in the power system. As newly advanced and developed technologies and the ongoing research efforts were made, awareness and interest were developed by pollution in the system created through harmonics. This research showed an advanced optimization solver called mixed-integer distributed ant colony optimization (MIDACO). The simulation is based on a single optimization of four objective functions, which minimizes the total of harmonic voltage distortions, maximizes the power factor, maximizes the transmission efficiency, and minimizes the resistor losses of the impedance of Thevenin. The presented solver is aimed at optimizing the size of a single-tuned passive filter based on harmonics elimination for all the objective functions involved. A completed and detailed study was conducted considering several significant constraints in designing the best filter. The global maximum and minimum can be achieved by maintaining the quality factor in a particular range, such as the harmonic resonance constraints and the capacitors that do not violate the limits of Institute of Electrical and Electronics Engineers (IEEE) standard 18-2012. The constraints also considered the design in which the harmonic voltage was below the standard limit, thus maintaining the desired range of power factor. The results are evaluated based on the system performance in four different cases in the study to be compared with the genetic algorithm (GA), particle swarm optimization (PSO), and previously published journals. In this research, a single-tuned filter design with and without damping resistor of the inductance was investigated. Then, the comparative study between the advantages and the disadvantages of damping over undamped filter was also presented in this research. Additionally, the numerical results were analyzed considering two objective functions simultaneously, which are total harmonic distortion (THD) and transmission losses. The results of the nondominated solution for the simulation of the multiobjective problem with different search effort on the part of the Pareto front have also been investigated. The advantages of multiobjective over single objective optimization are also presented and highlighted in this thesis.Government of Malaysia (Majlis Amanah Rakyat