Optimal Design of Passive Power Filter Using Multi-Objective Pareto-Based Firefly Algorithm and Analysis Under Background and Load-Side’s Nonlinearity

• Main Authors: Mohit Bajaj, Naveen Kumar Sharma, Mukesh Pushkarna, Hasmat Malik, Majed A. Alotaibi, Abdulaziz Almutairi
• Format: Article
• Language: English
• Published: IEEE 2021-01-01
• Series: IEEE Access
• Subjects: Background distortion; firefly algorithm; harmonic compensation; harmonic distortion; multi-objective optimization; passive power filters
• Online Access: https://ieeexplore.ieee.org/document/9343288/

In this paper, the optimal designing of passive power filter (PPF) is formulated as a multi-objective optimization (MOO) problem under several constraints of system's performance indices (PIs) such as individual as well as total harmonic distortion (THD) in the line current and the point of common coupling's (PCC) voltage, distribution line's ampacity under harmonic currents overloading, steady-state voltage profile, load power factor (PF) and a few associated with the filter itself. The optimal design parameters of a third-order damped filter are simultaneously determined for achieving maximum PF at the PCC while keeping system's other indices such as total demand distortion (TDD) in the line current, total voltage harmonic distortion (TVHD) at the PCC and total filter cost (FC) incurred at a minimum by obtaining a best-compromised solution using the newly proposed multi-objective Pareto-based firefly algorithm (pb-MOFA). A novel MOO approach inspired by the modified firefly algorithm and Pareto front is established in order to deal with PPF design problems. The extension of MOFA is considered for producing the Pareto optimal front and various conclusions are drawn by analysing the trade-offs among the objectives. The efficiency and accuracy of the proposed pb-MOFA, in solving the concerned MOO problem, is validated by comparing an obtained solution and three computed PIs viz. convergence metric (CM), generational distance (GD) and diversity metric (DM) with those obtained from popular multi-objective Pareto-based PSO (pb-MOPSO), non-dominated sorting genetic algorithm (NSGA-II) and recently introduced multi-objective slime mould algorithm (MOSMA). The need for true Pareto front (TPF) is served by the one obtained by Monte Carlo method. At last, the impacts of different background voltage distortion (BVD) levels and load-side's nonlinearity levels (NLLs) on filter performance are analysed.