Power System Harmonics Estimation: A new optimization technique-based implementation with African Vulture Optimization Algorithm based Least Square Method

• Published in: 2023 First International Conference on Advances in Electrical, Electronics and Computational Intelligence (ICAEECI) pp. 1 - 7
• Main Authors: Ali, Atif, Ahmed, Mushfiqur, Apon, Hasan Jamil, Rahman, Md. Samiur, Ahsan, Auhona, Prapti, Sumaiya Tasnim, Ahmed, Ashik
• Format: Conference Proceeding
• Language: English
• Published: IEEE 19.10.2023
• Subjects: African Vulture Optimization Algorithm; Estimation; Frequency estimation; Hardware; Harmonic analysis; Harmonics Estimation; Least Square Method; Optimization methods; Power Quality; Power system harmonics; Voltage
• DOI: 10.1109/ICAEECI58247.2023.10370852

In present days, estimating harmonic components has become an important research area for conventional power system. Several stochastic optimization methods have been employed to successfully estimate the amplitude and phase of a harmonic signal. A novel optimization strategy must be implemented to increase precision and reduce the computational time. This research provides a technique for harmonics estimation based on the African Vulture Optimization Algorithm (AVOA) that utilizes African vultures' natural hunting and navigational behavior. In this work, the amplitude of the harmonics of a test signal has been determined with the help of the traditional Least-Square (LS) approach, while the phase of the harmonics of that test signal has been determined using the African Vulture Optimization Algorithm (AVOA). The results demonstrate that this technique can estimate the harmonic components more precisely. Finally, the results have been compared with other three optimization techniques proposed in previous works, such as Particle Swarm Optimization with Passive Congregation (PSOPC) based least square method, Firefly Algorithm (FA) based least square method, and Archimedes Optimization Algorithm (AOA) based least square method. The proposed AVOA-based least-square method outperforms all three algorithms regarding estimation precision. In single frequency estimation, at a 40 dB signal-to-noise ratio (Gaussian noise), the percentage error result between the estimated signal and the original signal is 1.0173 e -4 , and the computational time is 0.0804 seconds. In case of multiple frequency estimation, with the same noise, the percentage error result is 1.07 e -4 and the computational time is 0.167 seconds.