The analysis of three level inverter circuit with regard to current harmonic distortion by using ANFIS

Abstract

The inverters are frequently used in power electronics applications in the industrial area. Research on multi-level inverter circuit designs in power systems interfaces has intensified in recent years. The design type and quality of the inverter circuit topology is very important for the output voltage to be closer to the sinusoidal waveform. In addition, artificial intelligence techniques for the control algorithms of the power switches of the inverter circuit provide useful information in terms of monitoring the estimated the voltage conditions of the inverter output parameters. In this work, an Adaptive-Network Based Fuzzy Inference Systems (ANFIS) model is proposed to estimate the Total Harmonic Distortion (THD) value of the output current of a three-phase three-level inverter circuit. By changing the switching frequency in the control circuit of the inverter, a data set for the THD values of the inverter current is obtained by Fast Fourier Transform (FFT) analysis in MATLAB. This data set is used in a training and testing phase of ANFIS artificial intelligence algorithm, and the THD value of the current is estimated. At the end of the estimation, the mean absolute error (MAE) values for training and testing are obtained as 0.1894% and 0.4009%, respectively, thereby, an ANFIS estimation example for parametric data set analysis in a power electronics circuit run with Matlab-Simulink software, and a parametric simulation study is presented to the literature for power electronics circuit designers.

ANFIS-based Parameter Estimation of a Single Phase Inverter Circuit with Isolation Transformer

“…Thus, it was established by taking advantage of the learning ability of artificial neural networks, considering the reasons such as the fuzzy logics inability to adapt easily to environmental conditions and its lack of learning ability. The uses of ANFIS can be listed as modeling nonlinear functions, linearly defining nonlinear components and estimating a chaotic time series [23,24].…”