| PhD Viva

Name of the Speaker: Ms. Jayashree Rajaram Yadav (EE16D008)
Guide: Prof. Krishna Vasudevan
Online meeting link: https://krishnavasudevan.my.webex.com/krishnavasudevan.my/j.php?MTID=mbecb03d07fb72007dece1e03eae37f7b
Date/Time: 28th February 2024 (Wednesday), 5:00 PM
Title: Frequency Domain Models of Power Electronic Interfaces for Use in Studies of Electrically Polluted Grids

Abstract

The grid scenario is changing with a significant expansion of renewable energy sources and non-linear loads with inbuilt power electronic systems. During this work, we have collected and analyzed extensive power quality data from various low voltage (LV) locations (including industrial, commercial, and residential) over a period of 3 years to understand the status of power quality (PQ) in the Indian grid in collaboration with the industrial partner. The locations exhibit different short circuit ratios and consist of a variety of power electronic-based loads, conventional and non-conventional energy sources. The findings shed light on the actual level of different PQ disturbances, including sags, swells, voltage unbalance, harmonics, interharmonics, flicker, and supraharmonics (2-150 kHz) present in the Indian grid, which had not been documented previously in detail. Additionally, we have carried out an international survey on unbalance in the harmonic content of grid voltage using data from 12 countries. Our findings would be useful to standards organizations to estimate the changing grid scenario and serve as input to equipment manufacturers to improve their design to bolster the immunity of their products.

The study throws light on the need to model and understand how power electronic interfaces affect power quality and are in turn affected by it. To further investigate this mutual impact, we have proposed a coupled Norton model designed for a commercial variable frequency drive (VFD) as motor drives are widely used in industry and represent a very widely used power electronic interface. The work explores different formats of depicting the power electronic interface and their relative strengths in arriving at the system behavior. This model accurately captures harmonic interactions, addressing both balanced and unbalanced conditions. The proposed models are 90-95% accurate for different input waveforms and have response times 10-20 times faster than conventional time-domain models. The accuracy of the model prediction has been compared with hardware experiments on a motor drive as well. We have further validated the models with actual field voltage waveforms and investigated their application for predicting resonance conditions and estimating network harmonic impedance in a power system. The obtained model is also used in three-phase system study having multiple linear and non-linear loads with and without neutral. The impact of change in reference conditions on the model due to presence of grid impedance and due to change in the power level is also discussed in this work. The models developed under this work have broad applications for power quality studies and show promise for future research.