Study Analysis and Simulation of Power System with Distributed Generation and Changing Load Profiles

Abstract

Modern power distribution systems are experiencing the integration of Distributed Generation newline(DG) and changing load profile, presenting significant challenges and opportunities of newlinemaintaining grid performance parameters. Of the several DG s, solar PV is an important player. newlineThe changing load profile is characterised by the increasing number of Electric Vehicles (EV) newlineconnected to the grid. These changing load profiles refer to the dynamic nature of power newlinedemand, which no longer remains static but varies significantly over time due to the newlineintermittent and unpredictable charging behaviour of EVs. This research focuses on analysing newlinethe impact of Solar DG and EV penetration levels on Grid performance parameters, including newlinevoltage profiles, power losses, ampere loading, MW and MVAr loading. Moreover, EVs often newlinebehave as dynamically changing loads due to their variable charging demands and usage newlinepatterns, further complicating grid performance parameters. The study employs advanced newlineanalytical methods, including Grey Relational Analysis (GRA) for optimal DG and EV newlineplacement and Iterative Parametric Analysis for DG sizing, ensuring an optimised and resilient newlinepower distribution network. newlineThe research methodology involves modelling and simulation using the balanced IEEE-33 bus newlinetest system in ETAP, incorporating EVs as both dynamic loads and energy storage sources newlinethrough Vehicle-to-Grid (V2G) and Grid-to-Vehicle (G2V) operations. A comprehensive load newlineflow analysis is conducted to evaluate power system parameters with solar DG and varying newlinelevels of EV penetration. Heatmap visualisations and t-Distributed Stochastic Neighbour newlineEmbedding (t-SNE) clustering techniques are applied to analyse spatial-temporal variations in newlinesystem performance. A comparative analysis of different DG and EV integration scenarios newline(0%, 25%, 50%, 75%, and 100%) is performed to assess their respective impacts on grid newlineperformance parameters. A predictive framework utilising the Random Forest Regressor newline(RFR) machine learning algorithm is developed to predict