Load forecasting for distribution network and electric vehicle charging optimisation to mitigate load fluctuations

• Main Author: Ng, Xin Yi
• Other Authors: Amer M. Y. M. Ghias
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2024
• Subjects: Engineering; Electric vehicle; Random charging; Distribution network; Monte Carlo simulation; Acceptance-Rejection sampling; Base load forecasting; Backpropagation algorithm in neural network; Rolling horizon; Electric vehicle charging strategies
• Online Access: https://hdl.handle.net/10356/176818
• Institution: Nanyang Technological University

Electric vehicles (EVs) offer significant advantages in energy conservation and emission reduction, leading to their growing popularity. However, the unpredictable nature of EV charging behaviours could lead to random charging patterns during peak electricity load hours, resulting in an “increasing peak” phenomenon that challenges grid reliability and stability. This project delves into the multifaceted challenges posed by the integration of EVs into the existing energy infrastructure. It begins with a comprehensive survey focusing on EV-related factors. Through Monte Carlo simulation, this project investigates the adverse effects of random EV charging demand based on real driving data, aiming to gain insights into its impacts on the grid. Additionally, this project employs a Backpropagation algorithm in neural network (BPNN), utilising historical data to accurately forecast base load demand. In response to these findings, this project proposes EV charging strategies aimed to optimise load fluctuations by minimizing peak-to-valley differences and enhance grid stability, with a focus on the distribution network in Singapore. Simulation results demonstrate the effectiveness of these strategies in improving load management, highlighting the importance of it in mitigating such challenges.