Data-Driven Electricity Demand Modeling for Electric Vehicles Using Machine Learning

A Master of Science thesis in Electrical Engineering by Reza Davoodi Far entitled, “Data-Driven Electricity Demand Modeling for Electric Vehicles Using Machine Learning”, submitted in July 2024. Thesis advisor is Dr. Mohamed Hassan and thesis co-advisor is Dr. Ahmed Osman-Ahmed. Soft copy is availab...

Full description

Saved in:
Bibliographic Details
Main Author: Far, Reza Davoodi (author)
Format: doctoralThesis
Published: 2024
Subjects:
Online Access:https://hdl.handle.net/11073/25621
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1864513431786225664
author Far, Reza Davoodi
author_facet Far, Reza Davoodi
author_role author
dc.contributor.none.fl_str_mv Hassan, Mohamed
Osman, Ahmed
dc.creator.none.fl_str_mv Far, Reza Davoodi
dc.date.none.fl_str_mv 2024-09-25T07:03:09Z
2024-09-25T07:03:09Z
2024-07
dc.format.none.fl_str_mv application/pdf
dc.identifier.none.fl_str_mv 35.232-2024.29
https://hdl.handle.net/11073/25621
dc.language.none.fl_str_mv en_US
dc.subject.none.fl_str_mv Electric vehicle
Energy demand prediction
OD matrix
Machine learning
Multiple linear regression
dc.title.none.fl_str_mv Data-Driven Electricity Demand Modeling for Electric Vehicles Using Machine Learning
dc.type.none.fl_str_mv info:eu-repo/semantics/publishedVersion
info:eu-repo/semantics/doctoralThesis
description A Master of Science thesis in Electrical Engineering by Reza Davoodi Far entitled, “Data-Driven Electricity Demand Modeling for Electric Vehicles Using Machine Learning”, submitted in July 2024. Thesis advisor is Dr. Mohamed Hassan and thesis co-advisor is Dr. Ahmed Osman-Ahmed. Soft copy is available (Thesis, Completion Certificate, Approval Signatures, and AUS Archives Consent Form).
format doctoralThesis
id aus_3f2f02b0651f4d6f612eaeb75b79ee08
identifier_str_mv 35.232-2024.29
language_invalid_str_mv en_US
network_acronym_str aus
network_name_str aus
oai_identifier_str oai:repository.aus.edu:11073/25621
publishDate 2024
repository.mail.fl_str_mv
repository.name.fl_str_mv
repository_id_str
spelling Data-Driven Electricity Demand Modeling for Electric Vehicles Using Machine LearningFar, Reza DavoodiElectric vehicleEnergy demand predictionOD matrixMachine learningMultiple linear regressionA Master of Science thesis in Electrical Engineering by Reza Davoodi Far entitled, “Data-Driven Electricity Demand Modeling for Electric Vehicles Using Machine Learning”, submitted in July 2024. Thesis advisor is Dr. Mohamed Hassan and thesis co-advisor is Dr. Ahmed Osman-Ahmed. Soft copy is available (Thesis, Completion Certificate, Approval Signatures, and AUS Archives Consent Form).To address the pressing concerns surrounding electric vehicle (EV) range anxiety and energy availability, accurate prediction and coordination of EV energy demand are vital. While existing EV demand predictors rely on traffic simulators or local charging datasets, their effectiveness may be limited by a lack of actual EV driving data, hindering their scalability across different geographic regions. This study proposes a data-driven, machine learning (ML) approach for EV demand prediction, leveraging vehicular traffic flow data based on data collection for 250 trips in Dubai, United Arab Emirates. This model develops an EV energy consumption equation using multiple linear regression with statistically significant parameters from the trip-wise data. This is then mapped to a sample of vehicular trip data between origins and destinations across Dubai. On the other hand, the 250-trip data is used to obtain an energy consumption model using the SAS Analytical Tool, resulting in the simplified model consisting of trip length variable, trip duration variable, and average altitude of the trip variable. Furthermore, the data is then used to train various ML models including extreme gradient boosting (XGBoost), random forests (RF), linear regression, and DL models, such as multilayer perceptron (MLP) and long short-term memory (LSTM) to predict the trip-wise energy consumption. The performance metrics for both training and testing of various machine learning (ML) and deep learning (DL) algorithms reveal a stark contrast, particularly in their R² and RMSE values. During training, models like XGBoost exhibit near-perfect R² (0.999) and minimal RMSE (9.37* 10-4 kWh), suggesting they fit the training data exceptionally well. However, in testing, these values drop significantly (R² of 0.487 and RMSE of 0.668 kWh for XGBoost), indicating poor generalization to unseen data. The decline in performance across all models from training to testing suggests issues such as overfitting and inherent noise in the dataset, where models learn the training data too closely, failing to perform well on new data. Acquiring firsthand EV data and utilizing secondary open-access data, this work lays the ground for further studies into UAE-specific EV electricity demand modeling.College of EngineeringDepartment of Electrical EngineeringMaster of Science in Electrical Engineering (MSEE)Hassan, MohamedOsman, Ahmed2024-09-25T07:03:09Z2024-09-25T07:03:09Z2024-07info:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/doctoralThesisapplication/pdf35.232-2024.29https://hdl.handle.net/11073/25621en_USoai:repository.aus.edu:11073/256212026-04-22T08:55:52Z
spellingShingle Data-Driven Electricity Demand Modeling for Electric Vehicles Using Machine Learning
Far, Reza Davoodi
Electric vehicle
Energy demand prediction
OD matrix
Machine learning
Multiple linear regression
status_str publishedVersion
title Data-Driven Electricity Demand Modeling for Electric Vehicles Using Machine Learning
title_full Data-Driven Electricity Demand Modeling for Electric Vehicles Using Machine Learning
title_fullStr Data-Driven Electricity Demand Modeling for Electric Vehicles Using Machine Learning
title_full_unstemmed Data-Driven Electricity Demand Modeling for Electric Vehicles Using Machine Learning
title_short Data-Driven Electricity Demand Modeling for Electric Vehicles Using Machine Learning
title_sort Data-Driven Electricity Demand Modeling for Electric Vehicles Using Machine Learning
topic Electric vehicle
Energy demand prediction
OD matrix
Machine learning
Multiple linear regression
url https://hdl.handle.net/11073/25621