Data-driven energy management in low-carbon microgrids : leveraging data analytics and customer behavior analysis | |
| Author | Nahid, Firuz Ahamed |
| Call Number | AIT Diss no.SE-23-03 |
| Subject(s) | Microgrids (Smart power grids) Smart power grids--Technological innovations Electric power--Management Consumer behavior |
| Note | A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Engineering in Energy |
| Publisher | Asian Institute of Technology |
| Abstract | The global imperative to achieve a sustainable energy future and mitigate the impacts of climate change necessitates the transition towards a low-carbon energy system. Microgrids have emerged as a promising solution in this setting but cannot be implemented without effective energy management strategies. This dissertation proposes a data-driven energy management strategy that optimizes energy scheduling and ensures reliable and sustainable energy supply using advanced computational methods and customer behavior analysis. Firstly, a short-term multi-step decomposition-based hybrid deep learning approach is proposed to forecast wind speed and solar radiation. The proposed approach is an Empirical Mode Decomposition (EMD) based Convolutional Long Short-Term Memory (CLSTM) neural network that can accurately forecast renewables like wind and solar, supporting a scalable power management system. Secondly, A customer centric hybrid load forecasting model is proposed, utilizing K Means clustering to group customers by energy usage patterns. An Ensemble Empirical Mode Decomposition (EEMD) based Convolutional Long Short-Tenn Memory (CLSTM) neural network optimized by Grey Wolf Optimizer (GWO) is used for energy consumption forecasting, enabling customers to contribute to the energy economy actively. Thirdly, a hybrid model is proposed for forecasting residential Electric Vehicle (EV) charging demand in microgrid scenarios. The model uses Ensemble Empirical Mode Decomposition and Convolutional Long Short-Term Memory neural networks. Then Gurobi Optimizer finds the optimal charging schedule under appropriate constraints. Finally, a data driven efficient energy management system is proposed considering demand response, anticipated renewables, and loads. The proposed model incorporates the data-driven forecasting models and employs the General Algebraic Modelling System (GAMS) to optimize energy schedules to reduce energy consumption expenses. The approach proposed exhibits the capability to effectively optimize the cost of energy consumption and enable customers to engage in decision-making processes related to energy management. Overall, the proposed data-driven, customer-focused energy management approach empowers customers and assists system operators in understanding the dynamic energy landscape. |
| Year | 2023 |
| Type | Dissertation |
| School | School of Environment, Resources, and Development |
| Department | Department of Energy and Climate Change (Former title: Department of Energy, Environment, and Climate Change (DEECC)) |
| Academic Program/FoS | Sustainable Energy Transition (SE) |
| Chairperson(s) | Weerakorn Ongsakul;Singh, Jai Govind (Co-Chairperson) |
| Examination Committee(s) | Roy, Joyashree;Salin, Krishna R. |
| Scholarship Donor(s) | Bangchak Petroleum Public Company Limited, Thailand;AIT Partial Scholarship |
| Degree | Thesis (Ph.D.) - Asian Institute of Technology, 2023 |