Energy AI

• This research is to develop a high-performance energy demand prediction model using data-driven analysis, machine learning, and ensemble models. Energy demand forecasting research is essential for efficient energy management and operation. Furthermore, accurate energy demand prediction enables future energy consumption planning.

• This study performs high-accuracy energy demand prediction based on actual data using missing data imputation method, data-driven analysis for extracting main consumption factors and constraints, pattern-based clustering analysis, machine learning, and ensemble model development.

• This study aims to develop a solar PV power generation prediction model using various machine learning algorithms, including ensemble-based techniques on public-open datasets. In order to plan a stable and efficient electricity supply and demand for smart grids, accurate and precise prediction technology is essential as the amount of solar PV generation fluctuates sensitively due to external environmental factors.

• This study develops an ensemble-based technique with a region of interest (ROI) to develop a hybrid Spatio-temporal solar PV power generation prediction model reflecting Spatio-temporal characteristics by combining observed numerical weather data and satellite imagery data.

A study on the optimal smart grid control methodology combining AI and energy big data

• In order to overcome the limitations of the previous energy management technologies in the Smartgrid, we aim to develop a highly reliable optimal operation model through integrative researches with deep learning technology.

• We aim to develop advanced core technologies for optimal operation of a truly energy-independent community including defining new research methodology based on construction ICT convergence, developing the new building simulation methodology, extracting potential impact factors based on passive and active components, analyzing data-driven optimization approach considering DR (Demand Response), proposing a novel deep learning technique considering uncertainty and time series characteristics (CNN, RNN, LSTM, ensemble-based model, etc.), and investigating a Deep Energy MPC model combining a MPC and AI technologies for energy management.

Development of AI-based Economic Evaluation Solution and Power Prediction of Renewable Energy System

• This study aims to develop dynamic and static data integrated algorithms for building energy demand forecasting models. We derive potential factors for building energy consumption, energy flow pattern from generation to consumption of renewable energy systems, and perform an environmental information collection and analysis.

• We propose novel energy demand (generation, storage, distribution, consumption) prediction technologies by using Digital Twin', 'Virtual Physical System (CPS)', and artificial intelligence (AI) of the key Industry 4.0 technologies.