Multi-stage Electric Grid Big Data Analytics Framework for Faster Processing and Forecasting

The proposed project will mainly focus on utilizing the historical energy consumption data and online delivered data for load forecasting and grid stability and efficiency enhancement. The project will effectively deal with the computational resources constraints. Large amounts of data will be analyzed and forecasting machine learning based models will be developed. The proposed framework utilizes a multi-stage approach to gain capabilities of processing large volumes of data within time constraints and simultaneously, generate highly accurate forecasting models. The proposed multi-stage framework utilizes the concept of data parallelism to handle data larger than single computational node memory/RAM and uses innovative clustering techniques to group similar data sources together and develop fewer models for the huge accumulated and continuously updated data. Parallel computing capabilities will be used. The goal is to reduce the overall execution time under high computational precision which allows responding to the large electric distribution network needs. Data clustering with transfer and adaptive learning procedures will be developed to generate aimed accurate machine learning models for load forecasting and grid stability enhancement. A Spark and non-Spark-based distributed processing will be interoperated with one another to process the available data.

Smart grids; Artificial Intelligence; Big Data Analytics and Management; Energy Forecasting