ρReveal: An AI-based Big Data Analytics Scheme for Energy Price Prediction and Load Reduction
Aparna Kumari, Sudeep Tanwar
Abstract
Nowadays, the prediction of energy prices play an important role for optimal energy management in advanced grid infrastructure, i.e., Smart Grid (SG). In this paper, we focus on predictive analytics of energy prices that are extremely large in nature and also difficult to handle using conventional computational methods. An accurate price prediction facilitates SG system to make efficient and profitable strategies to increase overall revenue. This paper proposes a novel and secure Big Data Analytics (BDA) scheme, i.e., ρReveal, which comprises an Artificial Intelligence (AI)-based model for prediction of energy prices using Bidirectional Long Short-Term Memory (BiLSTM). Then, Spark-based analytics is done on load reduction based on the predicted energy prices. Then, analytics reports are encrypted with a digital signature to handle security issues such as data modification attacks and data integrity attacks. The performance of ρReveal scheme is evaluated based on various prediction accuracy metrics such as Mean Absolute Error (MAE) and Root Mean Square Error (RMSE) compare to existing approaches.