A Comprehensive Review on Data-Driven Techniques in Smart Power Grids

As a promising vision toward obtaining high reliability and better energy management, nowadays power grid is transferring to the smart grid (SG). This process is changing continuously and needs advanced methods to process big data produced by different segments. Artificial intelligence methods can offer data-driven services by extracting valuable information which is produced by meter devices and sensors in smart grids. To this end, machine learning (ML), deep learning (DL), reinforcement learning (RL), and deep reinforcement learning (DRL) can be applied. These methods are able to process huge amounts of data and propose an appropriate solution to solve power industry complex problems. In this paper, the state-of-the-art approaches based on artificial intelligence used by smart power grids for applications and data sources are investigated. Also, the role of big data in smart power grids, and its features such life cycle, and efficient services such as forecast, predictive maintenance, and fault detection are discussed.