Deep Learning-Based Edge Real-Time Monitoring and Early Warning System for Smart Grids

The increasing needs of contemporary cities have led to a key challenge in today’s power grid: unreliable and inefficient power transportation. Because of this, power grid monitoring is one of the essential components of the power grid system and helps to avert serious safety incidents. Unfortunately, the old manual inspection method cannot effectively accomplish this goal because of its high cost and low efficiency. The smart grid, a new generation of power infrastructure, offers fresh perspectives on how to build an advanced information technology-based power system that is intelligent, dependable and efficient. Automatic monitoring in smart grids can be achieved by utilizing cutting-edge deep learning algorithms on robust cloud computing platforms in conjunction with Internet of Things (IoT) devices like smart cameras. It is challenging to completely satisfy the real-time needs of online power line monitoring due to the server’s constrained processing and storage capabilities. As a result, this study suggests an edge-based smart grid real-time monitoring and an early warning system that can assign computing jobs in a way that makes reasonable use of edge node resources. Additionally, it can gather and analyze data instantly and monitor the power grid online. This enhances the grid’s fault identification efficiency, lowers costs and lessens the strain on the online monitoring equipment on the edge computing (EC) server. Machine learning (ML) algorithms can also provide various forms of adaptive real-time control like load-balancing intelligent energy systems, real-time power price prediction, etc. Based on simulation research, the scheduling method presented in this work may effectively lower the power line monitoring system’s monitoring delay, enhance the system’s overall compatibility and adaptability, satisfy market demand and make promotion easier.