Cloud based data Analytics and Control of Substation Transformer and Transmission Lines using IoT

In the realm of advanced smart grids, real-time monitoring and control of substation equipment and transmission lines have become essential. This article focuses on cloud-based monitoring prediction and control of substation transformer and transmission lines using Artificial Intelligence (AI) and Internet of Things (IoT). 3D printed model of a transformer, transmission towers and lines equipped with various sensors are designed and developed to acquire data from the transformer and transmission lines. The main objective of this paper is to acquire data from installed sensors and run data analytics on the cloud. And most importantly, predict Transformer Health Index (HI) and identify fault conditions in the transmission lines. Cloud-based big data analytics are employed to train and test Machine Learning (ML) algorithms for HI prediction. Multiple regression ML algorithms are trained and tested on the cloud, with their results compared. The prototype is designed for 440 V, 10 A, 50 Hz, demonstrating the proposed cloud-based monitoring, control, and prediction approaches. The hardware setup is based on a Cortex ARM controller (ARDUINO MKR 1000) and incorporates a PT (Potential Transformer), CT (Current Transformer), PZEM 004T, MQ5, IR (Infrared), and ultrasonic sensors. Transformer data, such as temperature, oil level, and CH 4 levels, are sensed and logged onto the cloud, which is then used to train and test ML algorithms for transformer HI predictions. Additionally, 3-phase, 3-wire transmission lines are laid across three transmission towers, and line current data is also collected and logged onto the cloud for fault detection. HTTP (Hypertext Transfer Protocol) request is incorporated into a MATLAB analysis app on the cloud to send fault notification emails to technical personnel and resolve the fault by activating a relay. The results presented demonstrate the effectiveness of cloud monitoring, transformer health analytics, HI prediction, and the fault resolution approach using in-built HTTP functionality.