IoT based Fourth Generation SCADA System for High Voltage Networks Fault Diagnosis based on BSDT-ANN

Abstract

Smart Grid substations rely on conventional Supervisory Control and Data Acquisition (SCADA) systems for remote supervision and control. However, these systems are limited in the geographical area they cover. Recently, the Internet of Things (IoT) has paved the way for connecting a vast number of devices to the Internet, which would be effective and beneficial for power system automation and data acquisition. In this paper, an intelligent low-cost SCADA system based on IoT for transmission line fault diagnosis is proposed. In the first step of the proposed scheme, voltage and current signals at the relaying point are preprocessed and analyzed using Discrete Cosine Transform (DCT). Next, signal energy components are extracted and sent to a Boosted Decision Tree (BSDT), a reliable and fast ensemble classifier, to identify fault type and, accordingly, a specific ANN is activated to estimate fault location. The diagnosis data is sent to a microcontroller to be displayed, trip the load circuit, and allow communication with Cloud ThingSpeak platform via the ESP8266 wi-fi communication module which makes data available anywhere all over the world. The approach is applied on a real-world HV transmission line, located between Samalut and Cairo, Egypt, where more than 16000 faults cases are well tested. The results show the reliability, validity, and effectiveness of the proposed approach.