Multi- agent based realtime electric transformers monitoring system

Abstract

Vandalism of electrical equipment in Kenya has become a menace and a costly affair to the society in terms of safety and low economic activities. Particularly, transformer vandalism is on the rise year in year out and happens almost everywhere within the country. This project focused on a computer based mechanism of detecting and reporting transformer vandalism that was about to take place. The project aimed at sending out information or alerted security agents before actual vandalism took place. The project took advantage of motion detectors that were fitted in some of the transformers by Kenya Power and lighting company. These detectors sent different SMS alerts, depending on what was sensed/ detected. The SMS (alerts) were transmitted and received at a central point for further and quick analysis by computer algorithms developed in this project. Different programs in form of agents were used to receive and analyze different types of incoming alerts. There were different sample groups of people that were configured to receive these alerts. These sample groups represented The Kenya police, KPLC engineers and people living around the transformer. Vandalism eradication methods were not provided for in this project. The project gave computer based intelligent algorithms that made analysis, decision and reported vandalism to different persons. This was done without any human intervention in terms of analysis and decision making but through different computer programs in form of agents. This project also provided a means of capturing and storing different information and data of each transformer that was fitted with a motion detector for future consumption. Real-time monitoring of transformers project used Multi-agent based system methodology called PASSI because this methodology was simple to use, has got very clear steps and allows reusability of code. Testing and Evaluation of the project was carried out using a simulation method by use of predefined SMS as though they were being sent by a motion detector.