WEB-BASED PLATFORM FOR TRANSFORMER OIL HEALTH MONITORING AND ANALYSIS USING MACHINE LEARNING
Loading...
Files
Date
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Universiti Malaysia Sarawak
Abstract
Description
Traditional transformer oil maintenance relies on periodic manual testing, often using separate equipment for each oil parameter, which is costly, time‑consuming, and may miss early signs of degradation. To address these challenges, this project presents INSIGHT - Intelligent System for Grid Health and Transformer Oil Tracking, a web‑based platform that centralises oil‑sample data management and accelerates decision‑making. Users upload historical records, and the system streamlines data preprocessing, detects anomalies, diagnoses issues, and recommends maintenance actions. A key innovation is the use of machine‑learning classifiers to predict individual oil parameters from other oil parameters in the same sample, allowing rapid preliminary health assessments that help laboratories prioritise tests and resources. Interactive dashboards present historical and predicted oil samples within a unified interface, with clear separation to support comparison. Customisable reports facilitate clear communication with stakeholders. An administrator dashboard allows user management and activity auditing to ensure traceability and security. Developed using Agile’s Scrum framework, the architecture is designed to scale across diverse transformer types and datasets. By reducing manual testing requirements, centralising analyses, and highlighting the most critical parameters for laboratory confirmation, INSIGHT optimizes maintenance workflows, reduces costs and improves overall transformer reliability.