Sensor Network-based Intelligent System for Transformer Health Monitoring and Data Analysis

制造阿德里亚是专门为工业部门和采矿业设计和生产电力设备的公司。制造商Adria目前专注于通过使用在线监测、数据收集、事件检测和分析来改进其制造流程。制造商Adria正在寻求一种解决方案，将从传感器设备收集的无缝数据结合起来，然后执行事件检测和故障预测分析。该Engage赠款项目旨在研究和开发基于传感器的智能框架，该框架可以从高压设备收集数据，然后执行分布式和集中式数据分析。该项目的工作将包括以下三个目标：开发一个基于无线传感器网络(WSN)的系统来监控电力设备，然后连接到云服务器，开发一个阿帕奇Spark框架并执行数据分析。该研究具有创新性，因为它将开发一个基于传感器网络系统的新颖的物联网(IoT)架构，该架构在阿帕奇Spark框架上运行实时机器学习(ML)算法。该系统将利用云计算和雾计算来优化状态监测和特征提取过程。通过开发WSN试验台来从变压器收集数据，然后将WSN数据连接到云网关并将收集的数据发送到本地服务器，将达到本Engage应用程序中定义的目标。将使用基于ApacheSpark的ML库在传感器和云层进行数据分析。

Manufacture ADRIA is specialized in the design and production of electrical power equipment for theindustrial sectors and mining industry. Manufacture ADRIA is currently focusing on improving itsmanufacturing processes through the use of online monitoring, data collection, event detection, and analysis.Manufacture ADRIA is seeking a solution that combines seamless data collection from sensor devices and thenperforms analysis for event detection and failure prediction.This Engage grant project aims to investigate and develop a sensor-based intelligent framework that can collectdata from high voltage equipment and then perform distributed and centralized data analysis. The work on thisproject will include three objectives as follows; develop a Wireless Sensor Network (WSN)-based system tomonitor power equipment and then connect to a cloud server, develop an Apache Spark framework andperform data analysis.The proposed research is innovative since it will develop a novel Internet-of-Things (IoT) architecture based onthe sensor network system that runs real-time Machine Learning (ML) algorithms on an Apache Sparkframework. The system will utilize cloud and fog computing to optimize the condition monitoring and featureextraction process.The objectives defined in this Engage application will be achieved by developing a WSN testbed to collect datafrom a transformer, then connect the WSN data to a cloud gateway and send the collected data to a local server.Apache Spark-based ML libraries will be used to perform data analysis in the sensors and cloud levels.