Internet of Things Partial Discharge Detection (PD) System

许多研究人员面临的主要挑战之一是训练和测试机器学习(ML)模型所需的试验床和数据的可用性。我们在我的研究实验室中使用ML模型来评估和预测电力资产的健康状况，如高压变压器和发电机。我们对检测和预测高压变压器的局部放电(PD)活动特别感兴趣。过去，我们使用模拟和人工数据来进行局部放电分析和预测。然而，这些方法被认为是不现实的，因为它们依赖于合成数据。拟议的国际合作的主要目标是1)与布莱恩·斯图尔特博士在局部放电检测和预测系统方面建立国际研究伙伴关系，2)了解局部放电数据收集系统的要求和实际限制，以及3)设计局部放电数据采集原型。拟议的研究合作将有助于吸引加拿大电力行业和公用事业公司的兴趣，以支持为开发实时PD系统的研究项目提供资金。这将通过提交与合作者和行业合作伙伴的联合资助赠款来实现。这些系统将帮助当地工业在电力设备制造领域站稳脚跟，并帮助加拿大成为创新智能电网技术的全球领先者。

One of the main challenges many researchers face is the availability of testbeds and data for training and testing machine learning (ML) models. We use ML models in my research lab to assess and predict the health of electrical power assets such as high voltage (HV) transformers and generators. We are particularly interested in detecting and predicting partial discharge (PD) activities in HV transformers. In the past, we used simulations and artificial data to perform PD analysis and prediction. However, these approaches are considered unrealistic because they rely on synthetic data. The main objectives of the proposed international collaboration are to 1) Establish an international research partnership with Dr. Brian Stewart in PD detection and prediction systems, 2) Understand the requirements and practical limitations of PD data collection systems, and 3) Design a prototype PD data acquisition. The proposed research collaboration will help in attracting interest from Canadian power industries and utilities to support funding research projects to develop real-time PD systems. This will occur by submitting joint funding grants with the collaborator and the industrial partners. These systems will help local industries establish a presence in power equipment manufacturing and help establish Canada as a global leader in innovative smart grid technologies.