Knowledge Discovery from On-line Cable Condition Monitoring Systems - Insulation Degradation and Aging Diagnostics

这项建议涉及格拉斯哥喀里多尼亚大学（GCU）和斯特拉斯克莱德大学（UoS）学者之间的合作研究。该项目的主要目的是应用跨学科的方法来解决的问题，获取必要的信息诊断的基础上的局部放电（PD）活动的电缆绝缘在线状态监测。这将通过开发现代数据挖掘技术来实现，以直接从在线、数据丰富的状态监测系统中获取知识。应用系统的在线信息分析将通过在实验室环境和实际配电系统中进行的广泛的专用实验得到支持和验证。配电网络的故障对运营商来说代价高昂，对消费者来说也是一个严重的问题，他们在维修期间经历了停电和业务和社会活动的中断。如果没有开发出科学地确定电缆绝缘状况及其性能的技术，则纯粹根据年龄进行计划维护而不必要地更换电缆资产可能会导致类似的中断和过高的成本。这项拟议的研究计划将建立在三个领域，其中研究人员具有国际公认的专业知识：首先，测量和区分信号特征，从高发电厂，其次，确定在油/纸绝缘系统的退化，第三，应用软件来确定知识夹带在原始数据。这项研究计划将大大受益于最近资助的两个EPSRC项目在GCU和UoS（EP/D 048133和GR/86760）以及最近完成的工业资助项目所获得的知识。除了建议者的学术实力外，EDF Energy还提供了非常可观的工业贡献：30，000英镑的直接现金注入和大量实物捐助，即电缆样品，无限制地获取其在线状态监测系统的数据和其更换方案的故障/状态报告，以及获取其工作人员的实际专业知识。来自电缆制造公司普睿司曼电缆和系统有限公司（20，000实物），陶氏化学和基于PD的状态监测设备提供商IPEC Ltd（8，000实物）的进一步支持将确保研究的有效性，并通过调查一系列工厂类型和设置来扩大项目范围，并确保研究的更普遍适用性。

This proposal involves collaborative research between academics at Glasgow Caledonian University (GCU) and the University of Strathclyde (UoS). The primary aim of the project is to apply a cross-disciplinary approach to address the problem of acquiring essential information for diagnostics of on-line condition monitoring of cable insulation on the basis of partial discharge (PD) activity. This will be achieved by developing modern data mining techniques to acquire knowledge directly from on-line, data rich, condition monitoring systems. Analysis of on-line information from applied systems will be supported and validated through extensive, dedicated experiments carried out both in the laboratory environment as well as in practical power distribution systems. Failures in the power distribution network are costly to the operators and they are also a serious issue for consumers, who experience power cuts and disruption to their business and social activities during repairs. If techniques for establishing scientifically the condition of cable insulation and its performance are not developed, similar disruption and excessive cost can result from unnecessary replacement of cable assets on the basis of planned maintenance based purely on age. This proposed research programme will build on three areas in which the investigators have internationally recognised expertise: firstly, measuring and discriminating signal characteristics from high power plant, secondly, determining degradation in oil/paper insulation systems and, thirdly, applying software to determine knowledge entrained in raw data. This programme of research will significantly benefit from the knowledge gained from two recently funded EPSRC projects at GCU and UoS (EP/D048133 and GR/86760) as well as recently completed industrially funded projects. In addition to the academic strengths of the proposers, a very substantial industrial contribution is being provided by EDF Energy: a 30,000 direct cash injection and strong in-kind contribution, i.e. cable samples, unlimited access to data from its on-line condition monitoring systems and to fault/condition reports from its replacement programme as well as access to practical expertise of its staff. Further support from Cable manufacturing company Prysmian Cables and Systems Limited (20,000 in kind), Dow Chemical and PD based condition monitoring equipment provider IPEC Ltd (8,000 in-kind) will ensure breadth of validity of the research and broaden the scope of the project by investigating a range of plant types and set-ups and to ensure more general applicability of the research.