Abstract
The online monitoring of a high voltage apparatus is a crucial aspect for a predictive maintenance program. Partial discharges (PDs) phenomena affect the insulation system of an electrical machine and—in the long term—can lead to a breakdown, with a consequent, significant economic loss; wind turbines provide an excellent example. Embedded solutions are therefore required to monitor the insulation status. The paper presents an online system that adopts unsupervised methodologies for assessing the condition of the monitored machine in real time. The monitoring process does not rely on any prior knowledge about the apparatus; nonetheless, the method can identify the relevant drifts in the machine status. In addition, the system is specifically designed to run on low-cost embedded devices.
Highlights
A predictive maintenance apparatus monitors an electrical machine continuously to prevent a breakdown in the insulation system [1,2,3,4]
The experimental session involved a set of twisted pair specimens that underwent aging tests according to standard IEC 60851-5
An augmented monitoring system combining the two strategies could be able to start collecting q a (t) only when the transient has ended. Such data may be sent in real time to a central unit entitled to estimate the value of parameters K, n1 and n2 in (6) by exploiting the data acquired from different electrical machines
Summary
A predictive maintenance apparatus monitors an electrical machine continuously to prevent a breakdown in the insulation system [1,2,3,4]. The above-mentioned approaches all involve a data-driven classification systems, which in turn requires a training procedure Those methods do not provide either a run-time monitoring of the aging phenomena nor an estimation of the remaining life of a machine, a pair of essential requirements for a cost-effective maintenance scheduling. Template matching allowed the system both to classify the PD sources and to monitor changes over time This method, only applies if one knows in advance all possible defects (i.e., the PD pattern shapes) that may affect the observed device. The results proved that after an initial transient period, a linear relation existed between a specific feature and the degradation time of the specimens These approaches both showed that one could collect data from a set of available specimens, build an according model, and predict the lifespan of a new, unseen specimen. The paper illustrates an implementation of such strategy by applying the empirical model presented in [27]
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call