In Part 4 of our video series titled Essentials of Partial Discharge Testing, Caspar Steineke, partial discharge (PD) expert and co-creator of OMICRON MPD technology, discusses how disturbances, namely external noise typical in many onsite and offsite testing environments, can compromise the reliability of PD measurement data. He describes the various methods and source separation techniques used to reliably minimize or even eliminate the effects of external noise to enable highly accurate PD measurements, as well as the reliable analysis and interpretation of PD measurement data.
00:00 - Introduction 00:14 - How does external noise affect PD measurements? 00:43 - How can you minimize the effects of external noise? 02:12 - What tools are used for PD source separation? 03:15 - Conclusion and contact information
Partial discharge (PD) is a reliable indicator and major cause of deterioration in electrical insulation which can result in equipment failure and costly outages. Regular PD testing is a non-destructive and highly reliable method for assessing the insulation condition of electrical equipment, and it provides users with vital information to plan timely action that ensures their continued reliability and availability.
Hi, my name is Kasper Steinecke. I'm a co-creator of the original MPD technology
and I'm here today to share information on partial discharge testing with you.
Noise is always the enemy of measuring PD. Imagine doing a precision acoustic measurement
on a busy street. We have very small signals coming from our acid and yet at the same time
we might have large disturbances nearby. So measuring PD is very much about separating
noise from signal and there are numerous ways to do that. In a lab situation there are very clear
ways of dealing with noise like shielding and then it's all about cost. How much do you wish
to invest in a perfectly shielded lab versus applying techniques to remove existing noise?
On site it's a whole different ball game. It is all about being able to extract meaningful signal
in a very noisy environment. So we have a toolbox with multiple tools to deal with that noise
and to extract meaningful signals and perform valid measurements. Most real world noise
is spectrally limited. That is, it exists on the frequency spectrum within a confined
frequency range. What we can do then is we can freely select a measuring frequency after
analyzing that noise level and placing our measurement within a frequency range that is
least affected by noise. When that doesn't work because we have very wide band noise that occurs
spuriously, then we also have the means of gating noise. Gating is about detecting external
interference that constitutes noise and removing that from the measurement. So when we're taking
measurements on larger assets, particularly rotating machines and transformers,
we have multiple faces and we have large geographical dimensions. That is, we're looking
at potentially multiple sources of PD in one asset and then when we look at the PRPD, the
face resolved PD diagram, we're seeing all PD activity superimposed. Now, in order to make
sense of these measurements, it helps to separate out individual sources and look at them one by one.
We have two mechanisms that we invented for doing exactly that. One is three-part, a
multi-phase correlation that also speeds up measurement on multi-phase assets. The other one
is three-freq, a multi-frequency separation that separates out sources by looking at their
individual frequency signatures. I hope this information was of interest to you. Omicron
has over 20 years of experience in partial discharge testing and we'd be pleased to work
with you in this area.