Due to a changed situation on the energy market, assets are supposed to remain
in service as long as possible. To ensure a safe and reliable operation, condition based maintenance (CBM) is needed, for
which highly reliable diagnostic tools are required.
On-line monitoring has several advantages compared to off-line diagnostic measurement. Assessment based on real operating condition (load, temperature, vibration) will be available, and comparison does not depend on the person performing the tests. Monitoring concept can be applied during the entire life cycle of the asset , either via continuous or temporary monitoring systems
Partial discharges (PD) activity is a good indicator of the insulation condition. A successful PD measurement in stator windings requires successful separation of parallel active PD sources and distinction between harmful PD, normal PD occurrences and external noise inevitably present in industrial surroundings.
To achieve this, the following techniques are applied: Synchronous multi-channel data acquisition Multi-spectral evaluation
Advanced noise suppression ource separation techniques (3PARD, automated cluster separation)
3PARD visualizes the relation among amplitudes of a single PD pulse in one phase and its crosstalk generated signals in the other two phases. In the 3PARD diagram, several sources of PD are recognizable as different clusters. Back transformation of the chosen clusters to their correlated Phase-resolved PD Diagrams (PRPD) then shows the separated PD sources.
The monitoring software supports the automated cluster separation. A highly efficient hierarchical density- based clustering algorithm is applied for the automatic clustering of heterogeneous 3PARD data. For each cluster, the system automatically identifies the phase
of signal origin. For PD clusters the PD data set with all significant data is saved for additional expert analysis or future comparisons.
For HV assets with well-established PRPD pattern database such as rotating machines, the monitoring software automatically classifies the type and risk of defect. With this automatic pattern classification, it is possible to monitor
only harmful PD, separated from external
noise and normal PD occurrences.
For technical enquiries, please contact Seokhoon Hong (Regional Application Specialist – Partial Discharge) via email email@example.com or ENTEC A&T’s Email: firstname.lastname@example.org