Online UHF PD Monitoring for Transformers: Pulses Knowledge Acquisition

Partial discharges (PDs) of power transformers can be measured via ultra high frequency (UHF) technique. Under the integrated condition monitoring and diagnostics framework of HOMIS (Optimal Maintenance Information System for Hydro power stations), an UHF PD monitoring system was developed for power transformers. The system is composed of internal UHF sensor installed in the oil tank, an amplifier and envelope detector, an industrial computer for high speed data acquisition and real-time analysis. Due to the integration with SCADA and the condition monitoring system of other facilities in the station, the operating conditions and operating environments of the transformers can been synchronously acquired. Thus the correlative analysis of PDs with respect to the operating conditions can be performed. A large quantity of pulses from 4 transformers have been continuously collected and organized for years. To discover the comprehensive features of onsite pulses, an adaptive data storage strategy based on the operating conditions and online pulses recognition was employed. The identification method on features of the onsite pulses was studied with the aid of the platform of HOMIS. The pulses include periodic PD pulses, lightning pulses, switching pulses, and occasional discharge pulses caused by impulse voltages. Finally, the pulses knowledge base (PKB) was built using causal knowledge mode and IF-THEN rules. The PKB has made significant contributions to online pulses recognition and insulation condition assessment of transformers.