Statistical analysis of partial discharge characteristics for predictive maintenance of generator of geothermal power plant

In order to guarantee reliable industrial power supply, predictive maintenance for generator becomes highly necessary. This measure is expected to reduce disruptive generator due to degraded insulation of its stator winding. Partial discharge (PD) is responsible for this disadvantage condition. The objective of this paper is present partial discharge (PD) analysis using statistical methods for predictive maintenance of generator in geothermal power plant. TGA-B IRIS was used for online PD measurement. The measurement data is analysed statistically to forecast the time of critical condition of stator winding of generator using vector autoregressive (VAR) method. The analysis result became the basis of predictive maintenance.