A proposed method for establishing partial discharge acceptance limits on API 541 and 546 sacrificial test coils

Recent revisions to API standards have identified offline partial discharge analysis (PDA) as an appropriate factory acceptance test to evaluate the quality of certain new medium voltage motor and generator stator windings. As a starting point, API 541 recommends a partial discharge (PD) level of 100 pC (pico-Coulombs) for sacrificial stator coils measured with an IEC 60270 compliant instrument. The recommended PD level is subject to discussion between the purchaser and the manufacturer until more data are available. This paper provides PD data from on sacrificial coils and stator windings to support discussions on PD acceptance criteria. Three instruments were used for PD tests on resin-rich, press-cured sacrificial coils in laboratory conditions controlled to minimize noise, and in the typical environment of a factory acceptance test (FAT). These coils were then tested daily for a week, with identical measurement conditions of noise, time of day, operator and instrument. Also included are examples of offline PD data from two complete stator windings. The first was a new winding, tested in the factory on two different days. The second was an aged winding tested using the same methods but in different locations. Instrument selection, testing location, elapsed time between measurements and insulation system construction can influence offline PD measurements. The current evidence suggests that statistical distributions drawn from similar test conditions are more applicable than specific PD limits.