An approach to partial-discharge investigation by height-distribution analysis

Results of partial-discharge measurements, performed on insulation systems of rotating machines, are processed referring to the Weibull statistics. It is shown that the partial-discharge height distribution fit the two-parameter Weibull function when partial discharges having the same origin occur, whereas deviations from the Weibull plot are detected in the presence of various simultaneous partial-discharge phenomena. Probability functions based on the multiplicatively-mixed Weibull distribution used to fit data obtained in the presence of partial-discharge phenomena of two different types. The shape and the scale parameters of the Weibull functions derived as functions of test voltage and aging time prove to be characteristics of this kind of phenomenon. Pattern identification and clustering is thus achieved.