[Retracted] Research on Pattern Recognition Method of Transformer Partial Discharge Based on Artificial Neural Network

Table 2

Detection results of different types of defects.

Pr (%)	Metal protrusion defects	Oil-stained air gap defects	Surface discharge defects	Suspension potential defects
Metal protrusion defects	94	4	2	0
Oil-stained air gap defects	6	90	3	1
Surface discharge defects	5	6	89	0
Suspension potential defects	0	0	0	100