FUSION STRATEGY FOR SINGLE PHASE TO GROUND FAULT DETECTION IMPLEMENTED THROUGH FAULT MEASURES AND EVIDENCE THEORY

In order to fit the complex single phase to ground fault in neutral insulated or Peterson coil earthed power networks, this paper proposes an integrated fault detecting method with fusion of multi fault information. For flexible and reasonable fusion purpose, concept of fault measures is introduced to quantitatively evaluate the degree of each line tending to be the fault line. Fault measure algorithms for transient are established by means of wavelet analysis. Fault measure algorithms for magnitude and phase angle of steady quantities are formulated as well. D-S evidence theory is utilized to make the fusion of multi- fault information. In this application, the product of relative basic probability assignment and a confirmable belief coefficient practicably formulate the basic probability assignment function. Multi-information fusion by means of fault measures and D-S evidence theory is capable of focalizing fault detection, and therefore improves the detecting performance greatly. Field results show that the fusion strategy exhibits excellent property over the existing criteria.