Transformer winding defects identification based on a high frequency method

Abstract

The transformer diagnostic methods are systematically being improved and extended due to growing requirements for reliability of power systems in terms of uninterrupted power supply and avoidance of blackouts. Those methods are also driven by longer lifetime of transformers and demand for reduction of transmission and distribution costs. Hence, the detection of winding faults in transformers, both in exploitation or during transportation is an important aspect of power transformer failure prevention. The frequency response analysis method (FRA), more and more frequently used in electric power engineering, has been applied for investigations and signature analysis based on the admittance and transfer function. The paper presents a novel approach to the identification of typical transformer winding problems such as axial or radial movements or turn-to-turn faults. The proposed transfer function discrimination (TFD) criteria are based on the derived transfer function ratios, manifesting higher sensitivity.