基于电流 - 频域 - 振动参量的变压器绕组机械状态诊断方法

In order to diagnose abnormal looseness and serious deformation of transformer winding, the diagnosis method of transformer winding mechanical state based on current-frequency-vibration parameter was presented. Firstly, a mathematical model of transformer winding vibration was established to study the influence mechanism of winding vibration. The variation law of vibration signal was analyzed when the mechanical state of transformer winding is abnormal. Then, the diagnosis strategy was proposed: the winding abnormal state diagnosis method based on the mapping relationship of load current-frequency domain distribution-vibration amplitude was proposed, to identify whether the winding is in abnormal state. And then for the abnormal state winding, a fault diagnosis method for loose state and deformation of transformer winding based on complementary ensemble empirical mode decomposition (CEEMD) -energy entropy transformation was proposed, to diagnose the winding is in loose state or deformation fault. Finally, aiming at a 35 kV / 500 kVA transformer prototype, the experiment platform of normal winding, loose winding and seriously deformed fault winding was built, and the diagnosis experiment on mechanical state of transformer winding was carried out. The correlation distance between 3D surfaces of normal and abnormal windings is 0. 149 8, the energy entropy of loose and deformation windings are 0. 426 8, 0. 258 6 and 0. 146 8. The results indicate that the presented method can accurately diagnose the abnormal winding state, and distinguish loose and deformation faults.