A Method for Diagnosing Lower Transformer Faults Based on Current Signal Trajectories and Its Implementation

Abstract

In practical power systems, transformers are continuously involved in voltage conversion and electrical-energy transmission, so once abnormal operating conditions appear inside the equipment, the influence may gradually spread to the stability of the entire grid. Because early-stage transformer faults are usually weak and not easy to observe directly from conventional electrical quantities, this paper investigates a fault-monitoring method based on current-signal trajectories and discusses its application to transformer operating-state analysis. In the proposed approach, the primary-side and secondary-side current signals are taken as the basic data source, and the corresponding Lissajous trajectories are then constructed under different operating conditions. Instead of analyzing only waveform amplitude or phase separately, the method converts the current relationship into geometric trajectory characteristics, making the variation process easier to observe visually. To examine whether the method remains effective under different fault levels, a transformer equivalent-circuit model was built in Multisim, and several abnormal operating conditions were reproduced by changing equivalent impedance parameters and related electrical quantities. During the simulation analysis, several ellipse-related quantities were extracted from the generated trajectories, mainly including ellipse area together with the ratio between the major axis and minor axis. These geometric quantities were then compared under healthy and faulty operating conditions. The obtained results show that even under relatively weak parameter disturbances, such as ±3% variations, the trajectory shape had already changed slightly compared with the normal state, and the corresponding feature deviation approached about 5% in some simulation cases. After the disturbance level continued to increase, especially when the parameter variation reached around ±30%, the deformation of the ellipse became much more obvious, while part of the geometric feature variation exceeded 32%. From the overall simulation results, it can be seen that the trajectory characteristics still maintain visible differences under different transformer operating conditions, including relatively weak abnormal states appearing at an early stage. Compared with directly using conventional electrical signal quantities for analysis, the proposed method represents operating-state variation through geometric trajectories, so some small parameter changes can be reflected more intuitively. The study therefore suggests that current-signal trajectory analysis may provide another possible way for transformer online monitoring and operating-condition evaluation in practical engineering applications.