Premature failure detection of distribution transformer with unbalanced harmonic loads using hotspot temperature analysis

Abstract

Harmonics distortion is the most prominent power quality problem in an electrical power distribution network that interrupts a good quality of electric power to be drawn in the network. Additionally, the major impact of harmonics distortion is the risk of the distribution transformer failure due to the elevation of power losses and hotspot temperature (HST) in its three-phase low voltage (LV) winding cables. The major challenge is to identify the exact location and point where the premature failure could occur on the three-phase cables. This research proposes an HST mathematical expression to detect a premature failure at the three-phase LV winding along with the transformer loading. As most transformer failure cases were rooted in heat losses that require meticulous analysis, the best accuracy numerical method as Finite Element Method (FEM) is selected to analyse the HST of the thermal distribution transformer model. The HST is simulated by considering the three-phase unbalanced harmonic loads from three different group levels of THDI and under five different insulation temperature classes system. The simulation outputs are then verified with HST results from the HST mathematical model based on IEEE C57.110-2018 standard. Further analysis of the simulation results has been done to propose the HST mathematical expression, which will be assessed on the three-phase LV winding cables to detect premature failure. At the end of this research, it is found that the individual harmonic currents from the 7th until 19th order are the prominent harmonic orders that had exceeded the limit of MS 1555 (IEC 61000-3-4) standard. Other than that, based on the proposed HST mathematical expression, it is found out that if the transformer is being loaded with loading over 0.9 pu, the premature failure is expected to occur promptly in the group of THDI peak-level, prominently at 180, 200 and 220 insulation temperature classes. As for the lifetime expectancy of the distribution transformer, if the transformer is loaded with loading at 0.9 pu and above, the lifetime is approximated to drop by the minimum at 14.5% and maximum at 56% from its expectancy lifetime. Plus, it is also concluded that the possibility of the lifetime reduction to be happened at the premature failure point at average of 93.5%, 85.4% and 78% of the THDI peaklevel, THDI average-level and THDI low-level correspondingly. Hence, the findings have successfully shown the proposed method's effectiveness in vividly viewing the distribution transformer's current condition. Upon the early detection of the premature failure on the three-phase cables, the execution of the proposed HST mathematical expression is also able to identify the exact location and point where the premature failure shall happen. Thus, it outright protects the distribution transformer from any unwanted breakdown, next preserves its best performance and lifetime expectancy.