Energy Estimation Method for Power-Frequency Arc in Transformer Oil Based on Gap Length and Pressure

Abstract

Transformer's internal arc fault is one of its most serious faults, which might cause fire accidents. These accidents have caused widespread concern among scholars and indicated the importance of explosion-proof performance calibration. However, there is a lack of methods to estimate arc energy accurately. Within the gap length range of 5–100 mm, the current range of 1–10 kA, and the pressure range of 1–15 atm, this paper conducts 640 power-frequency arc experiments, the most numerous and well-organized experiments reported so far. An average arc voltage calculation method based on energy equivalence is proposed to avoid existing method bias and better fit the actual energy curve. Factors affecting the average arc voltage are analyzed. The relationship between the voltage and gap length approximates a linear function, and the relationship with pressure approximates a power function. An energy estimation method involving gap length and pressure is proposed. The errors of existing and proposed arc energy estimation methods are compared and reported for the first time. The proposed method dramatically reduces the average error from 43.6% to 11.5%. Repeatability experiments show that the discharge dispersion causes an average energy error of 9.1%, so the error of the proposed method is satisfactory.