Performance of Depressurization Devices to Dynamic Loads Generated by Arcing in Liquid Filled Transformers

Abstract

This study provides a methodology that can be used to evaluate the dynamic performance of fast depressurization devices used in liquid-filled oil transformers. Liquid-filled transformers are susceptible to explosions due to internal arcing if the dielectric insulation fails. The internal arc vaporizes a portion of the liquid and generates a sudden pressure wave. The first peak of the pressure wave has been measured to be as high as 13 bars, with time durations on the order of milliseconds [1]. Transformer tanks have a typical static withstand limit of approximately 1 bar gauge [2]. It is thus imperative that the tank be depressurized before the static pressure reaches such a threshold. One industry-accepted Fast Depressurization System [3] used to depressurize transformers after an internal arc is based on a patented rupture disk design [4]. This study compares the dynamic performance of this disk to results from a successful test campaign using a rupture disk as the depressurization device. Limiting loading rate values from the test campaign are then used to comment on the effectiveness of the design. The evaluation methodology is based on Pressure-Impulse (P-I) curves. The P-I curve was generated by running a series of Implicit Dynamic analysis using Code_Aster [5]. This iterative process first required establishing a failure mode that is consistent with actual observed failure in the field and observable in the Finite Element Analysis (FEA) model. The criteria were then used in interpreting the response of the Rupture Disk to a series of different half-sine wave pulse loading of varying amplitudes and time-periods. The generated P-I curve was then compared to loading rates observed in the test campaign [1] as well as three other higher loading rates (1.28 times, 2 times, 3.8 times, and 10.25 times the reported experimental rate) to qualitatively assess the effectiveness of the design. Results indicated that disk functions extremely effectively as a Fast Depressurization System as also corroborated by the test campaign. Although this methodology is used for the rupture disk, it is expected that this methodology can be extended to compare the dynamic performance of other depressurization devices.