Real time Sound Level-Based Load Estimation in Distribution Transformer

Abstract

The technique in-used to monitor loading limits of a transformer in distribution network could impact its performance and subsequently, maintenance and service life. This research focus on optimization of performance and service life of this device when connected in distribution network. The thermal-effect technique in-used has large delay response time on loading limits monitoring. This considerably affect measurement accuracy, reliability and performance. The contributing factors which includes ambient temperature, uneven heat distribution and slow rate of thermal conductivity in transformer’s metal casing. Thus, the method is inefficient for time critical monitoring of loading limits. In this paper, we proposed detecting loading limits by monitoring sound level produced from the transformer’s coil. Comparatively, the proposed technique further substantiates the theory; sound propagate faster in metal than heat conductivity in the same metal. An offline data-logger device was design and developed; it was used to collect experiment data from 500KVA(33KV/415V) and 7.5MVA(33KV/11KV) transformers in distribution network. The data collected were analyzed and respective results revealed 24-hours loading patterns and loading peaks. Steep raising and falling of the graph of the load waveforms indicates fast response to loading. Therefore, adopting sound-level-based loading limits monitoring will have recommended so as to improve the device performance and service life.