Modeling the Temperature of the Distribution Transformer Oil Using Transformer Body Temperature and Power Quality Parameters Based on Artificial Neural Network

Existing distribution transformers that supply electricity to customers are generally not equipped with temperature sensors. This causes these transformers to be prone to damages due to an extreme increase in their temperature. Such condition could be ameliorated by placing temperature sensor inside each transformer which is unfortunately not physically realizable. Hence, an alternate method is needed to perform the temperature monitoring function. This paper presents a proposed temperature sensor using an electronic device that models the internal temperature of a transformer without installing temperature sensor inside the transformer. The electronic device performs the temperature sensing function by modeling the temperature using Artificial Neural Network (ANN) in conjunction with measured power quality parameters on the secondary side of the transformer and the temperature on the outside of the transformer as measured by infrared sensors. Hardware implementation using the ARM STM32F4 embedded controller demonstrates the ability of the proposed device in producing accurate Intelligent Electronic Device (IED) which can further be developed to provide early online notification when a disturbance occurs.