Prediction of a Transformer’s Loading and Ambient Temperature based on SARIMA Approach for Hot-Spot Temperature and Loss-Of-Life Analyses

Abstract

Hot-Spot Temperature (HST) is among the important parameters that can be used to evaluate the Loss-Of-Life (LOL) of transformers. HST can be determined through thermal modeling of which loading is one of the important parameter that needs to be obtained. This paper presents the prediction transformer's loading of a 132/33 kV, 60 MVA Oil Natural Air Natural (ONAN) transformer and ambient temperature based on Seasonal Autoregressive Integrated Moving Average (SARIMA). First, the computed loading profile was validated with the measured data. Next, the loading profile was forecasted for 1 year to evaluate the HST and LOL of the transformer. Differential model in IEC60076-7 was used to determine the HST based on the forecasted data. Based on Mean Absolute Percentage Error (MAPE), Mean Absolute Error (MAE) and Root Mean Square Error (RMSE), the best fit of SARIMA model are used to represent the transformer loading. This leads to the prediction of transformer HST that fluctuates along the 365 days and the LOL increases linearly with multiple fluctuations. It is also found that the prediction estimates the maximum HST is 66.93°C and the corresponding LOL based on predicted 1 yearly data is 666 minutes.