Using a dynamic thermal model approach to estimate loss of life of mineral oil-immersed transformers

• Main Author: Sedehi Esfahani, Mohammad Tolou Askari
• Format: Thesis
• Language: English
• Published: 2010
• Online Access: http://psasir.upm.edu.my/id/eprint/26671/1/FK%202010%2089R.pdf; http://psasir.upm.edu.my/id/eprint/26671/

Transformers are among the most important equipment in electrical power systems, where their operations play the main roles in providing the reliable power systems. Because of the high cost of transformers and their permanent connections to the transmission and distribution systems, reliable measures are needed to extend the life span. Insulation life of the transformer and loading capabilities depend on several parameters, among which cooling mode, ambient temperature, oil viscosity are critical. Hot spot temperature is a major factor on the insulation life of the transformer. To predict the hot spot temperature and top oil temperature, many principal models have been proposed such as the classic thermal model, and thermal-electrical models. This research attempts to determine an accurate hot spot temperature, hence evaluate the loss of life of the transformer according to the numerical analysis method. Top oil temperature and hot spot temperature were determined by an accurate thermal model which takes into account the dynamic parameters such as ambient temperature, oil viscosity, winding losses, and loading profile. IEEE standards and thermal-electrical model, which are regarded as the two acceptable and reliable thermal models, have been used as underlying principles of this research. Both linear and non-linear regressions have been used for solving these thermal models. The Runge-Kutta numerical method is proposed in this study for solving the thermal model and for carrying out the work on different types of the transformer based on the data that have been collected from heat run tests, loading profile and electrical and physical characteristics of each transformer. An alternative solution for solving the thermal model is proposed in this study. Results are compared with the actual temperature and hot spot temperature in the IEEE standards. Since data has been collected from different types of electrical transformers, this thesis proposes a model which is based on the previous researches that are carried out on 187 MVA, 2.5 MVA, 30 MVA and 250 MVA transformers. The prediction of top oil temperature and hot spot temperature is very important for estimating the loss of life of the transformer in the system. Therefore, such an accurate technique is needed for solving the thermal models. The findings reveal that the uses of numerical methods are in good agreement with the measured values when comparing with the traditional methods. In addition, based on the theoretical definition and results obtained, the proposed fourth order Runge-Kutta has given reasonably good accuracy among other numerical methods such as the second order Runge-Kutta and Euler methods. This has led to an improvement in calculating the top oil temperature and hot spot temperature, as well as in predicting the loss of life of the transformer.