Developing Practical Methods for Ageing and Failure Probability Modeling of Mineral Oil Immersed Power Transformers Using Smart Utility Data
Abstract
Ageing of assets in the existing power networks infrastructure has turned into a big concern for the electric utilities. Electrical transmission systems built during the twenty years spanning the 1960 to 1980 period form the backbone and are still in service. The situation of such power networks being very different from the overdesigned networks of the former decades, the question arises, "should the asset owners break the precedents in lifecycle and replacement strategies"? The conventional time based methods used in developing failure probability evaluation for power transformers, a key factor considered for replacement planning, are challenged in this research. Methods are developed to factor in operational history of transformers in failure probability modeling. In the first method, health index, a proprietary number that represents transformer overall condition, is used to adjust failure probability. This is a novel application introduced for health index. In the second method, thermal modeling is used to assess the cellulose insulation condition and used for the first time as a basis for probability modeling. In the third method, a new approach is developed with Bayesian Networks to provide a more accurate estimation of cellulose insulation condition by considering both thermal model results and dissolved oil analysis. The third method is more complicated and requires more input data compared to the other methods. However, it is expected to provide more accurate value for failure probability. Data for actual transformers is collected from utilities and processed to validate the developed methods. Results indicate that the methods developed are more accurate than the conventional time based methods.
Description
Keywords
Engineering--Electronics and Electrical
Citation
Abbasi, E. (2017). Developing Practical Methods for Ageing and Failure Probability Modeling of Mineral Oil Immersed Power Transformers Using Smart Utility Data (Doctoral thesis, University of Calgary, Calgary, Canada). Retrieved from https://prism.ucalgary.ca. doi:10.11575/PRISM/25628