Cost-Benefit Analysis of Maintenance Measures for Power Transformers

The work described in this thesis was conducted in cooperation with SINTEF Energy Research and their collaborating partners in the research project Trafotiltak . The objective of the project is to develop a decision-making tool, based on both economic and technical data, to support asset management maintenance decisions on Norwegian power transformers. Deciding when to perform reinvestments and finding the right type of maintenance action that creates the most benefit is challenging. Therefore, this thesis explores the different maintenance measures available for improving the condition of mineral oil within the transformer. To analyse the performance of different measures, service data from power transformers owned by Norwegian companies, collaborating in the project, were collected. The condition of each transformer was calculated by using a health index score, shorted HI-score. This score allows the asset manager to quickly compare power transformers against each other and find the transformers with the greatest need of improvements. This data acquisition of the condition improvements from previous cases helps create a statistical overview on benefits gained from different measures, and could be used as the input in a cost-benefit analysis between different maintenance measures. The thesis has determined the estimated improvements for a standard reclamation of oil through the data acquisition, where also some of the other measures described in this thesis are included in the process. Due to lack of data from stand-alone measures e.g. recondition, oil change and drying of oil/paper it was not possible to make a corresponding statistical summary of expected improvements for other measures than reclamation of oil. To test the performance of the proposed Trafotiltak model and especially the cost-benefit model included in Trafotiltak , it was applied to two real power transformers in the need of condition improvements. By doing this it could calculate when maintenance was most beneficial and if reclamation of oil was a better alternative than to reinvest in a new transformer. The estimated improvements from reclamation found in this thesis are used as the benefit for the reclamation process in the model. The results from the case studies show that the model could give reasonable suggestions on which measure to choose and when measures should be performed. Even though the cost estimates and HI-scores that have been used are a little rough, they are still reliable and an indication that the model could be used in the industry to examine different maintenance alternatives