Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE credits
The Risk Index is a useful tool which allows quick conclusions regarding the possibility of something going wrong with a specific device or situation. In this work, the devices are high power transformers belonging to a specific company.
This company is EDP Renováveis (EDPR) which develops, manages and maintains wind farms with a total installed capacity of 9,036 MW (at the end of 2014), in ten countries located in Europe, North America and Brazil. Each of these wind farms has a substation with, at least, one high power transformer.
High power transformers are expensive devices and are crucial to the proper operation of the wind farm. As so, they present great importance to EDPR. In order to maintain them and make sure, as far as possible, that a break-down does not happen, the company performs several tests a year on these pieces of equipment, taking note of any symptom that might cause alarm. These tests use different methods, and are performed by various entities with diverse time schedules.
The amount of information to be taken into account for the Risk Index calculation is vast. From the location to the features of the transformer, going through the above mentioned tests, a sizable amount of data must be collected and processed. The Risk Index allows the ranking of the various conditions according to their real seriousness. For instance, if two transformers are in bad state, it should be able to determine which of them brings more negative consequences.
The Risk Index is obtained by predicting the possible scenarios resulting from the symptoms that the transformer is displaying. The consequences resulting from each symptom have to be determined with the information that was gathered in solid previous research work, and using the knowledge available from experts in the area. The consequences have also to be estimated, using formulas based on real cases that have occurred, and taking into account all the influencing parameters of that specific equipment. Testing the algorithms in several transformers with different problems and environments (location, rated power, manufacturer, etc.) and comparing the results among them is the ultimate method to improve the reliability of these formulas.
In order to complement the idea of the Risk Index, a Fail Index is also developed in this work. It simply pretends to illustrate the likelihood of failure of a given transformer. This index does not allow comparisons among different information, but it might be useful for a more down-to-earth analysis. It basically splits the transformer in its components. The state of each component is then decomposed in percentages, which are attributed to the results of the tests that contribute to evaluate the given component.
Coming up with a risk index is not useful just for itself: it must be followed by a structured program that can efficiently process a great deal of information, and display the results in an intuitive way for its users.