Data Mining Applied to the Instrumentation Data Analysis of a Large Dam

Itaipu dam is one of the biggest dams in the world. There are more than 2200 instruments installed, periodically monitored since 1986. Due to the large amount of collected and stored data, as well as to the possibility of useful but hidden information, it is necessary to use techniques that may allow us to transform data into knowledge. Itaipu company just finished to apply the automatic acquisition system in approximately 210 instruments. With this work we intend to confirm the choice of the instruments that shall be automated focusing on one type of them, namely extensometers, located at the dam's section F. This task was performed by applying mathematical models based on data mining techniques capable of analyzing and selecting the most relevant data. Among the techniques, we applied multivariate data analysis: principal components, factor and clustering analysis, in a variety of forms. The results were satisfactory as we show.