Electric Switch Machine Failure Detection Using Data-Mining Technique

Abstract

In order to detect equipment failures from measured values using equipment condition systems, it is necessary to establish an appropriate threshold. However, even when a threshold has been established, many cases have arisen where the equipment's installation environment was not given sufficient consideration and this has been one of the causes of erroneous detection. In this research, we examined a technique for failure detection without establishing a threshold. In addition, attempts were made to detect failures by statistical analysis, using one of the data-mining techniques, making use of switch data from an electric switch machine monitor.