hydroelectric plant signals

The file 'hidro con falla' contains 18,644 records regarding the behavior of 13 variables pertaining to the hydroelectric generation process at the 'Central Hidroeléctrica Amaime' plant, located in Palmira, Valle, Colombia. This file includes the following variables:TMRCLA Radial bearing metal temperature at the driven endTMA1LA Axial bearing metal temperature 1 driven endTMA2LA Axial bearing metal temperature 2 driven endTACCLA Temperature in oil sump driven endVEACLA Axial vibration at the driven-end bearingVERCLA Radial vibration at the driven-end bearingTMRCLOA Radial bearing metal temperature at the non-driven endTACCLOA Bearing oil sump temperature at the non-driven endVER Radial shaft vibrationTTA Oil tank temperaturePAA Oil accumulator pressureTAaT Oil outlet temperatureTAaC Oil supply temperature The values of the displayed variables can assume three distinct states (OK, Alarm, Trip), depending on the setting values. Consequently, an algorithm was applied to the 'hidro con falla' file to classify the signal states, using the signal settings as a reference, to generate the 'train' file. In this file, the signals were assigned qualitative values, and specific failure cases were defined using the KDD procedure.The 'modelo bayesiano' file was used to obtain the prior and conditional probability values of the analyzed variables. These probabilities were used to train the Bayesian model. Subsequently, the model was tested using data from the 'Test' file, which consists of 30% of the total data obtained from the failure cases in the original 'hidro con falla' file.