The aim of this paper is to review and compare diverse predictive analysis methods used for the inspection of the internal conditions of wind generator gear box bearings on wind turbines in order to determine the accuracy, deficiencies, and validity of these methods.
Thus, three different types of predictive analysis will be compared: visual analysis using boroscopy (that is an industrial type of endoscopy), oil analysis, and bearing condition unit vibrations analysis.
These analyses were carried out on a ten-year-old wind farm that has forty-eight 800 kW wind turbines; the results will allow other similar wind farms to determine the most appropriate predictive strategies.
In the studied gear boxes, damage is restricted to the bearings of the satellites. Therefore, the study has focused on this part of the gear box.
The study demonstrates that bearing condition unit vibration analysis can predict severe damage in all cases, so it is possible to predict bearing failure within 6 months; as a result, it is possible to establish the optimal moment to substitute bearings to avoid catastrophic failure in gear boxes. In addition, given that a boroscopy can detect all types of damage in the bearings of the satellite and thus can predict failure ahead of 6 months as well as detect low and moderate damage, it becomes the method of preference.
Keywords: wind, turbine, gear box, bearing, predictive analysis
PCN264 ESTIMATING THE BURDEN OF SALIVARY GLAND CANCER IN NORDIC COUNTRIES: A PREDICTIVE ANALYSIS USING TIME SERIES ANALYSIS METHODS
