Remaining Useful Life Prediction on Wind Turbine Gearbox

Abstract

This research proposes a methodology to estimate the reliability of gearbox using life data analysis and predict the Lifetime Use Estimation (LUE). Life data analysis involves collection of historical field replacements of gearbox and perform statistical analysis such as Weibull analysis to estimate the reliability. Remaining useful life is estimated by using Cumulative damage model and data-driven methods. The first approach is based on the physics of failure models of degradation and the second approach is based on the operational, environmental & loads data provided by the design team which is translated into a mathematical model that represent the behavior of the degradation. Data-driven method is used in this research, where the different performance data from components are exploited to model the degradation's behavior. LUE is used to make key business decisions such as planning of spares, service cost and increase availability of wind turbine. Gearbox is the heart of the wind turbine and it is made up of several stages of helical/planetary gears. Performance data is acquired separately for each of these stages and LUE is calculated individually. The individual LUE is then rolled up to estimate the overall Lifetime Use Estimation of gearbox. This will identify the weak link which is going to fail first and the failure mode which is driving the primary failure can be identified. Finally, corrective measures can be planned accordingly. The cumulated damage and LUE are estimated by using Inverse power law damage model along with Miner’s rule.