Predição de defeito por vibração em mancal de turbina hidrelétrica empregando vida útil remanescente

Abstract

The objective of this work is to forecast the useful life of the guide bearing of the bulb type turbine installed at the Jirau Hydroelectric Power Plant on the Madeira River in Rondônia to support the plant operator's assertive decision-making. The estimated useful life is calculated using the prediction called the remaining useful life. From the vibration measurements of a turbine bearing in the context of the Research Project Experimental Development of Digital Twin of the Jirau Hydroelectric Power Plant in Full Scope with Distributed Complex Event Processing Technology for Systemic Investigations (PD-06631-0007/2018). The remaining useful life of the turbine is calculated at each turbine rotation cycle (generating unit) to perform predictive maintenance. The forecast modeled in this work can be easily adapted to other turbine models. The Root Mean Square, (RMS) is used as the threshold of the monitoring signal. Spectral kurtosis analysis is used to make the decision for prominent turbine bearing failure. The result obtained for the remaining service life is compared with the actual results for a bearing of a wind generator unit with a long history of measuring vibration to failure. The results obtained in this work used a specific database with little data for analysis. Even with little data, the results show that the bearing appears to be in good health. With the model running on a larger database the results tend to have a better performance