TY  - EJOU
AU  - Chen, Wei 
AU  - Wei, Zhi 
AU  - Pei, Tingting 
AU  - Zhu, Jianghao 
AU  - Wu, Yang 

TI  - A Fusion Optimization Method for Remaining Useful Life Prediction of Wind Turbine Gearboxes
T2  - Energy Engineering

PY  - 
VL  - 
IS  - 
SN  - 1546-0118

AB  - Wind turbine gearboxes are critical components in large-scale power generation systems, and their unexpected failures often result in significant economic losses, long downtime, and decreased energy efficiency. Accurate prediction of their Remaining Useful Life (RUL) is therefore vital for enhancing operational reliability, implementing condition-based maintenance, and optimizing lifecycle management. However, existing approaches often neglect the memory effect in degradation processes and fail to establish an effective interaction between stochastic degradation modeling and <i>RUL</i> prediction. To address these challenges, this study proposes a novel fusion method that integrates a stochastic degradation model with an intelligent prediction framework. The degradation model employs Fractional Brownian Motion (FBM) to capture long-range dependence and memory effects in gearbox performance, while the prediction framework leverages an enhanced recurrent neural network optimized through evolutionary mechanisms. By linking degradation modeling with <i>RUL</i> prediction through parameter optimization, the proposed method strengthens the interaction between physical degradation and data-driven prediction. Simulation results based on gearbox datasets demonstrate that the proposed approach significantly improves <i>RUL</i> prediction performance, achieving a 23.2% reduction in <mml:math id="mml-ieqn-1"><mml:mover><mml:mrow><mml:mi>R</mml:mi><mml:mi>M</mml:mi><mml:mi>S</mml:mi><mml:mi>E</mml:mi></mml:mrow><mml:mo accent="false">¯</mml:mo></mml:mover></mml:math>, a 26.7% improvement in <mml:math id="mml-ieqn-2"><mml:mover><mml:mrow><mml:mi>S</mml:mi><mml:mi>F</mml:mi></mml:mrow><mml:mo accent="false">¯</mml:mo></mml:mover></mml:math>, and a 3.3% increase in <mml:math id="mml-ieqn-3"><mml:msup><mml:mi>R</mml:mi><mml:mrow><mml:mn>2</mml:mn></mml:mrow></mml:msup></mml:math> compared with traditional RNN and LSTM models, highlighting its potential for practical deployment in wind farm operations to support proactive maintenance scheduling and enhance system reliability.
KW  - Wind turbine gearbox; remaining useful life prediction; stochastic degradation modeling; fractional Brownian motion; neural networks; condition-based maintenance

DO  - 10.32604/ee.2025.073843