Remaining Useful Life Prediction Method For Different Types Of Rolling Bearings Based On Bi-Lstm Quantification
The remaining Useful Life (RUL) forecast for rolling bearings is still the crucial part of condition-based maintenance (CBM) for mechanical systems. To predict the RUL, the existing research utilized traditional Deep Learning techniques, however, it has trouble quantifying uncertainty. Therefore, this research suggested a novel DL model to improve RUL prediction. Initially, to define the degree of rolling bearing deterioration and comprehend the non-linear qualities, time domain features, frequency domain features, & time-frequency domain features are removed. Then, this study suggested using a Bi-LSTM - RF framework to predict the RUL, this framework has an LSTM layer in a combination of forward and backward motion, a fully connected layer, an RF classifier, & a dropout layer. As a result, our proposed deep learning-based RUL prediction obtains the Accuracy of 0.9845, Precision of 0.93, Recall of 1.0, & F1-score of 0.9656
Сведения
- Опубликовано
- 2024 · № 3, С. 96
- Язык
- Английский
- ISSN
- 2706-977X
- Издатель
- Karaganda Technical University named after Abylkas Saginov
- Аффилиации
- Koneru Lakshmaiah Education Foundation