đ
31 March 2025
Prediksi Remaining Useful Life Mesin Turbofan Menggunakan Metode Deep Layer Recurrent Neural Networks
Jurnal Informatika dan Rekayasa Perangkat Lunak
Universitas Wahid Hasyim
đ Abstract
In recent years, the development of prognostic and health management in engines and systems has grown rapidly, with one form of interaction being predicting the remaining useful life of aircraft turbofan engines. In this study, chose to use the Deep Learning RNN method with LSTM architecture. And in its limitation, the C-MAPSS FD001 dataset is used as a modeling dataset in which there is complex multivariate data. The main objective in this analysis and modeling process is to obtain the remaining useful life prediction results that can be a benchmark in analyzing the Remaining Useful Life (RUL) of the turbofan engine. The test results involved many criteria and parameters that were tested. Analysis of the results shows that the model with learning rate criteria of 0.001, number of epochs of 50, hidden unit of 12, Min-max normalization method, and Adam optimizer tested can understand quite accurately the RUL data pattern and follow the movement of the predetermined RUL target. Although there is a decrease and a slight spike in the unit level of machines 81 to 84, with an error value in RMSE of 26.48, but overall the model can be considered optimal in learning the prediction pattern of the RUL target. This research shows that the RNN-LSTM model has great potential in aircraft turbofan engine prognostic applications.
đ Keywords
#Deep learning; long short term memory; mesin turbofan; prognostik; recurrent neural networks; sisa berguna umur
âšī¸ Informasi Publikasi
Tanggal Publikasi
31 March 2025
Volume / Nomor / Tahun
Volume 7, Nomor 1, Tahun 2025
đ HOW TO CITE
Risky, Kevin Maulana; Insani, Fitri; Jasril, Jasril; Sanjaya, Suwanto; Handayani, Lestari, "Prediksi Remaining Useful Life Mesin Turbofan Menggunakan Metode Deep Layer Recurrent Neural Networks," Jurnal Informatika dan Rekayasa Perangkat Lunak, vol. 7, no. 1, Mar. 2025.
ACM
ACS
APA
ABNT
Chicago
Harvard
IEEE
MLA
Turabian
Vancouver
Download Citation
0
Sitasi
Cek Google Scholar