Category: CBM/PHM
-
Digital Twin ๊ธฐ๋ฐ ์์ง์ ๋น ์์คํ ๊ตฌ์ถ: ์ผ์ ์ตํฉ๊ณผ ์ค์๊ฐ RUL ์์ธก ํ์ดํ๋ผ์ธ ์ค๊ณ
Digital Twin predictive maintenance: fuse 12 sensor streams into real-time RUL predictions. Architecture diagram and failure detection code.
-
Transfer Learning์ผ๋ก ์๋ ๋ฐ์ดํฐ ๊ณ ์ฅ ์ง๋จ ๊ตฌํ
Transfer Learning on 50 images: fine-tune ImageNet CNNs for 94% fault detection accuracy. Beats training from scratch with 1000+ samples.
-
GAN์ผ๋ก ์๋ ๊ณ ์ฅ ๋ฐ์ดํฐ ์ฆ๊ฐํ๊ธฐ: CWRU ๋ฒ ์ด๋ง ๋ฐ์ดํฐ์ ์ด์ ํ์ง ์ ํ๋ ๊ฐ์ ์คํ
GAN data augmentation boosted bearing fault detection accuracy from 72% to 89% on CWRU dataset. Here's the exact preprocessing pipeline.
-
LSTM vs Transformer RUL ์์ธก: NASA CMAPSS ์คํ ๋น๊ต
Transformer beat LSTM by 18% RMSE on NASA turbofan RUL prediction. Attention maps reveal why โ and when LSTM still wins.
-
Wavelet Transform vs STFT: ๋ฒ ์ด๋ง ๊ณ ์ฅ ์ง๋จ์ ์ํ ์ต์ ์ ์๊ฐ-์ฃผํ์ ๋ถ์ ๊ธฐ๋ฒ ๋น๊ต
Wavelet vs STFT for bearing fault detection: frequency resolution comparison on 10,000 vibration signals. Clear winner emerges.