Optimizing Predictive Maintenance of Machines with Innovative DataAugmentation Strategies

This work explores the use of Wasserstein Generative Adversarial Networks (WGANs) to generate synthetic sensor data for Remaining Useful Life (RUL) prediction. The C-MAPSS FD001 dataset was first reduced using Principal Component Analysis (PCA) to retain key features. Synthetic data was created using WGAN and Conditional WGAN models and validated through statistical tests like the Kolmogorov-Smirnov test and Wasserstein distance. Machine learning models trained on the synthetic data showed performance close to those trained on real data. The results highlight that WGAN-based data generation can effectively support predictive maintenance by addressing data shortages.