Experimental results demonstrate that the hybrid architecture consistently outperforms the standalone TCN and Reservoir components, as well as other benchmark methods, achieving substantially improved PHM scores while retaining competitive RMSE performance.
Abstract
Accurate prediction of Remaining Useful Life (RUL) is essential for predictive maintenance in the aerospace industry, where unexpected failures pose significant safety risks and increase operational costs. Conventional deep learning models, such as Long Short-Term Memory (LSTM) networks and Convolutional Neural Networks (CNNs), have demonstrated strong predictive capabilities; however, they often incur high computational costs, are sensitive to noise, and struggle to capture long-term degradation patterns. To overcome these issues, this study presents a new hybrid deep learning model that combines Temporal Convolutional Networks (TCNs) with Reservoir Computing, leveraging the strengths of both architectures. The model is tested using the well-known NASA C-MAPSS dataset, a standard benchmark for RUL estimation. Performance is measured using both Root Mean Squared Error (RMSE) and a penalty-based PHM score that emphasizes timely failure prediction. The model attains test RMSE values of 14.81, 16.26, 15.57, and 17.97 on FD001, FD002, FD003, and FD004, respectively. Correspondingly, the PHM scores are reduced to 57.1, 204.46, 190.95, and 446.73 across the same subsets. Experimental results demonstrate that the hybrid architecture consistently outperforms the standalone TCN and Reservoir components, as well as other benchmark methods, achieving substantially improved PHM scores while retaining competitive RMSE performance. These results suggest that the proposed method provides a practical, real-time solution for predictive maintenance in aeroengine health monitoring, thereby improving reliability and reducing maintenance costs.
The study demonstrates that combining CNN–LSTM with an appropriate optimization strategy improves the reliability and accuracy of RUL prediction for turbofan engines and shows that Stochastic Gradient Descent provides the best convergence behaviour and prediction accuracy for the proposed architecture.
Rajneesh Kumar, Shivam Ojha, Amit Shelke et al.· Scientific Reports· 0 citations
To improve the operational readiness of combat aircraft engines and reduce unplanned maintenance costs, accurately estimating the remaining useful life (RUL) is critical. Traditional maintenance often proves insufficient under dynamic mission profiles. In this study, a deep learning-based predictive maintenance model capable of autonomously extracting features from multivariate sensor data was developed. Using the NASA C-MAPSS FD001 and FD004 datasets, data were converted into sequential blocks via 50- and 30-step sliding windows, respectively. The model's architectural superiority in autonomously extracting temporal degradation features was validated against RF, CNN-LSTM, and BiLSTM baselines. On FD001, it achieved an R-squared (R2) of 0.8901, a 13.28 RMSE, and a 320.34 NASA risk score, demonstrating generalizability on the multi-regime FD004 dataset with a 15.71 RMSE. The proposed maintenance protocol achieved a 0.9973 AUC at the critical 30-cycle threshold, ensuring high reliability. Additionally, a decision-support simulator has been developed to validate this protocol under aggressive combat flight profiles.
Accurate prediction of Remaining Useful Life (RUL) is critical for predictive maintenance and minimizing downtime in industrial systems. This paper presents a cross-domain deep learning framework based on a hybrid Convolutional Neural Network–Bidirectional Long Short-Term Memory (CNN–BiLSTM) architecture. Unlike domain-specific models that require handcrafted features, the proposed framework extracts local degradation features through CNN layers and captures long-term dependencies via BiLSTM networks. The model is evaluated on three heterogeneous datasets: construction machinery, continuous casting machines, and lithium-ion batteries. Experimental results show that CNN–BiLSTM consistently outperforms baselines, achieving up to 22% lower RMSE compared to GRU and 30–50% lower RMSE compared to traditional models. On the construction dataset, it achieves an MAE of 48.2 hours and RMSE of 67.1 hours (R2 = 0.88), outperforming GRU by 20%. For the casting dataset, the model attains an MAE of 87.6 tons and RMSE of 113.9 tons (R2 = 0.87), surpassing Random Forest by over 35%. On the battery dataset, CNN–BiLSTM reduces the MAE to 49.6 cycles and RMSE to 72.8 cycles (R2 = 0.89), while also achieving the lowest Timeliness Score (27.5) and PHM08 Score (192.4). Cross-domain experiments are evaluated under two settings: zero-shot transfer, where the model is trained on one source domain and directly tested on a different target domain without using labeled target-domain samples, and fine-tuned transfer, where 20% of labeled target-domain samples are used to update only the fully connected layers while keeping the CNN and BiLSTM layers frozen. The zero-shot results reflect the effect of domain shift, while the fine-tuned results show that lightweight transfer adaptation reduces RMSE by 25–40% across domains. These findings indicate cross-domain adaptability under limited target-domain supervision rather than fully unsupervised cross-domain generalization. These results highlight the feasibility of a unified CNN–BiLSTM framework for scalable, cross-domain RUL estimation and its suitability for real-world prognostic applications.
Sudip Saha, Muhammad Arslan Pervaiz, Muhammad Safwat Rahman et al.· PLoS ONE· 0 citations
A two-level stacking ensemble framework for Remaining Useful Life prediction of turbofan engines, evaluated on the NASA C-MAPSS benchmark using the FD001 and FD003 subsets, demonstrates the efficacy of stacking ensemble methods for prognostics and health management in safety-critical aerospace applications.
Limon Bin Hossain, Md. Salehin Seyam, Md. Rashedul Islam et al.· 0 citations
Unplanned equipment failure remains one of the costliest problems in modern manufacturing, driving research toward Predictive Maintenance (PdM) strategies that anticipate failures before they occur using real-time sensor data. This paper proposes a hybrid deep learning framework combining Convolutional Neural Networks (CNN) for local spatial feature extraction and Bidirectional Long Short-Term Memory (BiLSTM) networks with a temporal attention mechanism for capturing long-range degradation trends in multivariate time-series sensor data. The framework is designed to estimate Remaining Useful Life (RUL) and detect early fault signatures in Industry 4.0
environments equipped with IoT-based condition-monitoring sensors. We formulate the problem as a windowed sequence-to-value regression task, describe a sliding-window feature pipeline, and benchmark the framework's data
pipeline against classical machine learning regressors (Random Forest, Gradient Boosting, Support Vector
Regression, and a Multi-Layer Perceptron) on a run-to-failure sensor dataset. Experimental results show that
ensemble tree-based models achieve strong baseline performance (RMSE as low as 9.96 cycles, R2 = 0.946), establishing a validated pipeline and baseline for the full CNN-BiLSTM-Attention model. The proposed architecture, evaluation protocol, and ablation plan are presented in full so the framework can be reproduced and extended on industrial-scale datasets such as NASA C-MAPSS or live plant telemetry.
Renuka Surendra Deshpande· International Journal of All...· 0 citations
Accurate remaining useful life (RUL) prediction is essential for condition-based maintenance and safe aero-engine operation. To address non-stationary monitoring data, limited informative degradation samples, and the difficulty of jointly modeling local degradation patterns and temporal dependencies, this study proposes a degradation-aware dynamic masking augmentation method combined with a multiscale CNN–Transformer network. The 21 sensor variables in the NASA C-MAPSS dataset are first grouped by physical meaning and reconstructed into four-dimensional state features. A degradation-state score integrating local variance and trend slope is then used to adapt temporal masking probabilities across degradation stages, while feature masking probabilities are assigned according to feature importance. Masked positions are filled with adjacent unmasked observations, and invalid augmented samples are removed through trend consistency verification. A multiscale CNN with channel attention extracts local degradation features, and a Transformer encoder captures temporal dependencies. Bayesian optimization is used to determine key hyperparameters. On FD001, the proposed method achieves an MSE of 348.3970, an MAE of 8.3908, and an R2 of 0.9227, reducing MSE and MAE by 4.27% and 28.42%, respectively, compared with ML-RFR. It also achieves the highest R2 of 0.9369 on FD003. Cross-dataset and multi-seed experiments further confirm its applicability and stability.
Xudong Song, Guohua Wu, Meng-Dan Wang et al.· Machines· 0 citations
