Chunyan Hu

Aero-engine safety and prognostics and health management (PHM) rely on robust vibration-based fault diagnosis. However, many deep learning studies on rotating machinery are evaluated under random train–test splits that mix hardware instances and may obscure the domain shift faced in deployment. This paper presents a protocol-driven end-to-end baseline for multi-sensor aero-engine-relevant vibration diagnosis on the HIT inter-shaft bearing benchmark. Six synchronous vibration channels are segmented into fixed-length windows, standardized using source-domain statistics, and classified by a compact 1D CNN backbone with and without squeeze-and-excitation (SE) channel attention. A deeper ResNet1D baseline is further introduced to examine whether increasing backbone capacity improves cross-bearing generalization under the same source-only training protocol. We compare random segment-level splits with bearing-level cross-splits that hold out entire bearings as unseen target domains, and we report deployment-oriented indicators including balanced accuracy, false-alarm rate (FAR), and miss rate over five random seeds. Under random splits, the compact CNN baseline reaches near-ceiling test accuracy, confirming that the benchmark is readily separable under in-domain interpolation. In contrast, cross-bearing evaluation reveals severe degradation: in the representative split, the baseline CNN accuracy collapses to approximately 15% with near-zero normal-class recall, while ResNet1D improves fault sensitivity but still retains a high FAR above 88%. Additional cross-bearing permutations further show that this degradation is not attributable to a single unfavorable source–target split. These findings indicate that, under the tested source-only backbones and protocols, distribution mismatch is a dominant bottleneck for deployment-ready cross-bearing diagnosis. The results establish a reproducible baseline for protocol-driven evaluation in aero-engine PHM and motivate future work on domain adaptation, domain generalization, calibration, and sequential decision logic.

Online monitoring of rub-impact faults is essential for aero-engine safety, yet existing methods struggle with asynchronous multi-modal data, diverse sensor configurations, and the scarcity of reliable fault annotations. To address these challenges, this paper proposes RubNet, a domain generalization framework that first employs a physics-guided temporal alignment strategy to precisely align high-frequency vibration signals with low-frequency state parameters based on spectral centroid correlations. A multi-scale channel attention convolutional neural network is subsequently utilized to extract robust features, leveraging adversarial domain generalization to learn representations that remain invariant to shifts in sensor position and orientation. Furthermore, to mitigate label noise during temporal fine-tuning, the framework integrates an uncertainty-aware soft label aggregation mechanism that fuses conflicting annotations from human experts, vision language models, and pre-trained model predictions using the Dawid-Skene algorithm. Experimental validation on real-world ground test data demonstrates that RubNet achieves state-of-the-art performance with an F1-score of 93.26% and superior generalization capabilities across unseen sensor domains, while maintaining low computational complexity suitable for real-time deployment.

Aero engine surge diagnosis is a key technology in engine health management, and its diagnostic accuracy is of great significance for ensuring operational safety. Traditional threshold-based diagnostic methods are significantly affected by working conditions, which makes it difficult to achieve full working condition coverage. Moreover, due to issues such as varying feature thresholds across conditions, weak signal characteristics, and low identifiability, the diagnostic accuracy remains limited. To address these challenges, this paper proposes an STFF-CANet (Spatio-Temporal Feature Fusion Cross-Attentional Network) diagnosis model of aero engine surge based on spatio-temporal feature fusion. The model first employs a Convolutional Neural Network (CNN) to extract spatial features from the frequency domain of dynamic signals via Fast Fourier Transform (FFT). Simultaneously, a Bidirectional Long Short-Term Memory (BiLSTM) network is used to capture temporal features from signals optimized by Variational Mode Decomposition (VMD). A cross-attention mechanism is further introduced to achieve deep fusion of spatiotemporal features, thereby enhancing the capability to identify weak fault characteristics. In addition, the sliding window slice method is used to expand the sample size for the small sample fault data of the engine surge of an aero engine. This ensures both informational continuity between slices and statistical stability of features, effectively mitigating the difficulty of diagnosing early and weak surge characteristics under small-sample conditions. Experimental results demonstrate that the model achieves an F1-score, Recall, Precision, and Accuracy of 97.96%, 97.52%, 98.43%, and 99.01%, respectively, in surge fault classification. These outcomes meet the practical requirements for aero engine surge diagnosis and provide an effective solution for early fault warning in complex industrial equipment.

Safe transient operation is essential for turbofan engines, yet conventional min–max schedules are often conservative and require extensive manual tuning. This paper proposes a data-driven, safety-aware reinforcement learning framework for acceleration and deceleration scheduling. A multi-condition long short-term memory (LSTM) surrogate is trained using ground-test measurements augmented with trajectories generated by a calibrated high-fidelity component-level engine model. Similarity-consistent corrected variables are adopted to represent spool-speed dynamics across the flight envelope, and a physics-informed surrogate is included as a baseline. On the surrogate environment, a safety-aware Soft Actor–Critic (SAC) agent is trained offline to generate spool-speed reference trajectories while enforcing limits on surge margin, turbine inlet temperature, and spool overspeed. The learned scheduler is deployed at the feedforward scheduling layer and provides constraint-aware speed commands to a conventional inner-loop controller and protection logic, preserving a certifiable feedback structure. Proximal Policy Optimization (PPO) and Twin Delayed DDPG (TD3) are implemented as scalar-action baselines, whereas the SAC scheduler uses a progress-based two-dimensional action to jointly adjust the advancement along a nominal speed path and the fuel-flow profile. Simulations on the component-level model over multiple maneuvers and representative (H, Ma) points, with parameter perturbations and sensor noise, showthat the proposed scheduler reduces the mean rise and settling times of the high-pressure spool by about 31% versus a min–max schedule and achieves a further 3–4% speedup over PPO-, TD3-, and PINN-based baselines while satisfying all safety constraints.

For highly maneuverable aircraft, the afterburning engine serves as a core and critical component. Due to the complex structure of the afterburner and the strong coupling among parameters, mechanism-based modeling of afterburning engines remains extremely challenging. To address this problem, this paper proposes a data-driven hybrid algorithm modeling framework for a light-duty afterburning turbojet engine. Using test-bench data from the TWP220L light-duty afterburning turbojet, two hybrid algorithm models were developed: (i) PSO-DNN and (ii) NGO-LSSVM. Four models, DNN, PSO-DNN, LSSVM, and NGO-LSSVM, were compared by mapping engine input parameters (altitude, Mach number, rotor speed, and fuel flow rate) to two key performance outputs (thrust and turbine pressure ratio). Based on visual error analysis and regression evaluation metrics, it was found that the optimized algorithm significantly reduced the prediction error. The NGO-LSSVM model achieved the highest accuracy in both performance indicators, increasing R² by 5.3% for thrust, and increasing R² by 6.8% for turbine pressure ratio. This framework offers a practical and high-precision approach for light-duty afterburning engine performance prediction and lays a foundation for the development of model-based and data-driven onboard control strategies.

During high-altitude simulation tests, the accurate reproduction of environmental conditions directly affects the ability to reliably evaluate the engine’s performance under simulated high-altitude conditions. Traditional physical models, though interpretable, often fall short in handling nonlinear dynamics and time-delayed effects caused by disturbances such as abrupt flow rate fluctuations. For this problem, this work proposes a novel hybrid modeling framework that integrates physical principles with deep learning architectures. The proposed approach incorporates three key innovations. First, a physics-guided residual learning scheme is introduced, where the theoretical outlet temperature derived from energy conservation laws and assumptions of symmetric inlet mixing serves as a prior, and a data-driven model corrects residual deviations. Second, a multiscale feature extraction module is constructed by combining a Transformer with DWT, capturing both long-term regularities and short-term fluctuations that often exhibit quasi-symmetric patterns in structured systems. Third, a perturbation-aware memory structure is designed, fusing a Transformer branch for long-term dependency modeling with an extended LSTM (xLSTM) branch for short-term dynamic sensitivity. The experimental results on real-world test datasets demonstrate that the proposed hybrid model significantly outperforms traditional physical models. Specifically, it achieves an MAPE of 0.0156 and a R² of 0.9867, indicating high predictive accuracy. The model not only achieves superior prediction accuracy but also preserves physical interpretability, making it a promising solution for intelligent control in industrial mixing systems.

The health monitoring system has been the main technological approach to extending the life of gas turbine engines and reducing maintenance costs resulting from performance degradation caused by asymmetric factors like carbon deposition, damage, or deformation. One of the most critical techniques within the health monitoring system is performance degradation prediction. At present, most research on degradation prediction is carried out using NASA’s open dataset, C-MAPSS, without considering that monitoring measurements are not always available, as in the ideal dataset. This limitation makes fault diagnosis algorithms and remaining useful life prediction methods difficult to apply to real gas turbine engines. Therefore, to solve the problem of performance degradation prediction in light-duty gas turbine engines, a prediction diagram is proposed based on Long Short-Term Memory (LSTM). Various types of onboard signals are taken into consideration among the experimental data. Only accumulated usage time, total temperature and total pressure before the inlet, low-pressure rotor speed, high-pressure rotor speed, fuel flow rate, exhaust temperature, and thrust are used in the training process, which is indispensable for an aero-engine. A genetic algorithm (GA) is introduced into the training process to optimize the hyperparameters of LSTM. The performance degradation prediction modeled with the GA-LSTM method is validated using experimental data. The maximum prediction error of thrust is 70 daN, and the mean absolute percentage error (MAPE) is less than 0.04. This study provides a practical approach to implementing performance degradation prediction in health monitoring systems to improve gas turbine engine reliability, economy, and environmental performance.

In this paper, fixed-time consensus tracking for multiagent systems (MASs) with dynamics in the form of strict feedback affine nonlinearity is addressed. A fixed-time antidisturbance consensus tracking protocol is proposed, which consists of a distributed fixed-time observer, a fixed-time disturbance observer, a nonsmooth antidisturbance backstepping controller, and the fixed-time stability analysis is conducted by using the Lyapunov theory correspondingly. This paper includes three main improvements. First, a distributed fixed-time observer is developed for each follower to obtain an estimate of the leader's output by utilizing the topology of the communication network. Second, a fixed-time disturbance observer is given to estimate the lumped disturbances for feedforward compensation. Finally, a nonsmooth antidisturbance backstepping tracking controller with feedforward compensation for lumped disturbances is designed. In order to mitigate the “explosion of complexity” in the traditional backstepping approach, we have implemented a modified nonsmooth command filter to enhance the performance of the closed-loop system. The simulation results show that the proposed method is effective.

Modern electronic power systems rely heavily on analog circuits. The accurate detection of analog circuit faults, especially soft faults, is of great significance for the maintenance and inspection of electronic systems. This paper proposes the application of the Boruta feature selection method to the field of the soft fault diagnosis of analog circuits to screen out low-dimensional and efficient feature components from the high-dimensional time-domain statistical features and frequency-domain features of circuit responses. Then, the feature components are used as the input to train the LightGBM classification model, and the Bayesian optimization method is introduced to optimize the model’s hyperparameters. Finally, the trained fault diagnosis model is verified in two typical experimental circuits, and satisfactory accuracy is obtained.