In this paper, novel fault estimation and fault-tolerant control methods are proposed for dynamics of high-speed train based on descriptor systems with uncertainties in finite frequency domain. Dynamics of high-speed train is established based on multi-particle model considering that basic resistance is seen as the coefficient of state variables, and additive resistance and the operating noise are seen as multi-source disturbance. Concurrent actuator, sensor faults, and wind gust are considered simultaneously; wind gust is modeled as a disturbance generated by the exogenous system, and an uncertain descriptor system with actuator fault and the exogenous disturbance is established by seeing the sensor fault of high-speed train as the state variables. A robust disturbance-observer-based fault estimation method is proposed to decouple the non-linearity of the descriptor system, so that the combining estimation of the fault and wind gust is implemented. This observer has an unknown input structure, and its gain matrices are formulated as linear matrix inequalities. The observer not only guarantees the augmented state estimation error is asymptotic stable but also the actuator fault estimation and wind gust estimation errors are robust to the multi-source disturbance and the uncertainties. Based on the estimation results, the fault-tolerant controller associated with the state estimation, faults estimation, and wind gust estimation results is proposed to implement a stable close-loop fault-tolerant control for dynamics of high-speed train. Simulation examples are given to illustrate the effectiveness of this method.

Bearings are critical components of mechanical equipment, and predicting their remaining useful life (RUL) is important in industry. This paper proposes a RUL prediction method based on the assessment of a bearing’s health status. Features from the time, frequency, and time–frequency domains of the bearing’s vibration signal are extracted to construct a feature set. A multibranch encoder and restricted Boltzmann machine are used to improve the stacked autoencoder to reduce dimensionality. Local weights and health-sample means are introduced into the Mahalanobis distance to improve the health index. Subsequently, the weighted convolutional Euclidean distance serves as the distance metric in K -means clustering to achieve a more accurate health status assessment and provide historical data for RUL prediction. An improved self-attention (ISA) mechanism is proposed by incorporating depthwise separable convolutions and residual-like connections into self-attention mechanisms, enhancing the global and local dependencies of the temporal convolutional network (TCN). Thus, a more accurate RUL prediction is achieved. Comparative and ablation experiments confirm that the proposed ISA-TCN achieves superior predictive accuracy. Generalization experiments further demonstrate its strong adaptability, while anti-noise experiments demonstrate its strong robustness to uncertainties. Finally, experiments on multi-output RUL predictions validate the model’s effectiveness. This approach offers valuable insights for RUL prediction of rotating machinery under real-world scenarios involving variable operating conditions, noise interference, and multi-device environments.

This paper proposes a state estimation method for a sampled‐data descriptor system by the Kalman filtering method. The sampled‐data descriptor system is firstly discretized to obtain a discrete‐time nonsingular model. Based on the discretized nonsingular system, a strong tracking unscented Kalman filter (STUKF) algorithm is designed for the state estimation. Then, a defined suboptimal fading factor is proposed and added to the prediction covariance for decreasing the weight of the prior knowledge on the conventional UKF filtering solution. Finally, a simulation example is given to show the effectiveness of the proposed method.

In this study, an optimized long short-term memory (LSTM) network is proposed to predict the reliability and remaining useful life (RUL) of rolling bearings based on an improved whale-optimized algorithm (IWOA). The multi-domain features are extracted to construct the feature dataset because the single-domain features are difficult to characterize the performance degeneration of the rolling bearing. To provide covariates for reliability assessment, a kernel principal component analysis is used to reduce the dimensionality of the features. A Weibull distribution proportional hazard model (WPHM) is used for the reliability assessment of rolling bearing, and a beluga whale optimization (BWO) algorithm is combined with maximum likelihood estimation (MLE) to improve the estimation accuracy of the model parameters of the WPHM, which provides the data basis for predicting reliability. Considering the possible gradient explosion by training the rolling bearing lifetime data and the difficulties in selecting the key network parameters, an optimized LSTM network called the improved whale optimization algorithm-based long short-term memory (IWOA-LSTM) network is proposed. As IWOA better jumps out of the local optimization, the fitting and prediction accuracies of the network are correspondingly improved. The experimental results show that compared with the whale optimization algorithm-based long short-term memory (WOA-LSTM) network, the reliability prediction and RUL prediction accuracies of the rolling bearing are improved by the proposed IWOA-LSTM network.

In this paper, an optimized adaptive robust extended Kalman filter is proposed based on random weighting factors and an improved whale optimization algorithm for fault estimation of the dynamics of high-speed trains with constant time delays, drastically changing noise and stochastic uncertainties. Robust upper bounds are proposed to improve the performance of the extended Kalman filter by decreasing the influence of the linearization error on filtering for the dynamics of high-speed trains with constant time delays, and its robustness is proven to guarantee the feasibility of the proposed upper bounds. Furthermore, considering drastically changing noise with unknown statistics, a random weighting adaptive algorithm is proposed to implement unbiased noise estimation so that the robust extended Kalman filter can still be implemented well. In addition, a differential evolution algorithm and adaptive parameter are introduced to improve the performance of the whale optimization algorithm so that the stochastic uncertainties are optimized, and the influence of the stochastic uncertainties on filtering is further decreased. The simulation results in the three conditions show that, compared with the variational Bayes adaptive iterated extended Kalman filter, using the proposed method, the position, speed and fault estimation errors are decreased by 31.8%, 33.2% and 28.3%, respectively, on average, which depends on more accurate noise estimation.