The stable operation of the Francis turbine unit (FTU) determines the safety of the hydropower plant and the energy grid. The traditional FTU performance evaluation methods with a fixed threshold cannot avoid the influence of variable operating conditions. Meanwhile, anomaly samples and missing values in the low-quality on-site data distort the monitoring signals, which greatly affects the evaluation and prediction accuracy of the FTU. Therefore, an approach to the performance evaluation and prediction of the FTU considering low-quality data and variable operating conditions is proposed in this study. First, taking the variable operating conditions into consideration, a FTU on-site data-cleaning method based on DBSCAN is constructed to adaptively identify the anomaly samples. Second, the gate recurrent unit with decay mechanism (GRUD) and the Wasserstein generative adversarial network (WGAN) are combined to propose the GRUD–WGAN model for missing data imputation. Third, to reduce the impact of data randomness, the healthy-state probability model of the FTU is established based on the GPR. Fourth, the prediction model based on the temporal pattern attention–long short-term memory (TPA–LSTM) is constructed for accurate degradation trend forecasting. Ultimately, validity experiments were conducted with the on-site data set of a large FTU in production. The comparison experiments indicate that the proposed GRUD–WGAN has the highest accuracy at each data missing rate. In addition, since the cleaning and imputation improve the data quality, the TPA–LSTM-based performance indicator prediction model has great accuracy and generalization performance.

The prognostic is the key to the state-based maintenance of Francis turbine units (FTUs), which consists of performance state evaluation and degradation trend prediction. In practical engineering environments, there are three significant difficulties: low data quality, complex variable operation conditions, and prediction model parameter optimization. In order to effectively solve the above three problems, an ensemble prognostic method of FTUs using low-quality data under variable operation conditions is proposed in this study. Firstly, to consider the operation condition parameters, the running data set of the FTU is constructed by the water head, active power, and vibration amplitude of the top cover. Then, to improve the robustness of the proposed model against anomaly data, the density-based spatial clustering of applications with noise (DBSCAN) is introduced to clean outliers and singularities in the raw running data set. Next, considering the randomness of the monitoring data, the healthy state model based on the Gaussian mixture model is constructed, and the negative log-likelihood probability is calculated as the performance degradation indicator (PDI). Furthermore, to predict the trend of PDIs with confidence interval and automatically optimize the prediction model on both accuracy and certainty, the multiobjective prediction model is proposed based on the non-dominated sorting genetic algorithm and Gaussian process regression. Finally, monitoring data from an actual large FTU was used for effectiveness verification. The stability and smoothness of the PDI curve are improved by 3.2 times and 1.9 times, respectively, by DBSCAN compared with 3-sigma. The root-mean-squared error, the prediction interval normalized average, the prediction interval coverage probability, the mean absolute percentage error, and the R2 score of the proposed method achieved 0.223, 0.289, 1.000, 0.641%, and 0.974, respectively. The comparison experiments demonstrate that the proposed method is more robust to low-quality data and has better accuracy, certainty, and reliability for the prognostic of the FTU under complex operating conditions.

The reliability of turbo-generator bearings affects the stability and safety of power plants. To detect early symptoms of failure, prediction of the performance degradation of turbo-generator bearings is particularly important. At present, most related research relies on accelerated degradation test platforms; however, the industrial field environment and complex working conditions may affect practical application. In this study, a performance degradation prediction approach for turbo-generator bearings considering complex working conditions based on a clustering indicator and a self-optimized deep learning model (SODLM) is proposed. First, the Dirichlet process-Gaussian mixture clustering model is introduced to construct the performance degradation indicator (PDI) of the turbo-generator. Then through introducing the hyper-parameter constraint, the hyper-parameter influence significance comparison mechanism, and the minimum training epoch, an improved hyperband (IH) is established to realize the automatic tuning of the hyper-parameters in the prediction model. Furthermore, the SODLM is constructed based on the IH and a stacked one-dimensional convolutional neural network. Finally, the performance degradation trend of a turbo-generator bearing in a practical industrial field is successfully predicted through the proposed approach. The experimental analysis demonstrates that the proposed PDI avoids the influence of sampling frequency and complex working conditions. Compared with traditional optimization algorithms and prediction models, the proposed performance degradation prediction approach demonstrates better accuracy and stability.

Bearing fault diagnosis is an important task for condition monitoring and predictive maintenance of rotating machinery. Nevertheless, many existing deep learning-based methods have difficulty in jointly modeling multi-scale fault characteristics, adaptively highlighting informative features, and maintaining robustness under noisy measurement conditions. To address these issues, this study presents MDCAD-Net, a multi-dilated convolution attention denoising network that integrates multi-scale temporal feature extraction, attention-based feature refinement, and explicit noise suppression within an end-to-end learning framework. Parallel dilated convolutions with different dilation rates are employed to capture short-duration transient impulses as well as long-range periodic patterns in vibration signals. Channel-wise feature recalibration using squeeze-and-excitation networks and spatial-temporal attention via a convolutional block attention module are combined to enhance informative representations. In addition, a denoising block with gated attention and residual connections is introduced to reduce noise interference while retaining fault-related signal components. Experiments conducted on the Case Western Reserve University bearing dataset show that the proposed method achieves a classification accuracy of 98.93% and yields competitive performance compared with several commonly used deep learning models. Ablation studies and feature visualization results further illustrate the contributions of the individual components and the separability of the learned feature representations under noisy conditions. The results indicate the potential of the proposed framework for practical bearing fault diagnosis under noisy operating conditions.

A 2D microwave transmission cloak composed of periodically arranged dielectric cylinders is designed and investigated by using full-wave simulations. It is shown that manipulating the periodicity of the dielectric elements can be used to provide the refractive index distribution required for bending waves around the concealed object. A possibility for obtaining the desired curvilinear wave paths in the cloak is confirmed by the simulated field patterns. The proposed all-dielectric cloak design is expected to provide low loss and broadband operation, since its performance is not related to resonance phenomena.