Existing federated learning primarily focuses on problem setups where servers and clients engage in model training for one or multiple specific tasks. However, in real-world scenarios, the required diagnoses at clinical sites can vary due to the diversity of conditions, differing not only from each other but also from those at the server. In this study, we concentrate on the practical yet challenging Federated Agnostic Learning (FAL), where client-side diagnostic tasks are agnostic. We introduce a novel FedACT method to address this issue, which is composed of two components. The first component extracts shared features among multiple agnostic tasks using an end-to-end similarity layer based on contrastive learning to enhance generalizability. In the second component, we design personalized task-specific branches for comprehensive tasks, including classification and segmentation. The branches can accurately accomplish the respective tasks through knowledge transfer and enhanced discriminative capabilities across various classes and tasks. Moreover, to better accommodate the potential heterogeneity of data and unseen tasks, specialized updating and aggregation methods are devised for FedACT. The experimental results demonstrate the effectiveness of FedACT in various scenarios under the FAL setting.

Unmanned Aerial Vehicles (UAVs) offer a solution for remote inspection of wind turbines. However, in stormy weather conditions, the visual quality of UAV-taken images is affected by contaminated suspended atmospheric particles. To address this problem, a double-patch lightweight convolutional dehazing neural network (DPLDN) is proposed to reconstruct hazy images and enhance the image quality. Unlike other learning-based methods that measure transmission map and atmospheric light separately, the proposed DPLDN uses a transformed atmospheric scattering model to jointly transmission map and atmospheric light, employs depth-separable convolution instead of conventional convolution, and splits the image into double patches. In addition, a super-resolution reconstruction method is proposed to transform the processed low-resolution images into higher-quality images. Extensive experiments shows that our proposed method has better dehazing performance compared to other state-of-the-art image dehazing techniques. Meanwhile, the applicability of the method in wind turbine blade image segmentation is experimentally verified.

To address foreseeable challenges during the penetration of wind energy into the power grid, including accurate wind power forecasting and smart power generation scheduling, this study proposes a novel short-term wind speed forecasting model, named EMD-KM-SXL, which is based on empirical mode decomposition (EMD),$K $-means clustering and machine learning, and a new two-stage short-term wind power forecasting model based on wind speed forecasting and wind power curve (WPC) modeling. The former wind speed forecasting model regards historical wind speed observations as model input and the latter power forecasting model utilizes knowledge augmentation, introducing wind power conversion relationship, environmental factors as well as wind power system status parameters. In the proposed wind speed forecasting model, three machine learning models, including support vector regressor, XGBoost regressor, and Lasso regressor, are employed to forecast three types of frequency components that generated via EMD and$K $-means clustering. Then, the WPC model is utilized to compute potential outpower based on wind speed prediction value speed, which is regarded as the first stage of the proposed wind power forecasting model. In the second stage, environmental factors and wind power system status parameters are introduced and an artificial neural network model, considering preliminary predicted power, environmental factors, and wind power system status parameters as model input is built to make final power prediction. Computational results show that the proposed models achieve the best performance in terms of wind speed and power forecasting on different forecasting horizons ranging from 10 to 40 min, compared with benchmarking methods.

Nowadays, the meter reading is detached into two parts: 1) image collection and 2) image recognition. Most previous researchers perceived meter reading as an image classification problem and obtained impressive classification performance. However, the numerical is also a critical metric of meter measurement and has not been noticed in previous research. This article redefines the meter reading issue as a hybrid procedure of regression and classification and creates a specific model. The resulting algorithm bespeaks the performance of measurement and recognition. The model consists of a hybrid regression and classification loss function and multibranch convolutional neural networks. We construct two data sets to validate the model: 1) normal data set and 2) carry data set, corresponding to classification and numeric accuracy dividedly. The experiments show that the model establishes new state-of-the-art metrics and achieves 0.5312 mean square error (MSE) on numerical precision and 99.98% accuracy on classification accuracy for 3-min training, by over 70 times on MSE and 0.11% on accuracy than the best performing model proposed by a recent study. Furthermore, a production-ready meter-reading system was deployed in a genuine factory with hybrid regression and classification multibranch convolutional neural networks, smart meter shells, and a set of cloud servers.

This article proposes a generalized neural continual learning-based cybertwin (GNC) modeling framework to realize developing one wind turbine (WT) cybertwin serving multiple modeling tasks in the wind farm operations and maintenance (O&M). A generalized WT cybertwin modeling problem, which considers modeling one cybertwin for multiple tasks without additional computational burden, is studied for the first time. Fully connected neural networks are adopted as the backbone for developing the GNC model. The online elastic weight consolidation method is incorporated to mitigate the catastrophic forgetting phenomenon among different modeling tasks. Computational experiments are conducted to validate the effectiveness of the proposed GNC framework based on the supervisory control and data acquisition data. Modeling tasks in three important problems of the wind farm O&M, the WT gearbox failure detection, WT blade breakage detection, and wind power prediction, are considered in the experiment. Compared with other benchmarking models, such as the multiple neural cybertwins, neural cybertwin, and regularized neural cybertwin, the proposed GNC can achieve high accuracies on both new tasks and existing tasks, which further verifies the WT cybertwin generalization via the proposed GNC.

This paper investigates the multi-objective optimal operation problem of an integrated energy system (IES) which integrates grid-connected photovoltaic (PV) generator, gas boiler, battery energy storage system, and thermal storage to satisfy energy demand in forms of electricity and heat. To handle the changes from the system uncertainty (e.g., PV generation, electrical loads, thermal loads, etc.) and unknown thermal dynamic model for temperature control, deep reinforcement learning-based model-free optimization method is proposed to solve the multi-objective optimization problem in which the multi-objective optimization problem is firstly formulated as a multi-objective Markov decision process (MDP) problem. The multi-objective MDP problem is converted to many single-objective MDP problems by the sum technique which are solved by multi-agent deep deterministic policy gradient (DDPG) algorithm. To improve the performance of multi-agent DDPG algorithm, evolutionary computing-based parameter-tuning method is further proposed to fine-tune the policy parameters in DDPG algorithm. The proposed methods are verified on real data. Experiments results illustrate that the multi-agent DDPG algorithm can efficiently solve the multi-objective optimal operation problem of the IES while the evolutionary computing-based policy parameter-tuning method can further improve the approximation of Pareto frontier.

This article proposes a hybrid symbolic aggregate approximation and vector space model (SAX-VSM) method for automatically classifying soil water content cycles. In the proposed method, a novel similarity measure, the distance weighted cosine (DWC) similarity measure, is introduced to improve the classification performance of the SAX-VSM. The DWC similarity measure incorporates both direction and distance information of feature vectors. Meanwhile, a mixed-integer optimization problem is formulated to determine hyperparameters. An extended Rao-1 algorithm, I-Rao-1 algorithm, is developed to solve such optimization problems. To verify the feasibility and effectiveness of the proposed method, three soil moisture data sets collected from the Florida research trials are employed. Compared with state-of-the-art methods, the proposed method has achieved the best performance based on all data sets in terms of the highest accuracy, precision, and recall values. Therefore, it is promising to apply the proposed method into real applications in the smart irrigation system.

In this paper, a novel method is developed for day-ahead daily river-flow forecasting based on convolutional neural network (CNN). The proposed method incorporates both spatial and temporal information to improve the forecasting performance. A CNN model is usually trained by minimizing the mean squared error which is, however, sensitive to few particularly large errors. This character of squared error loss function will result in a poor estimator. To tackle the problem, a robust loss function is proposed to train the CNN. To facilitate the training of CNNs for multiple sites forecasting, transfer learning is also applied in this study. With transfer learning, a new CNN inherits the structure and partial learnable parameters from a well-trained CNN to reduce the training complexity. The forecasting performance of the proposed method is validated with real data of four rivers by comparing with widely used benchmarking models including the autoregressive model, multilayer perception network, kernel ridge regression, radial basis function neural network, and generic CNN. Numerical results show that the proposed method performs best in terms of the root mean squared error, mean absolute error, and mean absolute percentage error. The two-sample Kolmogorov-Smirnov test is further applied to assess the confidence on the conclusion.

In recent years, more and more wearable sensors have been employed in smart health applications. Wearable sensors not only can be used to collect valuable health-related data of their users, they can be also used in conjunction with other infrastructure-bound sensors, such as Microsoft Kinect sensor, to facilitate privacy-aware fine-grained activity tracking. This fusion of multimodal data promises a new type of smart health applications that coach a user to live a healthier life style by monitoring the user in realtime and reminding him or her when he or she engages in an unhealthy activity. In this paper, we investigate how to achieve fine-grained activity recognition in the context of such an application. In our scheme, the identification accuracy is improved by incorporating a nonlinear and local similarity measure, namely kernel risk-sensitive loss, into a novel multilayer neural network learning algorithm, called as stacked extreme learning machine. Furthermore, to achieve a good generalization performance with minimal human intervention, Jaya as a popular optimization algorithm, is also used to adjust key parameters in our proposed approach. The experiments are conducted to verify the effectiveness of the proposed scheme.

Recently, wind speed forecasting as an effective computing technique plays an important role in advancing industry informatics, while dealing with these issues of control and operation for renewable power systems. However, it is facing some increasing difficulties to handle the large-scale dataset generated in these forecasting applications, with the purpose of ensuring stable computing performance. In response to such limitation, this paper proposes a more practical approach through the combination of extreme-learning machine (ELM) method and deep-learning model. ELM is a novel computing paradigm that enables the neural network (NN) based learning to be achieved with fast training speed and good generalization performance. The stacked ELM (SELM) is an advanced ELM algorithm under deep-learning framework, which works efficiently on memory consumption decrease. In this paper, an enhanced SELM is accordingly developed via replacing the Euclidean norm of the mean square error (MSE) criterion in ELM with the generalized correntropy criterion to further improve the forecasting performance. The advantage of the enhanced SELM with generalized correntropy to achieve better forecasting performance mainly relies on the following aspect. Generalized correntropy is a stable and robust nonlinear similarity measure while employing machine learning method to forecast wind speed, where the outliers may exist in some industrially measured values. Specifically, the experimental results of short-term and ultra-short-term forecasting on real wind speed data show that the proposed approach can achieve better computing performance compared with other traditional and more recent methods.

The feasibility of monitoring the health of wind turbine (WT) gearboxes based on the lubricant pressure data in the supervisory control and data acquisition system is investigated in this paper. A deep neural network (DNN)-based framework is developed to monitor conditions of WT gearboxes and identify their impending failures. Six data-mining algorithms, thek-nearest neighbors, least absolute shrinkage and selection operator, ridge regression (Ridge), support vector machines, shallow neural network, as well as DNN, are applied to model the lubricant pressure. A comparative analysis of developed data-driven models is conducted and the DNN model is the most accurate. To prevent the overfitting of the DNN model, a dropout algorithm is applied into the DNN training process. Computational results show that the prediction error will shift before the occurrences of gearbox failures. An exponentially weighted moving average control chart is deployed to derive criteria for detecting the shifts. The effectiveness of the proposed monitoring approach is demonstrated by examining real cases from wind farms in China and benchmarked against the gearbox monitoring based on the oil temperature data.