Fault diagnosis is critical for mitigating potential risks and ensuring the operational safety of modern dynamic industrial systems. To address the scarcity of labeled data for unseen faults in practical industrial environments, Zero-Shot Learning (ZSL) has emerged as a promising paradigm. However, conventional ZSL methods are constrained by their limited capability in extracting dynamic time-series features and their susceptibility to the severe hubness problem during cross-modal alignment. This paper proposes a novel semantic consistency embedding model based on bidirectional LSTM (BiLSTM-SCE-Net) for robust zero-shot industrial fault diagnosis. First, bypassing redundant spatial convolutions, a BiLSTM architecture is utilized to directly extract deep temporal dependencies from raw sensor data. Second, a dual-reconstruction self-supervised mechanism with Tanh boundary constraints is constructed to map features into a joint latent space, thereby maximizing the preservation of underlying physical mechanism information. Furthermore, an adaptive bias-correction inference strategy is introduced to dynamically optimize the de-biasing coefficients without requiring additional training parameters, effectively eliminating the hubness effect from a geometric distance perspective. Experimental results on the Tennessee Eastman Process (TEP) dataset demonstrate that the proposed BiLSTM-SCE-Net achieves a superior average diagnostic accuracy of 80.86% across four challenging zero-shot scenarios. Compared with state-of-the-art methods, the proposed BiLSTM-SCE-Net exhibits exceptional robustness under extreme operating conditions characterized by significant semantic gaps.

We consider the problem of resilient control of linear time-varying cyber-physical systems (CPSs) over a finite horizon in the presence of hybrid threats via a game approach. Hybrid attacks, including denial of service (DoS) and false data injection (FDI) attacks, can restrict data messaging in CPSs by deliberately altering information. To mitigate damage caused by attacks, this paper constructs a control mechanism incorporating a compensation strategy. A game model is introduced to compensate for performance losses incurred by attacks, describing the interactive behaviour between attackers and defenders. An algorithm for designing resilient controllers is proposed to explore the system's optimal defence strategy, satisfying the $H_{\infty }$ performance condition from attack signals to controlled outputs. Finally, simulation examples are provided to validate the effectiveness of the proposed resilient controller design scheme.

Photovoltaic thermoregulating windows can actively regulate window temperature using on-site electricity generation, thereby improving indoor thermal conditions and reducing building energy demand. However, their coordinated operation with air-conditioning (AC) systems remains insufficiently studied. This paper proposes an MPC-based coordinated control framework for a PV thermoregulating window-AC system. A dynamic indoor temperature prediction model is established, and a multi-objective optimization problem is formulated to balance temperature tracking and energy consumption. The proposed strategy is validated through a TRNSYS-MATLAB co-simulation platform. Compared with conventional thermostatic on-off control, the proposed method reduces the mean absolute indoor temperature error by 47.6% and decreases energy consumption by 45.2% in winter and 34.9% in summer. The results demonstrate the effectiveness of coordinated PV window-AC control for energy-efficient and intelligent building thermal management.

This article studies the prescribed performance-based sampled-data control for air handling units formulated in fractional order dynamics with non linearities, system uncertainties, and extrinsic disturbances in the heating, ventilating, and air conditioning systems, which consumes about half of the energy of the electricity system for buildings, to achieve effective indoor temperature regulation. The unknown system nonlinearities are handled by approximation through the incorporation of fuzzy logic systems. Besides, taking the hereditary and infinite memory properties of fractional- order calculus into consideration, a refined control protocol is derived based on the integration of the prescribed performance bound technique and the adaptive sampled-data backstepping control scheme. With this control protocol, it can effectively lessen control-related resources while maintaining adherence to the specified indoor temperature tracking performance. Based on the Lyapunov stability analysis, it is confirmed that even under multiple time-varying system uncertainties and unknown disturbances in practice, the proposed method can assure the stability of the closed-loop systems with the indoor temperature tracking error to be constantly within the specified performance bound. Simulated and experimental results display the effectiveness of our developed scheme.

With the rising global energy consumption, the energy use of heating, ventilation, and air conditioning (HVAC) systems has become a critical concern. Existing deep reinforcement learning control methods for HVAC systems often exhibit slow convergence and poor adaptability to dynamic environments, resulting in significant indoor temperature fluctuations, inefficient temperature control strategies, and consequently, energy waste and failure to meet thermal comfort requirements. To address these challenges, this study proposes a novel HVAC control method based on the Soft Actor-Critic with Intrinsic Reward and Correlation-Aware CNN (SAC-IRCNet) model. The model incorporates cooling load prediction as a constraint to enable on-demand cooling supply. It integrates an Elliptical Dynamics Exploration intrinsic reward mechanism, which accelerates SAC convergence through elliptical exploration rewards and inverse dynamics models, thereby reducing energy waste. Additionally, the Correlation-Aware CNN enhances SAC's feature extraction capability by leveraging state correlations to better understand contextual information, enabling more accurate responses to environmental changes and improved thermal comfort. Experimental results show that SAC-IRCNet achieves a 3.94% faster reward convergence and a 5.78% higher maximum cumulative reward within five episodes compared to SAC. It reduces energy consumption by up to 17.25% (21,837.58 kW) and lowers thermal discomfort violations by up to 6.44%, demonstrating excellent generalization ability across two datasets. Note to Practitioners-This study proposes a new control method based on the SAC-IRCNet model to address the energy efficiency and thermal comfort issues of heating, ventilation, and air conditioning (HVAC) systems. The conventional deep reinforcement learning methods used for HVAC control, such as DQN, PPO, DDPG, and TD3, usually have some problems. First, value function-based methods like DQN cannot directly handle continuous action spaces. Second, mainstream on-policy algorithms such as PPO suffer from low sample efficiency, leading to excessively high training costs in computationally expensive building simulations. More critically, widely used off-policy algorithms for continuous control-DDPG and its improved version TD3-exhibit limited exploration capabilities due to their deterministic policies, hypersensitivity to hyperparameters, and unstable training. These limitations often result in slow convergence, suboptimal policies, and potentially unstable control commands in HVAC applications, which can induce indoor temperature fluctuations. The proposed SAC-IRCNet model enhances the feature extraction ability through a convolutional neural network with correlation perception, integrates the cooling load prediction constraint to achieve on-demand cooling, and introduces an elliptical intrinsic reward mechanism to accelerate convergence. This method not only reduces energy consumption by up to 17.25%, but also improves thermal comfort by minimizing discomfort violations. Due to its excellent generalization ability, this method can be applied to various HVAC systems to achieve more efficient energy utilization.

This work presents a novel adaptive sampled-data backstepping control method for the fractional-order air handling units (AHUs) in the building heating, ventilating, and air conditioning (HVAC) systems to achieve indoor temperature regulation. The control for fractional incommensurate AHUs, which relieve computational costs for implementing control and thus enhance building energy efficiency due to the concise and precise form of the system model in comparison to the integer-order case, are studied for the first time. By strictly considering the infinite-memory and hereditary characteristics of fractional-order systems, the temperature control scheme, which novelly combines the sampled-data scheme with the backstepping technique for reducing control and transmission resources, is proposed. It is proven on the basis of Lyapunov stability analysis to be effective with the indoor temperature being able to track the target temperature accurately while all the closed-loop signals remain globally bounded even with practically time-varying AHUs system uncertainties and external disturbances. Simulation studies verify the efficacy of the proposed strategy and validate the established results.

This paper presents a novel optimal tracking control strategy for a category of non-affine nonlinear systems that encounter both unmatched and matched disturbances. Traditional inverse optimal control methods typically require systems to be affine-in-control in order for explicit inversion and cost function determination, precluding direct application to non-affine nonlinear systems. In this study, a fast subsystem is built through a backstepping procedure to transform the original system into the standard singular perturbation model. Within the fast time scale, the boundary-layer subsystem is designed to stabilize the fast states around the desired manifold. Rather than acting as a pseudo controller, the desired manifold serves to simplify the original system in the slow time scale. Thereafter, an observer-based inverse optimal control law is employed for the resulting order-reduced slow subsystem. Finally, singular perturbation theory is leveraged to simultaneously achieve optimal performance and disturbance rejection while preserving non-affine structure, without introducing heavy computation overhead. The efficacy of the proposed method is demonstrated using two simulation examples.

With the rise in energy consumption in buildings, particularly for air conditioning systems, intelligent air conditioning control has become more important. Accurate prediction of central air conditioning system cooling load is crucial to optimizing energy efficiency. This paper proposes a Multi-Scale FourierGNN Crossformer (MFGformer) model based on Crossformer, which integrates Fourier Graph Neural Networks (FourierGNN) and Multi-scale Cross-axis Attention (MCA) mechanisms for multivariate cooling load time series forecasting of central air conditioning systems. The complexity of dynamic characteristics of multivariable time series data is often misunderstood by current forecasting models, which this model takes into account. The FourierGNN module maps time series data into the frequency domain through the Discrete Fourier Transform, effectively capturing periodic and trend features of the data. The MCA mechanism captures multi-scale characteristics and local detail information in time series data through dual cross-attention computation, enhancing the ability to capture dependencies between different variables. Validation on two real case datasets shows that the MFGformer model performs well in predicting the cooling loads of central air conditioning systems, especially when outputting 24-hour predictions at a 96-hour input time step, the model’s Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), Mean Arctangent Absolute Percentage Error (MAAPE), and Coefficient of Variation of Root Mean Square Error (CV-RMSE) are 130.7765 kW, 322.2957 kW, 25.2611% and 56.6846%. Compared with eight other typical models, the MAE, RMSE, MAAPE and CV-RMSE of the MFGformer model were reduced by 22.8869-156.1751 kW, 59.6766-336.4538 kW, 0.9422-51.109% and 8.6393-32.6965%, respectively, for different output step sizes. These results demonstrate the model’s ability to provide accurate predictions when dealing with data characterized by volatility and nonlinearity, while maintaining the ability to identify long-range dependencies.

The precise prediction of thermal load has consistently garnered attention owing to its significant impact on energy conservation in buildings. The methodologies employed primarily concentrate on modeling within steady-state conditions, utilizing time-series data or mechanism model on fixed parameters. However, given the pronounced time-varying and multifaceted disturbance characteristics associated with building loads, current approaches exhibit constrained efficacy in addressing abrupt fluctuations in demand load and managing data noise. This limitation consequently undermines the accuracy of predictions. This paper proposes a novel hybrid model and an error-trigger adjusting strategy for predicting the thermal load in super-high buildings. The model is constructed by combining a thermodynamic model and an error cancellation model. The former, derived from an examination of the variations of material and energy in buildings, is proposed in the form of an approximate resistance-capacitance structure. The latter is developed using a wavelet threshold denoising technique, in conjunction with a convolutional neural network and a long short-term memory network. A self-adaptive state transition algorithm has been proposed, which relies on dynamically adjusting factors within the feasible region to optimize the selection of unknown parameters in the thermodynamic model. To enhance the flexibility of the hybrid model in effectively respond to the intricacies and fluctuations within the thermal conditions of buildings, an error-trigger adaptive updating strategy and a parameter calibration method based on sensitivity analysis are established. The real-world application results demonstrate the effectiveness of the presented hybrid model and the adjusting strategy.

In compound fault diagnosis, the scarcity of samples leads to a low fault diagnosis rate. Existing zero-shot compound fault diagnosis methods lack the ability to simultaneously recognize both seen and unseen fault classes, especially when aligning attributes of unseen faults, where a dimensionality explosion occurs. To counter the deficiencies of traditional zero-shot compound fault diagnosis methods, this article introduces an innovative generalized zero-shot compound fault diagnosis approach. This approach pioneers the use of an attribute transformation strategy, establishing a knowledge bridge between single and compound faults through a semantic label definition module, thereby creating a fault attribute set. The fault generation module is then utilized to ingeniously convert the attribute set into training samples that are rich in fault characteristics. These samples are employed to train a multilabel classification module, enabling the model to identify compound faults, including those of unseen classes. Experiments conducted on two real-world bearing datasets have validated the effectiveness and strong generalization capabilities of this method, offering a novel solution and technical support for current zero-shot rotating machinery fault diagnosis.

This study introduces a new asynchronous sampled-data distributed consensus control protocol for nonlinear fractional-order multiagent systems (MASs) containing system uncertainties along with time-varying disturbances. With strict consideration of the hereditary and infinite-memory characteristics of fractional-order systems, a novel adaptive backstepping-based distributed sampled-data control scheme is developed for individual agents with asynchronous sampling mechanisms. Through Lyapunov stability analysis, it is demonstrated that the proposed strategy guarantees the stability of the entire closed-loop system, meaning that all signals will remain within bounds and each agent can achieve output consensus with the specified time-varying reference trajectory. The efficacy of the proposed approach is illustrated through simulation studies, which also serve to validate the results obtained.

Rotating machinery is an important part of modern industry, and bearings are one of the most important things. However, bearing fault data are difficult to collect, and bearing fault diagnosis under small samples has significant research potential. In this paper, we proposed a fault diagnosis framework that combines diffusion modeling and improved Vision Transformer. First, the short-time Fourier transform is applied to the original one-dimensional vibration signals to convert the data into time-frequency maps. Second, the conditional diffusion model was applied to generate the required samples and expand the dataset. Finally, the Long-patch Vision Transformer (LVT) proposed in this paper is used to classify the mixed samples. LVT designs a long-patch division method for time-frequency maps with dense transverse features. The LVT contains denser features in each patch, and this method is more suitable for time-frequency maps. Validating the method proposed in this paper on two datasets and comparing it with other methods, our method achieved the highest accuracy among the compared methods.

Electroluminescence (EL) imaging technology detects tiny defects at the production stage, which are small but have a significant impact on battery performance and life. In this paper, a CMLNet is proposed, which can effectively solve the problem of defect detection in EL images with complex backgrounds. First, the CSPHet network is introduced to replace the C2f structure in Yolov8. Second, a lightweight shared convolution detection head (LSCD-Head) is introduced. Then, a lightweight mixed local channel attention mechanism (MLCA) is integrated into the neck of the model to enhance the feature expression ability of the model. Finally view of Yolov8's sensitivity to CIoU, an Inner-ShapeIoU solution is introduced to enhance model generalisation ability. In the experiment of the PVEL-AD dataset, the improved model has a 42.4% reduction in the number of parameters compared with Yolov8n. At the same time, map@50 and map@50-95 of this model can reach 92.6% and 64.7%, respectively, which are 3.1% and 2.4% higher than those of the original Yolov8n network, achieving lightweight and high-precision optimisation. In addition, the generalisation capability of the model is verified on the PV-Defect dataset. The results show that CMLNet has remarkable advantages in various performance indexes while maintaining lightweight performance.

When wind turbines operate in complex environments, planetary gearboxes easily generate faults that will lead to increased wreck of the equipment or transmission failures. In this paper, a CNN model with adaptive parameters is proposed to realize the identification of planetary gearbox faults and improve the real-time performance and accuracy of fault diagnosis. Firstly, Ensemble Empirical Mode Decomposition (EEMD) is applied to scale the one-dimensional signal to solve the modal mixing problem of input data. Gramian Angular Difference Fields (GADF) is used to convert the processed data into images as the input of CNN model. Secondly, to capture more information and reduce the risk of overfitting, two convolutional neural networks incorporating different activation functions are connected in parallel to propose a multilayer CNN model structure. Additionally, Pied Kingfisher Optimization algorithm (PKO) is improved by integrating Tent chaotic mapping, second-order optimization and simulated annealing algorithm to optimize the multilayer CNN model automatically. Finally, the experimental results show that this improved model achieves real-time diagnosis due to the adaptive parameters, the diagnosis accuracy exceeds 95% under the same proportion of samples, and over 85% under the different proportions of samples. This approach significantly enhances planetary gearbox fault identification reliability.

With the increase in global shipping volumes and the complexity of maritime transport systems, vessel trajectory prediction serves an important tool in improving maritime safety. However, most existing vessel trajectory prediction methods focus on a single feature and unable fuse high-dimensional features. To solve these problems, CNN-GRU model with a hybrid attention mechanism (AM) is proposed based on Automatic Identification System (AIS) data. First convolutional neural network (CNN) is proposed to extract the spatio-temporal information of the trajectory data. Then a gated recurrent unit (GRU) is designed to extract the temporal relationship of the trajectories. Finally, AM is introduced to learn the deep-level features and predict the vessel trajectories. To validate the effectiveness of the model, experiments are conducted on three real AIS datasets. In comparison with other models, the method has a high trajectory prediction accuracy.

Pupil localization is one of the most critical and essential requirements for eye gaze estimation and eye movement tracking. Because pupil images contain monotonous and uncomplicated information, the dataset uses a single class of labels to describe the image content, and using convolutional neural networks can quickly and accurately identify the pupil position on the input image. On low-resolution images, traditional methods encounter issues of low accuracy and cumbersome design steps. A lightweight pupil localization algorithm is proposed in this paper, utilizing a convolutional neural network (CNN) with additional training samples. The experimental results demonstrate the algorithm’s significant effectiveness in identifying the pupil position within the training set, with the accuracy of pupil position in the test set reaching 97.78%. This provides an evidence of the algorithm’s feasibility for accurately localizing pupils in low-resolution images.

This study investigates hybrid-triggered and observer-based output feedback H-infinity control for Takagi-Sugeno (T-S) fuzzy cyber-physical systems (CPSs) under denial of service (DoS) attacks. To address the challenges posed by nonlinear discrete network physical systems, the T-S fuzzy model is adopted. The DoS attacks model and the hybrid trigger mechanism (HTM) are constructed using two independent Bernoulli distributions. A hybrid event-triggered cooperative control approach is proposed to overcome the limitation of directly measuring state variables based on observer output feedback. A well-established matrix decoupling technique is employed to ensure sufficient conditions for the controller and observer gain establishment. Systems stability is guaranteed by utilizing the Lyapunov function and Linear Matrix Inequality (LMI) techniques. From the numerical example, the method is proved to be effective.

Vacuum-photovoltaic glazing, renowned for exceptional thermal insulation and solar energy utilization, faces limitations in its adaptability to varying seasons. While it effectively reduces heat transmission into indoor spaces during summer, it becomes detrimental during winter. To address this challenge, this study introduces an innovative solution: vacuum-photovoltaic-thermoelectric (VPT) glazing, which integrates vacuum, photovoltaic and thermoelectric cooling/warming technologies. The theoretical models were developed and validated through WINDOW and Fluent simulations. A comparative analysis is conducted considering thermal performance under various environmental parameters. The results demonstrate that VPT glazing exhibits enhanced thermal performance, with interior glass temperature decreased by 3.0∼9.6oC in summer while increased by 2.5∼6.2oC in winter, accounting for ∼55.0% reduction in air-conditioning load. Compared to vacuum-photovoltaic glazing, VPT glazing reduces the coupling -value from 7.88 to 5.87 W·m-2·K-1 in summer and increases from -0.31 to 2.61 W·m-2·K-1 in winter. The solar heat gain coefficient decreases from 0.37 to 0.30 in summer and increases from 0.24 to 0.29 in winter. These results demonstrate the effectiveness of VPT glazing in adapting to different seasons and achieving better thermal environment performance. This study provides insights into VPT glazing design for environmental adaptation and offers implications for future research and applications in energy-efficient building technologies.

Addressing the problems of manual dependence and low accuracy of traditional building electrical system fault diagnosis, this paper proposes a novel method, which is based on random forest (RF) optimized by improved sparrow search algorithm (ISSA-RF). Firstly, the method utilizes a fault collection platform to acquire raw signals of various faults. Secondly, the features of these signals are extracted by time-domain and frequency-domain analysis. Furthermore, principal component analysis is employed to reduce the dimensionality of the extracted features. Finally, the reduced features are input into ISSA-RF for classification. In ISSA-RF, the ISSA is used to optimize the parameters of the RF. The parameters for ISSA optimization are n_estimators and min_samples_leaf. In this case, the accuracy of the proposed method can reach 98.61% through validation experiment. In addition, the proposed method also exhibits superior performance compared with traditional fault classification algorithms and the latest building electrical fault diagnosis algorithms.