Search

As the urgency to adopt renewable energy sources escalates, so does the need for accurate forecasting of power output, particularly for wind and solar power. Existing models often struggle with noise and temporal intricacies, necessitating more robust solutions. In response, our study presents the SL-Transformer, a novel method rooted in the deep learning paradigm tailored for green energy power forecasting. To ensure a reliable basis for further analysis and modeling, free from noise and outliers, we employed the SG filter and LOF algorithm for data cleansing. Moreover, we incorporated a self-attention mechanism, enhancing the model's ability to discern and dynamically fine-tune input data weights. When benchmarked against other premier deep learning models, the SL-Transformer distinctly outperforms them. Notably, it achieves a near-perfect R 2 value of 0.9989 and a significantly low SMAPE of 5.8507% in wind power predictions. For solar energy forecasting, the SL-Transformer has achieved a SMAPE of 4.2156%, signifying a commendable improvement of 15% over competing models. The experimental results demonstrate the efficacy of the SL-Transformer in wind and solar energy forecasting. [ABSTRACT FROM AUTHOR]

The health status of equipment is of paramount importance during the operation of nuclear power plants. The occurrence of faults not only leads to significant economic losses but also poses risks of casualties and even major accidents, with unimaginable consequences. This paper proposed a deep learning framework called PT-Informer for fault prediction, detection, and localization in order to address the challenges of online monitoring of the operating health of nuclear steam turbines. Unlike traditional approaches that involve separate design and execution of feature extraction for fault diagnosis, classification, and prediction, PT-Informer aims to extract fault features from the raw vibration signal and perform ultra-real-time fault prediction prior to their occurrence. Specifically, the encoding and decoding structure in PT-Informer ensures the capture of temporal dependencies between input features, enabling accurate time series prediction. Subsequently, the predicted data are utilized for fault detection using PCA in the PT-Informer framework, aiming to assess the likelihood of equipment failure in the near future. In the event of potential future failures, t-SNE is utilized to project high-dimensional data into a lower-dimensional space, facilitating the identification of clusters or groups associated with different fault types or operational conditions, thereby achieving precise fault localization. Experimental results on a nuclear steam turbine rotor demonstrate that PT-Informer outperformed the traditional GRU with a 4.94% improvement in R2 performance for prediction. Furthermore, compared to the conventional model, the proposed PT-Informer enhanced the fault classification accuracy of the nuclear steam turbine rotor from 97.4% to 99.6%. Various comparative experiments provide strong evidence for the effectiveness of PT-Informer framework in the diagnosis and prediction of nuclear steam turbine. [ABSTRACT FROM AUTHOR]

Construction projects face various constraints in terms of materials, labor, equipment, and documents, which can interrupt the scheduled work. Package‐based constraint management (PCM) is a state‐of‐the‐art graph‐based approach that follows the lean theory to effectively model, monitor, and remove constraints before the commencement of work, ensuring smooth construction and minimizing delay and waste. PCM relies on exploring and investigating project knowledge bases (KBs), formed by entity‐relation‐entity triples of constraints. However, most PCM KBs are incomplete and suffer from poor semantics, which hinders the PCM functions. Although many KB completion (KBC) methods exist in the field of artificial intelligence, they primarily focus on general knowledge and exclude the features of specific domains. Therefore, they cannot be directly applied to complete PCM KBs. To address the issue, this study proposes a novel deep learning model, referred to as the domain information enhanced graph neural network (D‐GNN). The features of the developed D‐GNN include (1) building a domain ontology to enrich semantics with rule reasoning, (2) applying the GNN to learn and encode embeddings of constraint entities and relations, and (3) employing a convolution neural network (CNN) for decoding and identifying missing triples. D‐GNN improves the existing KBC methods by integrating two types of domain information, namely, the ontological classes and working contexts into GNN and CNN, respectively. The experimental results verified that the D‐GNN reached an accuracy of 0.848–0.951, and the domain information integration increased the performance by up to 0.263. In practical testing, the D‐GNN significantly reduced the KBC time to 1/6–1/35 of the manual approach and reached higher accuracy. Therefore, the proposed D‐GNN can facilitate PCM by providing complete KBs and supporting downstream constraint monitoring and removal. [ABSTRACT FROM AUTHOR]

The effective use of wind energy is an essential part of the sustainable development of human society, in particular, at the recent unprecedented pressure in shaping a low carbon energy environment. Accurate wind resource and power forecasting play a key role in improving the wind penetration. However, it has not been well adopted in the real‐world applications due to the strong stochastic characteristics of wind energy. In recent years, the application boost of deep learning methods provides new effective tools in wind forecasting. This paper provides a comprehensive overview of the forecasting models based on deep learning in the field of wind energy. Featured approaches include time‐series‐based recurrent neural networks, restricted Boltzmann machines, convolutional neural networks as well as auto‐encoder‐based approaches. In addition, future development directions of deep‐learning‐based wind energy forecasting have also been discussed. [ABSTRACT FROM AUTHOR]

In this paper, a deep learning fault detection and prediction framework combining principal component analysis (PCA) and Informer is proposed to solve the problem of online monitoring of nuclear power valves which is hard to implement. More specifically, PCA plays the role of dimensionality reduction and fault feature extraction. It maps data with multi-dimensional space to low-dimensional space and extracts the main features. At the same time, the T-square and Q statistic thresholds are also provided to realize abnormal status monitoring. Meanwhile, Informer is a long-term series prediction method. It encrypts and decrypts data through the encoder and decoder to train a prediction model. Through the training of fault data, fault prediction can be realized. Experiments based on the sound waves collected from real valves can be continued, which also illustrates the effectiveness of the PCA–Informer model for fault diagnosis and fault prediction of nuclear power valves. Therefore, the online monitoring and maintenance of nuclear valves and other important equipment, without shutting down the nuclear station, can be achieved. [ABSTRACT FROM AUTHOR]

Short-term load forecasting is a key task to maintain the stable and effective operation of power systems, providing reasonable future load curve feeding to the unit commitment and economic load dispatch. In recent years, the boost of internal combustion engine (ICE) based vehicles leads to the fossil fuel shortage and environmental pollution, bringing significant contributions to the greenhouse gas emissions. One of the effective ways to solve problems is to use electric vehicles (EVs) to replace the ICE based vehicles. However, the mass rollout of EVs may cause severe problems to the power system due to the huge charging power and stochastic charging behaviors of the EVs drivers. The accurate model of EV charging load forecasting is, therefore, an emerging topic. In this paper, four featured deep learning approaches are employed and compared in forecasting the EVs charging load from the charging station perspective. Numerical results show that the gated recurrent units (GRU) model obtains the best performance on the hourly based historical data charging scenarios, and it, therefore, provides a useful tool of higher accuracy in terms of the hourly based short-term EVs load forecasting. [ABSTRACT FROM AUTHOR]

• A novel forecasting based reinforcement learning energy scheduling method is proposed. • A GC-LSTM algorithm is adopted in out-door temperature prediction. • A model-free During-DDQN algorithm is adopted to manage household energy. • The proposed method performs well in the household energy management system. The demand response (DR) strategy enables household users to actively optimize and dispatch the household energy management system (HEMS), which may significantly reduce the cost of user energy consumptions. However, the uncertainty of home users behaviors, the diversity of electrical equipment types, as well as the complexity of the working status of various devices have brought severe challenges to the home energy management system. This article proposes a new forecasting based optimization method to deliver the real-time scheduling considering the future environment trend. A recent proposed dueling based deep reinforcement learning approach is adopted to optimally dispatch the HEMS. In addition, due to the delay of heating, ventilation, and air Conditioning (HVAC) indoor and outdoor temperature data non-Gaussian, a new generalized corr-entropy assisted long short-term memory (GC-LSTM) neural network is proposed where the generalized correntropy (GC) loss function is adopted to predict the outdoor temperature. The proposed method is verified in a featured HEMS benchmark problem and the experimental results show that the user costs are effectively reduced while the user satisfaction is maintained utilizing the proposed method. [ABSTRACT FROM AUTHOR]

Load forecasting is one of the major challenges of power system operation and is crucial to the effective scheduling for economic dispatch at multiple time scales. Numerous load forecasting methods have been proposed for household and commercial demand, as well as for loads at various nodes in a power grid. However, compared with conventional loads, the uncoordinated charging of the large penetration of plug-in electric vehicles is different in terms of periodicity and fluctuation, which renders current load forecasting techniques ineffective. Deep learning methods, empowered by unprecedented learning ability from extensive data, provide novel approaches for solving challenging forecasting tasks. This research proposes a comparative study of deep learning approaches to forecast the super-short-term stochastic charging load of plug-in electric vehicles. Several popular and novel deep-learning based methods have been utilized in establishing the forecasting models using minute-level real-world data of a plug-in electric vehicle charging station to compare the forecasting performance. Numerical results of twelve cases on various time steps show that deep learning methods obtain high accuracy in super-short-term plug-in electric load forecasting. Among the various deep learning approaches, the long-short-term memory method performs the best by reducing over 30% forecasting error compared with the conventional artificial neural network model. [ABSTRACT FROM AUTHOR]

A large proportion of construction accidents are caused by workers' unsafe actions. Due to the complexity of the work environment and excessive demands of safety supervision on construction sites, many modern information technologies, such as computer vision (CV) technologies, have been increasingly applied and gradually replaced the traditional manual supervision to automatically identify unsafe behaviors from surveillance video data. Current models focus on high resolution (HR) video containing high quality features. However, challenges remain for more specific real-world construction scenarios, including but not limited to lack of high-quality data, far-field video recognition, privacy protection, and limited computing resources. To address the challenges, in this paper, a simple but effective method relying on the Teacher–Student Learning Architecture is proposed, where the teacher network assists the student network in effectively capturing useful features for downstream action recognition from extreme low resolution (eLR) video which can be easily transferred to other tasks. In addition, a Knowledge Learning Model and a similarity loss function are proposed to better guide the training of student network. In numerical experiments and case studies, our proposed training strategy achieves an top-1 accuracy of 28.53% and 71.81% on eLR HMDB dataset and eLR CMA dataset, 5.96% and 3.89% higher than the baseline model. The proposed strategy can make the automatic safety monitoring of construction works more accurate and reliable, thereby effectively reducing the accident rate and management cost. In addition, the performance of proposed strategy depends on efficient knowledge distillation methods, which may inspire future applications of CV-based deep learning models in construction management to achieve higher recognition results while maintaining low computational cost. [ABSTRACT FROM AUTHOR]

A large proportion of construction accidents are caused by unintentional and unsafe actions and behaviors. It is of significant difficulties and ineffectiveness to monitor unsafe behaviors using conventional manual supervision due to the complex and dynamic working conditions on construction sites. Recently, surveillance videos and computer vision (CV) techniques have been increasingly adopted to automatically identify risky behaviors. However, the challenge remains that spatial and temporal features in video clips cannot be effectively captured and fused by current CV models. To address this challenge, this paper describes a deep learning model named Spatial Temporal Relation Transformer (STR-Transformer), where spatial and temporal features of work behaviors are simultaneously extracted in paralleling video streams and then fused by a specially designed module. To verify the effectiveness of the STR-Transformer, a customized dataset is developed, including seven categories of construction worker behaviors and 1595 video clips. In numerical experiments and case studies, the STR-Transformer achieves an average precision of 88.7%, 4.0% higher than the baseline model. The STR-Transformer enables more accurate and reliable automatic safety surveillance on construction projects, and is expected to reduce accident rates and management costs. Moreover, the performance of STR-Transformer relies on efficient feature integration, which may inspire future studies to identify, extract, and fuse richer features when applying CV-based deep learning models in construction management. • A deep learning model is proposed for action recognition. • The model consists of two feature encoders and a feature fusion model. • The model identify unsafe actions of construction workers. • A novel dataset of construction worker actions is proposed. [ABSTRACT FROM AUTHOR]

Effective and timely fault detection and status monitoring of the industrial production process is essential to fully guarantee the operational safety. However, massive multi-source heterogeneous data analysis is facing many challenges. This paper proposes a process monitoring model combined with a parallel deep learning algorithm and Principal Component Analysis (PCA) method. Firstly, PCA is applied to realize fault diagnosis and extract fault characteristic variables, thus abnormal conditions can be revealed. In order to reduce the complexity of processing massive data, a parallel deep learning framework consisting of multi-models of convolutional neural network (CNN) and Long-Short Term Memory (LSTM) neural network is proposed to effectively predict the target variable status. Comprehensive experiments are taken under two real-system scenarios, and comparisons are made against four traditional neural network models to demonstrate its practicability and effectiveness. The generated results clearly show that the CNN-LSTM parallel model combined with PCA outperform other popular models due to its merged advantages of accurate time series prediction and effective fault feature extraction. [Display omitted] • A parallel deep learning approach with feature extraction and long time series prediction. • Multi-source unsupervised faulty variable extraction and detection can be realized. • Successfully solve the distribution power grid fault diagnosis in multiple locations. [ABSTRACT FROM AUTHOR]

Unstructured texts dominate data in construction projects. With the achievements of natural language processing (NLP) techniques, mining unstructured text data for smart construction has become increasingly significant. To understand state-of-the-art NLP for smart construction, uncover related issues, and propose potential improvements, this paper presents a comprehensive review of bottom-level techniques and mainstream applications of NLP in the industry. In total, 124 journal articles published in the last two decades are reviewed. NLP involves five core steps supported by various techniques, e.g., syntactic parsing, heuristic rules, machine learning, and deep learning. NLP has been applied for information extraction and exchanging and many downstream applications to facilitate management and decision-making. The role of NLP in smart construction and current challenges for fully reaping its benefits are discussed, and four research directions are identified, i.e., improving relation extraction, realising knowledge base auto-development, integrating multi-modal information, and achieving an accuracy-efficiency trade-off by developing an NLP application framework. It is envisioned that outcomes of this paper can assist both researchers and industrial practitioners with appreciating the research and practice frontier of NLP for smart construction and soliciting the latest NLP techniques. • An in-depth review of natural language processing (NLP) for smart construction is conducted. • 124 journal articles published in 2000–2021 are manually reviewed. • Bottom-level NLP techniques are thoroughly introduced. • NLP is used to extract and exchange information and to support downstream applications. • The role of NLP in smart construction, current challenges, and future directions are provided. [ABSTRACT FROM AUTHOR]

Recent advancements in computer vision and augmented reality (AR) technology have created a pathway for enhanced human–computer interaction. Despite the potential benefits, the integration of these technologies is not yet common in the industry. Object detection is the foundation of AR systems for inspection, and an efficient object detection model will enable accurate object localization in AR devices. However, the limited computing capabilities and memory usage of wearable AR devices pose a significant challenge to the deployment of state-of-the-art detectors on these devices in real work scenarios. Moreover, as the person wearing the AR device keeps moving, the constantly changing scale of the captured image often leads to significant information loss when feature maps of different scales are fused. To address these challenges, this paper proposes a new compatible detector YOLO-Master based on improved YOLOv5 through deep learning and AR. The paper introduces several improvements, such as using MobileNetv3 as the backbone to reduce memory overhead and computational costs and designing an attention module MCA based on coordinate to reduce information loss. Besides, 20 hyperparameters of YOLOv5 are optimized by the genetic algorithm(GA) before model training, which can help reduces the loss value generated. YOLO-Master was trained and tested on hydropower device images collected from a power plant in the China Southern Power Grid, and the experiments on a self-made dataset demonstrated that the mAP of the proposed method can reach 70.1%, which is 2.9% higher than the baseline of 67.2%. Additionally, the number of parameters of the model was reduced by 66.7%, enabling the model to ensure accuracy while being compatible with real-time performance, meeting the urgent needs of engineering deployment. • A Compatible detector is proposed for device detection in AR inspection system. • Computation cost is improved by a lightweight backbone network. • A coordinate-based attention module is designed for effective feature fusion. • Optimizing hyperparameters of the new architecture to make the detector robust. [ABSTRACT FROM AUTHOR]

Adaptive work packaging is paramount in helping reduce dynamic gaps between design and manufacturing in modular construction (MC), particularly in mass customization. However, current work packaging methods fail to automatically extract complex semantic relations among work package elements (e.g., products, tasks, and their dependencies) and dynamically reason the implicit semantic knowledge (e.g., the different granularity of semantics) as the project progresses. To address these issues, this study proposes a knowledge graph-enabled adaptive work packaging (K-GAWP) approach to dynamically form semantic-enriched work packages with different granularities. Thus far, this study first models the data of tasks, products, and their spatial relationships for MC production as graphs. Second, a novel multi-granularity knowledge reasoning method (product2task) is developed to map products to tasks in an adaptive manner. Third, a dedicated hierarchical clustering method (task2package) involving multiple features from the dependency structure matrix is proposed for work-package generation (i.e., task knowledge fusion). Finally, the K-GAWP's performance is evaluated through controlled experiments in a real MC project. The results indicate that the K-GAWP approach performs work packaging in an adaptive, accurate, and efficient manner, thereby improving the distributed planning and control of MC projects. [ABSTRACT FROM AUTHOR]

The development of the building energy management systems (BEMS) enable users to intelligently control Heating, Ventilation, Air-conditioning and Cooling (HVAC) systems based on digital information. In order to reduce the power consumption cost of the HVAC system while ensuring user satisfaction, a novel HVAC control system for building system based on a multi-step predictive deep reinforcement learning (MSP-DRL) algorithm is proposed in this paper. In the proposed method, the outdoor ambient temperature is predicted first by a featured deep learning method named GC-LSTM, where the Long Short-term Memory (LSTM) is enhanced by the generalized correntropy (GC) loss function to deal with the non-Gaussian characteristics of the collected outdoor temperature. In addition, the proposed temperature prediction model is combined with a reinforcement learning algorithm named Deep Deterministic Policy Gradient (DDPG) aiming to flexibly adjust the output power of the HVAC system under the dynamic changing of electricity prices. Finally, comprehensive simulation based on real world data is delivered. Numerical results show that the GC-LSTM algorithm is more accurate than other counterparts prediction algorithms, and the proposed HVAC control system based on the multi-step prediction deep reinforcement learning algorithm is effective and could save over 12% cost compared to other approaches, where the user comfort is maintained simultaneously. • A multi-step predictive deep reinforcement learning algorithm is proposed. • A GC-LSTM algorithm is adopted in multi-step outdoor temperature prediction. • A model-free DDPG algorithm is adopted to manage HVAC system for building system. • The proposed method performs well in the HVAC system. [ABSTRACT FROM AUTHOR]