Search

Rubidium (Rb) and strontium (Sr) contents and ratios were analyzed in 197 sediment samples collected from the CGS1 segment of the Chagelebulu Section in the Badain Jaran Desert of China to study millennial scale climate change during the Holocene. The results showed that the Rb and Sr contents and Rb/Sr ratios were low in the samples of dune sands and loess (Mz < 5.64Φ), and those values were high in the samples of loess (Mz > 5.64Φ) and paleosols, these data displayed 11 changing cycles in alternation of peaks and valleys in the chart. In addition, the Rb contents were positively correlated with Mz (mean particle diameter) and clay contents. While the correlations were weaker, Sr contents also showed a tendency to increase with increases in the Mz and clay contents. Based on a comprehensive analysis of the distribution of Rb and Sr within the CGS1 segment, it appears that the observed Rb and Sr contents and ratios varied in accordance with fluctuations of the East Asian winter and summer monsoons. In terms of timing and climate, low values (C1–C11) resulting from winter monsoons had a strong correlation with cold events in the North Atlantic: the period C1 corresponded to times ranging from 400 a to 1400 a and the periods C2, C3, C4, C7, C9, C10, and C11 corresponded to times of 3000 a, 4000 a, 5900 a, 8200 a, 9400 a, 10,300 a, and 11,000 a, respectively. The cold event C5 (6200 a) was also discovered in the North Atlantic in recent; and C6 (7100 a), C8 (8700 a) were discovered in some other places of China. These cold events indicated by Rb and Sr contents and ratios in the Chagelebulu Section of the Badain Jaran Desert recorded the regional response of global climate change during the Holocene. [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]

Abstract: By sensitizing with 514nm green light, 488nm blue light and 390nm ultraviolet light, respectively, recording with 633nm red light, effect of wavelength of sensitizing light on holographic storage properties in LiNbO3:Fe:Ni crystal is investigated in detail. It is shown that by shortening the wavelength of sensitizing light gradually, nonvolatile holographic recording properties of oxidized LiNbO3:Fe:Ni crystal is optimized gradually, 390nm ultraviolet light is the best as the sensitizing light. Considering the absorption of sensitizing light, to obtain the best performance in two-center holographic recording we must choose a sensitizing wavelength that is long enough to prevent unwanted absorptions (band-to-band, etc.) and short enough to result in efficient sensitization from the deep traps. So in practice a trade-off is always needed. Explanation is presented theoretically. [Copyright &y& Elsevier]

Accurate estimations of battery state-of-charge (SOC) for energy storage systems are popular research topics in recent years. Numerous challenges remain in several aspects, especially in dealing with the conflict of high model accuracy and complex model structure with heavy computational cost. This paper proposes a compact and optimized SOC estimation model, integrating a fast input selection algorithm to choose important terms as input variables, followed by a simple and efficient JAYA optimization scheme to tune the key parameters of neural network functions. From the real-system experiment results, it can be seen that the estimation model errors are greatly reduced by applying optimization method, and the model performance is validated through statistical error values including root mean square error, mean absolute error, mean absolute percentage error and SOC error. The experimental results demonstrate that the SOC estimations can be greatly improved after optimization of neural network parameters under different charging and discharging process. • A novel compact and optimized radial basis function neural network is proposed. • The method is adopted for the SOC estimation of energy storage systems. • Simulation results validate the excellent performance of the proposed method. [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]

Abstract Conventional unit commitment is a mixed integer optimization problem and has long been a key issue for power system operators. The complexity of this problem has increased in recent years given the emergence of new participants such as large penetration of plug-in electric vehicles. In this paper, a new model is established for simultaneously considering the day-ahead hourly based power system scheduling and a significant number of plug-in electric vehicles charging and discharging behaviours. For solving the problem, a novel hybrid mixed coding meta-heuristic algorithm is proposed, where V-shape symmetric transfer functions based binary particle swarm optimization are employed. The impact of transfer functions utilised in binary optimization on solving unit commitment and plug-in electric vehicle integration are investigated in a 10 unit power system with 50,000 plug-in electric vehicles. In addition, two unidirectional modes including grid to vehicle and vehicle to grid, as well as a bi-directional mode combining plug-in electric vehicle charging and discharging are comparatively examined. The numerical results show that the novel symmetric transfer function based optimization algorithm demonstrates competitive performance in reducing the fossil fuel cost and increasing the scheduling flexibility of plug-in electric vehicles in three intelligent scheduling modes. Highlights • A new unit commitment model considering electric vehicles is established. • A binary symmetric based hybrid meta-heuristic method is proposed. • The impact of transfer function on unit commitment problem is evaluated. • Three flexible scheduling modes are comparatively studied. [ABSTRACT FROM AUTHOR]

Highlights • A new compact teaching-learning based optimization method is proposed. • 32 well-known benchmark tests are conducted for algorithm evaluation. • The algorithm maintains high performance and reduce memory requirement. • The algorithm is used to train two typical function neural networks. Abstract The majority of embedded systems are designed for specific applications, often associated with limited hardware resources in order to meet various and sometime conflicting requirements such as cost, speed, size and performance. Advanced intelligent heuristic optimization algorithms have been widely used in solving engineering problems. However, they might not be applicable to embedded systems, which often have extremely limited memory size. In this paper, a new compact teaching-learning based optimization method for solving global continuous problems is proposed, particularly aiming for neural network training in portable artificial intelligent (AI) devices. Comprehensive numerical experiments on benchmark problems and the training of two popular neural network systems verify that the new compact algorithm is capable of maintaining the high performance while the memory requirement is significantly reduced. It offers a promising tool for continuous optimization problems including the training of neural networks for intelligent embedded systems with limited memory resources. [ABSTRACT FROM AUTHOR]

The optimization of maintenance scheduling and routing in offshore wind farms is crucial for the intelligent operation and maintenance of offshore wind energy. It involves determining the optimal timing, assigning vessels, planning routes, and efficiently completing maintenance tasks with maximum efficiency and minimal cost. Traditional branch-and-bound search methods, often used in multi-vessel cooperative maintenance scheduling models, face limitations in finding optimal solutions efficiently, while classic heuristic methods may compromise on global optimality. This paper introduces DivideMerge, a high-performance hybrid algorithm that synergizes heuristic and exact algorithms to address large-scale multi-vessel cooperative maintenance scheduling challenges in offshore wind farms. Initially, a maintenance task constraint decomposition heuristic method is utilized to break down the collaborative optimization scheduling problem into individual vessel scheduling sub-problems, ensuring adherence to the constraints of the original problem. Subsequently, the CP-SAT solver is employed to sequentially solve these sub-problems rapidly. The solutions for individual vessel scheduling are then merged to form a comprehensive solution for the multi-vessel cooperative maintenance scheduling problem. Computational results validate the effectiveness and robustness of DivideMerge, demonstrating a solution speed nearly 1000 times faster than the commercial Gurobi solver, thus offering a significant advancement in the field of offshore wind farm maintenance optimization. [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]

Abstract: Microarc oxidation coatings were fabricated on AZ31 magnesium alloy in the electrolyte of sodium phosphate. Potentiodynamic polarization and electrochemical impedance spectroscopy tests were employed to investigate the electrochemical corrosion behavior. The corroded surface was characterized by an optical microscope and X-ray diffraction. The influence of chloride ion concentration on the corrosion resistance of microarc oxidation coated AZ31 alloy is discussed. The corrosion current density enlarged ratio and the charge transfer resistance reduced ratio indicated that the extent of the corrosion damage of microarc oxidation coated AZ31 alloy is much higher when chloride ion concentration is greater than 5%. A corrosion mechanism related to the effect of chloride ion concentrations on the corrosion behavior is proposed. [Copyright &y& Elsevier]

• A new unit commitment framework with different levels of DSM for PEVs is proposed. • A novel binary competitive swarm optimization method is proposed and utilized. • The uncoordinated and coordinated charging is measured by a weighting factor. • Numerical studies on various unit scales and levels of load are investigated. Decreasing initial costs, the increased availability of charging infrastructure and favorable policy measures have resulted in the recent surge in plug-in electric vehicle (PEV) ownerships. PEV adoption increases electricity consumption from the grid that could either exacerbate electricity supply shortages or smooth demand curves. The optimal coordination and commitment of power generation units while ensuring wider access of PEVs to the grid are, therefore, important to reduce the cost and environmental pollution from thermal power generation systems, and to transition to a smarter grid. However, flexible demand side management (DSM) considering the stochastic charging behavior of PEVs adds new challenges to the complex power system optimization, and makes existing mathematical approaches ineffective. In this research, a novel parallel competitive swarm optimization algorithm is developed for solving large-scale unit commitment (UC) problems with mixed-integer variables and multiple constraints – typically found in PEV integrated grids. The parallel optimization framework combines binary and real-valued competitive swarm optimizers for solving the UC problem and demand side management of PEVs simultaneously. Numerical case studies have been conducted with multiple scales of unit numbers and various demand side management strategies of plug-in electric vehicles. The results show superior performance of proposed parallel competitive swarm optimization based method in successfully solving the proposed complex optimization problem. The flexible demand side management strategies of plug-in electric vehicles have shown large potentials in bringing considerable economic benefit. [ABSTRACT FROM AUTHOR]

• The sensor can detect H 2 S gas with a concentration as low as 0.5 ppm. • The frequency shift of sensor is derived from the change of mass loading on films. • The responses became much faster and stronger with the increase of RH. Surface acoustic wave (SAW) sensors based on sol–gel processed porous and nanoparticulate CuO films were explored for H 2 S detection operated at room temperature. The SAW sensor showed a negative frequency shift when exposed to H 2 S gas, due to the mass loading effect on the sensitive CuO film. It has a high sensitivity of H 2 S, a detection limit of 500 ppb and a good selectivity to H 2 S gas compared with H 2 , C 2 H 5 OH, acetone, NH 3 , CO, NO, NO 2 gases. Relative humidity (RH) was identified to have a critical influence on the sensing performance of the sensor. The responses became much faster and stronger with the increase of RH. The significant enhancement in the sensing performance is attributed to that the chemical reaction between H 2 S and CuO is hindered and the physical adsorption of CuO to H 2 S is enhanced. [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]

With the application of smart meters, more information is available from residential buildings for support heat load forecast. Yet, there is still a lack of an effective method to exploit the value of the high spatial granularity information, particularly for residential communities with high randomness in human behaviors. To fill this gap, this paper proposes a data-driven heat load forecast method based on the field measurements of smart meters A white-box machine learning algorithm, namely Dendritic Network, is employed to aggregate and analyze data obtained from different locations of district heating systems. An online correction mechanism is then proposed to regulate the forecast horizon and ensure the adaptiveness of the machine learning model under different scenarios. The results demonstrate that the dendritic network employed in the proposed machine learning method shows higher forecast accuracy, the root-mean-square error is 0.0029, in some research cases the coefficient of determination can reach 0.99–1, and wide scope of application in the area of heat load forecast. • Heat load prediction model is used for residential communities in central heating. • We used and compared 5machine learning prediction methods. • An online correction mechanism is proposed to ensure the accuracy. • The determination coefficient of modeling can reach more than 0.99 in some cases. [ABSTRACT FROM AUTHOR]

A Plate–fin heat exchanger (PFHE) is a compact and efficient thermal device, whose performance strongly depends on its structural design. However, the design optimization of a PFHE is a mixed-integer optimization problem with a strong nonlinear characteristic, which presents significant challenges for existing optimization algorithms. Meta-heuristic algorithms (MAs) are competitive for solving complex non-linear optimization problems. In this paper, an improved dynamic-opposite learning Jaya (DOLJaya) method, the goal is to make the algorithm adaptable to each problem. The results of eighteen unimodal and multi-modal benchmarks and nine hybrid benchmarks demonstrate that the proposed DOLJaya has competitive robustness, efficiency and effectiveness for solving complex nonlinear problems compared to its popular counterparts. At the same time, we selected the optimization of the plate–fin heat exchanger as the industrial test benchmark for optimization, and the results of DOLJaya algorithm have been improved by a maximum average of 108.29% and 7.60% compared with the original Jaya, which are also satisfactory. [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]

Abstract: This paper addresses the long-term corrosion behavior of microarc oxidation coated Mg alloys immersed in simulated body fluid for 28days. The coatings on AZ31 Mg alloys were produced in the electrolyte of sodium phosphate (Na3PO4) at the concentration of 20g/L, 30g/L and 40g/L, respectively. Scanning electron microscope (SEM) and optical micrograph were used to observe the microstructure of the samples before and after corrosion. The composition of the MAO coating and corrosion products were determined by X-Ray Diffraction (XRD). Corrosion product identification showed that hydroxyapatite (HA) was formed on the surface of the corroded samples. The ratio of Ca/P in HA determined by the X-ray Fluorescence (XRF) technique showed that HA is an acceptable biocompatible implant material. The potentiodynamic polarization and electrochemical impedance spectroscopy (EIS) were employed to characterize the corrosion rate and the electrochemical impedance. The corrosion resistance of the coated Mg alloys can be enhanced by optimizing the electrolyte concentrations for fabricating samples, and is enhanced after immersing the coated samples in simulated body fluid for more than 14days. The enhanced corrosion resistance after long-term immersion is attributed to a corrosion product layer formed on the sample surface. The inhibition mechanism of the corrosion process is discussed and presented with an animation. [Copyright &y& Elsevier]

Abstract: In this paper, we have investigated the grating erasure of a reduced LiNbO3:Fe crystal with different erasing wavelengths. The overall hologram evolution in the process of grating erasure is nonexponential due to strong absorption which is contrary to the mono-exponential law. The hologram in the rear part of the crystal can persist for a long time in the grating erasure due to weak erasing light intensity by strong absorption, which can enlarge the erasure time constant. From the erasure experiments, the global absorption can be taken as the optimum absorption to acquire a good trade-off between the sensitivity and hologram strength in the crystal. [Copyright &y& Elsevier]

Abstract: We have investigated ultraviolet (UV) photorefractive effect of lithium niobate doubly doped with Ce and Cu. It is found the diffraction efficiency shows oscillating behavior under UV-light-recording. A model in which electrons and holes can be excited from impurity centers in the UV region is proposed to study the oscillatory behavior of the diffraction efficiency. On the basis of the material equations and the coupled-wave equations, we found that the oscillatory behavior is due to the oscillation of the relative spatial phase shift Φ. And the electron–hole competition may cause the oscillation of the relative spatial phase shift. A switch point from electron grating to hole grating is chosen to realize nonvolatile readout by a red light with high sensitivity (0.4cm/J). [Copyright &y& Elsevier]