Search

Online anomaly detection is critical for industrial safety and security monitoring but is facing challenges due to the complexity of evolving data streams from working conditions and performance degradation. Unfortunately, existing approaches fall short of such challenges, and these models may be disabled, suffering from the evolving data distribution. The paper presents a framework for online anomaly detection of data streams, of which the baseline algorithm is the incremental learning method of Growing Neural Gas (GNG). It handles complex and evolving data streams via the proposed model Bayes-Optimized Adaptive Growing Neural Gas (BOA-GNG). Firstly, novel learning rate adjustment and neuron addition strategies are designed to enhance the model convergence and data presentation capability. Then, the Bayesian algorithm is adopted to realize the fine-grained search of BOA-GNG-based hyperparameters. Finally, comprehensive studies with six data sets verify the superiority of BOA-GNG in terms of detection accuracy and computational efficiency. [ABSTRACT FROM AUTHOR]

Irrigation with water containing a variety of microcystins (MCs) may pose a potential threat to the normal growth of agricultural plants. To investigate the phytotoxicity of MC-LR at environmental concentrations on rice (Oryza sativa L.), the characteristics of uptake and accumulation in plant tissues, as well as a series of key physio-biochemical process changes in leaves of rice seedlings, were measured at concentrations of 0.10, 1.0, 10.0, and 50.0 μg·L−1 in hydroponic nutrient solutions for 7, 15, 20, and 34 days. Results showed that MC-LR could be detected in rice leaves and roots in exposure groups; however, a significant accumulation trend of MC-LR in plants (BCF > 1) was only found in the 0.10 μg·L−1 group. The time-course study revealed a biphasic response of O2•− levels in rice leaves to the exposure of MC-LR, which could be attributed to the combined effects of the antioxidant system and detoxification reaction in rice. Exposure to 1.0–50.0 μg·L−1 MC-LR resulted in significant depletion of GSH and MDA contents in rice leaves at later exposure times (15–34 days). Low MC-LR concentrations promoted nitric oxide synthase (NOS) activity, whereas high concentrations inhibited NOS activity during the later exposure times. The reduced sucrose synthase (SS) activities in rice exposed to MC-LR for 34 days indicated a decrease in the carbon accumulation ability of plants, and therefore may be directly related to the inhibition of plant growth under MC exposure. These findings indicate that the normal physiological status would be disrupted in terrestrial plants, even under exposure to low concentrations of MC-LR. [ABSTRACT FROM AUTHOR]

Existing space application networks in space stations are mainly fiber-optic cable-based networks due to their low size, weight, and power (SWaP) values. While fiber networks in space stations offer data transmission at high speeds with minimal signal loss, their major disadvantage is the lack of flexibility and mobility when new and unplanned space scientific equipment is added to the network. To enhance the flexibility of space networks while increasing their throughput, this paper introduces the hybrid space network (HSN), a new space network architecture that incorporates an optical wireless link, to meet the ever-increasing demands for larger bandwidth and higher mobile access capabilities in space scientific experiments. To best utilize the HSN's system performance, we propose a multi-priority-based network scheduling scheme, which can dynamically adapt to the requirements of mass tasks and select the best transmission procedure. Through simulations, we find that by adding optical wireless communication (OWC) links to the state-of-the-art deterministic FC-AE-1553 space network, the HSN's bandwidth can be increased by 20 times with an average latency reduction of 87.3%. We believe that the proposed HSN's architecture may ultimately shape the future of space stations' wireless connectivity, and in the meantime, innovate many advanced space applications with larger data rates and mobility requirements. [ABSTRACT FROM AUTHOR]

Measuring the similarity between two trajectories is fundamental and essential for the similarity-based remaining useful life (RUL) prediction. Most previous methods do not adequately account for the epistemic uncertainty caused by asynchronous sampling, while others have strong assumption constraints, such as limiting the positional deviation of sampling points to a fixed threshold, which biases the results considerably. To address the issue, an uncertain ellipse model based on the uncertain theory is proposed to model the location of sampling points as an observation drawn from an uncertain distribution. Based on this, we propose a novel and effective similarity measure metric for any two degradation trajectories. Then, the Stacked Denoising Autoencoder (SDA) model is proposed for RUL prediction, in which the models can be first trained on the most similar degradation data and then fine-tuned by the target dataset. Experimental results show that the predictive performance of the new method is superior to prior methods based on edit distance on real sequence (EDR), longest common subsequence (LCSS), or dynamic time warping (DTW) and is more robust at different sampling rates. [ABSTRACT FROM AUTHOR]

The TESS Survey has released a large number of high-precision light curves of contact binaries. However, using the Phoebe program and Markov chain Monte Carlo (MCMC) algorithm to obtain the posterior distribution of contact binary parameters is a time-consuming process. In order to obtain the contact binary parameters from the TESS survey, we build neural network (NN) models and combine them with the MCMC algorithm to obtain the contact binary parameters and parameter errors quickly. NN model is used in place of the physical model, which can generate a light curve with a precision of less than a millimagnitude. The NN model is capable of generating light curves at a speed that is four orders of magnitude faster than Phoebe running on the same computing platform. In this study, we have determined the parameters of 318 contact binary systems exhibiting relatively symmetric light curves. Subsequently, a statistical analysis was conducted on the derived parameters of these 318 targets. The coefficient of determination (R 2) for 318 contact binaries between the light curves generated by Phoebe using these parameters obtained by the NN model and MCMC as inputs and the original light curves is greater than 0.99. Additionally, the distribution and correlation of the parameters for these 318 contact binary systems have been presented. [ABSTRACT FROM AUTHOR]

Asbestos has been confirmed as a major pollutant in asbestos-mining areas that are located in western China. In general, asbestos-fibre dust will is released into the environment due to the effect of intensive industrial activities and improper environmental management, such that the health of residents in and around mining areas is jeopardised. A typical asbestos mining area served as an example in this study to analyse the content and fibre morphology of asbestos in soil and air samples in the mining area. The effects of asbestos pollution in and around the mining areas on human health were also assessed based on the U.S. Superfund Risk Assessment Framework in this study. As indicated by the results, different degrees of asbestos pollutions were present in the soil and air, and they were mainly concentrated in the mining area, the ore-dressing area, and the waste pile. The concentration of asbestos in the soil ranged from 0.3% to 91.92%, and the concentration of asbestos fibres in the air reached 0.008–0.145 f·cc−1. The results of the scanning-electron microscope (SEM) energy suggested that the asbestos was primarily strip-shaped, short columnar, and granular, and the asbestos morphology of the soils with higher degrees of pollution exhibited irregular strip-shaped fibre agglomeration. The excess lifetime cancer risk (ELCR) associated with the asbestos fibres in the air of the mining area was at an acceptable level (10−4–10−6), and 40.6% of the monitoring sites were subjected to unacceptable non-carcinogenic risks (HQ > 1). Moreover, the waste pile was the area with the highest non-carcinogenic risk, followed by the ore dressing area, a residential area, and a bare-land area in descending order. In the three scenarios of adult offices or residences in the mining area, adults' outdoor activities in the peripheral residence areas, and children's outdoor activities, the carcinogenic-and non-carcinogenic-risk-control values in the air reached 0.1438, 0.2225 and 0.1540 f·cc−1, and 0.0084, 0.0090 and 0.0090 f·cc−1, respectively. The results of this study will lay a scientific basis for the environmental management and governance of asbestos polluted sites in China. [ABSTRACT FROM AUTHOR]

Methods for fault diagnosis based on metric learning, in which a query sample is classified by picking the closest prototype from the support set based on their feature similarities, have been the subject of many studies. In real-world applications of in-orbit products, such as circulating pumps, the computation of similarity between different pairs is prone to different degrees of inaccuracy, especially epistemic uncertainty. Knowing and considering the uncertainty of similarity may improve fault detection accuracy. This article provides a unique approach to fault diagnosis based on Prototypical Network (Pro-Net) and Uncertainty Theory. In particular, we use epistemic uncertainty by altering the representation of prototypes from a deterministic scalar to an uncertain representation. To assess the similarity between a query and the prototypes in a support set, we calculate the uncertain distance between the pairs using cross-entropy. Experiments with symmetrical structures reveal that our proposed method significantly enhances classification precision and achieves state-of-the-art performance. It improves the reliability of fault diagnosis and reduces the risk of making erroneous judgments in safety-critical systems, decreasing the possibility of adverse consequences. [ABSTRACT FROM AUTHOR]

As complex systems composed of physical and cyber components, mechanically pumped loop systems (MPLs) are vulnerable to both passive threats (e.g., physical failures) and active threats such as cyber-attacks launched on the network control systems. The impact of the aforementioned two threats on MPL operations is yet unknown, and there is no practical way to evaluate their severity. To assess the severity of the impact of physical failures and cyber-attacks on MPLs, a safety impact analysis framework based on Elman Neural Network (ENN) observers and the Gaussian Mixture Model (GMM) algorithm is suggested. The framework discusses three common attack and failure modes: sensor hard failure that occurs suddenly, sensor soft failure that occurs gradually over time, and denial-of-service (DoS) attacks that prevent communication between the controller and valve. Both sensor failures and DoS attacks render the system unsafe, according to simulation data. In comparison to DoS attacks, however, sensor failures, particularly soft failures, inflict the greatest harm to the MPLs. Furthermore, sensors engaged in global control, rather than those involved in local control, need additional protection. [ABSTRACT FROM AUTHOR]

Complicated factors in urban areas have been reported to impact the density, biomass, and diversity of earthworm fauna. Urban parks provide essential habitats for earthworm fauna. However, how earthworm fauna are impacted by park traits, such as construction age, distance to city center, visitor volumes, sizes of greenspaces/parks, and attractiveness, etc., still remains unknown. These traits are well characterized by the impacts of urbanization intensity and administration quality of parks in megacities. Therefore, 16 parks with gradients of construction ages and geographical locations in Beijing city were selected for investigation. Furthermore, an earthworm synthetic and simple index (ESI) for characterizing earthworm community has been developed to compensate for the lack of robustness by using single ecological indexes. The results showed that earthworm population density (38.6 ind/m2) and biomass (34.0 g/m2) in parks were comparable to those in other land use types in Beijing. Ecological groupings were dominated by disturbance-tolerant endogeic and deep soil-inhabiting anecic groups, and most of them were adults. The earthworm population was influenced by urbanization intensity, while the earthworm community composition, species biodiversity, and ESI were affected by administration quality of parks. The soil moisture and microbial biomass carbon were the key factors in shaping earthworm assemblages. ESI could be employed as an effective indicator in depicting character of earthworm fauna. This study highlighted the impacts of park traits on earthworms in urban parks. The variation in park traits that influence earthworm fauna was probably attributed to soil properties. [ABSTRACT FROM AUTHOR]

This paper investigates the interaction between groundwater and groundwater-dominated Crescent Spring in an aeolian sand environment to determine the drainage source of groundwater for the lake using hydraulic gradient and isotope analysis methods. The precipitation is found to have an insignificant effect on the lake in an arid climate environment. Hydrogen and oxygen isotopic evidence presents at least two aquifer contributions that recharge 87% and 13% water supply of the lake, respectively. The flow direction of groundwater is theoretically established based on the hydraulic gradient and isotope analysis methods. [ABSTRACT FROM PUBLISHER]

Feature point matching is a critical step in feature-based image registration. In this letter, a highly robust feature-point-matching algorithm is proposed, which is based on the feature point descriptor calculated by the triangle-area representation (TAR) of the K nearest neighbors (KNN-TAR). The affine invariant descriptor KNN-TAR is used to find the candidate outliers, and then, the real outliers will be removed by the local structure and global information. The experimental results show that the proposed method can remove the outliers from the initial matching result even when the outliers are of high proportion. Compared with graph transformation matching and restricted spatial-order constraints, KNN-TAR outperforms these methods with higher stability and precision. [ABSTRACT FROM PUBLISHER]

Despite hard sensors can be easily used in various condition monitoring of energy production process, soft sensors are confined to some specific scenarios due to difficulty installation requirements and complex work conditions. However, industrial process may refer to complex control and operation, the extraction of relevant information from abundant sensors data may be challenging, and description of complicated process data patterns is also becoming a hot topic in soft-sensor development. In this paper, a hybrid soft sensor model based mechanism analysis and data-driven is proposed, and ventilation sensing of coal mill in a power plant is conducted as a case study. Firstly, mechanism model of ventilation is established via mass and energy conservation law, and object-relevant features are identified as the inputs of data-driven method. Secondly, radial basis function neural network (RBFNN) is used for soft sensor modeling, and genetic algorithm (GA) is adopted for quick and accurate determination of the RBFNN hyper-parameters, thus self-adaptive RBFNN (SA-RBFNN) is proposed to improve the soft sensor performance in energy production process. Finally, effectiveness of the proposed method is verified on a real-world power plant dataset, taking coal mill ventilation soft sensing as a case study. [ABSTRACT FROM AUTHOR]