Search

Meningitis is a life-threatening disease. In order to reduce its threat to public health, the World Health Assembly indorsed a resolution in 2020 for urgent global action to prevent and control meningitis. Defeating Meningitis by 2030: a Global Roadmap was officially launched by the World Health Organization in 2021. We interpreted some key information of the roadmap from the aspects of coverage, objectives and pillar strategies, providing ideas for further strengthening the prevention and control of bacterial meningitis in China.脑膜炎是严重威胁生命的疾病。为降低脑膜炎对于全球公共卫生的威胁，2020年世界卫生大会通过了全球采取紧急行动预防和控制脑膜炎的决议（WHA73.9号决议）。2021年世界卫生组织在全球正式发布了《到 2030年战胜脑膜炎：全球路线图》。本文从涵盖范围、目标以及支柱性策略几方面对全球路线图的部分重点信息进行解读，为我国进一步加强对细菌性脑膜炎的防控提供思路。.

Effective fault diagnosis for power electronic converters is an essential and mandatory operation to reduce failures and unscheduled shutdowns. This paper presents a data-driven fault diagnosis method, using long short-term memory (LSTM) network, to detect multiple open-circuit switch faults of the back-to-back converter in doubly fed induction generator (DFIG)-based wind turbine systems. Twelve sensor signals of the back-to-back converter for different fault types are measured. Wind speed fluctuation and sensor bias faults are considered as interference factors. The method has been evaluated in a grid-connected DFIG simulink model. Simulation results have demonstrated that, compared with least squares support vector machine (LS-SVM), back-propagation artificial neural network (BPANN) and recurrent neural network (RNN), the proposed method excavates the deep information of the fault signal with the highest diagnosis accuracy and strongest robustness with short time delay. Experimental tests undertaken on a hardware-in-the-loop testing platform further validate the effectiveness of the LSTM-based network.

The accuracy of the simulation model has a profound impact on the optimal operation of the energy hubs (EHs). However, in many articles, the constant model of the efficiency of equipment is adopted to formulate the operation system, which would probably lead to a simplification of the simulation models. But, EHs are typically operated under off-design condition due to the fluctuations in cooling, heating, electricity requirement. Moreover, even though the off-design characteristics are considered, few studies have suggested comparing the differences between those two models by considering the operation cost. In order to assess the effect of the off-design characteristics of EH on the optimal operation accuracy in this paper, two test cases are performed on the fixed and variable load conditions, respectively. In addition, the individual effect of off-design characteristics of each equipment on the optimal operation cost of the EH is also investigated through four optimization runs. It is worth mentioning that the optimal operation problem of the EH considering the off-design characteristics and on-off status of the equipment is a mixed integer non-linear programming problem (MINLP). By testing the design and off-design models on the two cases, the results of simulation demonstrate that the optimal operation cost for the off-design model is larger than that for the design model. Nonetheless, in the aspect of the authenticity of the system operation strategy, the off-design model performs better than the design model. Furthermore, a larger relative error of the system operation cost between the two models can be observed when the EH is operated under a relatively lower load condition, revealing that the influence of off-design characteristic on the optimal operation of EHs is too significant to be neglected.

With the promotion and application of smart grid, the technology of non-intrusive load monitoring (NILM) has gained more and more attention in recent years. Different from direct device monitoring, it identifies the type of appliances using the aggregated load profile measured at a single metering point, which is more convenient and flexible, and thus has the potential to be extended to all households equipped with smart meters. This paper proposes an extensible and comprehensive model to solve the load disaggregation problem, which employs the additive factorial approximate maximum a posteriori (AFAMAP) based on iterative fuzzy ${c}$ -means (IFCM). To make sure that the model is adaptive to other households, the hidden Markov models (HMMs) are applied to obtain the independent load model of each appliance, and the IFCM is used to determine the number of hidden states adaptively. Finally, the AFAMAP is utilized to decompose the aggregated power consumption based on the independent load models built by HMM. Simulation studies are conducted on the open database of Almanac of Minutely Power dataset (AMPds), and the results have demonstrated that the proposed model is more accurate in comparison with the other models.

We report the control of Rashba spin-orbit interaction by tuning asymmetric hybridization between Ti orbitals at the LaAlO_{3}/SrTiO_{3} interface. This asymmetric orbital hybridization is modulated by introducing a LaFeO_{3} layer between LaAlO_{3} and SrTiO_{3}, which alters the Ti-O lattice polarization and traps interfacial charge carriers, resulting in a large Rashba spin-orbit effect at the interface in the absence of an external bias. This observation is verified through high-resolution electron microscopy, magnetotransport and first-principles calculations. Our results open hitherto unexplored avenues of controlling Rashba interaction to design next-generation spin orbitronics.

Mathematical expression recognition (MER) aims to convert an image of mathematical expressions into a Latex sequence. In practice, the task of MER is challenging, since 1) the images of mathematical expressions often contain complex structure relationships, e.g., fractions, matrixes, and subscripts; 2) the generated Latex sequences can be very complex and they have to satisfy strict syntax rules. Existing methods, however, often ignore the complex dependence among image regions, resulting in poor feature representation. In addition, they may fail to capture the rigorous relations among different formula symbols as they consider MER as a common language generation task. To address these issues, we propose a Structure-Aware Sequence-Level (SASL) model for MER. First, to better represent and recognize the visual content of formula images, we propose a structure-aware module to capture the relationship among different symbols. Meanwhile, the sequence-level modeling helps the model to concentrate on the generation of entire sequences. To make the problem feasible, we cast the generation problem into a Markov decision process (MDP) and seek to learn a Latex sequence generating policy. Based on MDP, we learn SASL by maximizing the matching score of each image-sequence pair to obtain the generation policy. Extensive experiments on the IM2LATEX-100K dataset verify the effectiveness and superiority of the proposed method.

One of the key steps in Neural Architecture Search (NAS) is to estimate the performance of candidate architectures. Existing methods either directly use the validation performance or learn a predictor to estimate the performance. However, these methods can be either computationally expensive or very inaccurate, which may severely affect the search efficiency and performance. Moreover, as it is very difficult to annotate architectures with accurate performance on specific tasks, learning a promising performance predictor is often non-trivial due to the lack of labeled data. In this paper, we argue that it may not be necessary to estimate the absolute performance for NAS. On the contrary, we may need only to understand whether an architecture is better than a baseline one. However, how to exploit this comparison information as the reward and how to well use the limited labeled data remains two great challenges. In this paper, we propose a novel Contrastive Neural Architecture Search (CTNAS) method which performs architecture search by taking the comparison results between architectures as the reward. Specifically, we design and learn a Neural Architecture Comparator (NAC) to compute the probability of candidate architectures being better than a baseline one. Moreover, we present a baseline updating scheme to improve the baseline iteratively in a curriculum learning manner. More critically, we theoretically show that learning NAC is equivalent to optimizing the ranking over architectures. Extensive experiments in three search spaces demonstrate the superiority of our CTNAS over existing methods., Accpeted by CVPR 2021. The code is available at https://github.com/chenyaofo/CTNAS

This paper proposes a forecast model for ultra-short-term prediction of wind speed and wind power, which is based on a morphological high-frequency filter (MHF) and a double similarity search (DSS) algorithm. The MHF is proposed to decompose the time series into two components: the mean trend, which reveals the non-stationary tendency of the time series, and the high frequency component, which depicts the fluctuations. The same strategy is employed to forecast the mean trend and the high frequency component, respectively. The two components are reconstructed in the phase space, respectively, where a non-uniform embedding strategy is proposed to better reveal their information. To select similar segments to be used for local forecast, the novel DSS algorithm is proposed for high frequency component, while the Euclidean distance is used for the mean trend. Finally, the least squares-support vector machine (LS-SVM) model is applied to forecast each component, respectively, and their sum composes the final prediction. Simulation studies are carried out using wind speed and wind power data obtained from four databases, and the results demonstrate that the MHF/DSS model provides more accurate and stable forecast compared to the other methods.

In order to ensure the safety of equipment and personnel in power system operation, power utilities make a lot of efforts to prevent the occurrence of accidents. It is very beneficial to investigate and learn from the past accident and near miss narratives. As the number of text records related to accidents keeps increasing every year, it will be difficult and time-consuming to analyse these data by manual. In this paper, an improved BiLSTM neural network based method is proposed to realize the automatic classification of the causes behind the accidents. The proposed model consists of embedding layer, encoder and decoder, which are further improved by introduced the mechanism of attention, dropout, layer normalization and residual connection. Comparison studies are conducted on different models, including textCNN, BiLSTM and BiLSTM-based improved models. The results proved the effectiveness and superiority of the proposed model.

The Gravitational Wave highly energetic Electromagnetic Counterpart All-sky Monitor (GECAM) is dedicated to detecting gravitational wave gamma-ray bursts. It is capable of all-sky monitoring over and discovering gamma-ray bursts and new radiation phenomena. GECAM consists of two microsatellites, each equipped with 8 charged particle detectors (CPDs) and 25 gamma-ray detectors (GRDs). The CPD is used to measure charged particles in the space environment, monitor energy and flow intensity changes, and identify between gamma-ray bursts and space charged particle events in conjunction with GRD. CPD uses plastic scintillator as the sensitive material for detection, silicon photomultiplier (SiPM) array as the optically readable device, and the inlaid Am-241 radioactive source as the onboard calibration means. In this paper, we will present the working principle, physical design, functional implementation and preliminary performance test results of the CPD., Comment: accepted to RDTM

With the integration of alternative energy and renewables, the issue of stability and resilience of the power network has received considerable attention. The basic step of fault diagnosis and isolation is fault identification and location. The conventional intelligent fault identification method is supervised, where the characteristics are labelled manually and its hunger to large amount of labelled data. In order to enhance the ability of the intelligent method and get rid of the dependence on a large amount of labelled data, this paper investigates a novel fault identification method based on deep reinforcement learning (DRL), which has not received enough attention in the field of fault identification. The proposed method considers different faults as a kind of parameters of the model so as to expand the scope of fault identification. In addition, the DRL algorithm is able to intelligently modify the fault parameters according to the observations obtained from the environment of power network rather than tuning the parameters manually and mechanically. The methodology is tested on the IEEE 14 bus for several scenarios and the performance of the proposed method is compared with that of population-based optimization methods and supervised learning methods. The obtained results have confirmed the feasibility and effectiveness of proposed method.

After years of operation of the power system, a large amount of text data has been accumulated, and it is particularly important to analyze them such as the violation data. In this context, this paper introduced a novel classification method, Improving Naive Bayes Based on Improving Term Frequency-Inverse Document Frequency (ITF-IDF), which aims to categorize the text and reduce the costs of labor analysis. The classification of the violation data which including personal behavior, instrument, security activities, supervision and two-ticket data. To increase the classification accuracy, the proposed method improved the weight of Naive Bayes, namely ITF-TDF. In the experimental studies, the Improving Naive Bayes is evaluated on the test data of spam message which is a binary classification and the violation data from the power system as multi-classification, and is compared with the classifiers based on conventional Naive Bayes, the Logistic Regression and the Support Vector Machine (SVM), respectively. The results demonstrate that the proposed method has a better performance than the other methods.

On March 11, 2020, WHO officially declared that COVID-19 had become Pandemic. As of March 31, the epidemic had affected more than 178 countries and regions, with more than 780 000 confirmed cases. The Pandemic Influenza Preparedness Framework for the sharing of influenza viruses and access to vaccines and other benefits (the 'PIP Framework' or 'Framework') is an international arrangement adopted by the World Health Organization in May 2011 to improve global pandemic influenza preparedness and response. Since the transmission route and transmission capacity of COVID-19 are similar to that of influenza A (H1N1) in 2009, which conforms to the basic elements of "human pandemic", and the epidemic scale has exceeded that of influenza A (H1N1)， it is probable to incorporate COVID-19 epidemic response into PIPF, and at the same time to verify and improve PIPF in practice. It is recommended that WHO, other international organizations and relevant countries make full use of the PIPF system to respond to the epidemic and better coordinate national actions at the global level. At the same time, China should also make the planning and deploy of domestic epidemic prevention and control and international epidemic cooperation under the framework.2020年3月11日，WHO正式宣布新型冠状病毒肺炎（简称新冠肺炎）疫情规模已形成大流行。截至3月31日，疫情波及超过178个国家和地区，累计确诊病例超过78万例。2011年，WHO已建立大流行性流感防范框架（PIPF），以平等、透明、公正地共享H5N1流感病毒及其他可能引起人间大流行的流感病毒，进而获取疫苗、共享其他相关利益，推动全球流感大流行防范和应对工作。由于新冠肺炎的传播途径、传播力与2009年甲型H1N1流感病毒相似，符合“引起人间大流行”的基本要素，且流行规模已超过甲型H1N1流感大流行，因此可以将新冠肺炎疫情应对纳入PIPF，同时实战验证并改进PIPF。建议WHO、其他国际组织和相关国家充分利用PIPF体系应对疫情，在全球层面更好地协调各个国家的行动；同时，我国也应在框架下全面做好国内疫情防控和国际疫情协同的规划与布局。.

Fourier transform infrared spectroscopy (FTIR) can provide abundant information for different organic functional groups and chemical bonds, and is therefore widely used in molecular structure analysis, qualitative analysis of mixtures, and quantitative analysis. However, it is difficult to decompose and extract the highly correlated component spectra from complex mixed spectrum, and existing methods for FTIR qualitative analysis has poor recognition effect. For this reason, a spectroscopic analysis model based on zero-phase component analysis (ZCA) whitening and particle swarm optimization (PSO) is proposed in this paper. The model uses wavelet threshold filter to de-noise the mixed spectrum, and uses ZCA whitening to remove the correlation of the component spectra. Finally, PSO is used to solve the component concentration. In order to verify the superiority of the ZCA whitening-based spectral analysis model, simulation studies are carried out on signals made up by twenty kinds of component spectra. The simulation results show that the model not only achieves effective qualitative analysis and quantitative analysis of the main components of the mixed spectrum, but also has better performance than the current mainstream spectrum analysis model.

We study the evolution of the temporal properties of MAXI 1820+070 during the 2018 outburst in its hard state from MJD 58190 to 58289 with Insight-HXMT in a broad energy band 1-150 keV. We find different behaviors of the hardness ratio, the fractional rms and time lag before and after MJD 58257, suggesting a transition occurred at around this point. The observed time lags between the soft photons in the 1-5 keV band and the hard photons in higher energy bands, up to 150 keV, are frequency-dependent: the time lags in the low-frequency range, 2-10 mHz, are both soft and hard lags with a timescale of dozens of seconds but without a clear trend along the outburst; the time lags in the high-frequency range, 1-10 Hz, are only hard lags with a timescale of tens of milliseconds; first increase until around MJD 58257 and decrease after this date. The high-frequency time lags are significantly correlated to the photon index derived from the fit to the quasi-simultaneous NICER spectrum in the 1-10 keV band. This result is qualitatively consistent with a model in which the high-frequency time lags are produced by Comptonization in a jet., Comment: Accepted for publication in ApJ

Objective To establish a method for determination of the azide ions in blood by gas chromatography-mass spectrometry （GC-MS） following pentafluorobenzyl derivatization. Methods A blood sample of 0.2 mL was placed into a 10 mL glass test tube, and the internal standard sodium cyanide, derivatization reagent pentafluorobenzyl bromide and catalyst tetradecyl benzyl dimethyl ammonium chloride were added in turn. After vortex mixing, the mixture was heated with low-power microwave for 3 min. After centrifugation, the organic phase was taken for GC-MS analysis. Results The azide ions in blood had a good linear relationship in the mass concentration range of 0.5 to 20 μg/mL. The lowest detection limit was 0.25 μg/mL and the relative recovery was 91.36%-94.58%. The method was successfully applied to a case of death from sodium azide poisoning. The mass concentration of azide ions in the blood of the dead was 11.11 μg/mL. Conclusion The method developed in this paper has strong specificity and is easy to operate, which is suitable for the rapid detection of azide ions in blood.五氟苄基衍生化GC-MS法检测血液中叠氮离子.目的 建立血液中叠氮离子的五氟苄基衍生化气相色谱-质谱（gas chromatography-mass spectrometry，GC-MS）法。 方法 取血液检材0.2 mL置于10 mL玻璃试管中，依次加入内标物氰化钠、衍生化试剂五氟苄基溴和催化剂十四烷基二甲基苄基氯化铵，涡旋混合后，微波炉低火档加热3 min，离心后取有机相进行GC-MS分析。 结果 血液中叠氮离子在0.5~20 μg/mL质量浓度范围内的线性关系良好，最低检测限为0.25 μg/mL，相对回收率为91.36%~94.58%，并成功应用于1例叠氮化钠中毒死亡案件，死者血液中叠氮离子质量浓度为11.11 μg/mL。 结论 本研究建立的方法专属性强，操作简便，适用于血液中叠氮离子的快速检测。.法医毒物化学；气相色谱-质谱法；五氟苄基衍生化；叠氮离子；氰离子.

No robust detection of prompt electromagnetic counterparts to fast radio bursts (FRBs) has yet been obtained, in spite of several multi-wavelength searches carried out so far. Specifically, X/gamma-ray counterparts are predicted by some models. We planned on searching for prompt gamma-ray counterparts in the Insight-Hard X-ray Modulation Telescope (Insight-HXMT) data, taking advantage of the unique combination of large effective area in the keV-MeV energy range and of sub-ms time resolution. We selected 39 FRBs that were promptly visible from the High-Energy (HE) instrument aboard Insight-HXMT. After calculating the expected arrival times at the location of the spacecraft, we searched for a significant excess in both individual and cumulative time profiles over a wide range of time resolutions, from several seconds down to sub-ms scales. Using the dispersion measures in excess of the Galactic terms, we estimated the upper limits on the redshifts. No convincing signal was found and for each FRB we constrained the gamma-ray isotropic-equivalent luminosity and the released energy as a function of emission timescale. For the nearest FRB source, the periodic repeater FRB180916.J0158+65, we find $L_{\gamma,iso}, Comment: 15 pages, 3 figures, 6 tables, accepted by A&A

We report on the observation of the accreting pulsar GRO J1008-57 performed by Insight-HXMT at the peak of the source's 2017 outburst. Pulsations are detected with a spin period of 93.283(1) s. The pulse profile shows double peaks at soft X-rays, and only one peak above 20 keV. The spectrum is well described by the phenomenological models of X-ray pulsars. A cyclotron resonant scattering feature is detected with very high statistical significance at a centroid energy of $E_{\rm cyc}=90.32_{-0.28}^{+0.32}$ keV, for the reference continuum and line models, HIGHECUT and GABS respectively. Detection is very robust with respect to different continuum models. The line energy is significantly higher than what suggested from previous observations, which provided very marginal evidence for the line. This establishes a new record for the centroid energy of a fundamental cyclotron resonant scattering feature observed in accreting pulsars. We also discuss the accretion regime of the source during the Insight-HXMT observation., Comment: 8 pages, 3 figures, accepted for publication in ApJL

Glitches correspond to sudden jumps of rotation frequency ($\nu$) and its derivative ($\dot{\nu}$) of pulsars, the origin of which remains not well understood yet, partly because the jump processes of most glitches are not well time-resolved. There are three large glitches of the Crab pulsar, detected in 1989, 1996 and 2017, which were found to have delayed spin-up processes before the normal recovery processes. Here we report two additional glitches of the Crab pulsar occurred in 2004 and 2011 for which we discovered delayed spin up processes, and present refined parameters of the largest glitch occurred in 2017. The initial rising time of the glitch is determined as $, Comment: 25 pages, 8 figures

With higher and higher penetration of wind power into power systems, dependence among the wind speeds of different wind farms should be considered when modeling the wind power outputs. In this paper, a novel pair-copula method is applied to formulate the dependence of multiple wind farms. A large number of stochastic scenarios, in which the complicated dependence of multiple wind farms are considered, are generated to represent the uncertainties of wind power based on quasi-Monte Carlo (QMC) simulations. To find an optimal dispatch solution, a risk constrained mean-variance (MV) model is constructed for the stochastic economic dispatch (SED) problem. The MV model considers economic cost and economic risk under the uncertainties of wind power simultaneously, among which economic risk is calculated by means of least variance of fuel cost. Moreover, with the probability density function (PDF) obtained for fuel cost, a predefined level of confidence interval is proposed to improve the MV model to acquire more practical dispatch solutions. For solving the multi-objective SED problem, group search optimizer with multiple producers (GSOMP) is employed in this paper. The effectiveness of the proposed pair-copula method and the improved MV model are validated via numerical simulations with a modified IEEE 30-bus system.

In order to improve the reliability of wind turbines, avoid serious accidents and reduce operation and maintenance (O&M) costs, it is important to effectively detect faults of wind turbines operating in harsh environment. This paper proposes a radically data-driven fault detection and diagnosis (FDD) method for wind turbines, which implements deep belief network (DBN). The DBN requires no knowledge of physical model, instead, it employs historical data without any pre-selection. The method has been evaluated in a wind turbine benchmark simulink model, in comparison with four model-based algorithms and four data-driven methods, and the results have shown that the proposed method achieves the highest accuracy. Moreover, extensive evaluation has been taken to analyse the robustness of proposed method, and the simulation results indicate the stable performance of proposed method in faults diagnosis of wind turbine.

Wind farms commonly cluster in regions rich in wind resources. Thus, correlation of wind speeds from different wind farms should not be ignored when modeling a power system with large wind energy penetration. This paper proposes a probabilistic optimal power flow (POPF) technique based on the quasi-Monte Carlo simulation (QMCS) considering the correlation of wind speeds using copula functions. In this paper, a copula function is used to model the dependent structure of random wind speeds and their forecast errors. QMCS is employed in the sampling procedure to reduce computation burden. The proposed method is applied in probabilistic power flow (PPF). Furthermore, the PPF is used in the POPF problem that aims at minimizing the expectation and downside risk of fuel cost simultaneously. Simulation studies are conducted on a modified IEEE 118-bus power system with wind farms integrated in two areas, and the results show that the accuracy and efficiency are improved by the proposed method.

Sympathetic interaction between transformers is a quite normal phenomenon in power systems. For the purpose of preventing transformer differential protection relays from malfunction, this paper proposes a morphological method for the identification of sympathetic inrush, which is the first time when mathematical morphology is applied in this field. Since the waveform of differential current is symmetrical in an internal fault case while asymmetrical in a sympathetic inrush case, the proposed method uses a morphological operator to extract the peaks and valleys of the differential current to distinguish sympathetic inrush. Considering the possible current-transformer (CT) saturation conditions, this paper combines a morphological gradient with a weighted mathematical morphological operator to improve the effectiveness of the proposed method. The proposed method is evaluated on data collected from simulation cases established in PSCAD/EMTDC and from laboratory experiments, respectively. Identification results have verified that by comparing with the traditional second harmonic restrain method, the proposed method can distinguish sympathetic inrush from an internal fault current more accurately and more effectively, even when the CT is fully saturated.

On-line monitoring and analysis of low frequency oscillation (LFO) are important for stability and security of a power system. This paper proposes a blind source separation (BSS) based method for LFO modal analysis of one-channel measured signal, which consists of a second-order blind identification (SOBI) algorithm and the Hilbert transform (HT) technique. This is the first time when a BSS technique is applied as a modal decomposition tool in this field. The proposed SOBI-HT based dominant mode identification method combines an iteration procedure with an identification procedure to determine the model order, and then identify dominant modes. The performance of the proposed method is evaluated on numerical simulation signals and a record of real data, verifying its effectiveness and accuracy.

Made available in DSpace on 2018-11-26T17:35:05Z (GMT). No. of bitstreams: 0 Previous issue date: 2017-09-01 Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) Copper tungstate (CuWO4) crystals were synthesized by the sonochemistry (SC) method, and then, heat treated in a conventional furnace at different temperatures for 1 h. The structural evolution, growth mechanism and photoluminescence (PL) properties of these crystals were thoroughly investigated. Xray diffraction patterns, micro-Raman spectra and Fourier transformed infrared spectra indicated that crystals heat treated and 100 C and 200 C have water molecules in their lattice (copper tungstate dihydrate (CuWO4.2H(2)O) with monoclinic structure), when the crystals are calcinated at 300 C have the presence of two phase (CuWO4.2H(2)O and CuWO4), while the others heat treated at 400 C and 500 C have a single CuWO4 triclinic structure. Field emission scanning electron microscopy revealed a change in the morphological features of these crystals with the increase of the heat treatment temperature. Transmission electron microscopy (TEM), high resolution-TEM images and selected area electron diffraction were employed to examine the shape, size and structure of these crystals. Ultraviolet-Visible spectra evidenced a decrease of band gap values with the increase of the temperature, which were correlated with the reduction of intermediary energy levels within the band gap. The intense photoluminescence (PL) emission was detected for the sample heat treat at 300 C for 1 h, which have a mixture of CuWO4.2H(2)O and CuWO4 phases. Therefore, there is a synergic effect between the intermediary energy levels arising from these two phases during the electronic transitions responsible for PL emissions. (C) 2017 Elsevier B.V. All rights reserved. Univ Estadual Piaui, PPGQ CCN GERATEC, Rua Joao Cabral,N 2231,POB 381, BR-64002150 Teresina, PI, Brazil Univ Fed Sao Carlos, DQ UFSCar, POB 676, BR-13565905 Sao Carlos, SP, Brazil ICE Univ Fed Amazonas, Av Rodrigo Otavio Japiim,POB 670, BR-69077000 Manaus, Amazonas, Brazil CCT Univ Fed Maranhao, POB 322, BR-65080805 Sao Luis, MA, Brazil Univ Estadual Paulista, Dept Fis Quim, BR-14800060 Araraquara, SP, Brazil IFSC Univ Sao Paulo, POB 369, BR-13560970 Sao Carlos, SP, Brazil Univ Estadual Paulista, Dept Fis Quim, BR-14800060 Araraquara, SP, Brazil CNPq: 479644/2012-8 CNPq: 304531/2013-8 FAPESP: 2012/14004-5 FAPESP: 2013/07296-2 CAPES: 1268069

Wind energy conversion system technology has attracted worldwide attention, and the condition monitoring and fault diagnosis for the system become significant issues. A data-driven fault diagnosis method is presented to detect and locate open-circuit switch faults of the back-to-back converter in permanent magnet synchronous generator (PMSG)-based wind generation system. Convolutional neural network (CNN)-based neural network is applied as a fault diagnosis method, and the dropout process is employed to deal with the over-fitting problem. Twelve sensor signals of current and voltage in the back-to-back converter in various conditions are measured. A grid-connected PMSG-based wind generation model has been built in MATLAB/Simulink to estimate the proposed algorithm. Least squares support vector machine (LSSVM) and back-propagation artificial neural network (BPANN) are applied as comparison methods. Simulation results reveals that the proposed theory has a decent performance regarding the detection and location of different faulty switches in an assembly of various operating conditions.

Non-intrusive load monitoring (NILM) is a task of estimating the contribution of individual appliance to the overall power consumption by using a set of electrical signals measured by a smart meter. In this paper, we propose a comprehensive and extensible framework based on DNNs. We employ denoising autoencoder (dAE) to reconstruct the power signal of individual appliance from aggregated power consumption, and we use long short term memory (LSTM) network to make sure which appliance the power signal belongs to. We select 5 appliances to validate our method, and the results have shown the advantages of the proposed framework in some aspects compared to hidden Markov models (HMMs) and premier dAE.