Feedback LSTM Network Based on Attention for Image Description Generator

Images are complex multimedia data which contain rich semantic information.Most of current image description generator algorithms only generate plain description,with the lack of distinction between primary and secondary object,leading to insufficient high-level semantic and accuracy under public evaluation criteria.The major issue is the lack of effective network on high-level semantic sentences generation,which contains detailed description for motion and state of the principal object.To address the issue,this paper proposes the Attention-based Feedback Long Short-Term Memory Network(AFLN).Based on existing codec framework,there are two independent sub tasks in our method:attention-based feedback LSTM network during decoding and the Convolutional Block Attention Module(CBAM)in the coding phase.First,we propose an attentionbased network to feedback the features corresponding to the generated word from the previous LSTM decoding unit.We implement feedback guidance through the related field mapping algorithm,which quantifies the correlation between previous word and latter word,so that the main object can be tracked with highlighted detailed description.Second,we exploit the attention idea and apply a lightweight and general module called CBAM after the last layer of VGG 16 pretraining network,which can enhance the expression of image coding features by combining channel and spatial dimension attention maps with negligible overheads.Extensive experiments on COCO dataset validate the superiority of our network over the state-of-the-art algorithms.Both scores and actual effects are proved.The BLEU 4 score increases from 0.291 to 0.301 while the CIDEr score rising from 0.912 to 0.952.

Study on Photocatalytic Degradation of Organic Dye by Titanium Dioxide Nanotube Array

Titanium based titanium dioxide （TiO2） nanotube arrays were prepared by electrochemical oxidation method, their microstructures were characterized, and the effects of sintering temperature and initial dye concentration and pH value on degradation performance of TiO2 nanotubc arrays wcrc investigated with methyl orange as a degradation object. The results showed that TiO2 nanotube arrays prepared by sintering at 500 ℃ exhibited good morphology and the highest photocata- lyric degradation efficiency; the degradation efficiency of the TiO2 nano material （500 ℃ ） to high concentration dye was higher than that to low concentration dye; the TiO2 nanotube array （500 ℃ ） exhibited higher degradation efficiency on dye solution at the pH of 3 than on that at the pH of 5.77 ; and the degradation efficien- cy of the TiO2 nanotube array （500 ℃） to 10 mg/L methyl orange solution （pH =3） reached 85.2%.

Simulation Study on Electromechanical Disturbance Propagation in Large Power System

Power systems are subjected to all kinds of random disturbances, showing the electromechanical dynamic. A lot of theoretical researches show that the disturbance power and frequency is propagating in the form of wave which is called electromechanical wave. But electromechanical wave theory is not widely used in actual power system. In this paper we focus on simulation study of electromechanical wave frequency sensitivity and propagation velocity and elaborating the simulation results with the electromechanical wave theory. Finally some summaries and expectations on electromechanical wave study are made.

A Pilot Protection Scheme of DC Lines for MMC-HVDC Grid Using Random Matrix

The over-current capacity of half-bridge modular multi-level converter(MMC)is quite weak,which requests protections to detect faults accurately and reliably in several milliseconds after DC faults.The sensitivity and reliability of the existing schemes are vulnerable to high resistance and data errors.To improve the insufficiencies,this paper proposes a pilot protection scheme by using the random matrix for DC lines in the symmetrical bipolar MMC high-voltage direct current(HVDC)grid.Firstly,the 1-mode voltage time-domain characteristics of the line end,DC bus,and adjacent line end are analyzed by the inverse Laplace transform to find indicators of fault direction.To combine the actual model with the data-driven method,the methods to construct the data expansion matrix and to calculate additional noise are proposed.Then,the mean spectral radiuses of two random matrices are used to detect fault directions,and a novel pilot protection criterion is proposed.The protection scheme only needs to transmit logic signals,decreasing the communication burden.It performs well in high-resistance faults,abnormal data errors,measurement errors,parameters errors,and different topology conditions.Numerous simulations in PSCAD/EMTDC confirm the effectiveness and reliability of the proposed protection scheme.

Delay-dependent Wide-area Damping Controller Synthesis Approach Using Jensen’s Inequality and Evolution Algorithm

A time-delay-dependent wide-area damping controller synthesis approach,based on Jensen’s integral inequality and evolution algorithm,is developed to suppress the adverse effect of time delay on the supplemental control of high-voltage direct current(DC)transmission systems.Initially,the state-space model of hybrid AC/DC systems with time delay is derived and the delay-dependent criteria for the stability of the closed-loop system are provided based on Jensen’s integral inequality.Subsequently,initial solutions are randomly generated to overcome the difficulty of solving the nonlinear matrix inequality.Finally,the time-delay stability upper bound of the controller is optimized using the differential evolution algorithm.In comparison to popular time-delay stable controller design methods,such as the free-weighting-matrix approach,the proposed method based on output feedback realization requires fewer decision variables and is more suitable for large-scale hybrid AC/DC systems.Three examples are introduced to verify the effectiveness of the proposed method.

Dynamic Phasor Modeling and Low-frequency Oscillation Analysis of Single-phase Vehicle-grid System

In practical operations,low-frequency oscillation(LFO)occurs and leads to converter blocking when multiple electrical rail vehicles at the platform are powered by the traction network.This paper proposes a small-signal model in state-space form for multiple vehicle-grid systems based on a dynamic phasor.This model uses the phasor amplitude and phase as variables to accurately describe the dynamics of the converter phase-domain control.An eigenvalue based-method is introduced to investigate the LFO with advantages of acquiring all oscillatory modes and analyzing participation factors.Two low-frequency dominant modes are identified by eigenvalues.Mode shape reveals that one of the modes involves the oscillations between the grid-connected converters and the traction network,and the other one involves the oscillations among these converters.Then the sensitivities of these two low-frequency modes to different system parameters are analyzed.Participation factors of system state variables,when the number of connected vehicle increases,are compared.Finally,the theoretical analysis is verified by nonlinear time-domain simulations and the modal analysis based on the estimation of signal parameters via the rotational invariance techniques(ESPRIT)method.

Measurement-based Frequency Dynamic Response Estimation Using Geometric Template Matching and Recurrent Artificial Neural Network

Understanding power system dynamics after an event occurs is essential for the purpose of online stability assessment and control applications.Wide area measurement systems(WAMS)based on synchrophasors make power system dynamics visible to system operators,delivering an accurate picture of overall operating conditions.However,in actual field implementations,some measurements can be inaccessible for various reasons,e.g.,most notably communication failure.To reconstruct these inaccessible measurements,in this paper,the radial basis function artificial neural network(RBF-ANN)is used to estimate the system dynamics.In order to find the best input features of the RBF-ANN model,geometric template matching(GeTeM)and quality-threshold(QT)clustering are employed from the time series analysis to compute the similarity of frequency dynamic responses in different locations of the power system.The proposed method is tested and verified on the Eastern Interconnection(EI)transmission system in the United States.The results obtained indicate that the proposed approach provides a compact and efficient RBF-ANN model that accurately estimates the inaccessible frequency dynamic responses under different operating conditions and with fewer inputs.

Development of an amperometric glucose biosensor based on the immobilization of glucose oxidase in an ormosil-PVA matrix onto a Prussian Blue modified electrode

An amperometric glucose biosensor was developed based on the immobilization of glucose oxidase in the organically modified silicate (ormosil)-polyvinyl acetate (PVA) matrix onto a Prussian Blue (PB)-modified glassy carbon electrode. A higher stability PB-modified electrode was prepared by the electrochemical deposition of FeCl3, K3[Fe(CN)6] and ethylenediamine tetraacetic acid (EDTA) under cyclic voltammetric (CV) conditions. The effects of the potential range of CV conditions, electrolyte cations, applied potential, pH, temperature and co-existing substances were investigated. The detection limit of the glucose biosensor was 8.1 μmol·L−1 (S/N = 3) with a linear range from 20 μmol·L−1 to 2 mmol·L−1 (R = 0.9965). The biosensor presented a fast response and good selectivity. Additionally, excellent reproducibility and stability of the biosensor were observed.

The Prevalence, Genetic Characterization, and Evolutionary Analysis of Porcine Pegivirus in Guangdong, China

Porcine pegivirus(PPgV)is a member of the Pegivirus genus in the Flaviviridae family.PPg V is an emerging virus that has been discovered in swine herds in Germany,the United States,China,Poland,Italy,and the United Kingdom,indicating a wide geographical distribution.In this retrospective study,339 pig serum samples were collected from 20 different commercial swine farms located in nine cities in Guangdong Province,China,from 2016 to 2018,to investigate the prevalence and genetic diversity of PPg V in this geographical region.PPg V was detected in 55%(11/20)of the farms using nested reverse transcription PCR,with 6.2%(21/339)of pigs testing positive for PPg V.The yearly PPg V-positive rate increased from 2.6%to 7.5%between 2016 and 2018.Sequencing of PPg V-positive samples identified two complete polyprotein genes and seven partial NS5 B genes from different farms.Comparative analysis of the polyprotein genes revealed that PPg V sequences obtained in this study showed 87.4%–97.2%similarity at the nucleotide level and 96.5%–99.4%similarity at the amino acid level with the reference sequences.Sequence alignment and phylogenetic analysis of the complete polyprotein gene and partial NS5 B and NS3 genes demonstrated a high genetic similarity with the samples from the USA.The finding of the wide distribution of PPg V in swine herds in Guangdong Province will contribute to the understanding of the epidemiological characteristics and genetic evolution of PPg V in China.

A measurement-based approach for power system instability early warning

Early warning of impending instability in a power system under disturbance conditions is important for preventing of system collapse.A measurement-based approach is proposed to assess the potential power system transient instability problem under cascading outages.Where a measurement-based index is obtained as the estimation accuracy of a linear autoregressive exogenous(ARX)model to estimate the dynamic response of the power system and indicate the system stability to some extent after a disturbance.The proposed approach was verified using a set of marginally stable cases in a 179-bus WECC equivalent power system.Then the instability early warning threshold for this system is obtained as 0.44.

Oscillation Characteristics Analysis and Equivalent Modeling of Multiple CIGs Connected to Weak Grid

The stability problem of weak grid connected converter interfaced generation(CIG)cannot be ignored.For multiple weak grid connected CIGs with different parameters,the system oscillation characteristics and equivalence methods still need to be further studied.This paper first discusses the oscillation characteristics when CIGs are perfectly coupled,perfectly decoupled and their general conditions respectively.Based on the Monte Caro simulation,the number of critical eigenvalues,the participation of each CIG to critical eigenvalues and the correlative parameters to participation are analyzed.Then the single-CIG and multi-CIG equivalence methods are proposed for stations containing nonidentical CIGs.The CIG parameters of a single-CIG equivalent model are identified based on the consistency of the output admittance characteristics.According to the different participations of CIGs with critical eigenvalues,the station is equivalent to a multi-CIG model.Results of large simulation samples show that the two equivalent models can both preserve the critical eigenvalues very well,and can be used for stability analysis.Furthermore,the multi-CIG equivalent model can also very well reflect the participation of CIGs in detailed models,and can be used for damping control study.