Search

With the growing focus on renewable energy, wind power is increasingly valued and advocated. In order to guarantee the stability of wind power system dispatch and management, reliable prediction of future wind speeds is essential. In this study, a short-term wind speed prediction model based on cross-channel data convolution, intelligent signal extension and attention mechanisms is proposed to enhance the prediction efficiency. The model first classifies the wind speed signal into IMFs (intrinsic mode functions) and residual data with the EMD (empirical mode decomposition) method, and then divides IMFs into rough prediction part and accurate prediction part according to the signal characteristics. CNN (convolutional neural network) modules are adopted for the rough prediction part to ensure a speedy process, whereas a CNN-AM (attentional mechanism)-LSTM (long short-term memory)-ECA (efficient channel attention) hybrid network is developed to for the accurate prediction part. Through the time-history prediction on measured 10 min average wind speed data, the results show that: (a) The channel-crossing one-dimensional (1D) convolution, intelligent signal extension, and attention mechanisms applied in the proposed model can effectively improve the accuracy of predictions; (b) The proposed prediction model is superior to the compared baseline models in precision and efficiency; and (c) The proposed model features strong migration learning ability for fast application on new datasets. • A computationally efficient wind speed prediction model is developed. • Intelligent padding from model predictions is considered to cope with end effect in signal analysis. • Spatial and channel attention mechanisms are applied to improve prediction accuracy. • The training, validation, and test sets are established by random sampling to avoid seasonal interference. [ABSTRACT FROM AUTHOR]

Tropical cyclones landing on the coast are usually associated with both strong winds and waves, posing a major threat to sea-crossing bridges. Strong waves riding on the elevated storm surge squeeze the narrow gap between the sea surface and bottom of the bridge deck, thus complicating the wind field and leading to different bridge nonlinear flutter responses as compared with the steady flow field. In this study, to evaluate the bridge safety under such circumstances, wind tunnel test was carried out for a typical sea-crossing bridge with streamlined box girder based on a developed dynamic response test system for simulating the coupling effect of wind, wave, and current on bridges, where the clearance, underneath the bridge deck, and wave speed can be systematically adjusted to investigate the wind field characteristics and bridge nonlinear flutter response. Numerical simulation with the LES (large eddy simulation) scheme was employed to successfully visualize the wind field around the bridge and the mechanism of the nonlinear flutter response was analyzed in detail. The results of this preliminary investigation show that: (a) In the wind tunnel tests, when the wave crest passes under the bridge, the acceleration effect due to the narrow gap can lead to an increase for the nonlinear flutter amplitude of the bridge, which threatens the structural safety; (b) The measured wind speed at the bridge site exhibits strong turbulence intensity and non-Gaussian distribution tendency; and (c) Based on the numerical simulation, the airflow between the wave crest and the main structure features a large vertical velocity component, leading to a large AOA (angle of attack) value of the incoming flow. Meanwhile, the air pressure around the girder is unevenly distributed, thus unavoidably weakening the aerodynamic stability of the bridge. • The non-linear bridge flutter stability is analyzed by simulating the wave-interfered wind field in the wind tunnel test. • Wave-interfered wind field characteristics are studied experimentally and numerically. • The bridge non-linear flutter response is closely related to the spatial relationship of wave and bridge. [ABSTRACT FROM AUTHOR]

Within industrial standards, effects due to wind–wave interaction on the marine atmospheric boundary layer (ABL) may be included via the Charnock sea-surface roughness parameter for open-sea and near-coastal waters. This roughness parameter does not accommodate the wide variety of wave states possible, nor does it modify the ABL to include speed-up effects resulting from an undulating wave profile. In an attempt to provide an updated definition of the general ABL profile for offshore wind engineering applications, an experimental and numerical study is performed to assess the interaction of the ABL on a fixed wave geometry. By extracting the mean velocity field during the wind–wave interaction, bespoke values of the velocity profile power exponent and wind risk factor can be obtained. A range of wave heights and wave lengths are considered from which a Kriging surrogate model is trained to supply the relevant wind profile parameters depending on the wave state. These newly garnered parameters enable the practitioner to define a reference wind velocity, and then adjust the velocity profile characteristics to contain the influence of the wave. This approach provides a new inlet velocity condition that can be used within computational wind engineering investigations without the need to explicitly model the wave surface, as well as flexibility in specifying the underlying wave conditions. Application of the re-calibrated velocity profile shows that wind forces are significantly greater throughout an offshore wind turbine (OWT) swept blade area when large (H = 15 m) and small (H = 1.5 m) wave heights are compared. • Wind-tunnel testing is performed for an ABL wind–wave interaction case study. • Numerical simulations are conducted for a range of wave heights and lengths. • New ABL velocity profile parameters are proposed for the mean ABL profile definition. • A Kriging surrogate model enables a bespoke calibration of the ABL parameters. • Wind–wave interaction has a significant effect on the average ABL velocity. [ABSTRACT FROM AUTHOR]

Coastal bridges are vulnerable to complicated hydrodynamics induced by hostile natural hazards, relevant research is thus required to ensure the safe operation of these critical infrastructure assets. Although coastal bridge hydrodynamic analyses can be carried out through commonly used approaches such as theoretical studies, numerical simulations, and experimental tests, they cannot fulfill the ever-increasing demand for applicability and computational efficiency. The recently advanced machine learning (ML) has emerged as a disruptive technology that fully revolutionized various scientific disciplines, resulting in a new paradigm to meet modern research needs. Aiming to provide the research community with holistic information on the key ingredients and current state-of-the-art of applying ML algorithms to coastal bridge hydrodynamics, this study presents a comprehensive review of the deployment of ML in coastal bridge hydrodynamics. The theoretical backgrounds of some representative ML algorithms are briefly introduced, and applications of ML to each of the research themes associated with coastal bridge hydrodynamics are systematically surveyed. Future research directions are also highlighted through the discussion of the current research limitations. According to this review, it is evident that ML models can be trained to learn and infer the intricate relationships between contributing parameters and responses of interest in coastal bridge hydrodynamics. In addition, it is envisioned that the research in coastal bridge hydrodynamics could be further advanced with the evolving ML technologies. [ABSTRACT FROM AUTHOR]

• HSP90 inhibitors 17-AAG and VER-82576 inhibit porcine deltacoronavirus replication at the early stage of infection. • HSP90 inhibitor KW-2478 shows no significant antiviral activity at any stage of infection. • 17-AAG and VER-82576 inhibit the expressions of TNF-α, IL-6 and IL-12, which are PDCoV-induced proinflammatory cytokines. • Our findings conclude that 17-AAG and VER-82576 are promising candidates for the treatment of PDCoV infection. Porcine deltacoronavirus (PDCoV) is highly pathogenic to piglets, and no specific drugs or vaccines are available for the prevention and treatment of PDCoV infection, the need for antiviral therapies is pressing. HSP90 inhibitors have potent inhibitory effects against the replication of numerous viruses, hence we evaluated three HSP90 inhibitors, 17-AAG, VER-82576, and KW-2478, for their effects on PDCoV infection in vitro. We evaluated their effectivenesses at suppressing PDCoV by qRT-PCR, western blot, and TCID 50 assay, and found that 17-AAG and VER-82576 inhibited PDCoV at the early stage of replication, while KW-2478 showed no significant antiviral activity at any stage of infection. These results indicated that the PDCoV-inhibitory effects of 17-AAG and VER-82576 might be exerted by targeting host cell factor HSP90AB1 but not HSP90AA1. Further study showed that HSP90AB1 mRNA and protein levels were not significantly different in 17-AAG and VER-82576-treated cells versus control cells. 17-AAG and VER-82576 were also evaluated for their effects on the expressions of TNF-α, IL-6, and IL-12, which are PDCoV-induced proinflammatory cytokines. We found that both 17-AAG and VER-82576 inhibited the expressions of TNF-α, IL-6, and IL-12 to varying degrees, but in a dose dependent manner. From our data we can conclude that the HSP90 inhibitors 17-AAG and VER-82576 are promising candidates for the treatment of PDCoV infection. [ABSTRACT FROM AUTHOR]