水下圆形重力式网箱缆绳张力对波、流的响应特性

通过模型试验研究了周长40 m、网高10 m的HDPE圆形升降式网箱沉降在水下时的缆绳张力。试验采用狄克逊准则,模型大比例尺λ=10和小比例尺λa=λd=2;网箱采用网格形式锚泊系统。试验时前主缆绳水平方向与波流夹角为0°,试验流速为0、5、10和15 cm/s,波浪周期范围1.3~2.2 s,波高范围10~30 cm。在各试验条件下对主缆绳和系框绳以及浮框绳张力进行了测试,以受力最大的缆绳为研究对象,分析了缆绳最大张力与波流因子之间的关系。结果显示:①在纯流作用下,前主缆绳受力最大;在纯波作用下,后侧主缆绳受力最大,最大张力随波高增大呈线性增长趋势,随周期变化的影响较小。②在波流作用下,高流速、低波高、高波周期条件时前主缆绳受力最大;在低流速、高波高、低波周期时后侧主缆绳受力最大;前主缆绳受水流作用影响较大,后侧主缆绳受波浪作用影响较大。③在波流作用下,各缆绳最大张力随波高、流速增大而增大,随周期增大而减小;后侧主缆绳最大张力无量纲量随波陡增大且基本呈线性增长趋势。

基于改进LSTM的儿童语音情感识别模型

为实现不同儿童情感需求状态下帧级语音特征的有效获取,建立一种基于改进长短时记忆(LSTM)网络的儿童语音情感识别模型。采用帧级语音特征代替传统统计特征以保留原始语音中的时序关系,通过引入注意力机制将传统遗忘门和输入门转换为注意力门,并根据自定义的深度策略计算得到深度注意力门,从而提高语音情感识别性能。实验结果表明,在Fau Aibo儿童情感数据语料库及婴儿哭声情感需求数据库上,该模型在召回率和F1分数上相比基于传统LSTM的识别模型分别提高了3.14%、5.50%和1.84%、5.49%,在CASIA中文情感数据库上,其相比基于传统LSTM和GRU的识别模型训练时间更短、儿童语音情感识别率更高。

Application of CFD Modeling to Hydrodynamics of CycloBio Fluidized Sand Bed in Recirculating Aquaculture Systems

To improve the efficiency of a CycloBio fluidized sand bed(CB FSB) in removal of dissolved wastes in recirculating aquaculture systems, the hydrodynamics of solid-liquid flow was investigated using computational fluid dynamics(CFD) modeling tools. The dynamic characteristics of silica sand within the CB FSB were determined using three-dimensional, unsteady-state simulations with the granular Eulerian multiphase approach and the RNG k-ε turbulence model, and the simulation results were validated using available lab-scale measurements. The bed expansion of CB FSB increased with the increase in water inflow rate in numerical simulations. Upon validation, the simulation involving 0.55 mm particles, the Gidaspow correlation for drag coefficient model and the Syamlal-O'Brien correlation for kinetic granular viscosity showed the closest match to the experimental results. The volume fraction of numerical simulations peaked as the wall was approached. The hydrodynamics of a pilot-scale CB FSB was simulated in order to predict the range of water flow to avoid the silica sand overflowing. The numerical simulations were in agreement with the experimental results qualitatively and quantitatively, and thus can be used to study the hydrodynamics of solid-liquid multiphase flow in CB FSB, which is of importance to the design, optimization, and amplification of CB FSBs.

A Comparative Study on Hydrodynamic Performance of Double Deflector Rectangular Cambered Otter Board

In the present work,the hydrodynamic performance of the double deflector rectangular cambered otter board was studied using wind tunnel experiment,flume tank experiment and numerical simulation.Results showed that the otter board had a good hydrodynamic performance with the maximum lift-to-drag ratio(K_(MAX) = 3.70).The flow separation occurred when the angle of attack(AOA) was at 45?,which revealed that the double deflector structure of the otter board can delay the flow separation.Numerical simulation results showed a good agreement with experiment ones,and could predict the critical AOA,which showed that it can be used to study the hydrodynamic performance of the otter board with the advantage of flow visualization.However,the drag coefficient in flume tank was much higher than that in wind tunnel,which resulted in a lower lift-to-drag ratio.These may be due to different fluid media between flume tank and wind tunnel,which result in the big difference of the vortexes around the otter board.Given the otter boards are operated in water,it was suggested to apply both flume tank experiment and numerical simulation to study the hydrodynamic performance of otter board.

Significant Effects of Fishing Gear Selectivity on Fish Life History

Over the past few decades, extreme changes have occurred in the characters of exploited fish populations. The majority of these changes have affected the growth traits of fish life history, which include a smaller size-at-age, an earlier age-at-maturation and among others. Currently, the causes of these life history traits changes still require systematic analyses and empirical studies. The explanations that have been cited are merely expressed in terms of fish phenotypic adaptation. It has been claimed that the original traits of fish can be recovered once the intensity of exploitation of the fish is controlled. Sustained environmental and fishing pressure will change the life history traits of most fish species, so the fish individual's traits are still in small size-at-age and at earlier age-at-maturation in exploited fish populations. In this paper, we expressed our view of points that fishing gear has imposed selectivity on fish populations and individuals as various other environmental factors have done and such changes are unrecoverable. According to the existing tend of exploited fish individual's life history traits, we suggested further researches in this field and provided better methods of fishery management and thereby fishery resources protection than those available early.

Experimental Study on the Hydrodynamic Characteristics of a Submersible Fish Cage at Various Depths in Waves

Submersible fish cages can be submerged under the water to mitigate the negative effects that arise from severe sea conditions and improve the growing environment for the farmed fish. Thus they are increasingly applied in offshore aquaculture. To ensure both safety and economic efficiency of submersible fish cages, it is important to determine the optimum submergence depth. In this study, a series of physical model experiments were conducted to investigate the hydrodynamic performance of a submersible fish cage at various submergence depths(1/6, 1/4, 1/3, and 1/2 of the water depth as well as the floating condition for reference) with a model scale of 1:20. The results of the physical model experiment for the different depths were compared to analyze the effects of submergence depths on the mooring line tension and the movement of the floating collar. The results showed that the mooring line tension and the floating collar movement significantly attenuated with increasing submergence depth. However, the attenuation tendency became stable when the fish cage reached a certain depth. According to the results, 1/3 of water depth was determined as the optimal submergence depth of the fish cages. Deeper submergence depths showed no significant advantage from a perspective of the hydrodynamic characteristics of the fish cage. The determination of the optimum submergence depth is beneficial for the structural design and operation safety of submersible net cages.

Spatial and temporal characteristics of trophic structure of the southern Huanghai Sea fish communities

The spatial and temporal characteristics of trophic structure of fish communities in the southern Huanghai Sea were examined based on the data sampled from bottom trawl surveys conducted during the autumn of 2000 and the spring of 2001. Hierarchical agglomerative cluster method and bootstrap randomization were used to identify significant trophic groups for each fish assemblage in the southern Huanghai Sea. A total of six major trophic groups were identified within this system, which classified predators based upon location in the water column or prey size （ i. e. , benthic to pelagic predators or fish to small invertebrate prey predators）. The similarity level used to identify significant trophic groups in each assemblage ranged from 24% to 34%. Although planktivores were the dominant trophic group in each assemblage （60% - 79% ）, there were spatial and temporal variations in the trophic structure, which reflected the differences in the abundance and availability of dominant preys. Simplified food webs were constructed to evaluate the most important trophic relationships between the dominant prey taxa and the fishes in each assemblage within this system. Although there were some differences in the key prey species among different food webs, pelagic prey items （mainly euphausiids and copepods） represent the most important energetic link between primary producers and higher trophic level predators. The trophic level for most fishes was between 3 and d, and the weighted mean trophic level for each assemblage ranged from 3.3 to 3.4. Compared with previous study in the mid-1980s, there was an obvious downward trend in the trophic level for most fish species, which resulted mainly from the fluctuation in key prey species in the Huanghai Sea. The decrease in the importance of Japanese anchovy seems to be offset by other abundant prey species such as Euphausia pacifica and copepods （ mainly Calanus sinicus ） .

Distribution Characteristics of Fat Greenling(Hexagrammos otakii) Inhabiting Artificial Reefs Around Qiansandao Island, Haizhou Bay, China

Chinese government has provided vigorous financial supports to artificial reef(AR)projects to repair marine ecosystem.Qiansandao AR,an important AR project of Fishery Resource Restoration Program in Shandong Province,was built during 2002 to 2007.It has enhanced the local fishery harvest and generated positive economic benefits.Fat greenling(Hexagrammos otakii)is a major commercial reef fish in Qiansandao with high economic values.Its distribution characteristics need to be further studied.In this research,we studied the potential influence of environmental variables and distance from ARs on catch per unit effort(CPUE)and body length of fat greenlings.The results showed that the average body length of fat greenlings distributed within 10 m from ARs was longer than that of fat greenlings distributed more than 600 m away from ARs.Body length also increased with the increased salinity,while it declined as temperature increased.The probability of the presence of fat greenling increased when temperature was below 20℃,then it slightly decreased.CPUE decreased as the distance increased.Our results elucidate the potential effects of environmental variables and distance from ARs on the distribution of fat greenling,which may also be helpful to the distribution research of other reef fish species.

高PVC外墙乳胶漆透水率的影响因素

《GB/T 9755-2014合成树脂乳液外墙涂料》规定了合成树脂乳液外墙涂料透水性的测试,此项性能规定的目的是规范控制低品质的高PVC外墙乳胶漆的使用。通过实验,探讨了不同填料种类、产品的PVC、遮盖聚合物、功能助剂以及实验板等因素对透水性测试结果的影响。

Comparison of 16S rRNA Gene Primers on Studying Microbial Community Composition in Bottom Water and Sediment of Artificial Reefs in Laoshan Bay, China

Marine microorganisms are indispensable regulators of nutrient cycling and energy flow,which are crucial for artificial reefs(ARs)ecosystems.However,little is known about the microbial communities in the bottom water and sediment of ARs.Studies of microbial diversities have greatly advanced due to the development of high-throughput sequencing technologies,whereas the results may vary significantly due to the primers’choice.This study investigated the influences of two 16S ribosomal RNA gene primer choices(V4 and V3-V4)on microbial community compositions and structures.The results showed that the taxonomic assignment de-tected by primer V3-V4 was higher compared with that obtained by primer V4,whereas microbial community compositions had strong correlations between the two primers.Microbial beta diversities of ARs can be uncovered by both primers,but the relation-ships between communities and environmental parameters were inconsistent.The performances of the two primers for water samples were highly consistent,but the inconformity was evident for sediment samples.Given the relatively lower taxonomic classification of primer V4 for sediment samples,primer V3-V4 was recommended for later studies.With the development and advancement of ARs in China,our findings provide a meaningful reference for ecologists focusing on the microbial diversities and ecological functions of these artificial habitats in the future.

Transfer learning with deep sparse auto-encoder for speech emotion recognition

In order to improve the efficiency of speech emotion recognition across corpora,a speech emotion transfer learning method based on the deep sparse auto-encoder is proposed.The algorithm first reconstructs a small amount of data in the target domain by training the deep sparse auto-encoder,so that the encoder can learn the low-dimensional structural representation of the target domain data.Then,the source domain data and the target domain data are coded by the trained deep sparse auto-encoder to obtain the reconstruction data of the low-dimensional structural representation close to the target domain.Finally,a part of the reconstructed tagged target domain data is mixed with the reconstructed source domain data to jointly train the classifier.This part of the target domain data is used to guide the source domain data.Experiments on the CASIA,SoutheastLab corpus show that the model recognition rate after a small amount of data transferred reached 89.2%and 72.4%on the DNN.Compared to the training results of the complete original corpus,it only decreased by 2%in the CASIA corpus,and only 3.4%in the SoutheastLab corpus.Experiments show that the algorithm can achieve the effect of labeling all data in the extreme case that the data set has only a small amount of data tagged.