At present, deep learning has achieved excellent performance in the field of network traffic classification. However, deep learning relies on massive data-driven classification models. When the data set is small, it is usually hindered. In order to solve this problem, few-shot network traffic classification technology based on deep learning has been gradually studied. In this paper, we conducted a comprehensive survey to fully understand the few-shot traffic classification techniques. We first define the few-shot traffic classification problem. Based on how to solve the contradiction between the few-shot data set and the large number of parameters to be trained in the model, we classify the current few-shot traffic classification method based on deep learning from two perspectives: (1) Data augmentation method, which uses the method of expanding the data set to enhance the supervision experience. In this paper, the methods based on data enhancement are divided into two categories based on GAN and feature transformation.(2) Model-based methods. Model-based methods are divided into three categories: transfer learning, metric learning and meta-learning. And discuss the advantages and disadvantages of each classification method. Finally, the results are summarized and the future development direction of few-shot network traffic classification technology based on deep learning is prospected.
Lightning is a kind of natural disaster that seriously threaten the safe and stable operation of power grids. This paper proposes a new kind of composite insulator combining the function of both arrester and insulation , which can meet the needs of lightning prevention of 35kV transmission lines. In this paper, 5-year lightning statistical data in Longhui, Hunan Province have been investigated. Then, the lightning trip rate of 35kV distribution network were analyzed by simulation, the influence of grounding resistance on the lightning trip rate of was discussed, and the lightning protection effect of the proposed composite insulator was evaluated.
Thanks to the development of 5G technology, the construction of edge data centers has been highly valued. In order to ensure the security of signal transmission in edge data centers, it is a promising measure to apply the quantum key distribution (QKD) technology to edge data centers. To solve the problem of noise interference in the process of applying QKD technology, this paper first proposes the edge-end quantum one-way edge data center architecture, that is, the edge data center that first distributes the data to the surrounding clusters, and then transmits the single data from the edge to the end. Directed quantum signals increase the safe communication distance to meet the requirements of information encryption. Secondly, we conduct research on the four-wave mixing (FWM)noise interference in the architecture and the secure key rate model of the QKD system and perform theoretical calculations of the noise power and the rate of the secure key. Finally, through simulation, we can evaluate that QKD has the characteristic of high degree of security and its performance is affected by channel location, classical signal power and QKD detection efficiency, etc. The FWM noise power increases first and then decreases with the increase of transmission distance, with the maximum distance of 15 km. When the distance between the classical channel and the quantum channel is more than 200 GHz, the effect of FWM noise on QKD is weak.
The test results of shear wave velocity can be applied to many aspects, such as the dynamic parameters of soil in seismic design, the possibility of liquefaction of sandy soil foundation, and the possibility of seismic subsidence of loess foundation; However, due to the influence of soil structure, genesis and other factors, as well as the artificial error in the process of wave velocity test, the shear wave velocity values of the same stratum or the same site also vary greatly. Based on the existing wave velocity data, with the help of statistical analysis tools and taking into account the actual engineering conditions in the project area, the mathematical expression of shear wave velocity can more accurately reflect the actual situation of the soil layer at the construction site. Therefore, whether from the practical work value or theoretical research, it is a very meaningful work to study the statistical law of shear wave velocity. Based on the shear wave velocity data of 170 boreholes (150 shallow holes and 20 deep holes) of multiple construction projects in the urban area of Beijing and the geotechnical test parameters, this paper studies the relationship between shear wave velocity and buried depth, density and other factors using the grey correlation theory, and establishes the empirical formula of shear wave velocity applicable to the urban area of Beijing using statistical tools.
The design of vibration damping controllers for uncertain mechanical systems with time-varying input delays is the main topic of this work. First, a state-space model based on matrix transformations is used to characterise the mechanical system. The uncertain system model is then created by adding parameters with linear variation. Second, by employing LMI technology, it is possible to derive LMI-based necessary requirements for system stability. A controller that is stable for the closed-loop system and has the desired amount of disturbance attenuation is realised by solving the obtained LMI. Finally, numerical examples are provided to show how well the suggested theorems work.
The factors of electricity consumption growth in China are increasing, such as "double carbon" constraint, new urbanization, rural revitalization, electrification, digitalization, new power system, power market, etc., which lead to a more complex influence mechanism of electricity consumption growth. This paper adopts a combination of quantitative and qualitative methods to build a spatial and temporal evolution system of power consumption influencing factors, and draws the following conclusions: economic and social factors are the basic support for power consumption growth, and the pulling effect of new technologies and new business models on power consumption growth is gradually increasing; there are some differences in power consumption influencing mechanisms among regions, and the influence of energy and electricity, climate and temperature factors on power consumption in east and central regions is increasing, while the influence of export factors on western and northeastern regions is increasing. Export factors on the western and northeastern regions to enhance the impact.
Communication engineering as a basic component of urban construction and development, project product quality directly determines the level of communication quality of social residents. Especially after the Chinese communication industry presents the characteristics of opening to the outside world, the international communication giants participate in the competition and cooperation in the Chinese communication field, which leads to the Chinese communication enterprises facing increasingly strong competition pressure. Facing the new market competition environment, Chinese communication enterprises want to occupy an important market position, improve the loyalty of enterprise users, obtain more economic benefits and social effects, to use big data technology to control the quality of communication engineering products, only in this way can achieve sustainable development goals. Based on the understanding of the development status of communication engineering industry and product quality control, this paper deeply discusses the communication engineering product quality control innovation channel with big data as the core according to the basic concept and unique characteristics of big data technology, in order to provide technical support for the innovation and development of communication industry.
With the maturation of experimental laser cooling and atom trapping techniques, various types of atomic optical devices developed by using cold atoms as media have been born in recent years, which play an important role in the field of atomic physics. In atomic optics, Bose-Einstein condensation (BEC) has been one of the hot topics in the study of scientists. It describes the phenomenon that bosons are concentrated in the lowest energy quantum state at extremely low temperatures. In general, atomic BEC can be expressed in the form of quantum wave packet. On the basis of previous studies, two kinds of motion behaviors of Bose-Einstein condensates passing through Gaussian laser field are investigated in this paper, namely the motion phenomena of atoms under the acceleration field and the S-wave scattering collision between atoms. The atomic motion equation is derived by establishing the theoretical model, and then the numerical simulation is carried out by MATLAB program. The results show that the acceleration field only affects the velocity of the wave packet. The attractive interatomic collision will cause the collapse of the wave packet. On the contrary, the repulsion will accelerate the diffusion of the wave packet. These results will provide more theoretical and feasible schemes for studying the coherent manipulation of atoms and laser in the future.
With the continuous increase in the number of Autonomous Maritime Radio Devices (AMRD), the international standards for AMRD are gradually improving, and China is also researching and developing the national standard for AMRD. On the basis of studying international standards, this paper compares the differences between international standards related to AMRD and China national standard, then studies and analyses AMRD key technology. Based on the technical characteristics and application requirements of AMRD, the application scenarios of various types of AMRD are studied and suggestions for further application of AMRD are proposed.
The cam curve was optimized and the bolt cam part was analyzed. it can be changed some materials and create an opening of the bolt, or the reduction of the height and the increase of length or width can lead to reducing friction. With the increase of the injection rate, the main roller pressure rapidly increased from more than 1000 to 10000N, nearly 10 times. The rate of fire is increased by 3 times and the pressure is close to 10 times.
This paper addresses the problems of high labour intensity, low work efficiency, poor product quality and lack of conveying equipment, as the harvesting and conveying work of striped seaweed is mainly done manually. In this paper, we designed and completed a post-harvest conveying equipment based on the biomechanical characteristics of striped seaweed and carried out experimental research on the conveying effect of striped seaweed. The post-harvest conveying equipment consists of an inlet, a conveying line, a flexible vane pump and an outlet. In this paper, pump speed, water to vegetable ratio and conveying height, which have an influence on the conveying effect of seaweed, were used as test factors, and conveying capacity, energy consumption per unit mass and residual rate were used as evaluation indicators. The results showed that the conveying capacity increased with the increase of pump speed, and the conveying capacity was 3122.684kg/h at 1390r/min; the energy consumption per unit mass decreased with the increase of vegetable-water ratio; the loss rate of purple cabbage increased with the increase of pump inlet conveying height, and the loss rate was 0.94% at 2.95m. 0.94% at 2.95m. In addition, the orthogonal test found that the main factors affecting the conveying capacity, energy consumption per unit mass and loss rate were the vegetable-to-water ratio, followed by the pump inlet conveying height and pump rotational speed. The best conveying effect of the seaweed conveying equipment was achieved with a maximum of 2445.744 kg of seaweed per hour, and the energy consumption and loss rate per unit mass were less than 0.0890 kW-h/t and 0.32% respectively, meeting the actual conveying requirements.
Database management systems introduce thousands of configuration options to provide flexibility. However, configuration complexity and configuration dependencies have inevitably affect the system reliability and become one of the major causes of system failure. Previous research neither specifically detects configuration-related bugs nor studies the characteristics of configuration-related code in DBMSs. In this paper, we undertake one of the first attempts to conduct a real-world configuration-related branch statements characteristic study in DBMSs. We first use a taint analysis tool to construct a data set of 347 real world configuration-related branch statements including 100 in MySQL, 98 in SQLite and 149 in PostgreSQL, which are very popular. Based on our study, we have obtained some findings: (1) More than half (53. 6%) of configuration-related branch statements contain only configuration variables, 93. 5% of which contain only one configuration variable. (2) 46. 4% of configuration-related branch statements contain not only configuration variables but also program variables. (3) Complex configuration-related branch statements (which contain function call) account for 17. 9%.
It has the advantages of low consumption of building materials, high safety performance and fast assembly, and it can also be fixed horizontally in a short time, which is a significant advantage in high-rise or large buildings. Therefore, in order to better promote the development of attached lift scaffolding technology, this study takes a large square project as an example, based on the analysis of the construction characteristics of attached scaffolding, construction principles and construction difficulties of the project, and the analysis and study of the scaffolding frame construction process and flow, aiming to sort out the control points of construction safety in the process of scaffolding assembly and erection, while ensuring the overall stability of the supporting structure, with a view to reduce the construction safety risks of large construction projects.
In view of the characteristics of non-stationary signals in bearings under abnormal conditions, a new diagnosis method for bearing fault detection was proposed based on TVF-EMD, kurtosis selection and envelope entropy as energy extraction method and random forest classification model. First, TVF-EMD method is used to decompose the initial signal, and the decomposed intrinsic modal components (IMFS) are obtained. Secondly, several natural modal components which can best represent the original signal are selected by the kurtosis values of the natural modal components. Secondly, the envelope entropy is used as the energy extraction method. Finally, the stochastic forest classification model is used to train and predict the original data. After comparing the test data with the original data, the accuracy rate reached 99.5 percent.
In order to solve the path planning problem of USV(unmanned surface vehicle), a global path planning method integrating global path and local path is proposed in this paper. First, the global path planning of grid electronic chart is carried out using two different reward modes of PPO(Proximal Policy Optimization Algorithms) algorithm, and the global optimal path is selected after comparison. Secondly, The inflection point on the global optimal path is selected as the key point of the DWA(dynamic window algorithm) to optimize the local path. The simulation results show that the method can take into account the requirements of high smoothness and short navigation distance, and plan a safe and reasonable route in the static environment.
In the current social construction and development, the computer network security architecture design is composed of software and hardware, which refers to the communication transmission activities with transmission control interconnection protocol as the core. With the continuous improvement of Internet information technology, a large number of Internet transaction modes, such as remote control and online payment, have emerged in the market. Although they provide convenient conditions for social residents, serious problems of cyberspace security have also arisen. In this paper, on the basis of understanding the development trend of modern science and technology innovation, according to the computer network security system framework design, clear network security service measures and management mode, in order to create a safe and comfortable network protection environment for social residents.
The construction industry has been faced with the problems of low labor productivity, labor shortage and high worker safety risk for a long time. However, building automation and robot technology bring new opportunities to solve these problems. After more than 40 years of development, the field of building automation and robot has accumulated a lot of research results, but it has not been widely used in practical engineering. Among them, high-rise buildings are large in quantity, wide in scope, complex in function, higher in engineering efficiency, quality and safety requirements, is an important direction of building automation and robot development. This study describes the development status of high-rise building automation and robot in a multi-directional way, which lays a theoretical foundation for the subsequent research. It clearly puts forward the priority development direction of high-rise building automation and robot, identifies the key factors of success, and provides important reference for relevant participants to grasp the key points and breakthrough the key, which is conducive to promoting the commercialization and practical application of the research results. Based on the research of robot prototype system, a more efficient and safe new method is provided for the safety inspection of glass curtain wall of high-rise buildings. The guarantee measures of automation and robot development in high-rise building are put forward systematically.
Big data technology as the foundation of the construction and development of modern society, according to the huge data information and diversified processing technology, technology innovation in each field provides technical support. Whether in the domestic environment or in the international environment, digital transformation has become the preferred strategic decision goal of the industry development, among which the unique advantages of large enterprises are very obvious. By understanding the management principles put forward by enterprise development and integrating big data technology to realize the goal of enterprise digital transformation, it can not only improve the business processing capacity and stabilize the basic assets, but also provide strong support for enterprise strategic planning. Therefore, based on understanding the concept of big data technology and the present situation of enterprise digital transformation, this article, based on the analysis of enterprise digital transformation path based on the big data, deeply studies how Chinese commercial banks apply the theory of big data technology in enterprise digital transformation.
Pile foundation is the most widely used type of foundation foundation in our country, and there are many forms of pile. Compared with other pile types, thread pile has high bearing capacity, strong adaptability, a wide range of application, environmental protection and remarkable economic benefits, so it has been widely used in our country. This paper studies the vertical bearing capacity of spiral anchor pile in silty soil by ABAQUS numerical simulation. The distribution of pile axial force, pile side resistance, pile soil interface stress and soil shear stress around the pile are analyzed. The influence of pile structural parameters is discussed, and the properties of grouting and ungrouting screw anchors in silty soil are studied. The results show that the pile side resistance is larger, the number of anchor disks is 4, and the bearing effect of anchor rod φ140mm is better, which significantly improves the ultimate bearing capacity of single pile.
It is necessary to establish three-dimensional digital twin model for the ultra-high voltage power transformers and bushings, and simulate and track the changes and trends of the physical parameters through digital twin model. Under high temperature conditions, tangent value of the loss angle and the real part value of dielectric constant of bushing core material of the converter transformer will suddenly increase, and the tangent value of the loss angle will remain between 0.004 and 0.022, while the real part value of dielectric constant will remain between 3.9 and 4.5. It can be seen that the amplitude of the fundamental wave is about 6000A, while amplitude of other harmonics shows a decreasing trend as the number of harmonics increases. Amplitude of the 50Hz fundamental current varies between -6000A and 6000A, and the eddy current heating power of the current carrying structure varies between 15W and 45W. However, during the first 1/4 cycle, the eddy current heating power of the current carrying structure gradually increases from 0W to around the peak of 45W.
