Search

By collecting the ripple current in the control coil, the self-sensing magnetic bearings can obtain rotor displacement information without a displacement sensor, which has the advantages of low cost and high integration. According to the problems of incomplete and inaccurate displacement information extraction caused by the traditional displacement estimation method using Fourier method to analyze the ripple current with non-stationary characteristics to estimate the rotor displacement with the Gibbs effect. Therefore, this paper proposes a rotor displacement estimation algorithm based on multiresolution filter bank biorthogonal spline wavelet for a magnetic bearing motor based on two-level switching power amplifier. This algorithm utilizes the biorthogonal spline wavelet with the characteristics of generalized linear phase and tight support, which can accurately demodulate the ripple current in the coil and extract the displacement information in the ripple current. Therefore, it is able to overcome the Gibbs effect in the traditional Fourier analysis estimation algorithm, to reduce the influence of ripple on rotor displacement estimation, and to improve the accuracy and stability of displacement estimation. In order to verify the correctness and effectiveness of the proposed algorithm, this paper establishes a simulation model of magnetic bearing in MATLAB Simulink and designs the proposed displacement estimator. Simulation results demonstrate that the proposed algorithm has higher accuracy and better stability than traditional displacement estimation algorithms. Experimental results show that the maximum deviation rate of the displacement estimation method proposed in this paper does not exceed 1% when the radial magnetic bearing gap is 0.5 mm, which can effectively estimate the rotor displacement.

In the vision task of a self-driving system, the use of visible light images to segment an object often loses its functionality at night or in harsh weather. The far-infrared image shows different pixel values according to the thermal radiation quantity of the object itself, so it can be adapted to perform well at night and in harsh weather conditions. However, at the same time, it has insufficient texture features, blurred object boundaries and temperature inversion, which has a great impact on the segmentation task of traditional algorithms. In response to the above problems, this article proposes a far-infrared object segmentation algorithm using deep learning. In the current popular encoding-decoding structure, multi-scale pooling layers are used to obtain receptive fields of different sizes. This is used to solve the effects caused by the blurring of infrared objects. The feature enhancement module is designed for the multi-receptive field feature map, which can filter out the most versatile and highly semantic feature channels to reduce the effect of temperature inversion on segmentation. The obtained high semantic feature map is guided into the decoding structure and is fused with the features obtained by the encoder and the decoder. This allows richer information to be obtained between different feature maps. Finally, we also release a new low-resolution far-infrared segmentation dataset. Experiments are performed on three datasets, and the segmentation result of the mIoU(mean Intersection over Union) reaches 70.59%, 30.98% and 60.67%. A large number of experiments confirm the effectiveness and robustness of the network in far-infrared images and verify that the dataset released in this article has strong reference significance.

Many researchers have used the map-matching algorithm to leverage inadequate traditional vehicle localization with HD maps for fast and efficient vehicle localization. The initialization process in the map-matching algorithm has always been problematic due to the nature of GNSS error, which might exceed the average lane width of approximately 3.5m. Thus, directly using the GNSS data to obtain the rough initial pose may cause the failure of the initialization process. The general solution is to randomly sample around the GNSS data and test the initialization with these hypotheses. However, this often leads to expensive computation and thus fails to run in realtime or online mode, as a dense sampling is required to achieve an acceptable level of initial estimation accuracy. As a viable alternative, we propose a multi-lane hypotheses approach to narrow down the search by limiting the sample pose to the number of lanes within the GNSS data's error radius. From these poses, we then perform an efficient map-matching to refine these pose candidates. Moreover, a novel belief function to evaluate the hypothesis is proposed to select the best hypothesis for system initialization robustly. Our evaluation result shows that we have outperformed the primary random sampling method in both accuracy and efficiency.

Magnetic coupled resonance wireless power transfer system has the advantages of long transfer distance and high efficiency. However, it needs a high-frequency AC power, and the influence of the circuit parasitic elements at high frequency cannot be ignored, which brings challenges to the research and the design of the high-frequency power for the wireless power transfer system. In this paper, the effect of the parasitic elements in the full-bridge inverter and the reasons of high-frequency ringing generated are analyzed in detail. The influences of device packaging and the printed circuit board (PCB) layout on circuit parasitic elements are studied. An optimized inverter PCB layout design that aims to reduce the parasitic elements and to provide a stable and high-quality AC power for the wireless power transfer system is presented. Finally, the performance of the conventional PCB layout design and the proposed optimized PCB layout design are compared through experiments. The experiment results show that the proposed design not only reduces the parasitic inductance in the circuit, but also reduces the electromagnetic interference noises and improves the stability of the inverter at 900kHz switching frequency.

Wind energy is a kind of sustainable energy with strong uncertainty. With a large amount of wind power injected into the power grid, it will inevitably affect the security, stability and economic operation of the power grid. High-precision wind power spot prediction and fluctuation interval information can provide more adequate decision-making support for grid scheduling and optimization. Hence, this paper proposes a K-Means-long short-term memory (K-Means-LSTM) network model for wind power spot prediction, and a nonparametric kernel density estimation (KDE) model with bandwidth optimization for wind power probabilistic interval prediction. The long short-term memory (LSTM) network has a strong memory function, which can establish the correlation between the data before and after, so as to improve the prediction accuracy. The K-Means clustering method forms different clusters of wind power impact factors to generate a new LSTM sub-prediction model. The optimization of the bandwidth in the nonparametric KDE is implemented by the mean integrated squared error criterion. In addition, a part of the dataset is deliberately demarcated from the wind power historical dataset to generate reasonable wind power prediction errors. The simulation results show that the proposed K-Means-LSTM network model has higher prediction accuracy than the back propagation (BP) neural networks, Elman neural networks, support vector regression (SVR) and LSTM network models. Compared with the KDE model with random bandwidth and the Gaussian distribution model, the bandwidth optimization model proposed in this paper has more narrow prediction intervals with higher interval coverage rates.

The emergence of the Internet of Things (IoT) is not only a global revolution in the information industry, but also brought tremendous changes to our lives. With the development of the technology and means of the IoT, information security issues have gradually emerged, and intrusion attacks have become one of the main problems of the IoT network security. The network layer of the IoT is the key connecting the platform and sensors or controllers of the IoT, and it is also the most standardized, the strongest and the most mature part of the whole physical network architecture. Its large-scale development has led to the network layer's security issues will receive more attention and face more challenges. This paper proposes an intrusion detection algorithm deployed on the network layer of the IoT, which uses the BPSO algorithm to extract features from the NSL-KDD dataset, and applies support vector machines (SVM) as the core model of the algorithm to detect and identify abnormal data, especially DoS attacks. Experimental results show that the model's detection rate of abnormal data and DoS attacks are significantly improved.

With the increase of the types of devices and he improvement of equipment capabilities, the Internet of Things is developing rapidly, which makes the applications in all fields of society and life more intelligent, but at the same time, the diversity and scale of devices lead to the heterogeneity issues in the system, which hinder the resource interoperability and information sharing in IoT application. In view of the fact that semantic technology provides machine-understandable descriptions, and enables better data interaction among machines, this paper uses the ontology to establish the semantic model of the Internet of Things mapping between the physical world and the information world. It not only supports resource description and solves interoperability problems, but also simplifies the difficulty of unified information representation and management realizing data integration and information sharing between different IoT applications.

Model Order Reduction (MOR) is a powerful tool to simulate very large scale system. To integrated circuit system, passive and stability are major criteria. The paper presents the Truncation Balanced Reduction (TBR) which can guarantee passive and stability algorithm based on the Hankel singular parameters. About how to select the truncated states, the paper presents the 102 rule with a example.

Traditional online monitoring methods are passive and less flexible. In this paper, a way of actively injecting tiny torque pulsation through the frequency conversion control system of the drive motor is presented, which implements the state perception of mechanical transmission system. At the steady state of the frequency conversion motor, a tiny swept-frequency signal is injected through the Flux Orientation Control strategy into the stator torque component current. The response data is calculated and analyzed to obtain the torsion frequency response curve of the system and extract each order natural frequency of the system to realize online monitoring. Finally, Carry out the comparative simulation verification and result analysis.

The frequency-coupling dynamics of the modular multilevel converter (MMC) complicate the stability analysis and controller design of MMC-based high-voltage direct current (HVDC) transmission systems. In this paper, a multi-frequency state-space model of MMC with the closed-loop control system is established, which allows to identify the critical oscillation modes and their sensitivities to the control parameters, and thus provides a design-oriented stability analysis. The mathematical relationships between the multi-frequency state-space models in different frames are established, which bridges the gap between the frequency-coupling dynamics of the MMC power stage and that of asymmetrical control loops. Design-oriented stability analysis is then performed based on a generic MMC-HVDC model, which confirms the effectiveness of the modeling techniques.

The rail can be competitive with car and air travel in terms of costs and journey time in countries with developed rail networks. However, it may potentially play a role in facilitating pandemic spread when encountering SARS-CoV-2. Mainly focusing on the role of rail traffic, this research constructs a 2-layer China rail network topological model based on collected China rail data, identifying 3 dynamic transmission scenarios. Then introduces the rail mobility rate into the traditional dynamic transmission model in epidemiology to fit the model for rail transportation. Finally, taking SARS-CoV-2 as an example, the transmission dynamics of sars-cov-2 is studied by simulation experiments, and the orbital movement rate control in different scenarios is measured as one of the potential interventions. Experiment outcomes lead to some managerial insights that (1) with an initial outbreak on the core layer, a faster response is required, while with an eruption on the periphery layer, more appropriate reactions and interventions are encouraged. (2) Adjusting the rail mobility rate of periphery transmission can be used to delay the local outbreak of SARS-CoV-2 in previously unaffected regions.

Job shop scheduling problem (JSP) is one of the most difficult scheduling problems, and flexible job shop scheduling problem (FJSP) is an extension of the classical JSP. In FJSP, a machine for each process should be selected from a given set, which introduces another decision element in the job path that makes FJSP more difficult than traditional JSP. In this paper, a recent proposed intelligence algorithm named grasshopper optimization algorithm (GOA) is used to solve FJSP. Numerical results with comparisons of other classic algorithm counterparts show that GOA has stronger global searching ability and performs better when solving FJSP.

Industrial scheduling problem has always been one of key problems in manufacturing and management planning enterprises. Flexible job shop scheduling problem is one of most difficult scheduling problems, which is also a classical NP hard problem. Compared to the JSP, each process in the FJSP should be processed on each available machine to minimize the total processing time. For this reason, it is far more complex than the classic JSP. In this paper, a recent proposed swarm intelligence algorithm named dragonfly algorithm (DA) is adopted to solve the FJSP. Some traditional algorithms are also used to evaluate the performance on several famous FJSP benchmark compared with DA. Numerical results show that DA has stronger global searching ability and can achieve better results in solving FJSP than other algorithms.

With the continuous access of a large number of diverse devices in the Internet of Things, the scope of application areas is gradually expanding. In order to solve the problem of heterogeneity and scale of device entities and data in the Internet of things, the paper combines web services and semantic technology to build an Internet of Things middleware oriented to semantic web services. The middleware is mainly to serve the Internet of things resources and data semantics, and shields the differences by using the standardized and unified description of the ontology model. Pub/sub communication is used to improve the efficiency of massive data distribution, and through a series of operations such as service registration, discovery and composition, combined with the quality of service to select the most appropriate service for the application to complete user demands. Finally, semantic reasoning can be carried out according to knowledge rules to improve the application of the Internet of things, making it more intelligent.

To relief the stress of global water shortage, a method to optimize the joint operation the coastal source-load-storage systems and the desalination plants is proposed. The objective function is established based on the minimization of the system total cost and the tie-line power fluctuation. Non-dominated Sorting Genetic Algorithm-III (NSGA-III) is selected as the optimization algorithm. A case study is conducted to verify the methodology. The optimization results can determine the optimal operation plan of the components in the system, and the Pareto-optimal fronts can be obtained for trading-off between different operation plans. Multi-objective Particle Swarm Optimization (MOPSO) and Multi-objective Artificial Bee Colony (MOABC) are selected as the comparing algorithms to solve the optimization. Results shows that the proposed methodology is able to achieve the optimal operation strategy of the coastal system, and NSGA-III has better performance on the operation optimization of the system.

The CPU-based deep learning seatbelt detection cannot satisfy the real-time requirements. The precision of seatbelt detection obviously decrease on the basis of deep learning due to the influence of some factors, like the light, rain and snow, window reflection etc. In order to improving detection precision under running CPU in real-time, the SSD MobileNet V2 network is simplified from network pruning and quantitative processing in this paper. The precision and speed of the detection have been improved according to the above optimization. The precision of seatbelt detection can be improved utilizing the regression constraint of seatbelt within the driver after integrating into one classifier. After that, the feature augmentation was added into the samples in order to improve the precision of seatbelt detection in deeply. Finally, the same driver was tracked by the particle filter algorithm. If the tracking failure, the obtained driver feature vector of deep learning will be used to instead of the histogram feature of particle filter and used for the following multi-frame image detection of seat belt. Finally, the detection results are 95.21% for detected precision, 97.19% for recall rate, and 21FPS for detected speed, respectively.

Flexible job shop scheduling problem (FJSP) has long been a complex problem due to the resource flexibility and strong constraints, generating a mixed-integer non-linear optimization problem. The problem becomes more complex with the increasing demand of energy reduction and the corresponding environmental impacts. Proper production scheduling is of significant potential in saving energy in the manufacturing system. In this paper, a multi-objective FJSP model is formulated with the objectives of minimizing the makespan and energy consumption considering strong transportation constraints. Two popular multi-objective optimization solver including Non-dominated Sorting Genetic Algorithm-II (NSGA-II) and A Multiobjective Evolutionary Algorithm Based on Decomposition (MOEA/D) are employed and compared in a real-world instance of the FJSP, associated with novel coding schemes. The results show that the proposed model is well solved by the two solvers and NSGA-II get the better solutions.

The Internet of Things seamlessly integrates physical space and information space, and realizes dynamic interaction through the devices sensing of the state changes in the physical environment real-time, thereby providing more intelligent services. Against the problem that the underlying multi-device data source could generate massive sensor data, this article implements the system to distribute perceptual data according to user interests on demand based on the asynchronous communication mechanism of pub/sub, and proposes a multi-level index subscription classification filtering algorithm. The algorithm mainly classifies the subscription set of the system with subscription size and the subscription frequency of attribute as tags, and completes subscription filtering according to the nature of event matching, to a certain extent reducing the number of subscriptions searched during the event-subscription matching process, accelerating the search speed of subscription.

With the complexity and large-scale of modern control systems, more and more weak faults threaten the safe operation of system equipment. Therefore, finding an effective method for diagnosing weak fault signals has become an urgent problem to be solved. However, the weak fault amplitude is small and the detection is difficult. Therefore, this paper proposes a semi-supervised weak fault symptom enhancement and diagnosis method Signal Enhancement Generative Adversarial Networks(SEGAN) to realize early warning and diagnosis of industrial equipment under complex conditions. The semi- supervised fault symptom enhancement architecture of Signal Enhancement Generative Networks(SEGAN) is constructed, and the weak fault symptom is enhanced by anti-encoding. For the enhanced signal, the fault feature identification method based on KL divergence is proposed to effectively realize timely and rapid fault diagnosis. In this paper, the effectiveness of the proposed method is verified by the natural vibration and percussive vibration data under the fault state of the power tower. Experiments show that the fault diagnosis of the natural vibration signal containing weak fault signs can be realized.

The adjustability of combined heat and power (CHP) plant is one of the main reasons for wind power consumption problem in northern China. During the heating season, a large number of CHP engines are limited by the operation mode of "electricity is determined by heat", and the regulating ability is not high, while the thermal energy storage is an effective way to increase peaking ability for CHP. Phase change heat storage is a new kind of heat storage method emerging in recent years, which has good peaking characteristics to break the thermoelectric constraints of CHP. The peaking strategy is generally based on the prediction of electrical load and thermal load, and the accuracy of the forecast directly affects the result of peaking. A rolling peaking scheme for power grid based on CHP system with phase change heat storage is proposed in this paper. Firstly, the operating characteristics of CHP system with phase change heat storage are analyzed. Secondly, taking the minimum standard deviation of daily load curve as the objective function and considering the constraints of CHP system with phase change heat storage comprehensively, the day-ahead peaking model and rolling peaking model are established. Finally, the differential evolution algorithm is used to solve the peaking model. Simulation results show that the rolling peaking model is more effective than the day-ahead peaking model.

The natural synchronous resonance issue is one of the key features of grid-forming converters. It is caused by the frequency-shifting nature of power calculation when adopting power based synchronization mechanism. This paper is a report of a 33kW power-hardware-in-the-loop test, whose results show that not only synchronous resonance but also near-synchronous resonance phenomenon can occur due to interaction between the active power control loops of grid-forming converters when operating in islanded mode. In addition to that, the test results also show that the different resonance modes are highly related to the cutting frequency design of the low-pass filter used in the power calculation process. In this paper, analytical model is established to figure out the root cause of such interaction. A preliminary design guideline for the filter design is also proposed to avoid such unstable interaction.

As electronic devices are becoming faster and incorporating more functions, they are simultaneously shrinking in size and weight. These factors suggest significant increases in the packaging densities and heat fluxes for the integrated circuits. Effective thermal management will be the key factor to ensure that these devices perform well with high efficiency and reliability. As the heat dissipation concentrates on tiny gate fingers, the operation of the GaN high-electron mobility transistor (HEMT) posts a huge challenge on thermal management. In order to enhance the heat dissipation capability of GaN power device, it is mostly soldered on the heat spreader with the AuSn solder. However, the thermal conductivity of the AuSn solder is only 57 W•K−1•m−1, which cannot fulfil the reliability requirements of future power electronic devices. New interconnection technologies have to be developed and one of them is a low-temperature pressure-less silver sintering paste with nano silver technology.The joining strength, thermal conduction, electric conductivity and long-term reliability of nano-silver soldering paste under low temperature without pressure are studied, and then compared with AuSn bonder. It forms a strong, highly electrically and thermally conductive bond. Chip shear tests show that 200 °C is already sufficient to generate bonds comparable to solder and high-strength welding interface if the remaining parameters (r, t and T, respectively) are set correctly. However, the strength of the welding interface is only a necessary criterion as chip performance comes into play. Therefore, reliability performance of thermal shock test, high temperature storage test and low temperature storage test are run, which return superior reliability of the sintered samples.In addition, the electrical performance of the GaN power chip is tested. The nano-silver bonded sample is compared with the AuSn sintered sample. The saturated signal output power of these two packaging processes can reach 41dBm, and the leakage current is in the normal range of 0.2A~0.3A. Through reliability test, there is no significant change in these samples, and they all achieve normal electrical performance of the power amplifier component.Besides, thermal performance is one of the important parameters affecting the reliability of high-power devices. In the field of electronic assembly, it is of vital importance to reduce the junction temperature. Ordinary conductive adhesives have lower thermal conductivity which generally below 10 W•K−1•m−1 compared with the AuSn solder. Considering that, it is not suitable to join GaN power devices with the AuSn solder and conductive adhesives. In order to verify the heat dissipation performance of the nano-sintered silver solder samples, 20 samples are prepared for thermal imaging experiments. By observing and comparing the temperature distribution on the surface of nano-silver solder paste sintered chips and the AuSn sintered chips. There are some fluctuations in the thermo-resistance of nano-silver solder paste sintered chips and the AuSn sintered chips. Although the thermo-resistance of both samples are concentrated in 1.00 °C•W−1to 1.20 °C•W−1 and heated by the same power, the nano-sintered silver solder achieves around 16.7% decrease in the maximum temperature of the test chip. With the attached heat spreader, 50 W power can be dissipated.

Fault feature extraction based on prior knowledge and raw data is increasingly becoming more challenging in energy Internet fault diagnosis due to complicated network topology and coupling disturbances introduced into the systems. Deep learning methods that have emerged in recent years, such as the convolutional neural networks (CNNs), have shown a number of advantages and great potentials in the field of feature extraction and image recognition. However, CNNs does not work well in fault diagnosis for industrial systems, due to the totally different data representations between images used in recognition and signals obtained from industrial processes. This paper tackles this problem by introducing a novel and generic fault diagnosis method for complicated system, namely, the Spearman rank correlation-based CNNs (SR-CNNs). By imposing the Spearman rank correlation image layer on the typical CNNs, the multiple time-series signals measured by the phasor measurement units (PMUs) is converted to appropriate data images, which are then fed to the CNNs. With the aid of this novel design, different fault features can be comprehensively extracted while the fault can be identified more quickly and precisely than other conventional approaches. To validate the efficacy of the proposed approach, an IEEE defined power gird with many new energy resources are used as the test platform. The experimental results confirm the effectiveness and superiority of the proposed method in energy Internet fault diagnosis over conventional methods.

State analysis is becoming increasingly challenging in energy system due to complicated network topology and coupling disturbances introduced into the systems. Deep learning method based neural network (DNN) that have emerged in recent years have shown a number of advantages and great potentials in the field of feature extraction and pattern recognition. However, the structure and parameters, which are the crucial factors for the performance of DNN, are hard to design and usually depend on the experience. In this paper, a design method of DNN using weight acceleration and information entropy is proposed for energy system state analysis. The Gaussian function and its integral are proposed for qualitative matching the ability of DNN and the complexity of data to tackle the analysis task effectively. Therefore, the under-fitting and over-fitting can be avoided and the performance of DNN is improved. The effectiveness and accuracy of the method are verified by simulating the fault data classification of the energy system.

In recent years, the scale of new energy represented by wind power and photovoltaic has been continuously developed at a high speed. Due to the strong randomness and volatility of wind energy, it poses a severe challenge for the safety, reliability, stability and economy of grid operation. The effective improvement of wind power prediction accuracy is of great significance to power grid operation. Along with the continuous development of artificial intelligence methods, the emerging wind power prediction methods have been greatly improved in terms of prediction accuracy and calculation rate. As a current predictive model for deep learning, Long Short-Term Memory (LSTM) is excellent for predicting long-term sequences. This paper proposes a wind power prediction method based on LSTM. In addition, Principal Component Analysis (PCA) method is used to achieve the dimension reduction of data. The simulation results demonstrate the effectiveness and advantages of the proposed PCA-LSTM method.

Multi-set query is a fundamental issue in data science. When the sizes of multi-sets are large, exact matching methods like hash tables need too much memory, and they cannot achieve high query speed. Bloom filters are recently used to handle big data query, but they cannot achieve high accuracy when the memory space is tight. In this paper, we propose a new data structure named coloring embedder, which is fast, accurate as well as memory efficient. The insight is to first map elements to a high dimensional space to almost eliminate hashing collisions, and then use a dimensional reduction representation, which is similar to coloring a graph, to save memory. Theoretical proofs and experimental results show that compared to the state-of-the[1]art, the error rate of the coloring embedder is thousands of times smaller even with much less memory usage, and the query speed of the coloring embedder is about 2 times faster. The source code of coloring embedder is released on Github.