Multiple Parallel Branches Research Articles

Modeling and Nonlinear Dynamic Analysis of a Photovoltaic System with Multiple Parallel Branches Based on Simplified Discrete Time Model

Modeling and Nonlinear Dynamic Analysis of a Photovoltaic System with Multiple Parallel Branches Based on Simplified Discrete Time Model

Improved detector in orchard via top-to-down texture enhancement and adaptive region-aware feature fusion

Accurate target detection in complex orchard environments is the basis for automatic picking and pollination. The characteristics of small, clustered and complex interference greatly increase the difficulty of detection. Toward this end, we explore a detector in the orchard and improve the detection ability of complex targets. Our model includes two core designs to make it suitable for reducing the risk of error detection due to small and camouflaged object features. Multi-scale texture enhancement design focuses on extracting and enhancing more distinguishable features for each level with multiple parallel branches. Our adaptive region-aware feature fusion module extracts the dependencies between locations and channels, potential cross-relations among different levels and multi-types information to build distinctive representations. By combining enhancement and fusion, experiments on various real-world datasets show that the proposed network can outperform previous state-of-the-art methods, especially for detection in complex conditions.

HS-YOLO: Small Object Detection for Power Operation Scenarios

Object detection methods are commonly employed in power safety monitoring systems to detect violations in surveillance scenes. However, traditional object detection methods are ineffective for small objects that are similar to the background information in the power monitoring scene, which consequently affects the performance of violation behavior detection. This paper proposed a small object detection algorithm named HS-YOLO, based on High-Resolution Network (HRNet) and sub-pixel convolution. First, to fully extract the microfeature information of the object, a small object feature extraction backbone network is proposed based on the HRNet structure. The feature maps of different scales are processed by multiple parallel branches and fused with each other in the network. Then, to fully retain the effective features of small objects, the sub-pixel convolution module is incorporated as the upsampling operator in the feature fusion network. The low-resolution feature map is upsampled to a higher resolution by reorganizing pixel values and performing padding operations in this module. On our self-constructed power operation dataset, the HS-YOLO algorithm achieved a mAP of 87.2%, which is a 3.5% improvement compared to YOLOv5. Particularly, the dataset’s AP for detecting small objects such as cuffs, necklines, and safety belts is improved by 10.7%, 5.8%, and 4.4%, respectively. These results demonstrate the effectiveness of our proposed method in detecting small objects in power operation scenarios.

Joint dual-stream interaction and multi-scale feature extraction network for multi-spectral pedestrian detection

Multispectral images can provide more information, so pedestrian detection based on multispectral images has received wide attention. Existing multispectral networks mainly focus on the misalignment of image pairs and the difference between modalities. However, these network structures lack effective information interaction between two feature streams and fail to consider the scale characteristics of pedestrian objects. To deal with this issue, we propose a high-performance network structure, which is called dual-stream interaction and multi-scale feature extraction network (DSI-MSE), and contains a dual-stream feature interaction (DSI) block, a multi-scale feature extraction (MSE) block and a detection (DET) block. The DSI block extracts the features through the dual-stream interaction of RGB images and thermal images, which fuses the intra-modal information and the inter-modal information. The MSE block is designed by multiple parallel branches for matching multiple scales of pedestrian, which enhances the expressiveness of features and refines richer feature expressions at different scales. Experimental results on KAIST and CVC-14 datasets demonstrate that the proposed DSI-MSE can obtain the state-of-the-art results on multi-spectral pedestrian detection tasks.

Image inpainting via progressive decoder and gradient guidance

Very recently, with the widespread research of deep learning, its achievements are increasingly evident in image inpainting tasks. However, many existing multi-stage methods fail to effectively inpainting the larger missing areas, their common drawback is that the result of each stage is easily misguided by the wrong content generated in the previous stage. To solve this issue, in this paper, a novel one-stage generative adversarial network based on the progressive decoding architecture and gradient guidance. Firstly, gradient priors are extracted at the encoder stage to be passed to the decoding branch, and multiscale attention fusion group is used to help the network understand the image features. Secondly, multiple parallel decoding branches fill and refine the missing regions by top-down passing the reconstructed priors. This progressively guided repair avoids the detrimental effects of inappropriate priors. The joint guidance of features and gradient priors helps the restoration results contain the correct structure and rich details. And the progressive guidance is achieved by our fusion strategy, combining reimage convolution and design channel coordinate attention to fuse and reweight the features of different branches. Finally, we use the multiscale fusion to merge the feature maps at different scales reconstructed by the last decoding branch and map them to the image space, which further improves the semantic plausibility of the restoration results. Experiments on multiple datasets show that the qualitative and quantitative results of our computationally efficient model are competitive with those of state-of-the-art methods.

Investigation of current balance methodologies for inductive power transfer coupler with multiple branches

AbstractIn the inductive power transfer (IPT) system, using multiple coil branches instead of a whole large coil can effectively solve the thickness problem and save space. In that case, there are small parameter differences in mutual inductance and self‐inductance between the coil branches in parallel. These differences are amplified in high‐frequency IPT systems, resulting in different currents flowing through the different coil branches. The current unbalance will create many hazards. Therefore, this paper analyzes the key factors that cause current to unbalance and studies the equivalent circuit model of an inductive power transfer system with secondary multiple parallel branches. Moreover, this paper designs the compensation circuit and studies two kinds of current balance methods. Furthermore, the parameter design criteria are detailed to obtain a nearly ideal current balance effect. The characteristic analysis, parameter design, and verification are described in detail.

ADE-CycleGAN: A Detail Enhanced Image Dehazing CycleGAN Network

The preservation of image details in the defogging process is still one key challenge in the field of deep learning. The network uses the generation of confrontation loss and cyclic consistency loss to ensure that the generated defog image is similar to the original image, but it cannot retain the details of the image. To this end, we propose a detail enhanced image CycleGAN to retain the detail information during the process of defogging. Firstly, the algorithm uses the CycleGAN network as the basic framework and combines the U-Net network's idea with this framework to extract visual information features in different spaces of the image in multiple parallel branches, and it introduces Dep residual blocks to learn deeper feature information. Secondly, a multi-head attention mechanism is introduced in the generator to strengthen the expressive ability of features and balance the deviation produced by the same attention mechanism. Finally, experiments are carried out on the public data set D-Hazy. Compared with the CycleGAN network, the network structure of this paper improves the SSIM and PSNR of the image dehazing effect by 12.2% and 8.1% compared with the network and can retain image dehazing details.

Inpainting larger missing regions via progressive guidance decoding network

AbstractFor images corrupted for various reasons, the size of the corrupted area is often arbitrary and it has been a challenge to inpainting the larger missing areas. Though popular multistage networks ease the inpainting difficulty by repairing damaged image from coarse to fine, their common drawback is that the result of each stage is easily misguided by the wrong content generated in the previous stage. To address this problem, we propose a novel progressive guidance decoding network. First, multiple parallel decoding branches fill and refine the missing regions by top–down passing the reconstructed priors. This inpainting way of progressive guidance avoids adverse effects of inappropriate premises, since the decoding branches can learn what priors can be utilized. And convolution layers of decoder with different locations would pass down the different priors. The joint guidance of features and gradient priors helps the inpainting result contains the correct structure and rich details. The second fold of progressive guidance is achieved by our fusing strategy, combining ghost convolution and the designed cascaded efficient channel attention (CECA) to fuse and reweight the features from different branches. CECA explores the dependencies among adjant and non-adjant channels more effectively than popular ones. Finally, we merges the different-scale feature maps reconstructed by the last decoding branch and mapping them to the image space, which further improves the semantic plausibility of the restoration results. Extensive experiments verify the effectiveness of our method in both subjective and objective evaluation.

Effective Point Cloud Analysis Using Multi-Scale Features.

Fully exploring the correlation of local features and their spatial distribution in point clouds is essential for feature modeling. This paper, inspired by convolutional neural networks (CNNs), explores the relationship between local patterns and point coordinates from a novel perspective and proposes a lightweight structure based on multi-scale features and a two-step fusion strategy. Specifically, local features of multi-scales and their spatial distribution can be regarded as independent features corresponding to different levels of geometric significance, which are extracted by multiple parallel branches and then merged on multiple levels. In this way, the proposed model generates a shape-level representation that contains rich local characteristics and the spatial relationship between them. Moreover, with the shared multi-layer perceptrons (MLPs) as basic operators, the proposed structure is so concise that it converges rapidly, and so we introduce the snapshot ensemble to improve performance further. The model is evaluated on classification and part segmentation tasks. The experiments prove that our model achieves on-par or better performance than previous state-of-the-art (SOTA) methods.

Quantifying the Effect of Multiphase Flow on Matrix Acidizing in Oil-Bearing Carbonate Formations

Summary It is common to inject acidic stimulation fluids into oil-bearing carbonate formations to enhance well productivity. This process of matrix acidizing is designed to maximize the propagation of wormholes into the formation by optimizing the injection parameters, including acid-injection rate and volume. Previous studies have suggested that saturation conditions, permeability, heterogeneity, temperature, and pressure can significantly affect the design of matrix-acidizing treatments. However, laboratory studies’ results are inconsistent in their conclusions and are mostly limited to water-saturated cores. In this work, we designed a systematic experimental study to evaluate the impact of multiphase flow on the acidizing process when injecting 15 wt% hydrochloric acid (HCl) into crude-oil-saturated Indiana Limestone cores. The results reveal the following: Contrary to published literature for water-saturated cores, acidizing in partially oil-saturated high-permeability cores at high pressure requires less acid volume than in low-permeability cores; lower-pressure acid injection results in more efficient wormhole propagation in low-permeability cores compared to high-pressure acid injection; acidizing in low- and high-permeability cores at low pressure leads to similar efficiency; and wormholing is more effective in partially oil-saturated cores, resulting in multiple parallel branches as compared to inefficient leakoff in water-saturated cores.

See more than once: Kernel-sharing atrous convolution for semantic segmentation

The state-of-the-art semantic segmentation solutions usually leverage different receptive fields via multiple parallel branches to handle objects of different sizes. However, employing separate kernels for individual branches may degrade the generalization of the network to objects with different scales, and the computational cost increases with the increase of the number of branches. To tackle this problem, we propose a novel network structure, namely Kernel-Sharing Atrous Convolution (KSAC), where branches with different receptive fields share the same kernel, i.e., let a single kernel ‘see’ the input feature maps more than once with different receptive fields. Experiments conducted on the benchmark PASCAL VOC 2012 dataset show that our proposed sharing strategy can not only boost the network’s generalization and representation abilities but also reduce the computational cost significantly. Specifically, on the validation set, when compared with DeepLabv3+, about 2.7G FLOPs and 12.7G FLOPs are saved for output stride = 16 and 8 respectively. In addition, different from the widely used ASPP structure, our proposed KSAC is able to further improve the mIOU by taking benefit of wider context with larger atrous rates. Finally, our KSAC achieves mIOUs of 88.1%, 45.47% and 80.7% on the PASCAL VOC 2012 test set (Everingham et al., 2009), ADE20K dataset (Zhou et al., 2017) and Cityscapes datasets (Marius et al., 2016), respectively. Our full code will be released on Github: https://github.com/edwardyehuang/iSeg.

Bayesian Estimation of Model Parameters of Equivalent Circuit Model for Detecting Degradation Parts of Lithium-Ion Battery

Nowadays, the use of electric vehicles is increasing leading to a growing demand for more efficient use of lithium-ion batteries. The state-of-charge (SOC) has been estimated in previous studies to optimize energy management of batteries. For more efficient battery utilization, detecting degradation is important. However, it is difficult for conventional methods to distinguish the effect of the model parameters including different time constants. Identifying model parameters of multiple RC parallel branches, which represent the impedance of wider frequency ranges, is a necessary requirement to detect the degradation of parts. In this study, we present a method for estimating the model parameters of multiple RC parallel branches. We designed the Markov Chain Monte Carlo algorithm by setting a search range limit and moving window, which enable estimation of the model parameters of parallel branches of different time constants. Through validation of the algorithm based on simulation, the model parameters of a third-order circuit were estimated to be within the error range of 15.2 %. In addition, impedance was calculated from the estimated model parameters in the test using a real battery dataset. The error of impedance was less than 10 % from 0.01 to 100 Hz which was sufficiently low to monitor the change of the parameters owing to degradation. As the impedance in the high-frequency band above 0.1 Hz is more likely to change because of degradation, the proposed method can be used to monitor the model parameters that change as a result of degradation.

Smart panel with piezoelectric patches connected to adaptive multi-resonant shunts for broadband vibration control

This paper presents a simulation study on the broadband control of flexural vibration of thin panels equipped with piezoelectric patches connected to adaptive multi-resonant shunts. The study considers the broadband flexural response of a thin rectangular aluminum panel exposed to a spatial and time stochastic excitation. The panel is equipped with thin square piezoelectric patches connect to shunts encompassing multiple parallel branches formed by RLC components connected in series. The capacitance C of each branch is kept fixed, whereas the inductance L and resistance R are adapted so as each branch maximizes the vibration power absorption from the resonant response of a specific flexural mode of the panel. The study proposes an on-line tuning approach, where the N-branches of each shunt are adapted sequentially to maximize the vibration absorption from the resonant responses of N neighbor flexural modes. The vibration power absorption of each branch is estimated from the electric power absorbed by the shunt. In this way a local tuning system can be implemented, which adapts on-line the shunts to control the flexural response produced by a group of flexural modes of the panel resonating in a given frequency band.

Improve YOLOv3 using dilated spatial pyramid module for multi-scale object detection

Effectively and efficiently recognizing multi-scale objects is one of the key challenges of utilizing deep convolutional neural network to the object detection field. YOLOv3 (You only look once v3) is the state-of-the-art object detector with good performance in both aspects of accuracy and speed; however, the scale variation is still the challenging problem which needs to be improved. Considering that the detection performances of multi-scale objects are related to the receptive fields of the network, in this work, we propose a novel dilated spatial pyramid module to integrate multi-scale information to effectively deal with scale variation problem. Firstly, the input of dilated spatial pyramid is fed into multiple parallel branches with different dilation rates to generate feature maps with different receptive fields. Then, the input of dilated spatial pyramid and outputs of different branches are concatenated to integrate multi-scale information. Moreover, dilated spatial pyramid is integrated with YOLOv3 in front of the first detection header to present dilated spatial pyramid-You only look once model. Experiment results on PASCAL VOC2007 demonstrate that dilated spatial pyramid-You only look once model outperforms other state-of-the-art methods in mean average precision, while it still keeps a satisfying real-time detection speed. For 416 × 416 input, dilated spatial pyramid-You only look once model achieves 82.2% mean average precision at 56 frames per second, 3.9% higher than YOLOv3 with only slight speed drops.

PMBANet: Progressive Multi-Branch Aggregation Network for Scene Depth Super-Resolution

Depth map super-resolution is an ill-posed inverse problem with many challenges. First, depth boundaries are generally hard to reconstruct particularly at large magnification factors. Second, depth regions on fine structures and tiny objects in the scene are destroyed seriously by downsampling degradation. To tackle these difficulties, we propose a progressive multi-branch aggregation network (PMBANet), which consists of stacked MBA blocks to fully address the above problems and progressively recover the degraded depth map. Specifically, each MBA block has multiple parallel branches: 1) The reconstruction branch is proposed based on the designed attention-based error feed-forward/-back modules, which iteratively exploits and compensates the downsampling errors to refine the depth map by imposing the attention mechanism on the module to gradually highlight the informative features at depth boundaries. 2) We formulate a separate guidance branch as prior knowledge to help to recover the depth details, in which the multi-scale branch is to learn a multi-scale representation that pays close attention at objects of different scales, while the color branch regularizes the depth map by using auxiliary color information. Then, a fusion block is introduced to adaptively fuse and select the discriminative features from all the branches. The design methodology of our whole network is well-founded, and extensive experiments on benchmark datasets demonstrate that our method achieves superior performance in comparison with the state-of-the-art methods. Our code and models are available at https://github.com/Sunbaoli/PMBANet_DSR/ .

High-Mobility Wideband Massive MIMO Communications: Doppler Compensation, Analysis and Scaling Laws

In this paper, we apply angle-domain Doppler compensation for high-mobility wideband massive multi-input multi-output (MIMO) uplink communications. The time-varying multipath channel is considered between high-speed terminal and static base station (BS), where multiple Doppler frequency offsets (DFOs) are associated with distinct angle of departures (AoDs). With the aid of large-scale uniform linear array (ULA) at the transmitter, we design a beamforming network to generate multiple parallel beamforming branches, each transmitting signal pointing to one particular angle. Then, the transmitted signal in each branch will experience only one dominant DFO when passing over the time-varying channel, which can be easily compensated before transmission starts. We theoretically analyze the Doppler spread of the equivalent uplink channel after angle-domain Doppler compensation, which takes into account both the mainlobe and sidelobes of the transmit beam in each branch. It is seen that the channel time-variation can be effectively suppressed if the number of transmit antennas is sufficiently large. Interestingly, the asymptotic scaling law of channel variation is obtained, which shows that the Doppler spread is proportional to the maximum DFO and decreases approximately as $1/\sqrt {M}$ ( $M$ is the number of transmit antennas) when $M$ is sufficiently large. The numerical results are provided to corroborate the proposed scheme.

Investigation of Flux-Coupling Type Superconducting Fault Current Limiter With Multiple Parallel Branches

The flux-coupling type superconducting fault-current limiter (FC-SFCL) is discussed in this paper. Considering the shortcomings of conventional inductive and resistive SFCLs, two novel FC-SFCLs with multiple parallel branches are carried out. Considering the characteristics of YBCO tapes in 77-K LN2 environment, the design scheme of the HTS module is studied according to the impedance parameters, the number of parallel branches, and insulation distance. Furthermore, the electromagnetic analysis and parameter calculation of the module are performed. In order to identify the status of the HTS module in normal operating, current limiting, and reclosing processes, the status identification method for HTS module is analyzed. Finally, the possible application of FC-SFCL in power grid is discussed.

Frequency‐dependent cable modelling for small‐signal stability analysis of VSC‐HVDC systems

State-of-the-art time-domain models of power cables account for the frequency dependency of physical parameters to enable accurate transient simulations of high-voltage transmission schemes. Owing to their formulation, these models cannot be directly converted into a state-space form as required for small-signal eigenvalue analysis. Thus, dc cables are commonly represented in high-voltage direct current (HVDC) power system stability studies by cascaded pi-section equivalents that neglect the frequency-dependent effects. This study demonstrates how the conventional cascaded pi-section model is unable to accurately represent the damping characteristic of the cable and how this can lead to incorrect stability assessments. Furthermore, an alternative model consisting of cascaded pi-sections with multiple parallel branches is explored, which allows for a state-space representation while accounting for the frequency dependency of the cable parameters. The performance of the proposed model is benchmarked against state-of-the-art cable models both in the frequency domain and in the time domain. Finally, the study provides a comparative example of the impact of the cable modelling on the small-signal dynamics of a point-to-point voltage-source converter (VSC) HVDC transmission scheme.

Modelling and Output Power Evaluation of Series-Parallel Photovoltaic Modules

Solar energy has received attention in the Middle East given the abundant and free irradiance and extended sunny weather. Although photovoltaic panels were introduced decades ago, they have recently become economical and gained traction. We present a mathematical model for series-parallel photovoltaic modules, evaluate the model, and present the I-V and P-V characteristic plots for various temperatures, irradiance, and diode ideality factors. The power performance results are then analyzed and recommendations are made. Unlike other related work, our evaluation uses standard spreadsheet software avoiding commercial simulation packages and application programming. Results indicate improved power performance with irradiance and parallel connections of cell series branches. Given the hot weather in our region, increasing the number of cells connected in series from 36 to 72 is not recommended. I. INTRODUCTION The push for renewable energy solutions is steady to reduce carbon emissions, protect the environment and population health, and effect climate and sea water level changes. Solar energy has received attention in the Middle East given the abundant and free irradiance and sunny weather. Although photovoltaic (PV) panels were introduced decades ago, they have recently become economical and gained traction. Saudi Arabia recently saw the successful installation of a 3.5 megawatt photovoltaic field, the largest solar power plant built in the country, and has plans to install 41 Gigawatts of solar power over the next 20 years. PV fields geographically distributed can be located near loads and cut down on fuel consumed by power generating plants. PV cells generate DC electricity during the day, which can be immediately consumed or stored in batteries for future consumption. An inverter is used to convert the generated power from DC to AC. The PV panel has a number of solar cells connected in series (typically 36 or 72) with the possibility to connect PV cell series branches in parallel. The solar cell is a p-n junction fabricated in a silicon (or other material) wafer or semiconductor layer. The photovoltaic effect causes the electromagnetic radiation of the solar energy hitting the PV cells to generate electricity. Photons with energy greater than the band gap energy of the wafer's semiconductor material are absorbed, creating electron-hole pairs, which in turn create a photocurrent in the presence of the p-n junction's internal electric field. The photocurrent's magnitude rises with the number of electron-hole pairs created, which is dependent on the irradiance level. Therefore the magnitude of the current generated by the PV module is proportional to the amount of incident solar radiation. PV module models have evolved to include detailed parameters and even multiple piecewise linear regions or better accuracy. However, most have focused on modeling a number of PV cells in series and stopped short of modeling multiple parallel branches of serially connected PV cells. In this paper, we present a mathematical model for series- parallel photovoltaic modules, evaluate the model, and present the I-V (module current vs. module voltage) and P-V (module power vs. voltage) characteristic plots for various temperatures, irradiance and diode ideality factors. The power performance results are then analyzed and recommendations are made. Unlike others' work, we model series-parallel PV cells with multiple parallel branches each consisting of a number of PV cells in series, and evaluate the model with a basic spreadsheet application, avoiding application programming and commercial simulators. This is facilitated by the mathematical technique (i.e. Newton-Raphson) used in the model evaluation which makes the model evaluation possible with a spreadsheet application. Thus, a key advantage of our approach is that our model can be reused by a larger population of researchers with no access to commercial simulator packages.

Blind channel identification from burst data using implicit matching of HOS

In this paper we present a new approach for blind identification of non-minimum phase FIR channels for single input single output (SISO) and multiple inputs single output (MISO) scenarios. The received record of samples is processed by multiple parallel branches, each comprising an FIR filter followed by a non-linear function and an accumulator. Using an appropriate cost function, the obtained vector of averages is then best fitted by a function of the system model. The non-linear function used in each branch is selected so as to obtain an implicit matching of all higher order statistics (HOS) of certain orders. Choice of the cumulants generating function (CGF) as this non-linear function is also possible and can be interpreted as matching the joint probability density function (PDF) for selected samples in the vector space of the channel. With this interpretation the coefficients of the FIR filters are actually sampling points in a multidimensional frequency domain. The main advantages of this approach are its high probability of identification success, its ability to obtain reliable channel estimation in low SNR using a short record of samples and its relative insensitivity to overestimation of the channel order. The method is suitable for single transmit and receive antennas and has easy extensions to multiple antenna scenarios. The validity of the proposed approach is confirmed by simulations for BPSK and QPSK modulations and suitable design guidelines for these modulations are given.