Parallel Structure Research Articles

Improving flood forecast accuracy based on explainable convolutional neural network by Grad-CAM method

The advent of deep learning techniques has shown promising improvement in flood forecasting accuracy which are crucial for operating reservoir and mitigating flood damages. However, the practical application of typical deep learning models, such as the Long Short-Term Memory network (LSTM), is hindered by their high complexity, computational cost, and lack of interpretability. This study introduces an Explainable Convolutional Neural Network (ECNN) model that can effectively capture inter-variable relationships and temporal dynamics to address these challenges. The proposed ECNN model was applied to the Lushui basin in China and compared with benchmark models. The interpretability of the ECNN model was analyzed using the Gradient-weighted Class Activation Mapping (Grad-CAM) method. Our findings demonstrated that the ECNN model exhibits exceptional performance in flood forecast tasks, particularly for long-term forecast horizons. Furthermore, precipitation and inflow play a decisive role in flood forecasting with the watershed-specific flood characteristics. The ECNN model is highly promising in real-time flood forecasting with ingenious parallel structure, remarkable accuracy, and interpretability.

Tradeoff analysis of the energy-harvesting vehicle suspension system employing inerter element

This paper explores the vibration isolation performance and vibration energy recovery performance of an energy-harvesting vehicle suspension system employing inerter element. The study takes into account the structural changes in the suspension caused by introducing the inerter as a new type of vibration isolation element. According to the parallel-series combination of an inerter and a damper, two different structures of energy-harvesting suspension dynamic models are constructed. The mechanism of the suspension parameter uptake on vehicle ride comfort and energy-harvesting characteristics is analyzed. A multi-objective optimal design method for the energy-harvesting vehicle suspension system employing inerter element is proposed, which considers both vehicle ride comfort and energy-harvesting characteristics. A trade-off was found between suspension isolation performance and energy-harvesting efficiency. The results indicate that various structures of energy-harvesting vehicle suspension systems employing inerter element exhibit different vibration isolation performances. Compared with the conventional energy-harvesting suspension, the series structure reduces body acceleration by 15.9 %, and the root mean square (RMS) of energy-harvesting power of the suspension is 22.2 W. The parallel structure reduces the RMS of body acceleration by 14.3 % and the RMS of energy-harvesting efficiency is 56.3 %, with the RMS energy-harvesting power of 84.9 W. The parallel structure demonstrates superior overall performance.

Shaping of the risk of a series-parallel manufacturing structure maintenance according to quasi-coherence and the [formula omitted]-th survival value

The paper presents the authors’ model for determining the probability value of the survival function of the difference of two random variables: TP – the amount of time of correct operation and TB – the amount of time of failure. For this purpose, the concept of K-th survival value is introduced. The K-th survival value defines the probability that the total time of correct operation of the production system will be longer by at least k time units (for k≥0) than the total amount of time of failures. The model was exemplified by a real production system with a series–parallel structure. In the nesting manufacturing structure, the availability of machines determines the number and the type of flow streams of the processed material. To formalise the chains of relations at a given Δt, this paper utilises the quasi-coherence property. The paper identifies four possible cases of system operation: 1) a fully operational system in which every machine is capable of meeting the assumed production schedule; 2) a system which has lost the property of quasi-coherence (certain system objects are in an inoperable state) but it is possible to meet the assumed production schedule; 3) a system which has lost the property of quasi-coherence but the assumed task schedule is possible to meet after redefining the material flow streams; 4) a system which has lost the property of quasi-coherence and it is impossible to meet the assumed production schedule. Based on the definition of the survival function [1] S of the random variable T: ST:=PT>τ=1-FTτ,τ∈R, the concept of K-th survival value was introduced. For possible cases, K-th survival value was defined as: KT1,Ti=PT1-Ti≥k=1-FT1-Tik (for i=2,3,4), where T1 and Ti are random variables with given probability distributions. In turn, k is a deterministic value that determines the total number of time units of correct system operation in relation to the total failure time. The application of the developed algorithm enables the system to be analysed in terms of three aspects: (1) to determine the probability of on-time execution of tasks, taking into account historical values of machine failure and correct operation durations; (2) to determine the stability of the system based on empirical data; (3) to monitor the level of risk in real time. The paper validates the model by determining the stability of the analysed system for real machine failure data. The introduced criterion of stochastic stability of the production system has been verified for data in which distributions of correct operation times and failure times are random variables from the family of Gamma distributions. The presented model can be included in the group of multifactorial risk assessment methods.

CETN: Contrast-enhanced Through Network for Click-Through Rate Prediction

Click-through rate (CTR) prediction is a crucial task in personalized information retrievals, such as industrial recommender systems, online advertising, and web search. Most existing CTR Prediction models utilize explicit feature interactions to overcome the performance bottleneck of implicit feature interactions. Hence, deep CTR models based on parallel structures (e.g., DCN, FinalMLP, xDeepFM) have been proposed to obtain joint information from different semantic spaces. However, these parallel subcomponents lack effective supervision and communication signals, making it challenging to efficiently capture valuable multi-views feature interaction information in different semantic spaces. To address these issues, we propose a simple yet effective novel CTR model: Contrast-enhanced Through Network (CETN). Drawing inspiration from sociology, CETN leverages the complementary nature of diversity and homogeneity to guide the model in acquiring higher-quality feature interaction information. Specifically, CETN employs product-based feature interactions and the augmentation (perturbation) concept from contrastive learning to segment different semantic spaces, each with distinct activation functions. This improves diversity in the feature interaction information captured by the model. Additionally, we introduce self-supervised signals and through connection within each semantic space to ensure the homogeneity of the captured feature interaction information. The experiments conduct on four real datasets demonstrate that our model consistently outperforms twenty baseline models in terms of AUC and Logloss.

Parallel detection structure exploiting the extrinsic information from the serial detection in the holographic data storage system

Parallel detection structure exploiting the extrinsic information from the serial detection in the holographic data storage system

Field-programmable gate array implementation of efficient deep neural network architecture

Deep neural network (DNN) comprises multiple stages of data processing sub-systems with one of the primary sub-systems is a fully connected neural network (FCNN) model. This fully connected neural network model has multiple layers of neurons that need to be implemented using arithmetic units with suitable number representation to optimize area, power, and speed. In this work, the network parameters are analyzed, and redundancy in weights is eliminated. A pipelined and parallel structure is designed for the fully connected network information. The proposed FCNN structure has 16 inputs, 3 hidden layers, and an output layer. Each hidden layer consists of 4 neurons and describes how the inputs are connected to hidden layer neurons to process the raw data. A hardware description language (HDL) model is developed for the proposed structure and the verified model is implemented on Xilinx field-programmable gate array (FPGA). The modified structure comprises registers, demultiplexers, weight registers, multipliers, adders, and read-only memory lookup table (ROM/LUT). The modified architecture implemented on FPGA is estimated to reduce area by 87.5% and improve timing by 3x compared with direct implementation methods.

Research on the biomimetic quadruped jumping robot based on an efficient energy storage structure

AbstractThe ability of quadruped robots to overcome obstacles is a critical factor that limits their practical application. Here, a design concept and a control algorithm are presented that aim at enhancing the explosive force of quadruped robots during jumping by utilizing elastic energy storage components. The hind legs of the quadruped robot are designed as energy storage units. Tension springs are utilized as components for storing energy and are installed in a parallel structure on the hind leg. Energy is stored during the compression process of the robot’s torso and released during the jumping phase. The optimal foot force is calculated using a single rigid body model. The mapping relationship between the force applied to the foot and the resulting joint torque is established by developing a dynamic model of the hind legs. Simulation experiments were conducted using the Webots physics engine to compare the impact of varying spring stiffness on joint torque during the jumping process. This study determined the optimal spring stiffness under specific conditions. The hind legs’ torque saving ratio reaches 19%, and the energy-saving ratio reaches 13%, which validates the effectiveness and feasibility of integrating elastic energy storage components.

A stepwise optimization method for topology structure of fluid machinery network

A stepwise optimization method for topology structure of fluid machinery network

Effect of temperature on the drag reduction characteristics of xanthan gum in the liquid ring vacuum pump

Xanthan gum (XG) solution has been proven to be an effective working fluid in the liquid ring vacuum (LRV) pump to achieve significant energy savings. As an important influencing factor of the LRV pump performance, the effect of temperature on the drag reduction performance, rheological properties, and microstructure of XG solution in the LRV pump were experimentally investigated in this paper. The results indicate that with the increase of temperature, the drag reduction rate of the XG solution in the LRV pump initially rise and subsequently decline, and XG solutions with different concentrations show different optimal drag reduction temperatures. The microscopic parallel double-chain structure of XG has the same change pattern as the drag reduction performance, while the steady-state and dynamic shear characteristics of the XG solution manifest a downtrend. For instance, as the temperature of 4000 ppm XG solution increases from 10 °C to 60 °C, the drag reduction rate decreases by 20 %; the parallel double-chain structure decreases from 91.3 % to 89.9 %; the yield stress and relaxation time decrease from 4.63993 Pa to 4477.56 s to 3.37808 Pa and 143.27 s, respectively; the drag reduction rate reaches the maximum value when the temperature is 40 °C.

How to optimize service-oriented cloud manufacturing

Within the context of Industry 4.0, service-oriented cloud manufacturing has emerged as a paradigm of significant interest, noted for enhancing manufacturing efficacy through integration, flexibility, and customization. The challenges of service selection and optimization, alongside transportation optimization, are pivotal for the effective implementation of service-oriented cloud manufacturing. This study explores an integrated problem that incorporates both sequential and parallel structures. To address this intricate issue, a binary-integer programming model is proposed, aiming to minimize the cumulative costs, including both manufacturing and transportation expenses. The validity and effectiveness of the proposed model are demonstrated through a real-world case study in mold manufacturing. Analysis of the experimental results provides managerial insights, which could inform the implementation and improvement of service-oriented cloud manufacturing strategies.

Frailty model for multiple repairable systems hierarchically represented subject to competing risks

In this paper, we propose a statistical model to describe the behaviour of failure times associated with groups of repairable systems hierarchically represented, under a competing risks framework, considering the existence of unobserved heterogeneity that acts individually on systems of each group, as well as the possibility of imperfect repairs whose initial failure rate is in the form of the power law. In this context for the unobserved heterogeneity in the groups, we consider the multiplicative frailty. To illustrate the use of the proposed model, we consider a database with the failures of 38 agricultural machines categorized into five different groups. We understand that the tractor fleet corresponds to the farm's agricultural system, therefore the need for intervention in this system occurs with the failure of any unit, individually, in a serial structure, of competing risks.e On the other hand, the understanding of the time in which all the machinery that makes up the fleet will have required some intervention is obtained by analysing the results under a parallel structure.

Hybrid modeling for improved extrapolation and transfer learning in the chemical processing industry

In the Chemical Processing Industry, shifts in market demand often require the implementation of new operational modes that balance economic advantages with the need to meet quality and sustainability targets. Hybrid modeling can support these process changes by combining physics-based and data-driven modeling principles to improve prediction accuracy and interpretation. In this work, the use of hybrid modeling for improved extrapolation and transfer learning activities is examined in the context of simulated biodiesel production. We study and compare different configurations of hybrid modeling, including the parallel and serial structures. We also compare hybrid modeling approaches with physics-based and data-driven models, and find that hybrid modeling consistently outperforms these benchmarks in both extrapolation and transfer learning tasks. Hybrid modeling also requires fewer samples than other benchmarks for the transfer learning task. These results suggest that hybrid modeling is an effective approach for supporting decision-making and optimizing process changes in the CPI.

Band-Pass and Band-Stop Filter Frequency Selective Surface with Harmonic Suppression

This study investigated a frequency-selective surface (FSS) that is used to suppress harmonics affecting the performance and accuracy of radar systems. One side of the FSS features a metal grid structure, and when converted into an equivalent circuit model, it exhibits the characteristics of a band-pass filter with an L and C parallel structure. The other side of the FSS features a metallic loop structure and when represented as an equivalent circuit model, it exhibits the characteristics of a band-stop filter with an L and C series structure. The reflection coefficient (S11) and transmission coefficient (S21) of the FSS designed based on theory are compared using a CST studio suite and Keysight’s Advanced Design System. In addition, the transmission coefficients are verified through actual measurements, wherein the measured transmission coefficient is −0.1 dB at 3.0 GHz and approximately −50 dB at the harmonic frequency of 6.0 GHz. The designed FSS is attached to an actual radar system, and the 2D radiation pattern and maximum gain are measured during steering in boresight, azimuth (30∘) and elevation (30∘) directions. At 3.0 GHz, the maximum gain in boresight is 17.25 dB without the FSS and 17.12 dB with the FSS. At 6.0 GHz, the maximum gain is 12.79 dB without the FSS and 2.69 dB with the FSS. At 3.0 GHz, the maximum gain during azimuth steering is 16.13 dB without the FSS and 16.68 dB with the FSS, and the maximum gain during elevation steering is 15.74 dB without the FSS and 15.90 dB with the FSS.

PHCDTI: A multichannel parallel high-order feature crossover model for DTIs prediction

The exploration of non-covalent interactions between drugs and proteins (NCIs) has significantly improved the performance of drug–target interactions (DTIs) prediction models. However, existing methods for simulating NCIs are limited to single or multiple interactions, ignoring the combinatorial effects (i.e., high-order crossover between drugs and proteins (HOC-D&P)) that can arise when multiple NCIs coexist. To overcome this limitation, a multi-channel parallel high-order prediction model (PHCDTI) is proposed, aiming to simulate the complex interactions between drugs and proteins, thereby modeling HOC-D&P. Firstly, a tri-channel parallel model structure is constructed, allowing each channel to independently learn the interaction patterns between drug molecules and amino acid groups. Secondly, the three channels are designed from low-order to high-order, incorporating linear, inner-product, and pair-interaction feature crossover methods to simulate the complex interactions between drug molecules and amino acid groups. The stacking of residual connections enables the model to effectively model HOC-D&P. Additionally, by utilizing multi-head attention mechanisms for inner product and squeeze-excitation networks (SENET) for pair-interaction, the interpretability of the NCIs modeling process and the specific representation of drug molecules or amino acid groups are achieved. PHCDTI was evaluated on three benchmark datasets in four experimental settings, and the results indicated that it outperformed the most recent baseline model. In all cases, the average accuracy, precision, recall, AUC, and AUPR of PHCDTI are improved over the best baseline model by 3.56%, 6.06%, 3.99%, 2.50%, and 8.11%, respectively. Moreover, strong information filtering capabilities were exhibited by PHCDTI, with the effectiveness of the three-channel design confirmed by ablation studies. Finally, a case study on the human γ-aminobutyric acid receptors (GABARs) demonstrated that PHCDTI can serve as a powerful tool for real-world drug screening and design.

Comparison of Double-Stranded DNA at the 5' and 3' Ends of the G-Triplex and Its Application in the Detection of Hg(II).

Leveraging the fluorescence enhancement effect of the G-triplex (G3)/thioflavin T (ThT) catalyzed by the adjacent double-stranded DNA positioned at the 5' terminus of the G3, the G3-specific oligonucleotide (G3MB6) was utilized to facilitate the rapid detection of mercury (Hg(II)) through thymine-Hg(II)-thymine (T-Hg(II)-T) interactions. G3MB6 adopted a hairpin structure in which partially complementary strands could be disrupted with the presence of Hg(II). It prompted the formation of double-stranded DNA by T-Hg(II)-T, inducing the unbound single strand of G3MB6 to spontaneously form a parallel G3 structure, producing a solid fluorescence signal by ThT. Conversely, fluorescence was absent without Hg(II), since no double strand and formation of G3 occurred. The fluorescence intensity of G3MB6 exhibited a positive correlation with Hg(II) concentrations from 17.72 to 300 nM (R2 = 0.9954), boasting a notably low quality of limitation (LOQ) of 17.72 nM. Additionally, it demonstrated remarkable selectivity for detecting Hg(II). Upon application to detect Hg(II) in milk samples, the recovery rates went from 100.3% to 103.2%.

Non-linear torsional Alfv\'en waves evolving in stratified viscous plasmas: Coronal hole plumes

We model solar atmospheric structures characterised by parallel structuring. We focus on Alfv\'en waves in the weakly non-linear regime to highlight the efficiency of non-linear wave steepening when dissipative effects are prominent. We also consider the local and equilibrium conditions involved in shock formation and the shock's contributions to coronal seismology. Coronal plumes were modelled analytically by implementing the magnetohydrodynamic (MHD) theory in cylindrical geometry. Here, the stratification and viscosity are present internal to the plume, whilst effects of the external medium, together with equilibrium conditions, are implied where the magnetic fields are parallel to the plume axis. We implemented a second-order thin flux tube approximation to obtain a wave equation that points to effects tied to non-linear, dissipative, and stratification terms, as well as terms representing atmospheric conditions. The impact of shear viscosity on non-linear Alfv\'en waves extracted by the Cohen-Kulsrud-Burgers-type equation proves more efficient when propagated to higher altitudes. The dissipative effects linked to the dimensionless viscosity indicate that the dissipative effects are not linear. Meanwhile, the delay in shock formation enables energy conversions at higher altitudes, thereby maintaining coronal heating at higher levels. The efficiency of parallel structuring and viscous damping is enhanced by such transverse structuring, as it is directly proportional to the external plasma-beta . It is observed that Alfv\'en pulses may undergo a backward shock, either in the lower levels of coronal plasma or as they propagate toward higher regions, implying a conversion of energy occurring at various altitudes. A peak was observed, indicating that the interplay reverses at heights around $1.5$ solar radii. Such effects are shown to play a key role in the context of coronal seismology.

Contextual Enhancement–Interaction and Multi-Scale Weighted Fusion Network for Aerial Tracking

Siamese-based trackers have been widely utilized in UAV visual tracking due to their outstanding performance. However, UAV visual tracking encounters numerous challenges, such as similar targets, scale variations, and background clutter. Existing Siamese trackers face two significant issues: firstly, they rely on single-branch features, limiting their ability to achieve long-term and accurate aerial tracking. Secondly, current tracking algorithms treat multi-level similarity responses equally, making it difficult to ensure tracking accuracy in complex airborne environments. To tackle these challenges, we propose a novel UAV tracking Siamese network named the contextual enhancement–interaction and multi-scale weighted fusion network, which is designed to improve aerial tracking performance. Firstly, we designed a contextual enhancement–interaction module to improve feature representation. This module effectively facilitates the interaction between the template and search branches and strengthens the features of each branch in parallel. Specifically, a cross-attention mechanism within the module integrates the branch information effectively. The parallel Transformer-based enhancement structure improves the feature saliency significantly. Additionally, we designed an efficient multi-scale weighted fusion module that adaptively weights the correlation response maps across different feature scales. This module fully utilizes the global similarity response between the template and the search area, enhancing feature distinctiveness and improving tracking results. We conducted experiments using several state-of-the-art trackers on aerial tracking benchmarks, including DTB70, UAV123, UAV20L, and UAV123@10fps, to validate the efficacy of the proposed network. The experimental results demonstrate that our tracker performs effectively in complex aerial tracking scenarios and competes well with state-of-the-art trackers.

Modelling a chemical plant using grey‐box models employing the support vector regression and artificial neural network

AbstractIn this work, the performances of a nonlinear dynamic industrial process are examined using grey‐box (GB) models. To understand the dynamics of the system, the transient state is targeted. A white‐box (WB) model holds the prevailing knowledge using some assumptions. The performance of this model is limited. Artificial neural network (ANN) and support vector regression (SVR), which are techniques employed in numerous chemical engineering applications, are employed to construct the associated black‐box (BB) models. GA is used to optimize the SVR parameters. Dimensional and range extrapolations of different manipulated inputs, feed concentrations, feed temperatures, and cooling temperatures of the GB model and BB model are discussed. The different inputs extrapolation has different results because each input's effectiveness in the system is different. The results are compared, and the best model is suggested among the models, ANN, SVR, first principle (FP)‐ANN serial structure, FP‐ANN parallel structure, FP‐SVR serial structure, and FP‐SVR parallel structure.

Cepharanthine triggers ferroptosis through inhibition of NRF2 for robust ER stress against lung cancer

BackgroundSevere endoplasmic reticulum (ER) stress elicits apoptosis to suppress lung cancer. Our previous research identified that Cepharanthine (CEP), a kind of phytomedicine, possessed powerful anti-cancer efficacy, for which the underlying mechanism was still uncovered. Herein, we investigated how CEP induced ER stress and worked against lung cancer. MethodsThe differential expression genes (DEGs) and enrichment were detected by RNA-sequence. The affinity of CEP and NRF2 was analyzed by cellular thermal shift assay (CETSA) and molecular docking. The function assay of lung cancer cells was measured by western blots, flow cytometry, immunofluorescence staining, and ferroptosis inhibitors. ResultsCEP treatment enriched DEGs in ferroptosis and ER stress. Further analysis demonstrated the target was NRF2. In vitro and in vivo experiments showed that CEP induced obvious ferroptosis, as characterized by the elevated iron ions, ROS, COX-2 expression, down-regulation of GPX4, and atrophic mitochondria. Moreover, enhanced Grp78, CHOP expression, β-amyloid mass, and disappearing parallel stacked structures of ER were observed in CEP group, suggesting ER stress was aroused. CEP exhibited excellent anti-lung cancer efficacy, as evidenced by the increased apoptosis, reduced proliferation, diminished cell stemness, and prominent inhibition of tumor grafts in animal models. Furthermore, the addition of ferroptosis inhibitors weakened CEP-induced ER stress and apoptosis. ConclusionIn summary, our findings proved CEP drives ferroptosis through inhibition of NRF2 for induction of robust ER stress, thereby leading to apoptosis and attenuated stemness of lung cancer cells. The current work presents a novel mechanism for the anti-tumor efficacy of the natural compound CEP.

Performance evaluation of an automatic resonant switched capacitor-based voltage balancing circuits for series connected batteries

Electrical energy storage (EES) plays a crucial role in various power applications. Voltage imbalance is a common issue that can negatively affect the efficiency, reliability, and safety of EESs. Several types of voltage balancing (VB) circuits have been proposed in much of the literature. Among these VB circuits, switched capacitor (SC)-based circuits have attracted significant interest due to their efficiency, cost-effectiveness, compact size, and ease of control, but their balancing performance is not yet satisfactory. As a result, structural modifications in SC-based circuits have been widely proposed to improve balancing performance. However, not all of these circuit structures have been implemented into a resonant switched capacitor (RSC)-based voltage balancing, which has higher efficiency. Hence, this study aims to assess the efficiency of RSC-based VB circuits by conducting analog simulation using the Matlab Simulink software. This research evaluates the performance of the VB circuit not only in terms of its speed and efficiency, but also in terms of its energy distribution. The results show that the delta structure is the fastest in terms of balancing speed when completing the balancing process, followed by the mesh structure and the parallel structure. The best energy distribution is produced by a parallel structure, as indicated by the change in voltage of all battery cells always moving towards a convergent value, regardless of the variations in initial imbalance conditions. Meanwhile, other circuit structures distribute energy randomly, allowing the voltage of the battery cells to change not directly towards a convergent value. Lastly, the paper summarizes the balancing speed, efficiency, circuit complexity, and quality of energy distribution.