Network Modeling Framework Research Articles

Research on Integrated Energy Distributed Sharing in Distribution Network Considering AC Power Flow and Demand Response

Under the constraint of the AC power flow architecture considering reactive power regulation, the examination of integrated energy distributed transactions among multiple subsystems can promote the research in the field of energy sharing. It is difficult to fully cover the consideration of AC power flow, demand response, integrated energy, and other factors in traditional related research. In response, a study is therefore conducted in this paper on integrated energy sharing in the distribution network. First, this paper introduces the system operation framework of the proposed distribution network model, and explains the interaction between all the players. Secondly, a distribution network power flow model and an integrated energy subsystem model are respectively. In particular, the subsystem model specifically considers new energy, demand response, integrated energy, and other factors. Then, a cooperative game model is constructed based on the cooperative relationship among subsystems in the distribution network system, followed by the analysis of the benefits brought by cooperation to the distribution network and the subsystems themselves. Finally, a distributed solution flow is established for the model based on the Alternating Direction Method of Multipliers (ADMM) algorithm. The results of the example analysis reveal the effectiveness of the model proposed in increasing the degree of energy utilization and further absorbing new energy in the distribution network system, each subsystem can generate up to 12% more absorption capacity than it would otherwise operate separately to accommodate more renewable energy in the distribution system.

Design a New Neural Network Architecture Using a Layer of Neurons

Nowadays there are many different models of artificial neural networks. The difference between these models lies in the learning methods only, that is, in the rules for changing the parameters of the algorithm tuning, with or without links and side comments. While studying the general framework of network models, we see that the rules for obtaining the result and the mechanism for calculating the error can differ. For example, a multilayer realization might produce a threshold function when used as a classifier, or a linear function if used as an internal typeface.
 Where this research came to discuss the possibilities of the standard representation of some models of artificial neural networks, which clarify and treat some of the characteristics of that representation. Which can be considered an essential element in the process of typical representation of these networks, where a new proposal is made during this representation by using a “layer” of neurons, in other words, using a group of neurons that work in parallel and perform the same functions for which they were set .

The virtual aging brain: Causal inference supports interhemispheric dedifferentiation in healthy aging

The mechanisms of cognitive decline and its variability during healthy aging are not fully understood, but have been associated with reorganization of white matter tracts and functional brain networks. Here, we built a brain network modeling framework to infer the causal link between structural connectivity and functional architecture and the consequent cognitive decline in aging. By applying in-silico interhemispheric degradation of structural connectivity, we reproduced the process of functional dedifferentiation during aging. Thereby, we found the global modulation of brain dynamics by structural connectivity to increase with age, which was steeper in older adults with poor cognitive performance. We validated our causal hypothesis via a deep-learning Bayesian approach. Our results might be the first mechanistic demonstration of dedifferentiation during aging leading to cognitive decline.

Experimental and numerical investigations of forced response of multi-element lean-premixed hydrogen flames

Understanding the distinctive thermoacoustic properties of multi-element lean-premixed hydrogen flames is crucial for the development of hydrogen-capable gas turbines. Extending our earlier investigations of the self-excited dynamics of a cluster of lean-premixed pure hydrogen-air flames, here we explore the dynamic response to harmonic velocity perturbations, using an integrative approach combining loudspeaker-forced direct measurements and LES-based numerical simulations. Electronically-excited OH intensity distribution, generally assumed to be equivalent to the flame's heat release rate, shows an anchored conical reaction zone. Numerical simulations of heat release rate contours, however, reveal the formation of a more concentrated thin annulus region resembling an inverted V shaped layer, created by preferential diffusion of hydrogen molecules. This difference between OH-intensity distribution and heat release rate contours suggests consequential uncertainties in surrogate-dependent flame transfer function evaluations; the transfer function gains from measured data are well defined by a fast-decaying low-pass filter behavior, but the simulation result exhibits slowly-decaying large gain with slight overshoot over the entire range of frequency. Because the systematic uncertainties associated with the estimation of the flame's heat release rate cannot be accurately quantified, we rely on a reduced-order network modeling framework combined with direct measurement to test the accuracy of the prediction of the growth of self-sustained pressure fluctuations. We demonstrate that LES-based transfer functions predict the system's stability more accurately, and this suggests that the calculated heat release rate data are closer to the hydrogen flame's true heat release response than the wavelength-specific light intensity measurement. This assessment enables us to identify structurally simple moderate transverse oscillations of local regions containing atypically concentrated energy as pivotal processes controlling high-frequency hydrogen combustion dynamics.

Flood risk cascade analysis and vulnerability assessment of watershed based on Bayesian network

Under the dual impact of climate change and accelerated urbanization, urban flood disasters are frequent, resulting in severe losses, and there has been widespread concern. Through constructing the topological structure of flood control network and collecting the hydrological information from hydrological stations, this study proposes a Bayesian flood network model of Haihe river basin (Beisanhe and Yongdinghe river basin) to assess the flood risk. Flood Probability (FP) and Flood Network Cascade (flood network cascade) are proposed to characterize the vulnerability and flood cascade of flood control networks respectively. The accuracy of the flood cascade simulation of the Beisanhe river basin exceeds 80%, and the accuracy of the Yongdinghe river basin exceeds 85%. It is proved that the Bayesian network model proposed in this study has good simulation performance. The derived vulnerability assessment indicates that the central and southern regions of the Beisanhe River and Yongdinghe River basin are easily affected by floods at other nodes. The proposed Bayesian network modeling framework can also simulate the flood cascade in flood control infrastructure, so it can be used for scenario planning and real-time flood forecasting to provide information for flood management and operation.

Starting from the Bottom: Coupling a Genetic Algorithm and a Pore Network Model for Porous Electrode Design

Porous electrodes are performance- and cost-defining components in modern electrochemical systems as they provide surfaces for electrochemical reactions, facilitate mass transport, conduct electrons and heat, and determine the hydraulic resistance [1]. Hence, electrode engineering is a valid approach to increase power density and improve const competitiveness. In convection-enhanced technologies, currently used porous materials are fibrous carbonaceous electrodes developed for low temperature fuel cells; yet their microstructure and surface chemistry limits the performance of emerging electrochemical systems such as redox flow batteries. Moreover, the empirical design of these electrodes is time- and resource-intensive which limits exploration of the wider design space. Microstructure-informed multiphysics simulations can be leveraged to aid the theoretical understanding and design of advanced electrode architectures [2]. However, while these simulations have improved our understanding, they have only recently [3] been deployed to realize the bottom-up design of electrode microstructures. The combination of microstructure-informed multiphysics with evolutionary algorithms could accelerate progress in the optimization of porous electrodes for specific applications. In this work, we deploy three-dimensional simulations in combination with a genetic algorithm for the bottom-up design of porous electrodes for redox flow batteries.In the first part of the talk, the coupling of a pore network modeling framework [4] with an evolutionary algorithm [5] is described. The pore network model is a microstructure-informed, electrolyte-agnostic simulation framework for flow batteries utilizing a network-in-series approach to account for species depletion over the entire length of the electrode, developed using an open-access platform (OpenPNM) [6]. The electrochemical model is solved for the electrolyte fluid transport and couples both half-cells by iteratively solving the species and charge transport at a low computational cost. In the genetic algorithm, the electrode microstructure evolves driven by a fitness function that minimizes pumping power requirements and maximizes electrochemical power output, where the optimization only relies on the electrolyte chemistry and initial electrode morphology and flow field geometries as inputs. The analyzed proof-of-concept employs a flow-through cubic lattice structure with fixed pore positions and shows significant improvement of the fitness function over 1000 generations, where the fitness improved by 75% predominantly driven by reducing the pumping requirements by 73%. The evolutionary design resulted in a bimodal pore size distribution containing longitudinal electrolyte flow pathways of larges pores. Additionally, we found an increase in surface area at the membrane-electrode interface resulting in a 42% enhancement of the electrochemical performance.In the second part of the talk, I will discuss our latest progress in the development of the topology optimization by implementing commercial fibrous electrodes as offspring networks, integrated flow field geometries, and extended evolutionary freedom during the optimization. Coupling the genetic optimization to the desired flow field geometry affects the evolution of the fitness function, shifting the balance between electrochemical and hydraulic performance, highlighting the importance of the coupled optimization of flow fields and electrodes. By including more evolutionary freedom in the algorithm (i.e., by allowing merging and splitting of pores outside fixed coordinates), existing fibrous electrodes can be enhanced by for example reducing their pumping losses. The presented genetic algorithm offers great potential for the predictive design of electrode microstructures tailored for specific reactor architectures and redox chemistries. Hence, this framework can be used to accelerate and broaden the design and fabrication process of advanced electrode structures. Although applied to flow batteries in this study, the methodology can be leveraged to advance electrode microstructures in other electrochemical systems by adapting the relevant physics.

Uncertainty quantification of the virial black hole mass with conformal prediction

ABSTRACT Precise measurements of the black hole mass are essential to gain insight on the black hole and host galaxy co-evolution. A direct measure of the black hole mass is often restricted to nearest galaxies and instead, an indirect method using the single-epoch virial black hole mass estimation is used for objects at high redshifts. However, this method is subjected to biases and uncertainties as it is reliant on the scaling relation from a small sample of local active galactic nuclei. In this study, we propose the application of conformalized quantile regression (CQR) to quantify the uncertainties of the black hole predictions in a machine learning setting. We compare CQR with various prediction interval techniques and demonstrated that CQR can provide a more useful prediction interval indicator. In contrast to baseline approaches for prediction interval estimation, we show that the CQR method provides prediction intervals that adjust to the black hole mass and its related properties. That is it yields a tighter constraint on the prediction interval (hence more certain) for a larger black hole mass, and accordingly, bright and broad spectral line width source. Using a combination of neural network model and CQR framework, the recovered virial black hole mass predictions and uncertainties are comparable to those measured from the Sloan Digital Sky Survey. The code is publicly available.

Utilizing Pore Network Modeling for Performance Analysis of Multi-Layer Electrodes in Vanadium Redox Flow Batteries

Vanadium redox flow batteries (VRFBs) are promising energy storage devices. The microstructure of the porous electrode affects the performance of VRFBs. Therefore, identifying optimized electrode structures is an active research area. However, designing optimal microstructures requires studying varieties of structural parameters and design cases using a modeling tool with low computational cost. In this study, a pore network modeling (PNM) framework was developed to study the effects of multi-layer electrodes on VRFB electrode performance. In contrast to previous experimental works that were focused on multi-layer structure of the same material, this study explored the effect of using different microstructures in each layer. Using an image generation algorithm, fibrous materials were generated from which pore networks were extracted. The developed PNM included a modification by adding throat nodes in the geometry to accommodate a velocity dependent mass transfer coefficient. The results showed that putting a highly permeable layer near the membrane provides an alternative preferential path for fluid to distribute and supply those regions with reactive species, resulting in 57% increase in limiting current density in contrast to the opposite order. However, selection of the desired structures must be based on a trade-off between the current/power density and pressure drop.

Contrastive Learning Based Graph Convolution Network for Social Recommendation

Exploiting social networks is expected to enhance the performance of recommender systems when interaction information is sparse. Existing social recommendation models focus on modeling multi-graph structures and then aggregating the information from these multiple graphs to learn potential user preferences. However, these methods often employ complex models and redundant parameters to get a slight performance improvement. Contrastive learning has been widely researched as an effective paradigm in the area of recommendation. Most existing contrastive learning-based models usually focus on constructing multi-graph structures to perform graph augmentation for contrastive learning. However, the effect of graph augmentation on contrastive learning is inconclusive. In view of these challenges, in this work, we propose a contrastive learning based graph convolution network for social recommendation (CLSR), which integrates information from both the social graph and the interaction graph. First, we propose a fusion-simplified method to combine the social graph and the interaction graph. Technically, on the basis of exploring users’ interests by interaction graph, we further exploit social connections to alleviate data sparsity. By combining the user embeddings learned through two graphs in a certain proportion, we can obtain user representation at a finer granularity. Meanwhile, we introduce a contrastive learning framework for multi-graph network modeling, where we explore the feasibility of constructing positive and negative samples of contrastive learning by conducting data augmentation on embedding representations. Extensive experiments verify the superiority of CLSR’s contrastive learning framework and fusion-simplified method of integrating social relations.

Use of a Bayesian network as a decision support tool for watershed management: a case study in a highly managed river-dominated estuary

Decision making in water resource management has many dimensions including water supply, flood protection, and meeting ecological needs, therefore, is complex, full of uncertainties, and often contentious due to competing needs and distrust among stakeholders. It benefits from robust tools for supporting the decision-making process and for communicating with stakeholders. This paper presents a Bayesian network (BN) modeling framework for analyzing various management interventions regulating freshwater discharges to an estuary. This BN was constructed using empirical data from 98months of monitoring the Caloosahatchee River Estuary in south Florida during the period 2008-2021 as a case study to illustrate the potential advantages of the BN approach. Results from three different management scenarios and their implications on down-estuary conditions as they affected eastern oysters (Crassostrea virginica) and seagrass (Halodule wrightii) are presented and discussed. Finally, the directions for future applications of the BN modeling framework to support management in similar systems are offered.

Land carbon metabolism in Xiamen-Zhangzhou-Quanzhou region based on ecological network analysis.

Discovering the underlying mechanisms between carbon metabolism and carbon balance of human-natural system is of important theoretical and practical significance for reducing regional carbon emissions and promoting low-carbon development. Taking Xiamen-Zhangzhou-Quanzhou region from 2000 to 2020 as an example, we constructed the framework of spatial network model of land carbon metabolism based on carbon flow, and inquired into the spatial and temporal heterogeneity in the carbon metabolic structure, function and ecological relationships by using the ecological network analysis. The results showed that the dominant negative carbon transitions related with land use changes came from the conversion of cultivated land to industrial and transportation land, and that the high value areas of negative carbon flow were mainly distributed in the areas with relatively developed industries in the middle and east parts of Xiamen-Zhangzhou-Quanzhou region. Competition relationships were the dominant type and the spatial expansion was obvious, which led to the decreases of the integral ecological utility index and affected the regional carbon metabolic balance. The ecological network hierarchy of driving weight changed from a pyramid structure to a relatively more regular structure, with the producer contributing the most. The ecological network hie-rarchy of pull weight changed from a pyramid structure to an inverted pyramid structure, mainly due to the excessive increase of industrial and transportation land weights. Low-carbon development should focus on the sources of negative carbon transitions caused by land use conversion and its comprehensive impacts on carbon metabolic balance, to formulate differentiated low-carbon land use patterns and carbon emission reduction policies.

LogBTF: gene regulatory network inference using Boolean threshold network model from single-cell gene expression data.

From a systematic perspective, it is crucial to infer and analyze gene regulatory network (GRN) from high-throughput single-cell RNA sequencing data. However, most existing GRN inference methods mainly focus on the network topology, only few of them consider how to explicitly describe the updated logic rules of regulation in GRNs to obtain their dynamics. Moreover, some inference methods also fail to deal with the over-fitting problem caused by the noise in time series data. In this article, we propose a novel embedded Boolean threshold network method called LogBTF, which effectively infers GRN by integrating regularized logistic regression and Boolean threshold function. First, the continuous gene expression values are converted into Boolean values and the elastic net regression model is adopted to fit the binarized time series data. Then, the estimated regression coefficients are applied to represent the unknown Boolean threshold function of the candidate Boolean threshold network as the dynamical equations. To overcome the multi-collinearity and over-fitting problems, a new and effective approach is designed to optimize the network topology by adding a perturbation design matrix to the input data and thereafter setting sufficiently small elements of the output coefficient vector to zeros. In addition, the cross-validation procedure is implemented into the Boolean threshold network model framework to strengthen the inference capability. Finally, extensive experiments on one simulated Boolean value dataset, dozens of simulation datasets, and three real single-cell RNA sequencing datasets demonstrate that the LogBTF method can infer GRNs from time series data more accurately than some other alternative methods for GRN inference. The source data and code are available at https://github.com/zpliulab/LogBTF.

Unusual Water Flow in Ultra-Tight Porous Media: Integration of Profession and Innovation

Hydraulic fracturing is an effective method for stimulating reservoirs, making the economic development of ultra-tight shale gas and coalbed methane reservoirs possible. These formations are rich in nanopores, in which the fracturing fluid, such as fresh water, the flow, and the behavior of this flow differ significantly from those described in the classic Navier-Stokes formula. In bulk space, the interaction force exerted by the solid phase can be ignored, but the solid–fluid interaction plays a dominant role in nanoconfinement spaces in which the pore size is comparable to the molecular diameter. Nanoconfined water molecules tend to approach the water-wet pore surface, enhancing the water viscosity, which is a key parameter affecting the water flow capacity. Conversely, water molecules tend to stay in the middle of nanopores when subjected to a hydrophobic surface, leading to a decrease in viscosity. Thus, nanoconfined water viscosity is a function of the strength of the surface–fluid interaction, rather than a constant parameter, in classic theory. However, the influence of varying the viscosity on the nanoscale water flow behavior is still not fully understood. In this research, we incorporate wettability-dependent viscosity into a pore network modeling framework for stable flow for the first time. Our results show that: (a) the increase in viscosity under hydrophilic nanoconfinement could reduce the water flow capacity by as much as 11.3%; (b) the boundary slip is the primary mechanism for boosting the water flow in hydrophobic nanopores, as opposed to the slight enhancement contributed by a viscosity decline; and (c) water flow characterization in nanoscale porous media must consider both the pore size and surface wettability. Revealing the varying viscosity of water flow confined in nanopores can advance our microscopic understanding of water behavior and lay a solid theoretical foundation for fracturing-water invasion or flowback simulation.

Direct and indirect viral associations predict coexistence in wild plant virus communities

Viruses are a vastly underestimated component of biodiversity that occur as diverse communities across hierarchical scales from the landscape level to individual hosts. The integration of community ecology with disease biology is a powerful, novel approach that can yield unprecedented insights into the abiotic and biotic drivers of pathogen community assembly. Here, we sampled wild plant populations to characterize and analyze the diversity and co-occurrence structure of within-host virus communities and their predictors. Our results show that these virus communities are characterized by diverse, non-random coinfections. Using a novel graphical network modeling framework, we demonstrate how environmental heterogeneity influences the network of virus taxa and how the virus co-occurrence patterns can be attributed to non-random, direct statistical virus-virus associations. Moreover, we show that environmental heterogeneity changed virus association networks, especially through their indirect effects. Our results highlight a previously underestimated mechanism of how environmental variability can influence disease risks by changing associations between viruses that are conditional on their environment.

A hybrid framework of artificial intelligence-based neural network model (ANN) and central composite design (CCD) in quality by design formulation development of orodispersible moxifloxacin tablets: Physicochemical evaluation, compaction analysis, and its in-silico PBPK modeling

A hybrid framework of artificial intelligence-based neural network model (ANN) and central composite design (CCD) in quality by design formulation development of orodispersible moxifloxacin tablets: Physicochemical evaluation, compaction analysis, and its in-silico PBPK modeling

Model predictive control of agro‐hydrological systems based on a two‐layer neural network modeling framework

SummaryWater scarcity is an urgent issue to be resolved and improving irrigation water‐use efficiency through closed‐loop control is essential. The complex agro‐hydrological system dynamics, however, often pose challenges in closed‐loop control applications. In this work, we propose a two‐layer neural network (NN) framework to approximate the dynamics of the agro‐hydrological system. To minimize the prediction error, a linear bias correction is added to the proposed model. The model is employed by a model predictive controller with zone tracking (ZMPC), which aims to keep the root zone soil moisture in the target zone while minimizing the total amount of irrigation. The performance of the proposed approximation model framework is shown to be better compared to a benchmark long‐short‐term‐memory model for both open‐loop and closed‐loop applications. Significant computational cost reduction of the ZMPC is achieved with the proposed framework. To handle the tracking offset caused by the plant‐model‐mismatch of the proposed NN framework, a shrinking target zone is proposed for the ZMPC. Different hyper‐parameters of the shrinking zone in the presence of noise and weather disturbances are investigated, of which the control performance is compared to a ZMPC with a time‐invariant target zone.

Click-through Rate Prediction and Uncertainty Quantification Based on Bayesian Deep Learning.

Click-through rate (CTR) prediction is a research point for measuring recommendation systems and calculating AD traffic. Existing studies have proved that deep learning performs very well in prediction tasks, but most of the existing studies are based on deterministic models, and there is a big gap in capturing uncertainty. Modeling uncertainty is a major challenge when using machine learning solutions to solve real-world problems in various domains. In order to quantify the uncertainty of the model and achieve accurate and reliable prediction results. This paper designs a CTR prediction framework combining feature selection and feature interaction. In this framework, a CTR prediction model based on Bayesian deep learning is proposed to quantify the uncertainty in the prediction model. On the squeeze network and DNN parallel prediction model framework, the approximate posterior parameter distribution of the model is obtained using the Monte Carlo dropout, and obtains the integrated prediction results. Epistemic and aleatoric uncertainty are defined and adopt information entropy to calculate the sum of the two kinds of uncertainties. Epistemic uncertainty could be measured by mutual information. Experimental results show that the model proposed is superior to other models in terms of prediction performance and has the ability to quantify uncertainty.

Use of an artificial neural network model for estimation of unfrozen water content in frozen soils

The variation of unfrozen water content (UWC) has a significant influence on the physical and mechanical behaviors of frozen soils. Several empirical, semi-empirical, physical, and theoretical models are available in the literature to estimate the UWC in frozen soils. However, these models have limitations due to the complex interactions of various influencing factors that are not well understood or fully established. For this reason, in the present study, an artificial neural network (ANN) modeling framework is proposed and the PyTorch package is used for predicting UWC. Extensive UWC data of various types of soils tested under various conditions were collected through an extensive search of the literature. The developed ANN model showed good performance for the testing dataset. Its performance was further compared with two traditional statistical models on four soils and found to outperform these traditional models. Detailed discussions on the developed ANN model, and its strengths and limitations in comparison to different other models are provided. The study demonstrates that the proposed ANN model is simple yet reliable for estimating the UWC of various soils. In addition, the summarized UWC data and the proposed machine learning modeling framework are valuable for future studies related to frozen soils.

A methodology framework for bipartite network modeling

The graph-theoretic based studies employing bipartite network approach mostly focus on surveying the statistical properties of the structure and behavior of the network systems under the domain of complex network analysis. They aim to provide the big-picture-view insights of a networked system by looking into the dynamic interaction and relationship among the vertices. Nonetheless, incorporating the features of individual vertex and capturing the dynamic interaction of the heterogeneous local rules governing each of them in the studies is lacking. The methodology in achieving this could hardly be found. Consequently, this study intends to propose a methodology framework that considers the influence of heterogeneous features of each node to the overall network behavior in modeling real-world bipartite network system. The proposed framework consists of three main stages with principal processes detailed in each stage, and three libraries of techniques to guide the modeling activities. It is iterative and process-oriented in nature and allows future network expansion. Two case studies from the domain of communicable disease in epidemiology and habitat suitability in ecology employing this framework are also presented. The results obtained suggest that the methodology could serve as a generic framework in advancing the current state of the art of bipartite network approach.Graphical

Enhancing remote sensing satellite image classification through parameter optimization in hybrid CNN-LSTM model

Satellite imagery is a vital resource in Earth observation, offering valuable insights for applications ranging from environmental monitoring to disaster response. In the pursuit of accurate and robust classification, we present a study focused on parameter optimization within the framework of hybrid Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) models. This research seeks to maximize the potential of hybrid CNN-LSTM architecture, which inherently combine spatial feature extraction capabilities of CNNs with LSTM’s ability to model temporal dependencies. We delve into a systematic exploration of hyper-parameters, including the number of CNN layers, LSTM units, learning rates, and dropout rates, fine-tuning them to achieve peak classification performance. These parameters are meticulously adjusted and validated on a diverse and representative satellite image dataset. Beyond hyper-parameter optimization, we address the challenge of class imbalance frequently encountered in satellite image datasets. By incorporating specialized techniques during the optimization process, we mitigate class imbalance effects and bolster the model’s accuracy. The study’s findings not only unveil optimal parameter configurations but also shed light on the intricate interplay between architecture complexity and model performance. The resulting optimized hybrid model demonstrates exceptional classification accuracy, validating its applicability to real-world remote sensing tasks.