Reference Model Research Articles

Customizing pyfao56 for evapotranspiration estimation and irrigation scheduling at the Limited Irrigation Research Farm (LIRF), Greeley, Colorado

Estimation of evapotranspiration (ET, the water used by soil evaporation and plant transpiration) and soil water depletion (Dr, the amount of water to bring the soil water in the root zone to field capacity) are critical in irrigation water management. The open-source Python-based crop ET and water balance modeling package named “pyfao56” was originally developed based on the dual crop coefficient approach as described in the Food and Agricultural Organization of the United Nations, Irrigation and Drainage Paper No. 56 (FAO-56). The package later expanded on the seminal FAO-56 document to consider various options relevant to ET and water balance modeling, including both short (grass) and tall (alfalfa) reference crops, optional discretization of variable soil layers based on field capacity, and interpretation of readily available water (RAW) and Dr based on both a dynamic (growing) root zone and maximum root zone. The package requires two input data objects; the “Parameters” class defines variables affecting soil water balance and ET and the “Weather” class specifies relevant meteorological data. Other optional input data objects include the “Irrigation” class for irrigation events, the “Soil_Profile: class for defining stratified soil layer data, and the “Update” class for assimilation of measured data. Additional tools are available for estimating and forecasting standardized reference ET (ETref), providing seasonal water balance summaries, computing goodness-of-fit statistics between measured and modeled values, and visualizing time series plots of daily Dr, ET, and crop coefficients. The current pyfao56 release (v1.2.1) was incorporated into a customized workflow for specific use at the USDA-ARS Limited Irrigation Research Farm (LIRF) in Greeley, Colorado, and named the LIRF Implementation of Pyfao56 (LIRFIP). Field data from 2023 full and limited irrigation field trials were used to demonstrate the functionality of LIRFIP and its customizations and integration of pyfao56 into the LIRF workflow. Specific customizations included use of pyfao56 “customload” functions to input data from various sources, including weather from an on-site micrometerological station (using the application programming interface (API) developed by the maintainers of the Colorado Agricultural Meteorological (CoAgMet) weather station network), irrigation events (via a shared Google docs sheet), field capacity values by plot and soil layer (via an Excel spreadsheet), and measured soil water content as well as basal crop coefficient (Kcb) calculated from fractional canopy cover (fc) (via an SQLite database). To evaluate multiple research plots, LIRFIP was designed to iterate simulations from multiple instances of pyfao56 and to produce customized output summary files and interactive hypertext markup language (html) graphs as an aid for irrigation management decisions for LIRF field trials. This study demonstrated pyfao56 as a useful, flexible, customizable, and repeatable ET-based water balance model by showcasing its integration within a specific computational workflow for irrigation management field research at LIRF. The approach serves as an example for pyfao56 integration in other water management tools as conceived by water managers, researchers, and practitioners worldwide.

Sensorless speed controller of an induction motor with MRAS-based model predictive control

Speed sensors play critical role in induction motor (IM) control systems. However, the use of these sensors cause several drawbacks such as increasing the overall cost, decreased reliability on control and increased noise problems in data acquisition. The sensorless control methods also have been proposed in control systems of IM to increase the overall performance. On the other hand, model predictive controllers (MPCs) which is one of the most recently used controller methods require the speed measurement to perform speed and torque control of an IM. In addition to control method, the topology of voltage source inverter (VSI) affects the overall efficiency of the IM drive because of its losses and total harmonic distortion (THD) ratios delivered by the topology. The main proposal of this study is to implement an IM drive built with a three-level active neutral point clamped (3L-ANPC) voltage source inverter (VSI) and to propose an MPC that eliminates sensor requirement in speed control of IM. The proposed IM drive takes benefits of multilevel inverter topology, which reduces the torque ripple, enables low switching frequency and provides decreased harmonic distortion. The switching pulses of ANPC VSI are generated by the proposed sensorless MPC method for controlling the torque of IM. The modelling and experimental verification of the ANPC based IM drive have been carried out and the sensorless MPC controller has been validated in this study. There two types of model reference adaptive system (MRAS) have been proposed to ensure precise estimation of IM speed that should be supplied to MPC controller. These MRAS methods are titled as the rotor flux based model reference adaptive system (MRASF), and the stator current based model reference adaptive system (MRASCC) according to estimation references. The proposed MPC-based IM drive has been analyzed under variable speed and torque circumstances to detect the efficiency and reliability of MRASF and MRASCC algorithms. The efficiency indicators such as flux and torque responses, and speed control of the suggested MPC IM drive are validated with experimental analysis. The results of revealed that the IM effectively tracks the reference speed and produces the required torque depending to the precise control of proposed MPC algorithm.

Supervised Learning in Model Reference Adaptive Sliding Mode Control

Supervised Learning in Model Reference Adaptive Sliding Mode Control

Head model dataset for mixed reality navigation in neurosurgical interventions for intracranial lesions

Mixed reality navigation (MRN) technology is emerging as an increasingly significant and interesting topic in neurosurgery. MRN enables neurosurgeons to “see through” the head with an interactive, hybrid visualization environment that merges virtual- and physical-world elements. Offering immersive, intuitive, and reliable guidance for preoperative and intraoperative intervention of intracranial lesions, MRN showcases its potential as an economically efficient and user-friendly alternative to standard neuronavigation systems. However, the clinical research and development of MRN systems present challenges: recruiting a sufficient number of patients within a limited timeframe is difficult, and acquiring low-cost, commercially available, medically significant head phantoms is equally challenging. To accelerate the development of novel MRN systems and surmount these obstacles, the study presents a dataset designed for MRN system development and testing in neurosurgery. It includes CT and MRI data from 19 patients with intracranial lesions and derived 3D models of anatomical structures and validation references. The models are available in Wavefront object (OBJ) and Stereolithography (STL) formats, supporting the creation and assessment of neurosurgical MRN applications.

Research on Stability Control of Distributed Drive Vehicle with Four-Wheel Steering

The four-wheel steering distributed drive vehicle is a novel type of vehicle with independent control over the four-wheel angle and wheel torque. A method for jointly controlling the distribution of the wheel angle and torque is proposed based on this characteristic. Firstly, the two-degrees-of-freedom model and ideal reference model of four-wheel steering vehicle are established; then, the four-wheel steering controller and torque distribution controller are designed. The rear wheel angle is controlled by the feedforward controller and the feedback controller. The feedforward controller takes the side slip angle of the center of mass as the control target, and the feedback controller takes the yaw angle as the control target. Torque is controlled by two control layers, the additional yaw moment of the upper layer is calculated by the vehicle motion state and fuzzy control theory, and the lower layer distributes wheel torque through the road adhesion coefficient and wheel load. Finally, a simulation platform is established to verify the effectiveness of the proposed control algorithm.

A Novel Rotor Harmonic Winding Configuration for the Brushless Wound Rotor Synchronous Machine

In the last decade, permanent magnet (PM)-free or hybrid PM machines have been extensively researched to find an alternative for high cost rare-earth PM machines. Brushless wound rotor synchronous machines (BL-WRSMs) are one of the alternatives to these PM machines. BL-WRSMs have a lower torque density compared to PM machines. In this paper, a new topology is introduced to improve the torque producing capability of the existing BL-WRSM by utilizing the vacant spaces in the rotor slots. The new topology has two harmonic windings placed on the rotor which induce separate currents. A capacitor is used between the two harmonic windings to bring the currents in phase with each other. The harmonic winding currents are fed to the rectifier which is also placed on the rotor. Due to additional harmonic winding, the overall field current fed to the rotor field winding has been increased and hence the average torque has also increased. Finite element analysis (FEA)-based simulations are performed using ANSYS Maxwell to validate the proposed topology. The results show that the average torque of the machine has been significantly increased compared to the reference model. The detailed comparison results are provided in this paper.

The Representational Challenge for Designing and Managing 5P Medicine Ecosystems.

Health and social care systems around the globe currently undergo a transformation towards personalized, preventive, predictive, participative precision medicine (5PM), considering the individual health status, conditions, genetic and genomic dispositions, etc., in personal, social, occupational, environmental and behavioral context. This transformation is strongly supported by technologies such as micro- and nanotechnologies, advanced computing, artificial intelligence, edge computing, etc. For enabling communication and cooperation between actors from different domains using different methodologies, languages and ontologies based on different education, experiences, etc., we have to understand the transformed health ecosystems and all its components in structure, function and relationships in the necessary detail ranging from elementary particles up to the universe. That way, we advance design and management of the complex and highly dynamic ecosystem from data to knowledge level. The challenge is the consistent, correct and formalized representation of the transformed health ecosystem from the perspectives of all domains involved, representing and managing them based on related ontologies. The resulting business view of the real-world ecosystem must be interrelated using the ISO/IEC 21838 Top Level Ontologies standard. Thereafter, the outcome can be transformed into implementable solutions using the ISO/IEC 10746 Open Distributed Processing Reference Model. Model and framework for this system-oriented, architecture-centric, ontology-based, policy-driven approach have been developed by the first author and meanwhile standardized as ISO 23903 Interoperability and Integration Reference Architecture.

A Combustion Kinetic Model Based on the Fast Chemistry Assumption: Application to Non-Sooty and Sooty Laminar Diffusion Flames

ABSTRACT In this article, a combustion kinetic model based on the fast chemistry (FC) assumption is proposed to compute the finite reaction rate of gaseous species in laminar diffusion flames. The algorithm developed for this purpose is numerically robust and relies on the adiabatic flame temperature to limit the fuel consumption rate. While there is no FC model specifically developed for laminar flames, the Eddy Dissipation Model (EDM) correction scheme for laminar regimes is taken as the reference model. The proposed model addresses the shortcomings of the reference model, i.e. the sensitivity to the transport time scale and the cell size and the need to calibrate a numerical constant repeatedly. The new model formulation does not rely on any numerical constant and completely removes the dependency on the cell size and the species transport time scale, as shown for multiple counterflow and axisymmetric laminar diffusion flames. Furthermore, the proposed model provides consistent temperature and species mass fraction predictions across the mentioned flames. Finally, a sooty axisymmetric flame is simulated using the mentioned models to investigate the numerical interaction between combustion and soot chemistry. The results show that the reference model requires recalibration to capture the temperature and soot volume fraction distributions while the proposed model yields reasonably accurate temperature and soot predictions.

Multi-Step Design Optimization for the Improvement of an Outer-Rotor Brushless Direct Current Motor

Brushless Direct Current (BLDC) motors have seen significant improvements across various electrical applications. The growing focus on motor design research highlights the BLDC motor’s superior efficiency compared to traditional motors, which consume more power. BLDC motors are compact, lightweight, energy-efficient, and easy to control, making them ideal for modern applications. This study aims to enhance BLDC motor design and performance by employing the Taguchi method, Response Surface Methodology (RSM), and Finite Element Method (FEM) for multi-stage optimization. A 26-watt BLDC electric fan motor is the reference model for this study. The Taguchi method helps identify optimization points, guiding further enhancements in the second stage. The study proposes a design with improved output power, torque, and efficiency. The final design achieves a 15% higher energy efficiency than the reference model, with a 10 W increase in output power and a 0.032 Nm increase in maximum torque. The FEM analysis using JMAG software v 21.2 validates the proposed design, which shows improved configurations compared to the reference model, demonstrating the efficacy of the optimization techniques for BLDC motor design.

Calibrated intraoral scan protocol (CISP) for full-arch implant impressions: An in vitro comparison to conventional impression, intraoral scan, and intraoral scan with scan-aid.

To assess a newly developed intraoral scan protocol in enhancing the accuracy of complete-arch implant impressions. Four impression approaches were applied to the same maxillary edentulous model with 6 implants: (1) intraoral scan (IOS), (2) intraoral scan with scan aid (IOS-SA), (3) calibrated intraoral scan protocol (CISP), and (4) conventional splinted open-tray impression (CONV). Each approach was repeated 10 times, and a direct scan of the model with a desktop scanner was used as a reference model. The alignment of scans and the reference model was conducted by two methods: (a) aligning all scan bodies to evaluate the overall fit, and (b) aligning the first and second scan bodies to simulate the Sheffield fit test for passive fitting of multiple implant-supported prostheses. Linear deviations from the reference model (trueness) and within each group (precision) were analyzed using Python scripts. When aligned by all scan bodies, the CISP group exhibited comparable mean trueness (38.33 μm) and precision (45.97 μm) to the CONV group (44.30 and 47.92 μm respectively), both of which significantly outperformed the IOS group (86.82 and 83.17 μm, respectively). Furthermore, in the virtual Sheffield fit test, the CISP group achieved the highest levels of mean trueness at the end span (121.7 μm), making a linear deviation reduction of 36.7%, 60%, and 41.4% when compared to the CONV, the IOS, and the IOS-SA groups, respectively. Moreover, the CISP group (104.3 μm) displayed a remarkable 65, 182, and 86 μm advantage in precision over the CONV, IOS, and IOS-SA groups, respectively. CISP demonstrated comparable accuracy to the gold standard, the conventional splinted open-tray impression. Furthermore, it excelledin the virtual passive fitting test.

Self-Supervised Machine Learning to Characterise Step Counts from Wrist-Worn Accelerometers in the UK Biobank.

Step count is an intuitive measure of physical activity frequently quantified in health-related studies; however, accurate step counting is difficult in the free-living environment, with error routinely above 20% in wrist-worn devices against camera-annotated ground truth. This study aims to describe the development and validation of step count derived from a wrist-worn accelerometer and assess its association with cardiovascular and all-cause mortality in a large prospective cohort. We developed and externally validated a self-supervised machine learning step detection model, trained on an open-source and step-annotated free-living dataset. 39 individuals will free-living ground-truth annotated step counts were used for model development. An open-source dataset with 30 individuals was used for external validation. Epidemiological analysis was performed using 75,263 UK Biobank participants without prevalent cardiovascular disease (CVD) or cancer. Cox regression was used to test the association of daily step count with fatal CVD and all-cause mortality after adjustment for potential confounders. The algorithm substantially outperformed reference models (free-living mean absolute percent error of 12.5%, versus 65-231%). Our data indicate an inverse dose-response association, where taking 6,430-8,277 daily steps was associated with 37% [25-48%] and 28% [20-35%] lower risk of fatal CVD and all-cause mortality up to seven years later, compared to those taking fewer steps each day. We have developed an open and transparent method that markedly improves the measurement of steps in large-scale wrist-worn accelerometer datasets. The application of this method demonstrated expected associations with CVD and all-cause mortality, indicating excellent face validity. This reinforces public health messaging for increasing physical activity and can help lay the groundwork for the inclusion of target step counts in future public health guidelines.

Atmospheric Chemistry Surrogate Modeling With Sparse Identification of Nonlinear Dynamics

AbstractModeling atmospheric chemistry is computationally expensive and limits the widespread use of chemical transport models. This computational cost arises from solving high‐dimensional systems of stiff differential equations. Previous work has demonstrated the promise of machine learning (ML) to accelerate air quality model simulations but has suffered from numerical instability during long‐term simulations. This may be because previous ML‐based efforts have relied on explicit Euler time integration—which is known to be unstable for stiff systems—and have used neural networks which are prone to overfitting. We hypothesize that parsimonious models combined with modern numerical integration techniques can overcome this limitation. Using a small‐scale mechanism to explore the potential of these methods, we have created a machine‐learned surrogate by (a) reducing dimensionality using singular value decomposition to create an interpretably‐compressed low‐dimensional latent space and (b) using Sparse Identification of Nonlinear Dynamics (SINDy) to create a differential‐equation‐based representation of the underlying dynamics in the compressed latent space with reduced stiffness. The root mean square error of ML model prediction for ozone concentration over 9 days is 37.8% of the root mean square concentration across all simulations in our testing data set. The surrogate model is 10× faster with 12× fewer integration timesteps compared to reference model and is numerically stable in all tested simulations. Overall, we find that SINDy can be used to create fast, stable, and accurate surrogates of a simple photochemical mechanism. In future work, we will explore the application of this method to more detailed mechanisms and their use in large‐scale simulations.

Implementation of a Universal Framework Using Design Patterns for Application Development on Microcontrollers.

This article focuses on the area of software development for microcontrollers and details the implementation of modern programming practices and principles in embedded systems and IoT applications. This article explains how we implemented previously unimplemented principles and applied design patterns for quality software design on microcontrollers, which are currently only used for developing applications on the higher layers of the IoT reference model. A custom modular framework for microcontrollers is presented, based on applying SOLID principles and adapting design patterns specific to the microcontrollers' application development needs. The implemented framework enables independent communication between modules and flexible integration of hardware components. It is designed with platform independence in mind, contributing to its wide adaptability and ease of use in diverse development environments. By applying these technological approaches, we can create applications that are not only testable and extensible in terms of application logic but also allow for easy adaptation to changes in these hardware resources. Utilizing these capabilities represents an innovative approach to development for microcontrollers that fundamentally improves the long-term sustainability and scalability of applications.

Influence of Concrete Compressive Strength on L-Shaped Shear Wall Performance in Buildings within High-Seismicity Zones

Reinforced concrete mid-rise structures are supported by shear walls to withstand lateral forces resulting from earthquakes. In this context, the presence of openings in these walls is essential both functionally and aesthetically. This study employed L-shaped shear walls with varying vertical opening ratios, ranging from 15% to 50%, and compressive strengths ranging from 20MPa to 40MPa in a ten-story building. The study results demonstrate the significant role played by openings and the compressive strength of concrete in reinforced concrete walls to ensure buildings withstand earthquake forces. In this regard, we have developed reference models that adhere to seismic design principles and utilize optimized dimensions for openings based on the compressive strength of the concrete.

ОДНОВИМІРНА ЗГОРТКОВА МОДЕЛЬ НЕЙРОННОЇ МЕРЕЖІ ДЛЯ ОБРОБЛЕННЯ СИГНАЛІВ З АМПЛІТУДНОЮ МОДУЛЯЦІЄЮ БАГАТЬОХ СКЛАДОВИХ

The processing of radio signals using artificial neural networks (ANNs) has great potential for research, which can be explained by the adaptability of ANNs to various transmission conditions and the ability to detect abstract patterns of changes in signal parameters. The article reviews the works of other authors devoted to different ways of using ANNs for processing radio signals. Taking into account the information in the reviewed works, the research task was formed, which consists in developing an optimized ANN model for radio signal processing. Signals with amplitude modulation of many components (AMMC) were chosen to form training samples for ANN. The choice of modulation type is justified by greater energy efficiency compared to other widely used digital modulation types, such as quadrature amplitude modulation. Mathematic basis of AMMC signal generation is described. The process of finding the coordinates of three component 8-AMMC signal constellation is explained, the formation of signals in the time plane based on the found coordinates is explained as well as their discretization and the addition of white noise. An iterative algorithm for generating initial data for ANN based on the described ratios is proposed. The general structure of one-dimensional convolutional neural network is considered. Functions of individual neurons, connections between them, the formation of layers and the convolution operation are described mathematically. On the basis of the previously given ratios, a final display of the network was formed. Specific dimensions and activation functions for layers are selected. The use of convolutional layers is justified by time invariance. Based on the reviewed mathematical models, selected activation functions and dimensions, a neural model was formed. The process of validating the effectiveness of the formed neural model is described, which is based on comparing the symbolic error probabilities of the proposed and reference models at different signal-to-noise ratios. The validation results are presented. The advantages of the obtained model over the previously proposed purely recurrent model and the AMMC reference receiver are explained.

A new analytical model for transient heat transfer modeling of geothermal monotube heat exchangers

In this study, we present a new analytical model for the analysis of transient heat transfer (the model is solved after Laplace transformation) in geothermal monotube heat exchangers (GHE). The model accounts for various factors, including the heat exchanger's geometric parameters, the nature of the fluid (air and water), flow characteristics (laminar and turbulent), and heat transfer boundary conditions. Initially, the model's performance is validated by comparing its results with those obtained from computational fluid dynamics simulations using OpenFOAM. This validation process ensures the reliability and accuracy of the model. Additionally, an analysis of the heat transfer mechanisms and key modeling parameters identified from the comparison is provided. In the subsequent section, the analysis is extended by comparing the proposed model to four existing analytical models available in the literature. These reference models employ distinct approaches to simulate heat transfer to the ground. By examining a range of cases, the performance of the model is systematically evaluated against these conventional methods. This work closes by an analysis of the analytical models and the impact on the accuracy of the results highlighting the advantages of the proposed method in geothermal applications.

Supporting load frequency control process by using adaptive model reference virtual inertia controller when connecting renewable energy plants

Virtual inertia control can be used in combination with conventional load frequency controllers to support the system's inertia against the attenuation caused by renewables. As fractional-order controllers add more degrees of freedom, leading to a better response and improved controller robustness, this work proposes the implementation of a developed fractional-order model reference adaptive virtual inertia controller when integrating renewables in combination with fractional-order ID-P with filter controller as secondary controllers. To validate the introduced controllers, their results have been compared to those of using integer and fractional-order PID controllers as secondary controllers, in combination with adaptive and fixed parameters virtual inertia controllers. The Artificial Rabbits heuristic algorithm is used for the purpose of tuning secondary controllers and fixed parameters virtual inertia controllers. The proposed combination of adaptive model reference virtual inertia control with fractional-order ID-P with a filter achieves the best response among all the studied combinations. Using the proposed combination shows a relative percentage improvement in the cost function value of 26% to 160% compared to all the studied combinations. Additionally, the proposed combination shows a maximum frequency overshoot value of 0.04 Hz and a minimum undershoot value of -0.06 Hz.

Total-body imaging of mu-opioid receptors with [11C]carfentanil in non-human primates.

Mu-opioid receptors (MORs) are widely expressed in the central nervous system (CNS), peripheral organs, and immune system. This study measured the whole body distribution of MORs in rhesus macaques using the MOR selective radioligand [11C]carfentanil ([11C]CFN) on the PennPET Explorer. Both baseline and blocking studies were conducted using either naloxone or GSK1521498 to measure the effect of the antagonists on MOR binding in both CNS and peripheral organs. The PennPET Explorer was used for MOR total-body PET imaging in four rhesus macaques using [11C]CFN under baseline, naloxone pretreatment, and naloxone or GSK1521498 displacement conditions. Logan distribution volume ratio (DVR) was calculated by using a reference model to quantitate brain regions, and the standard uptake value ratios (SUVRs) were calculated for peripheral organs. The percent receptor occupancy (%RO) was calculated to establish the blocking effect of 0.14mg/kg naloxone or GSK1521498. The %RO in MOR-abundant brain regions was 75-90% for naloxone and 72-84% for GSK1521498 in blocking studies. A higher than 90% of %RO were observed in cervical spinal cord for both naloxone and GSK1521498. It took approximately 4-6min for naloxone or GSK1521498 to distribute to CNS and displace [11C]CFN from the MOR. A smaller effect was observed in heart wall in the naloxone and GSK1521498 blocking studies. [11C]CFN total-body PET scans could be a useful approach for studying mechanism of action of MOR drugs used in the treatment of acute and chronic opioid use disorder and their effect on the biodistribution of synthetic opioids such as CFN. GSK1521498 could be a potential naloxone alternative to reverse opioid overdose.

Evaluating Existing Ocean Glider Sampling Strategies for Submesoscale Dynamics

Abstract Mixing in the upper ocean is important for biological production and the transfer of heat and carbon between the atmosphere and deep ocean, properties commonly targeted by observational campaigns using ocean gliders. We assess the reliability of ocean gliders to obtain a robust statistical representation of submesocale variability in the ocean mixed layer of the Weddell Sea. A 1/48° regional simulation of the Southern Ocean is sampled with virtual ‘bow-tie’ glider deployments, which are then compared against the reference model output. Sampling biases of lateral buoyancy gradients associated with the arbitrary alignment between glider paths and fronts are formally quantified, and the magnitude of the biases are comparable to observational estimates, with a mean error of 52%. The sampling bias leaves errors in the retrieved distribution of buoyancy gradients largely insensitive to deployment length and the deployment of additional gliders. Notable sensitivity to these choices emerges when the biases are removed by sampling perpendicular to fronts at all times. Detecting seasonal change in the magnitude of buoyancy gradients is sensitive to the glider-orientation sampling bias but the change in variance is not. We evaluate the impact of reducing the number of dives and climbs in an observational campaign and find small reductions in the number of dive-climb pairs have a limited effect on the results. Lastly, examining the sensitivity of the sampling bias to path orientation indicates the bias is not dependent on the direction of travel, in our deep ocean study site.

Single-nucleus RNA transcriptome profiling reveals murine adipose tissue endothelial cell proliferation gene networks involved in obesity development

Endothelial cells play an important role in the metabolism of adipose tissue (AT). This study aimed to analyze the changes that adipose tissue in AT endothelial cells undergo during the development of obesity, using single-nucleus RNA sequence (snRNA-seq). Mouse paraepididymal AT cells were subjected to snRNA-seq with the 10X Genomics platform. The cell types were then clustered using t-distributed stochastic neighbor embedding and unbiased computational informatics analyses. Protein–protein interactions network was established using the STRING database and visualized using Cytoscape. The dataset was subjected to differential gene enrichment analysis. In total, 21,333 cells acquired from 24 mouse paraepididymal AT samples were analyzed using snRNA-seq. This study identified 18 distinct clusters and annotated macrophages, fibroblasts, epithelial cells, T cells, endothelial cells, stem cells, neutrophil cells, and neutrophil cell types based on representative markers. Cluster 12 was defined as endothelial cells. The proportion of endothelial cells decreased with the development of obesity. Inflammatory factors, such as Vegfa and Prdm16 were upregulated in the medium obesity group but downregulated in the obesity group. Genes, such as Prox1, Erg, Flt4, Kdr, Flt1, and Pecam1 promoted the proliferation of AT endothelial cells and maintained the internal environment of AT. This study established a reference model and general framework for studying the mechanisms, biomarkers, and therapeutic targets of endothelial cell dysfunction-related diseases at the single-cell level.