Simulation Experiments Research Articles

One-step inverse generation network for sparse-view dual-energy CT reconstruction and material imaging

Objective. Sparse-view dual-energy spectral computed tomography (DECT) imaging is a challenging inverse problem. Due to the incompleteness of the collected data, the presence of streak artifacts can result in the degradation of reconstructed spectral images. The subsequent material decomposition task in DECT can further lead to the amplification of artifacts and noise. Approach. To address this problem, we propose a novel one-step inverse generation network (OIGN) for sparse-view dual-energy CT imaging, which can achieve simultaneous imaging of spectral images and materials. The entire OIGN consists of five sub-networks that form four modules, including the pre-reconstruction module, the pre-decomposition module, and the following residual filtering module and residual decomposition module. The residual feedback mechanism is introduced to synchronize the optimization of spectral CT images and materials. Main results. Numerical simulation experiments show that the OIGN has better performance on both reconstruction and material decomposition than other state-of-the-art spectral CT imaging algorithms. OIGN also demonstrates high imaging efficiency by completing two high-quality imaging tasks in just 50 seconds. Additionally, anti-noise testing is conducted to evaluate the robustness of OIGN. Significance. These findings have great potential in high-quality multi-task spectral CT imaging in clinical diagnosis.

Pyrolytic hydrocarbon generation characteristics of the Chang 7 shale based on different experimental methods: Implications for shale oil and gas in the Ordos Basin

Shale oil and gas resources are abundant in the Chang 7 shale of the Yanchang Formation in Ordos Basin. To determine the characteristics and influencing factors of hydrocarbon generation evolution of the Chang 7 shale, a series of thermal simulation experiments were conducted on low‐maturity shale and kerogen samples. The results indicate that the maximum yield of shale oil are 294.5 and 304.3 mg/g TOC for kerogen sample at heating rates of 20 and 2°C/h, and the corresponding experimental temperatures are 360.2°C and 408.0°C, respectively. The utilization of lower heating rates is favourable for shale oil generation and it is recommended to employ a lower heating rate during in situ heating processes to maximize the economic benefits. The formation of crude oil cracking gas begins when simulating temperature exceeds 528.0°C (Easy Ro 2.6%) at a heating rate of 20°C/h and 480.0°C (Easy Ro 2.5%) at a heating rate of 2°C/h, as indicated by the carbon isotopic composition of gaseous hydrocarbons. The maximum oil production rate of the rock powder sample is 159.8 mg/g TOC, which is lower than that of the kerogen sample. It suggests that certain minerals in the Chang 7 shale may impede hydrocarbon generation. After the addition of pyrite, the highest yield of shale oil is 213.96 mg/g TOC, 33.9% higher than the yield of the original rock powder sample, reflecting the positive catalytic effect of pyrite on hydrocarbon generation of Chang 7 shale. Under geologic conditions, pyrite catalytic hydrocarbon generation may act primarily on the migration of organic matter by macromolecules, which considerably increases the probability of direct contact between pyrite and organic matter. Therefore, the organic‐rich shale with high pyrite content in Chang 7 member is the preferred target for in situ conversion of shale oil and gas in the Ordos Basin.

Impact of Varying Number of Radio Occultation Observations on Regional Weather Prediction Over India During the Summer Monsoon Season

AbstractObserving system simulation experiments are carried out to investigate the added value of radio occultation (RO) refractivity observations with various spatial sampling scenarios on regional weather predictions across the Indian region. A full summer monsoon season (June through September 2020) was used to demonstrate how varying the numbers and horizontal resolutions of RO data impacted regional‐scale weather forecasting. The MPAS (Model for Prediction Across Scales) was used to produce a nature run at a maximum horizontal resolution of 10 km. Then the WRF model with 12‐km horizontal resolution was used to carry out assimilation/forecast experiments with varying number of simulated RO observations. When the performance of the experiments is taken into account for moisture, temperature, winds and rainfall, as well as prediction lengths, the results show that RO observations with 50‐km resolution assimilated every 6 hr would provide the best results. Increasing the horizontal resolution to 25 km per 6 hr shows little overall improvement. Furthermore, RO data with horizontal resolutions lower than 100 km per 6 hr have only a small impact on the regional numerical weather prediction system. The number of low Earth orbit satellites in low‐inclination orbits required to achieve occultations every 6 hr with 50‐km resolution based on the COSMIC‐2 mission is approximately 700. This work is relevant for the deployment of the cost‐effective RO observing system for improved weather forecasting over the Indian region.

A complete lean system design for team-oriented product development projects

Purpose The inherent variability on process times and demand are the factors that prevent the efficient application of lean philosophy in multi-project product development (PD) environments. Considering this variability, a hybrid push–pull project control system is developed, and value stream costing (VSC) analysis is performed to reflect the relation between project lead time, capacity and project cost. The assessment of the push/pull project control on lead time improvement and long-term savings on capacity have been aimed with the proposed complete design structure. Design/methodology/approach In a team-based structure, formed through clustering, push control techniques for planning tasks within cross-functional teams and pull control techniques for planning tasks between cross-functional teams are developed. The final step evaluates the proposed structure through VSC and long-term savings have been pointed out, especially in terms of freed-up capacity. For the validation of the proposed methodology, an office furniture manufacturing firm’s PD department has been considered and the performance of the hybrid system has been observed through simulation experiments and based on the simulation results, the lean system is evaluated by VSC. Findings The results of simulation experiments show a superior performance of the proposed hybrid push/pull project control mechanism under different settings of cycle time between projects or shortly project cycle time, dispatching rules within teams and variability levels. The results of the Box-Score (tool to apply VSC) indicate increased capacity in the long term to add extra projects during the planning period with the same project lead time and without additional cost. Research limitations/implications Although extensive simulation experiments have been performed to quantify the effect of project control structure and positive results have been reported on lead time and cost, the proposed design structure has not been tested in all existing PD environments. Originality/value To the best of authors’ knowledge, the quantification of the effect of hybrid project control with VSC is the first attempt to be applied in lean PD projects.

State of energy estimation of lithium-ion battery based on long short-term memory optimization Adaptive Cubature Kalman filter.

State of energy (SOE) is an important parameter to ensure the safety and reliability of lithium-ion battery (LIB) system. The safety of LIBs, the development of artificial intelligence, and the increase in computing power have provided possibilities for big data computing. This article studies SOE estimation problem of LIBs, aiming to improve the accuracy and adaptability of the estimation. Firstly, in the SOE estimation process, adaptive correction is performed by iteratively updating the observation noise equation and process noise equation of the Adaptive Cubature Kalman Filter (ACKF) to enhance the adaptive capability. Meanwhile, the adoption of high-order equivalent models further improves the accuracy and adaptive ability of SOE estimation. Secondly, Long Short-term Memory (LSTM) is introduced to optimize Ohmic internal resistance (OIR) and actual energy (AE), further improving the accuracy of SOE estimation. Once again, in the process of OIR and AE estimation, the iterative updating of the observation noise equation and process noise equation of ACKF were also adopted to perform adaptive correction and enhance the adaptive ability. Finally, this article establishes a SOE estimation method based on LSTM optimized ACKF. Validate the LSTM optimized ACKF method through simulation experiments and compare it with individual ACKF methods. The results show that the ACKF estimation method based on LSTM optimization has an SOE estimation error of less than 0.90% for LIB, regardless of the SOE at 100%, 65%, and 30%, which is more accurate than the SOE estimation error of ACKF alone. It can be seen that this study has improved the accuracy and adaptability of LIB's SOE estimation, providing more accurate data support for ensuring the safety and reliability of lithium batteries.

Study on the Effect of Magnesium Chloride-Modified Straw Waste Biochar on Acidic Soil Properties.

Soil biochar is a kind of organic matter rich in carbon, which is of great significance in soil fertility improvement, fertilizer type innovation and greenhouse gas emission reduction. In this paper, Mg-modified biochar was prepared by thermal cracking using rice straw and corn straw as raw materials. The Mg-modified biochar and unmodified biochar were fully mixed with prepared soil samples at the addition amounts of 0.5% (w/w), 1% (w/w) and 2% (w/w), respectively, and then simulated indoor soil cultivation experiments were carried out. The effects of magnesium ion-modified biochar and non-modified biochar on soil chemical properties and the effects of different amounts of biochar on soil properties were studied. The results showed that the yield of Mg-modified biochar from rice straw and corn straw, prepared by pyrolysis, was 65%, and the ash content was large. The pH of MG-modified corn stalk biochar (MCBC) is weakly basic (8.55), while the pH of MG-modified rice stalk biochar (MRBC) is basic (10.1), and their internal structures are slightly different. After the application of biochar prepared from rice straw and maize stover, soil indicators were determined. Compared to the control, the chemical properties of the treated soil samples were significantly improved, with an increase in soil pH, an increase in the content of effective nutrients, such as fast-acting potassium, fast-acting phosphorus and alkaline dissolved nitrogen, and an increase in the content of the total phosphorus and total nitrogen, as well as an increase in the content of organic matter. The Mg-modified biochar was generally superior to the unmodified biochar in improving soil fertility, at the same addition level. It was also found that the rice-straw biochar performed better than the corn-stover biochar and had a more obvious effect on soil improvement in terms of fast-acting potassium, ammonium nitrogen, nitrate nitrogen, total phosphorus and total nitrogen contents.

A Reconstruction Methodology of Dynamic Construction Site Activities in 3D Digital Twin Models Based on Camera Information

Digital twin technology significantly enhances construction site management efficiency; however, dynamically reconstructing site activities presents a considerable challenge. This study introduces a methodology that leverages camera data for the 3D reconstruction of construction site activities. The methodology was initiated using 3D scanning to meticulously reconstruct the construction scene and dynamic elements, forming a model base. It further integrates deep learning algorithms to precisely identify static and dynamic elements in obstructed environments. An enhanced semi-global block-matching algorithm was then applied to derive depth information from the imagery, facilitating accurate element localization. Finally, a near-real-time projection method was introduced that utilizes the spatial relationships among elements to dynamically incorporate models into a 3D base, enabling a multi-perspective view of site activities. Validated by simulated construction site experiments, this methodology showcased an impressive reconstruction accuracy reaching up to 95%, this underscores its significant potential in enhancing the efficiency of creating a dynamic digital twin model.

Design of Human–Computer Interaction Gesture Recognition System Based on a Flexible Biosensor

The continuous development of high-speed Internet technology has made the application of robots increasingly widespread. Current robots and human–computer interaction systems mostly use rigid materials, such as metals and semiconductors, which have limitations in terms of deformability and flexibility. In addition, the biocompatibility and user comfort of these materials are also an issue. Therefore, research into new flexible biosensors is essential to improve the flexibility, comfort, and interactivity of these systems. This research will select polymer hydrogel as the electrode material of the sensor and polydimethylsiloxane as the base material of the sensor to design a resistance flexible biosensor to solve the poor flexibility. The research will use a template-matching method to verify the feasibility of gesture recognition of the flexible sensor. The remote control system of the robot finger is designed by a proportional-integral differential controller tuned by aradial basis function neural network. The feasibility of the research system is verified by simulation and scene experiments. The flexible sensor studied and prepared had a sensitivity of 0.7269, a tensile limit of 300%, and a thickness of 0.16 mm, showing good sensitivity and stability. The recognition accuracy of the sensor designed in the study was 92.8%, which was 8.1% higher than that of the data glove. Compared with traditional proportional-integral derivative (PID) controllers, the improved controller system error was within 10 to 3 rad, which had better adaptability and stability. Key information includes the design method of the flexible biosensor, its high sensitivity and stability under multiple stretches, and the proposal and validation of a new RBFNN–PID control model. These results showed that using this new sensor and control model significantly improved the control accuracy of mechanical fingers and the effect of gesture recognition. These results have important implications for the development of more advanced human–computer interaction systems. They not only improve the performance and reliability of the system, but also improve the user's interactive experience. These technologies are particularly promising in the fields of prosthetics for disabled people, advanced game controllers, and remotely controlled robots operating in hazardous environments. The research results are expected to lead to the development of advanced prosthetics, augmented reality devices, advanced game controllers, and automated robots. The main contribution of the research is to design a resistive flexible biosensor, which improves the traditional sensor's poor flexibility and large size and improves the sensor's ability to sense small changes. Future research may focus on further improving the sensor's long-term stability and performance under a variety of environmental conditions. In addition, commercializing these technologies and making them universal is also an important direction for the future.

Evaluation of low-temperature oxidation analysis and the development effect of high-pressure air injection in low-permeability reservoirs

In order to solve the problems of conventional water injection development difficulties and low recovery factor in low-permeability reservoirs, the method of high-pressure air drive is adopted to achieve the purpose of reservoir energy enhancement and efficiency improvement. This paper conducted an experimental study on the mechanism of low-temperature oxidation (LTO) for crude oil in the process of high-pressure air flooding, elaborated the relationship between the LTO properties of crude oil and the temperature, pressure, and water saturation of the reservoir, and analyzed the differences in LTO oxygen consumption and oil components under different reaction conditions. In addition, combined with the air flooding physical simulation experiment, the dynamic evolution law of recovery rate in the air flooding process was revealed. Findings from this inquiry indicate that an escalation in the oxidation temperature significantly amplifies the oxygen incorporation reaction within the crude oil matrix. This augmentation in oxidative conditions leads to an uptick in oxygen consumption, which subsequently precipitates a reduction in the lighter fractions of the oxidized oil while enriching its heavier components. Elevated pressures were found to enhance the propensity for the amalgamation of unstable hydrocarbons with oxygen, fostering comprehensive and heterogeneous oxidation reactions. Notably, an excessive presence of water was observed to detrimentally affect the thermal efficacy of crude oil oxidation processes. In the context of low-permeability reservoirs, air injection techniques have emerged as superior in effectuating oil displacement, although an increase in injection pressures has been associated with the phenomenon of gas channeling. Interestingly, adopting a sequential strategy of initiating water flooding before air flooding facilitated the conveyance of high-pressure air via established flushing channels, although it appeared to attenuate the intensity of crude oil oxidation, culminating in an oil recovery efficiency peaking at 51%.

Effect of bridge height on airflow and aeolian sand flux near surface along the Qinghai-Tibet Railway, China

In this work, we studied the near-surface flow field structure of railway bridges with different heights through field investigation and wind tunnel simulation experiments. Meanwhile, we simulated the distribution of sand accumulation around a bridge via CFD software based on the sand accumulation around the Basuoqu bridge in the Cuona Lake section of the Qinghai–Tibet Railway. Results show that the sand around this railway bridge is mainly from the lake sediment on the west side of the railway and the weathered detritus on the east side. The height of the railway bridge in the sandy area affects the distribution of the near-surface flow field and the variation in speed on both sides of the bridge. The wind speed trough on both sides of the 6 m high bridge is higher, and the horizontal distance between the wind speed trough and the bridge section is 1.5 times that of the 3 m high bridge. Wind speed attenuates in a certain range on the windward and leeward sides of the bridge, forming an aeolian area; under the beam body, it is affected by the narrow tube effect, forming a wind erosion area. The height of the bridge determines its sand transport capacity. Under certain wind conditions, the overhead area at the bottom of the 3 m high bridge and its two sides do not have the sand transport capacity, so sand accumulates easily. Nevertheless, the sand accumulation phenomenon gradually disappears with the increase in bridge clearance height. The objectives of this study are to reveal the formation mechanism of sand damage for railway bridges, provide theoretical support for the scientific design of railway bridges in sandy areas, and formulate reasonable railway sand prevention measures to ensure the safety of railway running, which have certain theoretical significance and practical value.

Leading the Lorenz 63 system toward the prescribed regime by model predictive control coupled with data assimilation

Abstract. Recently, concerns have been growing about the intensification and increase in extreme weather events, including torrential rainfall and typhoons. For mitigating the damage caused by weather-induced disasters, recent studies have started developing weather control technologies to lead the weather to a desirable direction with feasible manipulations. This study proposes introducing the model predictive control (MPC), an advanced control method explored in control engineering, into the framework of the control simulation experiment (CSE). In contrast to previous CSE studies, the proposed method explicitly considers physical constraints, such as the maximum allowable manipulations, within the cost function of the MPC. As the first step toward applying the MPC to real weather control, this study performed a series of MPC experiments with the Lorenz 63 model. Our results showed that the Lorenz 63 system can be led to the positive regime with control inputs determined by the MPC. Furthermore, the MPC significantly reduced necessary forecast length compared to earlier CSE studies. It was beneficial to select a member that showed a larger regime shift for the initial state when dealing with uncertainty in initial states.

Linear Motor System Identification and Simulation Experiments Based on LabVIEW

Linear Motor System Identification and Simulation Experiments Based on LabVIEW

Recognition of marine navigators’ workload based on eye movement features using bridge simulation

The human factor accounts for 80% of marine traffic accidents, either directly or indirectly. Workload significantly impacts the reliability of marine navigators’ behavior. Employing eye movement measurement as an objective assessment tool can effectively evaluate seafarers’ workload. This study conducts bridge simulation experiments to analyze eye movement features and investigate marine navigators’ workload levels across different tasks. By categorizing marine navigators’ workload into high and low level based on specific navigation duties. The study employs an independent sample T-test to select blink frequency, fixation time percentage, and pupil area standard deviation as key eye movement features for workload recognition. A marine navigators’ workload recognition model is developed utilizing the support vector machine (SVM) algorithm, which is subsequently trained and evaluated with experimental data. Results demonstrate a high recognition rate of 90.91% for marine navigators’ workload, indicating the SVM method’s accuracy in workload assessment based on eye movement features. These findings lay the foundation for the advancement of intelligent monitoring systems for marine navigators’ workload and the implementation of human–computer interaction active early warning technology and equipment.

Future changes in the vertical structure of severe convective storm environments over the U.S. central Great Plains

Abstract The effect of warming on severe convective storm potential is commonly explained in terms of changes in vertically-integrated (“bulk”) environmental parameters, such as CAPE and 0–6-km shear. However, such events are known to depend on details of the vertical structure of the thermodynamic and kinematic environment that can change independently of these bulk parameters. This work examines how warming may affect the complete vertical structure of these environments for fixed ranges of values of high CAPE and bulk shear, using data over the central Great Plains from two high-performing climate models (CNRM and MPI). To first order, projected changes in the vertical sounding structure is consistent between the two models: the environment warms approximately uniformly with height at constant relative humidity and the shear profile remains relatively constant. The boundary layer becomes slightly drier (−2–6% relative humidity) while the free troposphere becomes slightly moister (+1–3%), with a slight increase in moist static energy deficit aloft with stronger magnitude in CNRM. CNRM indicates enhanced low-level shear and storm-relative helicity associated with stronger hodograph curvature in the lowest 2 km, whereas MPI shows near zero change. Both models strongly underestimate shear below 1 km compared to ERA5, indicating large uncertainty in projecting subtle changes in the low-level flow structure in climate models. Evaluation of the net effect of these modest thermodynamic and kinematic changes on severe convective storm outcomes cannot be ascertained here but could be explored in simulation experiments.

Marginalized LASSO in the low-dimensional difference-based partially linear model for variable selection

The difference-based partially linear model is an appropriate regression model when both linear and nonlinear predictors are present in the data. However, when we want to optimize the weights using the difference-based method, the problem of variable selection can be difficult since low-variance predictors present a challenge. Therefore, this study aims to establish a novel methodology based on marginal theory to tackle such mixed relationships successfully, emphasizing variable selection in low dimensions. We suggest using a marginalized LASSO estimator with a penalty term that is not as severe and related to the difference order. As part of our numerical analysis of small sample performance, we undertake comprehensive simulation experiments to numerically demonstrate the strength of our proposed technique in estimation and prediction compared to the LASSO under a low-dimensional setup. This is done so that we can numerically demonstrate the strength of our proposed method in estimation and prediction. The bootstrapped method is utilized to evaluate how well our proposed prediction method performs when examining the King House dataset.

A noise‐tolerant fuzzy‐type zeroing neural network for robust synchronization of chaotic systems

SummaryAs a significant research issue in control and science field, chaos synchronization has attracted wide attention in recent years. However, it is difficult for traditional control methods to realize synchronization in predefined time and resist external interference effectively. Inspired by the excellent performance of zeroing neural network (ZNN) and the wide application of fuzzy logic system (FLS), a noise‐tolerant fuzzy‐type zeroing neural network (NTFTZNN) with fuzzy time‐varying convergent parameter is proposed for the synchronization of chaotic systems in this paper. Notably the fuzzy parameter generated from FLS combined with traditional convergent parameter embedded into this NTFTZNN can adjust the convergence rate according to the synchronization errors. For the sake of emphasizing the advantages of NTFTZNN model, other three sets of contrast models (FTZNN, VPZNN, and PTZNN) are constructed for the purpose of comparison. Besides, the predefined‐time convergence and noise‐tolerant ability of NTFTZNN model are distinctly demonstrated by detailed theoretical analysis. Furthermore, synchronization simulation experiments including two chaotic systems with different dimensions are provided to verify the related mathematical theories. Finally, the schematic of NTFTZNN model for chaos synchronization is accomplished completely through Simulink, further accentuating its effectiveness and potentials in practical applications.

A Data and Model-Driven Clutter Suppression Method for Airborne Bistatic Radar Based on Deep Unfolding

Space–time adaptive processing (STAP) based on sparse recovery achieves excellent clutter suppression and target detection performance, even with a limited number of available training samples. However, most of these methods face performance degradation due to grid mismatch, which impedes their application in bistatic clutter suppression. Some gridless methods, such as atomic norm minimization (ANM), can effectively address grid mismatch issues, yet they are sensitive to parameter settings and array errors. In this article, the authors propose a data and model-driven algorithm that unfolds the iterative process of atomic norm minimization into a deep network. This approach establishes a concrete and systematic link between iterative algorithms, extensively utilized in signal processing, and deep neural networks. This methodology not only addresses the challenges associated with parameter settings in traditional optimization algorithms, but also mitigates the lack of interpretability issues commonly found in deep neural networks. Moreover, due to more rational parameter settings, the proposed algorithm achieves effective clutter suppression with fewer iterations, thereby reducing computational time. Finally, extensive simulation experiments demonstrate the effectiveness of the proposed algorithm in clutter suppression for airborne bistatic radar.

DOA Estimation Based on Virtual Array Aperture Expansion Using Covariance Fitting Criterion

Providing higher precision Direction of Arrival (DOA) estimation has become a hot topic in the field of array signal processing for parameter estimation in recent years. However, when the physical aperture of the actual array is small, its aperture limitation means that even with super-resolution estimation algorithms, the achievable estimation precision is limited. This paper takes a novel approach by constructing an optimization algorithm using the covariance fitting criterion based on the array output’s covariance matrix to fit and obtain the covariance matrix of a large aperture virtual array, thereby providing high-precision angular resolution through virtual aperture expansion. The covariance fitting expansion analysis and discussion are unfolded for both uniform linear arrays (ULAs) and sparse linear arrays (SLAs) under four different scenarios. Theoretical analysis and simulation experiments demonstrate that these methods can enhance the effective performance of angle estimation, especially in low signal-to-noise ratios (SNRs) and at small angular intervals by fitting virtual extended aperture data.

Analysis of driving left-turn errors of young and older drivers at urban intersections

Numerous traffic crashes occur at intersections in urban areas. Understanding the mechanism of driving errors at intersections and providing effective solutions will significantly reduce traffic injuries. The objective of this study is to examine the age and hazard level’s effects on driver left-turn errors at urban intersections. In order to achieve this goal, this study customized the cognitive reliability and error analysis method (CREAM) for left-turn driving task at intersections, and driver error profiles were investigated using a driving simulation experiment as a case study. The results showed that age did not significantly affect drivers’ left turn error probability, while the error probability of both groups of drivers increased with the hazard level. Drivers were most likely to have inadequate monitoring and misjudgment in left-turn maneuvers, more so in the older group. Young drivers were more likely to make planning, decision-making, and adjustment errors compared to older drivers. In addition, observation errors were the most likely cognitive (74.5%). This study achieves the common and characteristic performance of human errors under the three hazardous intersections for the two age groups, and provides a basis for the design of effective training modules and driver assistance systems.

Software-Defined Virtual Private Network for SD-WAN

Software-Defined Wide Area Networks (SD-WANs) are an emerging Software-Defined Network (SDN) technology to reinvent Wide Area Networks (WANs) for ubiquitous network interconnections in cloud computing, edge computing, and the Internet of Everything. The state-of-the-art overlay-based SD-WANs are simply conjunctions of Virtual Private Network (VPN) and SDN architecture to leverage the controllability and programmability of SDN, which are only applicable for specific platforms and do not comply with the extensibility of SDN. This paper motivates us to refactor traditional VPNs with SDN architecture by proposing an overlay-based SD-WAN solution named Software-Defined Virtual Private Network (SD-VPN). An SDN-based auto-constructed VPN model and its evaluating metrics are put forward to automatically construct overlay WANs by node placement and service orchestration of SD-VPN. Therefore, a joint placement algorithm of VPN nodes and algorithms for overlay WAN service loading and offloading are proposed for SD-VPN controllers. Finally, a three-layer SD-VPN system is implemented and deployed in actual network environments. Simulation experiments and system tests are conducted to prove the high-efficiency controllability, real-time programmability, and auto-constructed deployability of the proposed SD-VPN. Performance trade-off between SD-VPN control channels and data channels is evaluated, and SD-VPN controllers are proven to be extensible for other VPN protocols and advanced services.