Simulation Runtime Research Articles

Multistructured Accretion Flow of Sgr A*. II. Signatures of a Cool Accretion Disk in Hydrodynamic Simulations of Stellar Winds

Hydrodynamic simulations of the stellar winds from Wolf–Rayet stars within the Galactic center can provide predictions for the X-ray spectrum of the supermassive black hole Sgr A*. Herein, we present results from updated smooth particle hydrodynamics simulations, building on the architecture of Cuadra et al. and Russell et al., and find that a “cold” (104 K) gas disk forms around Sgr A* with a simulation runtime of 3500 yr. This result is consistent with previous grid-based simulations, demonstrating that a cold disk can form regardless of numerical method. We examine the plasma scenarios arising from an environment with and without this cold disk, by generating synthetic spectra for comparison to the quiescent Fe Kα Sgr A* spectrum from Chandra HETGS, taken through the Chandra X-ray Visionary Program. We find that current and future X-ray missions are unlikely to distinguish between the kinematic signatures in the plasma in these two scenarios. Nonetheless, the stellar wind plasma model presents a good fit to the dispersed Chandra spectra within 1.″5 of Sgr A*. We compare our results to the radiatively inefficient accretion flow (RIAF) model fit to the HETGS spectrum presented in Paper I and find that the Bayesian model evidence does not strongly favor either model. With 9″ angular resolution and high spectral resolution of the X-IFU, NewAthena will offer a clearer differentiation between the RIAF plasma model and hydrodynamic simulations, but only a future X-ray mission with arcsecond resolution will significantly advance our understanding of Sgr A*’s accretion flow in X-rays.

Enhancing Indoor Localization Accuracy in Wireless Sensor Networks: A Hybrid Approach Integrating Hierarchical Structure Poly Particle Swarm Optimization and Teaching–Learning-Based Optimization (HSPPSO-TLBO)

Background and Objective: Enhancing localization accuracy while minimizing development costs poses significant challenges in deploying and managing wireless sensor networks (WSNs). This paper presents an advanced algorithm for node localization in indoor environments, integrating sophisticated optimization techniques. The hybrid algorithm, HSPPSO-TLBO, combines Hierarchical Structure Poly Particle Swarm Optimization (HSPPSO) with Teaching– Learning-Based Optimization (TLBO). Methods: The proposed algorithm HSPPSO-TLBO aims to minimize the mean squared range error (MSRE) resulted by calculating internal distances between nodes using Received Signal Strength Indicator (RSSI). TLBO, with its robust global search capabilities, complements HSPPSO’s local search, preventing convergence to inappropriate local optima. HSPPSO-TLBO offers easy implementation and leverages the cost-free feature of RSSI, making it an attractive choice for enhancing localization precision. Results: Simulation results demonstrate the superior performance of HSPPSO-TLBO compared to other algorithms using different meta-heuristic optimization techniques. The outstanding performance of HSPPSO-TLBO is evident across various evaluation metrics, including localization error, localization rate, and simulation runtime. Conclusion: The proposed algorithm utilizing HSPPSO and TLBO is exceptionally effective in improving localization precision in indoor WSNs due to several key characteristics. These include the seamless integration and easy implementation of both HSPPSO and TLBO, along with the costfree advantage of using the RSSI technique. This combination makes the algorithm a highly functional solution for improving localization accuracy. other: N/A

Toroidal modified Miller-Turner CME model in EUHFORIA: Validation and comparison with flux rope and spheromak

Rising concerns about the impact of space-weather-related disruptions demand modelling and reliable forecasting of coronal mass ejection (CME) impacts. In this study, we demonstrate the application of the modified Miller-Turner (mMT) model implemented within EUropean Heliospheric FORecasting Information Asset (EUHFORIA) in forecasting the geo-effectiveness of observed coronal mass ejection (CME) events in the heliosphere. Our goal is to develop a model that not only has a global geometry, in order to improve overall forecasting, but is also fast enough for operational space-weather forecasting. We test the original full torus implementation and introduce a new three-fourths Torus version called the Horseshoe CME model. This new model has a more realistic CME geometry, and overcomes the inaccuracies of the full torus geometry. We constrain the torus geometrical and magnetic field parameters using observed signatures of the CMEs before, during, and after the eruption. We perform EUHFORIA simulations for two validation cases -- the isolated CME event of 12 July 2012 and the CME--CME interaction event of 8-10 September 2014. We performed an assessment of the model's capability to predict the most important $B_z$ component using the advanced dynamic time-warping (DTW) technique. The Horseshoe model predictions of CME arrival time and geo-effectiveness for both validation events compare well with the observations and are weighed against the results obtained with the spheromak and FRi3D models, which were already available in EUHFORIA. The runtime of the Horseshoe model simulations is close to that of the spheromak model, which is suitable for operational space weather forecasting. However, the capability of the magnetic field prediction at 1 AU of the Horseshoe model is close to that of the FRi3D model. In addition, we demonstrate that the Horseshoe CME model can be used for simulating successive CMEs in EUHFORIA, overcoming a limitation of the FRi3D model.

Thermal–Electrical–Mechanical Coupled Finite Element Models for Battery Electric Vehicle

The safety of lithium-ion batteries is critical to the safety of battery electric vehicles (BEVs). The purpose of this work is to develop a method to predict battery thermal runaway in full electric vehicle crash simulation. The thermal–electrical–mechanical-coupled finite element analysis is used to model an individual lithium-ion battery cell, a battery module, a battery pack, and a battery electric vehicle with 24 battery modules in a live circuit connection. The lithium-ion battery is modeled using a representative approach, with each internal battery component individually modeled to represent its geometric shape and realistic thermal, mechanical, and electrical properties. A resistance heating solver and Randles circuit model built with a generalized voltage source are used to simulate the electrical behavior of the battery. The thermal simulation of the battery considers the heat capacity and thermal conductivity of different cell components, as well as heat conduction, radiation, and convection at their interfaces. The mechanical property of battery cell and battery module models is validated using spherical punch tests. The electrical property of the battery cell and battery module models is verified against CircuitLab simulation in an external short-circuit test. The simulation results for the battery module’s internal resistance are consistent with both experimental data and literature values. The multi-physics coupling phenomenon is demonstrated with a cylindrical compression simulation on the battery module. The multi-physics BEV model with 24 live battery modules is used to simulate the external short-circuit test and the side pole impact test. The simulation run time is less than 24 h. The results demonstrated the feasibility of using a representative battery model and multi-physics analysis to predict battery thermal runaway in full electric vehicle crash analysis.

Hyper-reduced-order model for estimating convection heat transfer coefficients of turbine rotors

Forced convection heat transfer occurs inside the steam turbine, and it is crucial for thermal and strength analysis to accurately estimate the convection heat transfer coefficients of the turbine rotor. This paper presents a hyper-reduced-order model for efficiently estimating the forced convection heat transfer coefficients of a steam turbine rotor. Unlike the full-order finite element method, the hyper-reduced-order model uses the discrete empirical interpolation method to approximate the domain integration of transient heat conduction simulations. This approach greatly reduces the degrees of freedom of numerical integration, thus substantially improving the computational efficiency of forward heat conduction problems. The hyper-reduced-order model and Levenberg–Marquardt algorithm are combined to minimize temperature errors between the numerical simulation results and remote sensor measurements and iteratively estimate the heat transfer coefficients of forced convection in the rotor. The accuracy and efficiency of the proposed model are verified through a transient heat conduction simulation. Moreover, the effects of the initial guess values and measurement errors are investigated to demonstrate the universality and robustness of the proposed approach in determining the convection heat transfer coefficient. Compared with the full-order finite element model, the proposed hyper-reduced-order model can increase the efficiency of forward heat conduction simulations of the turbine rotor by more than 10 times, and it only takes 0.87 s computer runtime in single transient step simulation, and then rapidly estimate the forced convection heat transfer coefficients. Furthermore, the performance of the proposed approach is insensitive to the initial guess values, and it exhibits great robustness under measurement error. The mean errors are 4.78 %, 2.67 % and 10.51 % when the initial values are 0.015 mW·mm−3·°C−1, 0.05 mW·mm−3·°C−1 and 0.15 mW·mm−3·°C−1 respectively. When there is no measurement error, the estimated error of the convection heat transfer coefficients is 2.67 %, and even if the measurement error reaches 5 %, and the accuracy loss is only 5.57 %. Consequently, this approach is highly suitable for estimating convection heat transfer coefficients during the dynamic operation of steam turbines, and this research provides reliable guidance for the optimization design and safe operation of steam turbines.

Preheating with deep learning

We apply deep learning techniques to the late-time turbulent regime in a post-inflationary model where a real scalar inflaton field and the standard model Higgs doublet interact with renormalizable couplings between them. After inflation, the inflaton decays into the Higgs through a trilinear coupling and the Higgs field subsequently thermalizes with gauge bosons via its SU(2)×U(1) gauge interaction. Depending on the strength of the trilinear interaction and the Higgs self-coupling, the effective mass squared of Higgs can become negative, leading to the tachyonic production of Higgs particles. These produced Higgs particles would then share their energy with gauge bosons, potentially indicating thermalization. Since the model entails different non-perturbative effects, it is necessary to resort to numerical and semi-classical techniques. However, simulations require significant costs in terms of time and computational resources depending on the model used. Particularly, when SU(2) gauge interactions are introduced, this becomes evident as the gauge field redistributes particle energies through rescattering processes, leading to an abundance of UV modes that disrupt simulation stability. This necessitates very small lattice spacings, resulting in exceedingly long simulation runtimes. Furthermore, the late-time behavior of preheating dynamics exhibits a universal form by wave kinetic theory. Therefore, we analyze patterns in the flow of particle numbers and predict future behavior using CNN-LSTM (Convolutional Neural Network combined with Long Short-Term Memory) time series analysis. In this way, we can reduce our dependence on simulations by orders of magnitude in terms of time and computational resources.

IMMISCIBLE VISCOUS FINGERING MODELING OF TERTIARY POLYMER FLOODING BASED ON REAL CASE OF HEAVY OIL RESERVOIR MODEL

In immiscible displacements, lower viscosity injected fluids with higher mobility than crude oil can create viscous fingers, affecting displacement efficiency. The Buckley–Leverett approach for relative permeabilities (kr) may not represent accurately 2D features like increased water saturation in viscous fingering. Based on the physics issue, this work applies Sorbie's 4-Steps methodology to a 3D simulation of an offshore heavy oil reservoir focusing on waterflooding and tertiary polymer flooding, assessing their impact on oil production forecasts. It also explores the application of this methodology to coarse grid simulation models, employing pseudo kr functions by data assimilation. During tertiary polymer injection, two processes were identified in oil displacement: viscous crossflow mechanism and oil bank mobilization by a second finger. This combination resulted in earlier and increased oil production. For both strategies, refining the grid increased simulation runtime from minutes to days compared to coarse grids, making it impractical for intensive processes. From data assimilation, the best solution with matched field indicators reduced runtime from days to minutes. This study expanded the 4-Steps methodology for 3D reservoir simulation, proposing kr as uncertainties. Data assimilation enhances the methodology, generating pseudo kr for coarser grid simulations, reducing computational costs, and capturing small-scale phenomena.

Utilizing well-reservoir pseudo-connections for multi-stage hydraulic fracturing modeling in tight gas saturated formations

Purpose. Research is aimed at integrating multi-stage hydraulic fracturing in horizontal wells with hydrodynamic simulation as a mandatory part of planning the mining of any shale oil or gas reservoir. Methods. Geological and hydrodynamic reservoir modeling is part of the research. The properties and geometries of the hydraulic fracture network and its representation in the dynamic reservoir model were assessed. The comparative characterization was carried out based on the two methods of fracture modeling: cell dimension reduction for explicit fracture modeling (LGR – local grid refinement) and implicit fracture modeling method, presented in this paper, with additional pseudo-connections between well and reservoir. Findings. A hydrodynamic model for low-permeable reservoir, produced by horizontal well, hydraulically fractured with 5 stages, has been generated. This model is calibrated to the production history and flowing bottom hole pressure by applying two methods of fracture modeling. Modeling results show that it is possible to replicate historical well production by using both methods. However, the proposed method with pseudo connections has several advantages compared to the generally accepted, local grid refinement (LGR) method. Originality. For the first time, a system of pseudo connections between well and reservoir was constructed to model a multi-stage hydraulic fracturing for a hydrodynamic model of tight reservoir. Hydrodynamic simulation results were refined and calibrated to the history of hydrocarbon production and flowing bottom hole pressure data using the pseudo-connections and LGR methods. The similarity of the results by applying LGR and pseudo-connections methods was revealed. Practical implications. The use of pseudo connections for hydraulic fracturing modeling can reduce simulation run time for cases where multi-stage hydraulic fracturing has already been carried out or is planned in the future. Additionally, the use of this method allows testing a larger number of realizations and scenarios, including hydraulic fracturing design (number of stages, size and conductivity of resulted fracture systems, fracture orientation, etc.), well placement and fracture growth relative to well trajectory. Also, there is no need to rebuild a model every time for each realization, as is the case with the LGR method.

High-Resolution Modeling Without Computation Slowdown for PETALE in CROCUS

In a collaboration between Ecole Polytechnique Fédérale de Lausanne (EPFL) and CEA, in the fall of 2020, the experimental Programme d’Étude en Transmission de l’Acier Lourd et ses Eléments (PETALE) was successfully carried out in the CROCUS reactor of EPFL. This article presents and compares the methods tested in the modeling of the experiments, specifically focusing on the metal reflectors installed at the periphery of CROCUS. A basic design model consisting of a few cuboids was refined to a fully detailed version, without impacting the run time of simulations. Notably, each reflector sheet of PETALE was segmented into 121 voxels based on topological measurements. This detailed voxelization did not affect calculation times, thanks to the use of three-dimensional lattices as available in Serpent 2. Profiling of the simulations revealed the high computational surface transformations associated with Serpent 2 and highlighted the efficiency benefits of factorizing these into universe transformations. As the CROCUS simulations were carried out using a modified build of Serpent 2, additional simulations were also performed using a standard version of Serpent 2 with a GODIVA model as a neutron source to ensure that the findings are generalizable. These additional tests confirmed the initial results, with significant performance variations observed between the models, particularly larger in surface-tracking mode than in delta-tracking mode. Consequently, the modeling method may therefore be applied to future high-fidelity modeling of neutron transmission and shielding experiments.

A journey in unraveling the enantiorecognition mechanism of 3,5-dinitrobenzoyl-amino acids with two Cinchona alkaloid-based chiral stationary phases: The power of molecular dynamic simulations

BackgroundInnovations in computer hardware and software capabilities have paved the way for advances in molecular modelling techniques and methods, leading to an unprecedented expansion of their potential applications. In contrast to the docking technique, which usually identifies the most stable selector-selectand (SO-SA) complex for each enantiomer, the molecular dynamics (MD) technique enables the consideration of a distribution of the SO-SA complexes based on their energy profile. This approach provides a more truthful representation of the processes occurring within the column. However, benchmark procedures and focused guidelines for computational treatment of enantioselectivity at the molecular level are still missing. ResultsTwenty-eight molecular dynamics simulations were performed to study the enantiorecognition mechanisms of seven N-3,5-dinitrobenzoylated α- and β-amino acids (DNB-AAs), occurring with the two quinine- and quinidine-based (QN-AX and QD-AX) chiral stationary phases (CSPs), under polar-ionic conditions. The MD protocol was optimized in terms of box size, simulation run time, and frame recording frequency. Subsequently, all the trajectories were analyzed by calculating both the type and amount of the interactions engaged by the selectands (SAs) with the two chiral selectors (SOs), as well as the conformational and interaction energy profiles of the formed SA-SO associates. All the MDs were in strict agreement with the experimental enantiomeric elution order and allowed to establish (i) that salt-bridge and H-bond interactions play a pivotal role in the enantiorecognition mechanisms, and (ii) that the π-cation and π-π interactions are the discriminant chemical features between the two SOs in ruling the chiral recognition mechanism. SignificanceThe results of this work clearly demonstrate the high contribution given by MD simulations in the comprehension of the enantiorecognition mechanism with Cinchona alkaloid-based CSPs. However, from this research endeavor it clearly emerged that the MD protocol optimization is crucial for the quality of the produced results.

Reducing the Runtime of Fault-Tolerant Quantum Simulations in Chemistry through Symmetry-Compressed Double Factorization.

Quantum phase estimation based on qubitization is the state-of-the-art fault-tolerant quantum algorithm for computing ground-state energies in chemical applications. In this context, the 1-norm of the Hamiltonian plays a fundamental role in determining the total number of required iterations and also the overall computational cost. In this work, we introduce the symmetry-compressed double factorization (SCDF) approach, which combines a CDF of the Hamiltonian with the symmetry shift technique, significantly reducing the 1-norm value. The effectiveness of this approach is demonstrated numerically by considering various benchmark systems, including the FeMoco molecule, cytochrome P450, and hydrogen chains of different sizes. To compare the efficiency of SCDF to other methods in absolute terms, we estimate Toffoli gate requirements, which dominate the execution time on fault-tolerant quantum computers. For the systems considered here, SCDF leads to a sizable reduction of the Toffoli gate count in comparison to other variants of DF or even tensor hypercontraction, which is usually regarded as the most efficient approach for qubitization.

Automation and optimization of stopping and range of ions in matter simulation runtime

Prior to every ion implantation experiment a simulation of the ion range and other relevant parameters is performed using Monte-Carlo based codes. Although increasing computing power has improved the speed of these calculations, the demands on Monte-Carlo codes are also increasing, requiring evaluation of the optimal number of simulations while ensuring accuracy within threshold bounds. We evaluate the “Stopping and Range of Ions in Matter” (SRIM) code due to its widespread usage. We show how dividing simulations into multiple parallel simulations with different random seeds can lead to calculation speedup and find lower bounds for the required number of ion traces simulated based on an exemplar system of a Ga focused ion beam and a high energy C beam as used in high linear energy transfer testing. Our results indicate simulations can yield results within the underlying data accuracy of SRIM at 10X and 100X shorter simulation time than the SRIM default values.

MODEL-BASED ANALYSIS TO EVALUATE THE EFFECT OF POLYMER PROPERTIES AND PHYSICAL PHENOMENA ON POLYMER FLOODING OPERATIONS

Polymer flooding, known for enhancing heavy oil displacement, may encounter efficiency-limiting physical phenomena. This work assesses the impact of key factors like viscosity, shear rate, and adsorption in field applications using a model-based approach on the EPIC001 case, a real offshore Brazilian heavy oil reservoir. In simulation results, increasing the displacing fluid viscosity using apparent or zero-shear functions, oil production and economic returns show improvements over waterflooding. The shear rate effect slightly increases oil production and enhances injectivity loss due to shear-thinning in polymer flow. However, modeling it increases computation costs, as it extends simulation run time from minutes to days, making it impractical for intensive processes like production optimization. An analysis of the method’s effectiveness shows that it varies based on the adsorption level considered. At its highest value, even with a higher oil recovery factor, the economic return was lower than using waterflooding. Combining shear rate and adsorption has a minimal impact on field indicators when compared to adsorption alone. This work enhances the comprehension of physical phenomena and non-Newtonian behavior in tertiary polymer flooding on heavy oil reservoirs and its impact on the production forecast. It also highlights important considerations for modeling-based procedures.

Efficient simulation of stochastic interactions among representative Monte Carlo particles

Context. Interaction processes between discrete particles are often modelled with stochastic methods such as the Representative Particle Monte Carlo (RPMC) method which simulate mutual interactions (e.g. chemical reactions, collisions, gravitational stirring) only for a representative subset of n particles instead of all N particles in the system. However, in the traditionally employed computational scheme the memory requirements and the simulation runtime scale quadratically with the number of representative particles. Aims. We want to develop a computational scheme that has significantly lower memory requirements and computational costs than the traditional scheme, so that highly resolved simulations with stochastic processes such as the RPMC method become feasible. Results. In this paper we propose the bucketing scheme, a hybrid sampling scheme that groups similar particles together and combines rejection sampling with a coarsened variant of the traditional discrete inverse transform sampling. For a v-partite bucket grouping, the storage requirements scale with n and v2, and the computational cost per fixed time increment scales with n ⋅ v, both thus being much less sensitive to the number of representative particles n. Extensive performance testing demonstrates the higher efficiency and the favourable scaling characteristics of the bucketing scheme compared to the traditional approach, while being statistically equivalent and not introducing any new requirements or approximations. With this improvement, the RPMC method can be efficiently applied not only with very high resolution but also in scenarios where the number of representative particles increases over time, and the simulation of high-frequency interactions (such as gravitational stirring) as a Monte Carlo process becomes viable.

Fast urban inundation simulation with RIM2D for flood risk assessment and forecasting

IntroductionUrban pluvial flooding is a growing concern worldwide as consequence of rising urban population and climate change induced increases in heavy rainfall. Easy-to-implement and fast simulation tools are needed to cope with this challenge.MethodsThis study describes the development of the parsimonious, GPU-accelerated hydraulic model RIM2D for urban pluvial flood simulations. This is achieved by considering the built-up urban area as flow obstacles, and by introducing capacity-based approaches to consider urban drainage by infiltration on pervious surfaces and sewer drainage from roofs and sealed surfaces. The model performance was analyzed by simulating 8 heavy rainfall events in a test area in the city of Dresden, Germany. For these events detailed discharge measurements of sewer discharge are available, providing a unique dataset for evaluating the sewer drainage simulation, which is of high importance for realistic pluvial inundation simulations in urban areas.Results and discussionWe show that the model simulates the temporal dynamics of the sewer discharge and the sewer volume within acceptable ranges. Moreover, the erratic variation of the simulated to measured sewer discharge suggests that the deviations from the measurements are caused by the precipitation input rather than the model simplifications. We conclude that RIM2D is a valid tool for urban inundation simulation. Its short simulation runtimes allow probabilistic flood risk assessments and operational flood forecasts.

Deterministic and stochastic tendency adjustments derived from data assimilation and nudging

AbstractWe develop and compare model‐error representation schemes derived from data assimilation increments and nudging tendencies in multidecadal simulations of the Community Atmosphere Model, version 6. Each scheme applies a bias correction during simulation runtime to the zonal and meridional winds. We quantify the extent to which such online adjustment schemes improve the model climatology and variability on daily to seasonal timescales. Generally, we observe about a 30% improvement to annual upper‐level zonal winds, with largest improvements in boreal spring (around 35%) and winter (around 47%). Despite only adjusting the wind fields, we additionally observe around 20% improvement to annual precipitation over land, with the largest improvements in boreal fall (around 36%) and winter (around 25%), and around 50% improvement to annual sea‐level pressure, globally. With mean‐state adjustments alone, the dominant pattern of boreal low‐frequency variability over the Atlantic (the North Atlantic Oscillation) is significantly improved. Additional stochasticity increases the modal explained variances further, which brings the variability closer to the observed value. A streamfunction tendency decomposition reveals that the improvement is due to an adjustment to the high‐ and low‐frequency eddy–eddy interaction terms. In the Pacific, the mean‐state adjustment alone led to an erroneous deepening of the Aleutian low, but this was remedied with the addition of stochastically selected tendencies. Finally, from a practical standpoint, we discuss the performance of using data assimilation increments versus nudging tendencies for an online model‐error representation.

Data-driven prediction of laminar burning velocity for ternary ammonia/hydrogen/methane/air premixed flames

Zero-carbon fuels such as hydrogen and ammonia play a pivotal role in the energy transition by offering cleaner alternatives to natural gas (methane), especially in industrial combustion systems. Binary and ternary blends of these fuels offer a transitionary, low-carbon solution in the near future. Laminar burning velocity (LBV), as a fundamental combustion property, is significantly different for ammonia, hydrogen, and methane. Although the LBV of binary blends of these fuels is well-studied, ternary blends have not been extensively studied. In this study, the primary objective is to employ a simple ensemble learning method to predict the LBV of ternary ammonia/hydrogen/methane/air mixtures. The training dataset consists of experimental data sourced from a large number of publications (3,846 data points), as well as synthetic data generated by 1D freely propagating premixed flame simulations in Cantera using a detailed chemical kinetic model. Three machine learning algorithms, namely artificial neural networks, gaussian process regression, and extreme gradient boosting trees are trained and optimised. Then, a simple ensemble averaging method is used to reduce overfitting and improve robustness. The ensemble model achieves coefficient of determination (R2) of 0.991 on the test set with an inference time that is approximately 8,000 times faster than the 1D simulation run time. The ensemble model is capable of predicting LBVs of ammonia/hydrogen/methane/air mixtures for T=[295–756K], P=[1–10bar], ϕ=[0.5–1.8] across all possible blending ratios.

A design TCADAS tool for semiconductor devices and case study of 65 nm conventional floating-gate MOS transistor

An automatic programming tool has become an essential component in virtual fabrication in recent years. This paper aims to propose a methodology of virtual fabrication for semiconductor devices and design a tool called Technology Computer-Aided Design Automatic Simulation (TCADAS) which can perform a completely virtual fabrication, device simulation, process variation, and output extraction. Especially, the TCADAS tool eliminates drudgery when studying semiconductor devices such as complexity in setting inputs, substantial manual work, and long run time of simulations. This work investigates the completed flow for a 65 nm conventional Floating-gate MOS transistor as a case study, which is widely considered a vital determinant of the non-volatile memory field. A detailed automatic process from the entry to a completed three-dimension structure, device characterization, extract important output parameters such as memory window, ION/IOFF, Gate Coupling Ratio (GCR), and speed operations are presented. The TCADAS tool was designed by Python language and the utilities of TCAD Silvaco tools including Athena, Atlas, and DevEdit3D.

Supervised Local Training With Backward Links for Deep Neural Networks

The restricted training pattern in the standard BP requires end-to-end error propagation, causing large memory costs and prohibiting model parallelization. Existing local training methods aim to resolve the training obstacles by completely cutting off the backward path between modules and isolating their gradients. These methods prevent information exchange between modules and result in inferior performance. This work proposes a novel local training algorithm, BackLink, which introduces inter-module backward dependency and facilitates information to flow backward along with the network. To preserve the computational advantage of local training, BackLink restricts the error propagation length within the module. Extensive experiments performed in various deep convolutional neural networks demonstrate that our method consistently improves the classification performance of local training algorithms over other methods. For example, our method can surpass the conventional greedy local training method by 6.45% in accuracy in ResNet32 classifying CIFAR100 and recent work by 2.58% in ResNet110 classifying STL-10 with much lower complexity, respectively. Analysis of computational costs reveals that small overheads are incurred in GPU memory costs and runtime on multiple GPUs. Our method can lead up to a 79% reduction in memory cost and 52% in simulation runtime in ResNet110 compared to the standard BP. Therefore, our method could create new opportunities for improving training algorithms towards better efficiency for real-time learning applications.

Power sharing enhancement through a decentralized droop-based control strategy in an islanded microgrid

A decentralized intelligent droop-based control strategy is proposed in this paper for the enhancement of equal active and reactive power sharing in an islanded inverter-based microgrid. Droop control is mostly preferred because it does not need communication facilities for its implementation. However, the presence of different feeder impedances for the different distributed generators (DGs) in a microgrid make reactive power sharing inaccurate. Furthermore, as a result of considerable load changes and different droop characteristics for the DGs, the inverter's output voltage and frequency are altered and these in turn alter the reactive and active power sharing respectively. With these conditions, the generalized droop control (GDC) strategy fails to effectively share load reactive and active power among the DGs. In order to extenuate these challenges, this paper presents a nonlinear autoregressive exogenous neural network droop-based control (NARX-NN DBC) strategy which does not depend on the varying line impedances, droop gains for the different DGs and is less affected by fluctuating loads in the grid. A microgrid, made up of two DGs and a load is modelled in MATLAB/Simulink environment and validation of the proposed control strategy is done through many simulations. Within the simulation runtime, NARX-NN DBC yielded a maximum frequency percentage deviation of 0.46% from the nominal value of 50 Hz whereas GDC yielded 0.62%. Regarding the voltage, NARX-NN DBC gave maximum deviation of 0.026% meanwhile GDC gave 0.079% from the nominal value for 380 V. In addition, during 0.57–0.64 s with load active power demand of 4.6 kW, NARX-NN DBC registered 0.43% power sharing error whereas GDC registered 6.5%. On the other hand, during 0.57–0.64 s, with 5kVAr load power demand, NARX-NN DBC registered 0.2% whereas, GDC registered 2%. These obtained results clearly show that NARX-NN DBC strategy has a better performance compared to GDC strategy with respect to power sharing in an autonomous microgrid.