Gated Neural Network-Based Unsteady Aerodynamic Modeling for Large Angles of Attack

Modeling of unsteady aerodynamic loads at high angles of attack using a small amount of experimental or simulation data to construct predictive models for unknown states can greatly improve the efficiency of aircraft unsteady aerodynamic design and flight dynamics analysis.In this paper,aiming at the problems of poor generalization of traditional aerodynamic models and intelligent models,an intelligent aerodynamic modeling method based on gated neural units is proposed.The time memory characteristics of the gated neural unit is fully utilized,thus the nonlinear flow field characterization ability of the learning and training process is enhanced,and the generalization ability of the whole prediction model is improved.The prediction and verification of the model are carried out under the maneuvering flight condition of NACA0015 airfoil.The results show that the model has good adaptability.In the interpolation prediction,the maximum prediction error of the lift and drag coefficients and the moment coefficient does not exceed 10%,which can basically represent the variation characteristics of the entire flow field.In the construction of extrapolation models,the training model based on the strong nonlinear data has good accuracy for weak nonlinear prediction.Furthermore,the error is larger,even exceeding 20%,which indicates that the extrapolation and generalization capabilities need to be further optimized by integrating physical models.Compared with the conventional state space equation model,the proposed method can improve the extrapolation accuracy and efficiency by 78%and 60%,respectively,which demonstrates the applied potential of this method in aerodynamic modeling.

Comparison of Two Recurrent Neural Networks for Rainfall-Runoff Modeling in the Zou River Basin at Atchérigbé (Bénin)

Hydrological models are developed to simulate river flows over a watershed for many practical applications in the field of water resource management. The present paper compares the performance of two recurrent neural networks for rainfall-runoff modeling in the Zou River basin at Atchérigbé outlet. To this end, we used daily precipitation data over the period 1988-2010 as input of the models, such as the Long Short-Term Memory (LSTM) and Recurrent Gate Networks (GRU) to simulate river discharge in the study area. The investigated models give good results in calibration (R2 = 0.888, NSE = 0.886, and RMSE = 0.42 for LSTM;R2 = 0.9, NSE = 0.9 and RMSE = 0.397 for GRU) and in validation (R2 = 0.865, NSE = 0.851, and RMSE = 0.329 for LSTM;R2 = 0.9, NSE = 0.865 and RMSE = 0.301 for GRU). This good performance of LSTM and GRU models confirms the importance of models based on machine learning in modeling hydrological phenomena for better decision-making.

Modeling of Gate Tunneling Current for Nanoscale MOSFETs with High-k Gate Stacks

A quantum model based on solutions to the Schrodinger-Poisson equations is developed to investigate the device behavior related togate tunneling current for nanoscale MOSFETs with high-k gate stacks. This model can model various MOS device structures with combinations of high-k dielectric materials and multilayer gate stacks,revealing quantum effects on the device performance. Comparisons are made for gate current behavior between nMOSFET and pMOSFET high- k gate stack structures. The results presented are consistent with experimental data, whereas a new finding for an optimum nitrogen content in HfSiON gate dielectric requires further experimental verifications.

Combined Novel Gate Level Model and Critical Primary Input Sharing for Genetic Algorithm Based Maximum Power Supply Noise Estimation

A gate level maximum power supply noise （PSN） model is defined that captures both IR drop and di/dt noise effects. Experimental results show that this model improves PSN estimation by 5.3% on average and reduces computation time by 10.7% compared with previous methods. Furthermore,a primary input critical factor model that captures the extent of primary inputs＇ PSN contribution is formulated. Based on these models,a novel niche genetic algorithm is proposed to estimate PSN more effectively. Compared with general genetic algorithms, this novel method can achieve up to 19.0% improvement on PSN estimation with a much higher convergence speed.

Quasi-two-dimensional threshold voltage model for junctionless cylindrical surrounding gate metal-oxide-semiconductor field-effect transistor with dual-material gate

Based on the quasi-two-dimensional （2D） solution of Poisson＇s equation in two continuous channel regions, an an- alytical threshold voltage model for short-channel junctionless dual-material cylindrical surrounding-gate （JLDMCSG） metal-oxide-semiconductor field-effect transistor （MOSFET） is developed. Using the derived model, channel potential dis- tribu6o~, h~riz~atal electrical ~eld distributign, a~d threshold v~ltage roll-off of ~LDMCSG MOSFET are in,instigated. Compared with junctionless single-material CSG （JLSGCSG） MOSFET~ JLDMCSG MOSFET can effectively suppress short-channel effects and simultaneously improve carrier transport efficiency. It is also revealed that threshold voltage roll- off of JLDMCSG can be significantly reduced by adopting both a small oxide thickness and a small silicon channel radius. The model is verified by comparing its calculated results with that obtained from three-dimensional （3D） numerical device simulator ISE.

Performance and analytical modelling of halo-doped surrounding gate MOSFETs

Halo structure is added to sub-100 nm surrounding-gate metal-oxide-semiconductor fieldeffect-transistors （MOS- FETs） to suppress short channel effect. This paper develops the analytical surface potential and threshold voltage models based on the solution of Poisson＇s equation in fully depleted condition for symmetric halo-doped cylindrical surrounding gate MOSFETs. The performance of the halo-doped device is studied and the validity of the analytical models is verified by comparing the analytical results with the simulated data by three dimensional numerical device simulator Davinci. It shows that the halo doping profile exhibits better performance in suppressing threshold voltage roll-off and drain-induced barrier lowering, and increasing carrier transport efficiency. The derived analytical models are in good agreement with Davinci.

A two-dimensional analytical subthreshold behavior model for junctionless dual-material cylindrical surrounding-gate MOSFETs

A two-dimensional analytical subthreshold behavior model for junctionless dual-material cylindrical surrounding- gate （JLDMCSG） metal-oxide-semiconductor field-effect transistors （MOSFETs） is proposed. It is derived by solving the two-dimensional Poisson＇s equation in two continuous cylindrical regions with any simplifying assumption. Using this analytical model, the subthreshold characteristics of JLDMCSG MOSFETs are investigated in terms of channel electro- static potential, horizontal electric field, and subthreshold current. Compared to junctionless single-material cylindrical surrounding-gate MOSFETs, JLDMCSG MOSFETs can effectively suppress short-channel effects and simultaneously im- prove carrier transport efficiency. It is found that the subthreshold current of JLDMCSG MOSFETs can be significantly reduced by adopting both a thin oxide and thin silicon channel. The accuracy of the analytical model is verified by its good agreement with the three-dimensional numerical simulator ISE TCAD.

Research on Flight First Service Model and Algorithms for the Gate Assignment Problem

Aiming at the problem of gate allocation of transit flights,a flight first service model is established.Under the constraints of maximizing the utilization rate of gates and minimizing the transit time,the idea of“first flight serving first”is used to allocate the first time,and then the hybrid algorithm of artificial fish swarm and simulated annealing is used to find the optimal solution.That means the fish swarm algorithm with the swallowing behavior is employed to find the optimal solution quickly,and the simulated annealing algorithm is used to obtain a global optimal allocation scheme for the optimal local region.The experimental data show that the maximum utilization of the gate is 27.81%higher than that of the“first come first serve”method when the apron is not limited,and the hybrid algorithm has fewer iterations than the simulated annealing algorithm alone,with the overall passenger transfer tension reducing by 1.615;the hybrid algorithm has faster convergence and better performance than the artificial fish swarm algorithm alone.The experimental results indicate that the hybrid algorithm of fish swarm and simulated annealing can achieve higher utilization rate of gates and lower passenger transfer tension under the idea of“first flight serving first”.

Building up a Mathematical Model of Shaftes System and CAD for Linked Tetra-Shaft and Double-Cantilever Flat Flap Gate

Linked tetra shaft and double cantilever flat flap gate is a new type of structure in water conservancy projects,but the traditional method is now adopted in its design.In order to the application and dissemination of the type of sluice,this paper researches the difficult points of design advance,through researching the motion locus & stress coundition of linked tetra shaft system.The writer will build up the mathmatical model and handle it with the computer.Thus,we can achieve the modern desing on the basis of the software of linked tetra shaft system development.

Modified model of gate leakage currents in AlGaN/GaN HEMTs

It has been reported that the gate leakage currents are described by the Frenkel-Poole emission(FPE) model,at temperatures higher than 250 K.However,the gate leakage currents of our passivated devices do not accord with the FPE model.Therefore,a modified FPE model is developed in which an additional leakage current,besides the gate(ⅠⅡ),is added.Based on the samples with different passivations,the ⅠⅡcaused by a large number of surface traps is separated from total gate currents,and is found to be linear with respect to(φB-Vg)0.5.Compared with these from the FPE model,the calculated results from the modified model agree well with the Ig-Vgmeasurements at temperatures ranging from 295 K to 475 K.

Temperature- and voltage-dependent trap generation model in high-k metal gate MOS device with percolation simulation

High-k metal gate stacks are being used to suppress the gate leakage due to tunneling for sub-45 nm technology nodes.The reliability of thin dielectric films becomes a limitation to device manufacturing,especially to the breakdown characteristic.In this work,a breakdown simulator based on a percolation model and the kinetic Monte Carlo method is set up,and the intrinsic relation between time to breakdown and trap generation rate R is studied by TDDB simulation.It is found that all degradation factors,such as trap generation rate time exponent m,Weibull slope β and percolation factor s,each could be expressed as a function of trap density time exponent α.Based on the percolation relation and power law lifetime projection,a temperature related trap generation model is proposed.The validity of this model is confirmed by comparing with experiment results.For other device and material conditions,the percolation relation provides a new way to study the relationship between trap generation and lifetime projection.

Analytical model including the fringing-induced barrier lowering effect for a dual-material surrounding-gate MOSFET with a high-κ gate dielectric

By solving Poisson＇s equation in both semiconductor and gate insulator regions in the cylindrical coordinates, an analytical model for a dual-material surrounding-gate （DMSG） metal-oxide semiconductor field-effect transistor （MOSFET） with a high-k gate dielectric has been developed. Using the derived model, the influences of fringing-induced barrier lowering （FIBL） on surface potential, subthreshold current, DIBL, and subthreshold swing are investigated. It is found that for the same equivalent oxide thickness, the gate insulator with high-k dielectric degrades the short-channel performance of the DMSG MOSFET. The accuracy of the analytical model is verified by the good agreement of its results with that obtained from the ISE three-dimensional numerical device simulator.

Double-gate tunnel field-effect transistor:Gate threshold voltage modeling and extraction

The tunnel field-effect transistor(TFET) is a potential candidate for the post-CMOS era.As one of the most important electrical parameters of a device,double gate TFET(DG-TFET) gate threshold voltage was studied.First,a numerical simulation study of transfer characteristic and gate threshold voltage in DG-TFET was reported.Then,a simple analytical model for DG-TFET gate threshold voltage VTG was built by solving quasi-two-dimensional Poisson equation in Si film.The model as a function of the drain voltage,the Si layer thickness,the gate length and the gate dielectric was discussed.It is shown that the proposed model is consistent with the simulation results.This model should be useful for further investigation of performance of circuits containing TFETs.

Statistical Modeling of Gate Capacitance Variations Induced by Random Dopants in Nanometer MOSFETs Reserving Correlations

We consider intrinsic gate capacitance variations due to random dopants in the nanometer metal oxide semi- conductor field effect transistor （MOSFET） channel. The variations of total gate capacitance and gate transcapacitances are investigated and the strong correlations between the trans-capacitance variations are discovered. A simple statistical model is proposed for accurately capturing total gate capacitance variability based on the correlations. The model fits very well with the Monte Carlo simulations and the average errors are -0.033% for n-type metal-oxide semiconductor and -0.012% for p-type metal-oxide semiconductor, respectively. Our simulation studies also indicate that, owing to these correlations, the total gate capacitance variability will not dominate in gate capacitance variations.

An improvement to computational efficiency of the drain current model for double-gate MOSFET

As a connection between the process and the circuit design, the device model is greatly desired for emerging devices, such as the double-gate MOSFET. Time efficiency is one of the most important requirements for device modeling. In this paper, an improvement to the computational efficiency of the drain current model for double-gate MOSFETs is extended, and different calculation methods are compared and discussed. The results show that the calculation speed of the improved model is substantially enhanced. A two-dimensional device simulation is performed to verify the improved model. Furthermore, the model is implemented into the HSPICE circuit simulator in Verilog-A for practical application.

An analytic model for gate-all-around silicon nanowire tunneling field effect transistors

An analytical model of gate-all-around （GAA） silicon nanowire tunneling field effect transistors （NW-TFETs） is developted based on the surface potential solutions in the channel direction and considering the band to band tunneling （BTBT） efficiency. The three-dimensional Poisson equation is solved to obtain the surface potential distributions in the partition regions along the channel direction for the NW-TFET, and a tunneling current model using Kane＇s expression is developed. The validity of the developed model is shown by the good agreement between the model predictions and the TCAD simulation results.

Capacitive Model and S-Parameters of Double-Pole Four-Throw Double-Gate RF CMOS Switch

In this paper, we have analyzed the Double-Pole Four-Throw Double-Gate Radio-Frequency Complementary Metal-Oxide-Semiconductor (DP4T DG RF CMOS) switch using S-parameters for 1 GHz to 60 GHz of frequency range. DP4T DG RF CMOS switch for operation at high frequency is also analyzed with its capacitive model. The re-sults for the development of this proposed switch include the basics of the circuit elements in terms of capacitance, re-sistance, impedance, admittance, series equivalent and parallel equivalent of this network at different frequencies which are present in this switch whatever they are ON or OFF.

基于CNN-GRU-ISSA-XGBoost的短期光伏功率预测

针对光伏功率随机性及波动性大,单一预测模型往往难以准确分析历史数据波动规律,从而导致预测精度不高的问题,提出一种基于卷积神经网络-门控循环单元(CNN-GRU)和改进麻雀搜索算法(ISSA)优化的极限梯度提升(XGBoost)模型的短期光伏功率预测组合模型.首先去除历史数据中的异常值并对其进行归一化处理,利用主成分分析法(PCA)进行特征选取,以便更好地识别影响光伏功率的关键因素.然后采用CNN网络提取数据的空间特征,再经过GRU网络提取时间特征,针对XGBoost模型手动配置参数困难、随机性大的问题,利用ISSA对模型超参数寻优.最后对两种方法预测的结果用误差倒数法减小误差的同时对权重进行更新,得到新的预测值,从而完成对光伏功率的预测.实验结果表明,所提出的CNN-GRU-ISSA-XGBoost组合模型具有更强的适应性和更高的精度.

基于ICEEMDAN和时变权重集成预测模型的变压器油中溶解气体含量预测

为了实现对变压器油中溶解气体体积分数的精确预测,同时克服仅使用单一预测模型导致预测精度及泛化能力不足的局限,提出了一种基于改进完全自适应噪声集合经验模态分解(improved complete ensemble empirical mode decomposition,ICEEMDAN)和灰色关联系数时变权重集成预测模型的变压器油中溶解气体预测方法。首先将溶解气体含量序列模态分解为一系列具有不同时间尺度的子序列。然后,使用门控循环神经网络和麻雀搜索算法优化支持向量机对各子序列进行训练,组合为一个集成预测模型;并比较不同预测方法的预测精度,计算灰色关联系数时变权重,形成各子系列的预测结果。最后将各子序列的预测结果叠加重构,得到最终预测结果。算例分析结果显示:该方法单步预测的均方根误差、平均绝对误差和相关系数分别为0.593、0.422和0.768,相比其他算法在预测精度上有明显提升,同时具有很强的泛化性能,可以为油浸式变压器内部状态监测提供依据。

基于模态分解及GRU-XGBoost短期电力负荷预测

精确的短期电力负荷预测能有效提高电力系统运营水平。针对电力负荷数据受多种因素影响,波动性和随机性强等问题,提出了一种基于模态分解及混合模型的负荷预测方法。首先,采用主成分分析法(principal component analysis,PCA)对负荷特征向量进行处理,去掉冗余信息,再用完全自适应噪声集合经验模态分解(complete ensemble empirical mode decomposition with adaptive noise,CEEMDAN)将历史负荷分解为简化的几个子序列;其次,选择引入样本熵(sample entropy,SE)来计算子序列熵值,将相近的子序列重构得到随机、细节、低频和趋势分量后选用不同结构门控循环单元(gate recurrent unit,GRU)对不同分量类型进行预测,再使用极致梯度提升模型(extreme gradient boosting,XGBoost)对各分量残差进行拟合,各重组序列的预测值为GRU预测值与XBGoost拟合值之和,重组各序列得到最终预测值。选取3年时电力负荷数据进行实验,结果表明,所提模型的均方根误差(root mean square error,RMSE)、平均绝对百分比误差(mean absolutepercentage error,MAPE)和平均绝对误差(mean absolute error,MAE)分别为370.676 MW、99.07%和246.89 MW,与单一模型和混合模型相比,实现了评价指标的明显减少。