A sub-grid scale model for Burgers turbulence based on the artificial neural network method

The present study proposes a sub-grid scale model for the one-dimensional Burgers turbulence based on the neuralnetwork and deep learning method.The filtered data of the direct numerical simulation is used to establish thetraining data set,the validation data set,and the test data set.The artificial neural network(ANN)methodand Back Propagation method are employed to train parameters in the ANN.The developed ANN is applied toconstruct the sub-grid scale model for the large eddy simulation of the Burgers turbulence in the one-dimensionalspace.The proposed model well predicts the time correlation and the space correlation of the Burgers turbulence.

Remote Control and Telemetry System for Large-scale Model Test at Sea

Physical testing of large-scale ship models at sea is a new experimental method.It is a cheap and reliable way to research the environment adaptability of a ship in complex and extreme wave conditions.It is necessary to have a stable experimental system for the test.Since the experimental area is large, a remote control system and a telemetry system are essential, and were designed by the authors.An experiment was conducted on the Songhuajiang River to test the systems.The relationship between the model's speed and its electromotor's revolutions was also measured during the model test.The results showed that the two systems make it possible to carry out large-scale model tests at sea.

Evaluation of Subgrid-scale Models in Large-eddy Simulations of Turbulent Flow in a Centrifugal Pump Impeller

The current research of large eddy simulation （LES） of turbulent flow in pumps mainly concentrates in applying conventional subgrid-scale （SGS） model to simulate turbulent flow, which aims at obtaining the flow field in pump. The selection of SGS model is usually not considered seriously, so the accuracy and efficiency of the simulation cannot be ensured. Three SGS models including Smagorinsky-Lilly model, dynamic Smagorinsky model and dynamic mixed model are comparably studied by using the commercial CFD code Fluent combined with its user define function. The simulations are performed for the turbulent flow in a centrifugal pump impeller. The simulation results indicate that the mean flows predicted by the three SGS models agree well with the experimental data obtained from the test that detailed measurements of the flow inside the rotating passages of a six-bladed shrouded centrifugal pump impeller performed using particle image velocimetry （PIV） and laser Doppler velocimetry （LDV）. The comparable results show that dynamic mixed model gives the most accurate results for mean flow in the centrifugal pump impeller. The SGS stress of dynamic mixed model is decompose into the scale similar part and the eddy viscous part. The scale similar part of SGS stress plays a significant role in high curvature regions, such as the leading edge and training edge of pump blade. It is also found that the dynamic mixed model is more adaptive to compute turbulence in the pump impeller. The research results presented is useful to improve the computational accuracy and efficiency of LES for centrifugal pumps, and provide important reference for carrying out simulation in similar fluid machineries.

Validating the Runoff from the PRECIS Model Using a Large-Scale Routing Model

The streamflow over the Yellow River basin is simulated using the PRECIS （Providing REgional Climates for Impacts Studies） regional climate model driven by 15-year （1979-1993） ECMWF reanalysis data as the initial and lateral boundary conditions and an off-line large-scale routing model （LRM）. The LRM uses physical catchment and river channel information and allows streamflow to be predicted for large continental rivers with a 1°×1° spatial resolution. The results show that the PRECIS model can reproduce the general southeast to northwest gradient distribution of the precipitation over the Yellow River basin, The PRECIS- LRM model combination has the capability to simulate the seasonal and annual streamflow over the Yellow River basin. The simulated streamflow is generally coincident with the naturalized streamflow both in timing and in magnitude.

Long Short-Term Memory Recurrent Neural Network-Based Acoustic Model Using Connectionist Temporal Classification on a Large-Scale Training Corpus

A Long Short-Term Memory(LSTM) Recurrent Neural Network(RNN) has driven tremendous improvements on an acoustic model based on Gaussian Mixture Model(GMM). However, these models based on a hybrid method require a forced aligned Hidden Markov Model(HMM) state sequence obtained from the GMM-based acoustic model. Therefore, it requires a long computation time for training both the GMM-based acoustic model and a deep learning-based acoustic model. In order to solve this problem, an acoustic model using CTC algorithm is proposed. CTC algorithm does not require the GMM-based acoustic model because it does not use the forced aligned HMM state sequence. However, previous works on a LSTM RNN-based acoustic model using CTC used a small-scale training corpus. In this paper, the LSTM RNN-based acoustic model using CTC is trained on a large-scale training corpus and its performance is evaluated. The implemented acoustic model has a performance of 6.18% and 15.01% in terms of Word Error Rate(WER) for clean speech and noisy speech, respectively. This is similar to a performance of the acoustic model based on the hybrid method.

Two-Dimensional Modeling of the NAPL Dissolution in Porous Media: Heterogeneities Effects on the Large Scale Permeabilities and Mass Exchange Coefficient

In this paper, we are interested by the dissolution of NAPL (Non-Aqueous Phase Liquid) contaminants in heterogeneous soils or aquifers. The volume averaging technique is applied to 2D systems with Darcy-scale heterogeneities. A large-scale model is derived from a Darcy-scale dissolution model in the case of small and large Damkholer numbers, i.e., for smooth or sharp dissolution fronts. The resulting models in both cases have the mathematical structure of a non-equilibrium dissolution model. It is shown how to calculate the resulting mass exchange and relative permeability terms from the Darcy-scale heterogeneities and other fluid properties. One of the important finding is that the obtained values have a very different behavior compared to the Darcy-scale usual correlations. The large scale correlations are also very different between the two limit cases. The resulting large-scale models are compared favorably to Darcy-scale direct simulations.

Advances of Model Order Reduction Research in Large-scale System Simulation

Model Order Reduction (MOR) plays more and more imp or tant role in complex system simulation, design and control recently. For example , for the large-size space structures, VLSI and MEMS (Micro-ElectroMechanical Systems) etc., in order to shorten the development cost, increase the system co ntrolling accuracy and reduce the complexity of controllers, the reduced order model must be constructed. Even in Virtual Reality (VR), the simulation and d isplay must be in real-time, the model order must be reduced too. The recent advances of MOR research are overviewed in the article. The MOR theor y and methods may be classified as Singular Value decomposition (SVD) based, the Krylov subspace based and others. The merits and demerits of the different meth ods are analyzed, and the existed problems are pointed out. Moreover, the applic ation’s fields are overviewed, and the potential applications are forecaste d. After the existed problems analyzed, the future work is described. There are som e problems in the traditional methods such as SVD and Krylov subspace, they are that it’s difficult to (1)guarantee the stability of the original system, (2) b e adaptive to nonlinear system, and (3) control the modeling accuracy. The f uture works may be solving the above problems on the foundation of the tradition al methods, and applying other methods such as wavelet or signal compression.

A new mixed subgrid-scale model for large eddy simulation of turbulent drag-reducing flows of viscoelastic fluids

A mixed subgrid-scale（SGS） model based on coherent structures and temporal approximate deconvolution（MCT） is proposed for turbulent drag-reducing flows of viscoelastic fluids. The main idea of the MCT SGS model is to perform spatial filtering for the momentum equation and temporal filtering for the conformation tensor transport equation of turbulent flow of viscoelastic fluid, respectively. The MCT model is suitable for large eddy simulation（LES） of turbulent dragreducing flows of viscoelastic fluids in engineering applications since the model parameters can be easily obtained. The LES of forced homogeneous isotropic turbulence（FHIT） with polymer additives and turbulent channel flow with surfactant additives based on MCT SGS model shows excellent agreements with direct numerical simulation（DNS） results. Compared with the LES results using the temporal approximate deconvolution model（TADM） for FHIT with polymer additives, this mixed SGS model MCT behaves better, regarding the enhancement of calculating parameters such as the Reynolds number.For scientific and engineering research, turbulent flows at high Reynolds numbers are expected, so the MCT model can be a more suitable model for the LES of turbulent drag-reducing flows of viscoelastic fluid with polymer or surfactant additives.

Sample-data Decentralized Reliable H∞ Hyperbolic Control for Uncertain Fuzzy Large-scale Systems with Time-varying Delay

这份报纸学习样品数据的问题为有变化时间的延期的不明确的连续时间的模糊大规模系统的可靠 H 夸张控制。第一，模糊夸张模型( FHM )被用来为某些复杂大规模系统建立模型，然后根据 Lyapunov 指导方法和大规模系统的分散的控制理论，线性 matrixine 质量( LMI )基于条件 arederived toguarantee H 性能不仅当所有控制部件正在操作很好时，而且面对一些可能的致动器失败。而且，致动器的精确失败参数没被要求，并且要求仅仅是失败参数的更低、上面的界限。条件依赖于时间延期的上面的界限，并且不依赖于变化时间的延期的衍生物。因此，获得的结果是不太保守的。最后，二个例子被提供说明设计过程和它的有效性。

Decentralized Stabilization of Large-scale Uncertain Systems with State-Delays——LMIs Approach

This paper deals with the problem of decentralized robustcontrol for a class of interconnected uncertain systemswith state delays.The parameter uncertainties are un-known but norm-bounded.A new sufficient condition isobtained for each subsystem and overall system to be sta-bilizable via linear memoryless state feedback robust de-centralized controllers.The results depend on the size of the delays and are given in terms of linear matrix ine-qualities,so they are less conservative than those of delay-independent.Moreover,matching condition is not a necessary condition.Finally,an example is presented to illustrative the effectiveness of the proposed method.

Large Eddy Simulation of Stratified Turbulent Channel Flow with a Dynamic Subgrid Scale Model

In this paper, large eddy simulation coupled with a dynamic subgrid scale (SGS) model is used to study turbulent channel flows with heat transfer. Some fundamental flow behaviors are discussed. Based on our calculated results, the dynamic SGS model can reasonably predict some main characteristics of stratified turbulent channel flows. Our results are also in good agreement with theoretical analyses and previous calculated results.

Utilization Survey of Livestock Manure Resources in Large-scale Farms of Yangzhou

Based on surveying the conditions of large -scale farms and commercial manure in the each county of Yangzhou city, the situations and problems for utilization of livestock manure resources were grasped. After an analysis of the potential value of livestock manure, the suggestion and strategy for utilization of livestock manure resources were proposed based on the actual conditions in Yangzhou city.

Trend Analysis of Large-Scale Twitter Data Based on Witnesses during a Hazardous Event: A Case Study on California Wildfire Evacuation

Social media data created a paradigm shift in assessing situational awareness during a natural disaster or emergencies such as wildfire, hurricane, tropical storm etc. Twitter as an emerging data source is an effective and innovative digital platform to observe trend from social media users’ perspective who are direct or indirect witnesses of the calamitous event. This paper aims to collect and analyze twitter data related to the recent wildfire in California to perform a trend analysis by classifying firsthand and credible information from Twitter users. This work investigates tweets on the recent wildfire in California and classifies them based on witnesses into two types: 1) direct witnesses and 2) indirect witnesses. The collected and analyzed information can be useful for law enforcement agencies and humanitarian organizations for communication and verification of the situational awareness during wildfire hazards. Trend analysis is an aggregated approach that includes sentimental analysis and topic modeling performed through domain-expert manual annotation and machine learning. Trend analysis ultimately builds a fine-grained analysis to assess evacuation routes and provide valuable information to the firsthand emergency responders.

Stability Analysis of Continuous-Time Fuzzy Large-Scale System

A continuous-time fuzzy large-scale system F consists of some interconnected Takagi-Sugeno fuzzy subsystems. Two sufficient conditions for the asymptotic stability of this system (namely, theorem 1 and theorem 2) are derived via a multiple Lyapunov function approach. In theorem 1, the information of membership functions of fuzzy rules should be known in order to analyze the stability of F. But in general this information is not easy to be acquired for their time-varying property. So theorem 2 is provided to judge the asymptotic stability of F, based on which there is no need to know the information of membership functions in stability analysis. Finally, a numerical example is given to show the utility of the method proposed in this paper.

Robust non-fragile control for discrete-delay nonlinear large-scale systems

Time-delays,due to the information transmission between subsystems,naturally exist in large-scale systems and the existence of the delay is frequently a source of instability. This paper considers the problems of robust non-fragile fuzzy control for a class of uncertain discrete nonlinear large-scale systems with time-delay and controller gain perturbations described by T-S fuzzy model. An equivalent T-S fuzzy model is represented for discrete-delay nonlinear large-scale systems. A sufficient condition for the existence of such non-fragile controllers is further derived via the Lyapunov function and the linear matrix inequality( LMI) approach. Simulation results demonstrate the feasibility and the effectiveness of the proposed design and the proper stabilization of the system in spite of controller gain variations and uncertainties.

Modified Shifting Bottleneck Heuristic for Scheduling Problems of Large-Scale Job Shops

A modified shifting bottleneck algorithm was proposed to solve scheduling problems of a large-scale job shop.Firstly,a new structured algorithm was employed for sub-problems so as to reduce the computational burden and suit for large-scale instances more effectively.The modified cycle avoidance method,incorporating with the disjunctive graph model and topological sort algorithm,was applied to guaranteeing the feasibility of solutions with considering delayed precedence constraints.Finally,simulation experiments were carried out to verify the feasibility and effectiveness of the modified method.The results demonstrate that the proposed algorithm can solve the large-scale job shop scheduling problems(JSSPs) within a reasonable period of time and obtaining satisfactory solutions simultaneously.

Large Eddy Simulations of Flow Instabilities in a Stirred Tank Gen-erated by a Rushton Turbine

The aim of this work is to investigate the flow instabilities in a baffled, stirred tank generated by a single Rushton turbine by means of large eddy simulation （LES）. The sliding mesh method was used for the coupling between the rotating and the stationary frame of references. The calculations were carried out on the ＂Shengcao-21C＂ supercomputer using a computational fluid dynamics （CFD） code CFX5. The flow fields predicted by the LES simulation and the simulation using standard κ-ε model were compared to the results from particle image velocimetry （PIV） measurements. It is shown that the CFD simulations using the LES approach and the standard κ-ε model agree well with the PIV measurements. Fluctuations of the radial and axial velocity are predicted at different frequencies by the LES simulation. Velocity fluctuations of high frequencies are seen in the impeller region, while low frequencies velocity fluctuations are observed in the bulk flow. A low frequency velocity fluctuation with a nondimensional frequency of 0.027Hz is predicted by the LES simulation, which agrees with experimental investigations in the literature. Flow circulation patterns predicted by the LES simulation are asymmetric, stochastic and complex, spanning a large portion of the tanks and varying with time, while circulation patterns calculated by the simulation using the standard κ-ε model are symmetric. The results of the present work give better understanding to the flow instabilities in the mechanically agitated tank. However, further analysis of the LES calculated velocity series by means of fast Fourier transform （FFT） and/or spectra analysis are recommended in future work in order to gain more knowledge of the complicated flow phenomena.

Formation conditions and exploration direction of large natural gas reservoirs in the oil-prone Bohai Bay Basin, East China

The Bohai Bay Basin is a typical oil-prone basin, in which natural gas geological reserves have a small proportion. In this basin, the gas source rock is largely medium-deep lake mudstone with oil-prone type Ⅱ2-Ⅱ1 kerogens, and natural gas preservation conditions are poor due to active late tectonic movements. The formation conditions of large natural gas fields in the Bohai Bay Basin have been elusive. Based on the exploration results of Bohai Bay Basin and comparison with large gas fields in China and abroad, the formation conditions of conventional large-scale natural gas reservoirs in the Bohai Bay Basin were examined from accumulation dynamics, structure and sedimentation. The results show that the formation conditions of conventional large natural gas reservoirs in Bohai Bay Basin mainly include one core element and two key elements. The core factor is the strong sealing of Paleogene "quilt-like" overpressure mudstone. The two key factors include the rapid maturation and high-intensity gas generation of source rock in the late stage and large scale reservoir. On this basis, large-scale nature gas accumulation models in the Bohai Bay Basin have been worked out, including regional overpressure mudstone enriching model, local overpressure mudstone depleting model, sand-rich sedimentary subsag depleting model and late strongly-developed fault depleting model. It is found that Bozhong sag, northern Liaozhong sag and Banqiao sag have favorable conditions for the formation of large-scale natural gas reservoirs, and are worth exploring. The study results have important guidance for exploration of large scale natural gas reservoirs in the Bohai Bay Basin.

An improved large eddy simulation of two-phase flows in a pump impeller

An improved large eddy simulation using a dynamic second-order sub-grid-scale （SGS） stress model has been developed to model the governing equations of dense turbulent particle-liquid two-phase flows in a rotating coordinate system, and continuity is conserved by a mass-weighted method to solve the filtered governing equations. In the cur- rent second-order SGS model, the SGS stress is a function of both the resolved strain-rate and rotation-rate tensors, and the model parameters are obtained from the dimensional consistency and the invariants of the strain-rate and the rotation-rate tensors. In the numerical calculation, the finite volume method is used to discretize the governing equations with a staggered grid system. The SIMPLEC algorithm is applied for the solution of the discretized governing equations. Body- fitted coordinates are used to simulate the two-phase flows in complex geometries. Finally the second-order dynamic SGS model is successfully applied to simulate the dense turbu-lent particle-liquid two-phase flows in a centrifugal impeller. The predicted pressure and velocity distributions are in good agreement with experimental results.

PCP-tuning:面向小样本学习的个性化连续提示调优

随着“提示学习”的兴起,预训练语言模型在少样本学习中取得了显著的表现,其中的关键问题是如何为每个训练样本构建合适的提示.近年来研究人员提出了一系列提示构造方法,有的构造离散型的提示,有的构造连续型的提示,但通常都是将一个提示应用到整个数据集上.然而,实验结果表明,很难找到一个能够适用于任务中所有样本的提示.为此,提出了一种用于小样本学习的个性化连续型提示调优方法(PCP-tuning),其目的是根据数据集中每个样本的语义来生成个性化的连续型提示.同时,还提出了两种校准技术来控制生成的连续型提示的分布,以获得更好的下游任务表现.最后在10个基准任务上进行大量实验,证明了新方法的优越性能.