Coseismic effects recorded by Fujian subsurface fluid network and its meaning to earthquake prediction

Based on analyses of the spatio-temporal evolutionary characteristics of teleseismic response recorded by Fujian subsurface fluid network and in combination with earthquakes happened in Fujian province during the same period, this paper points out that the step-like rising of water level after distant earthquakes may include some regional stress field information, and the area where water level step-like rises could be the position that the stress concentrated on and where the future earthquakes would occur. If combined with other impending precursors, the location of the events may be predicted to a certain degree.

Retrofit design of composite cooling structure of a turbine blade by fluid networks and conjugate heat transfer methods

This article discusses the development of the numerical methods of gas flow coupled with heat transfer,and introduces the fluid net-works method for rapid prediction of the performance of the composite cooling structures in turbine blade.The reliability of these methods is verified by comparing experimental data.For a HPT rotor blade,a rapid prediction on the internal cooling structures is first made by using the fluid network analysis,then an assessment of aerodynamic and heat transfer characteristics is conducted.Based on the network analysis results,three ways to improve the design of the cooling structures are tested,i.e.,adjusting the cooling gas flow mass ratios for different inner cooling cavities,reducing the flow resistances of the channel turning structures,and improving the local internal cooling structure geometries with high temperature distribution.Through the verification of full three-dimensional gas/solid/coolant conjugate heat transfer calculation,we conclude that the modified design can make the overall temperature distribution more even by significantly reducing the highest temperature of the blade surface,and reasonably matching the parameters of different coolant inlets.The results show that the proposed calculation methods can remarkably reduce the design cycle of complex turbine blade cooling structure.

A New Algorithm of Auto-Modelling for Fluid Network

The analysis of the solution of fluid network model is carried out to match the need of graphicallymodular autor-modelling for power plant simulators. Because of the symmetry and sparsity of thelinear system of equations, a new method of improved Gauss elimination is presented for the solutionof large scale sparse matrices. Comparison of the new method with the classical Gauss eliminationmethod, the Gauss-Seidel iterative method are given. The results show that the algorithm provided isbetter than the others and is suitable for auto-modelling of fluid networks of power plants.

Synthesis of the fluid machinery network in a circulating water system

Energy consumption of the fluid machinery network in a circulating water system takes up a large part of energy consumption in the process industry, so optimization on the network will enhance the economic and environmental performance of the industry. In this paper, a synthesis approach is proposed to obtain the optimal network structure. The effective height curves are used as tools to perform energy analysis, so that the potential placement of water turbines and auxiliary pumps can be determined with energy benefit. Then economic optimization is carried out, by the mathematical model with the total cost as the objective function, to identify the branches for water turbines and auxiliary pumps with economic benefit. In this way, the optimal fluid machinery network structure can be obtained. The results of case study indicate that the proposed synthesis approach to optimize the fluid machinery network will obtain more remarkable benefits on economy, compared to optimizing only the water turbine network or pump network. The results under different flowrates of circulating water reveal that using a water turbine to recover power or adding an auxiliary pump to save energy in branches are only suitable to the flowrate in a certain range.

Dynamic Characteristics Analysis on MHTGR Plant’s Secondary Side Fluid Flow Network

Multipe NSSS (Nuclear Steam Supply System) modules use the common feeding-water system to drive the common turbine power generation set. The SSFFN (secondary side fluid flow network) of MHTGR plant has features i.e. strong-coupling and nonlinearity. A wide range of power switching operation will cause unsteady flow, which may destroy the working elements and will be a threat for normal operation. To overcome those problems, a differential-algebraic model and PI controllers are designed for the SSFFN. In MATLAB\SIMULINK environment, a simulation platform is established and used to make a simulation of SSFFN of a MHTGR plant with two NSSS modules, which uses feedwater valves to control the mass flow rate in each module instead of feedwater pump. Results reflect good robustness of controllers.

Dynamic Shortest Path Algorithm in Stochastic Traffic Networks Using PSO Based on Fluid Neural Network

The shortest path planning issure is critical for dynamic traffic assignment and route guidance in intelligent transportation systems. In this paper, a Particle Swarm Optimization (PSO) algorithm with priority-based encoding scheme based on fluid neural network (FNN) to search for the shortest path in stochastic traffic networks is introduced. The proposed algorithm overcomes the weight coefficient symmetry restrictions of the traditional FNN and disadvantage of easily getting into a local optimum for PSO. Simulation experiments have been carried out on different traffic network topologies consisting of 15-65 nodes and the results showed that the proposed approach can find the optimal path and closer sub-optimal paths with good success ratio. At the same time, the algorithms greatly improve the convergence efficiency of fluid neuron network.

Segmentation of retinal fluid based on deep learning:application of three-dimensional fully convolutional neural networks in optical coherence tomography images

AIM: To explore a segmentation algorithm based on deep learning to achieve accurate diagnosis and treatment of patients with retinal fluid.METHODS: A two-dimensional(2D) fully convolutional network for retinal segmentation was employed. In order to solve the category imbalance in retinal optical coherence tomography(OCT) images, the network parameters and loss function based on the 2D fully convolutional network were modified. For this network, the correlations of corresponding positions among adjacent images in space are ignored. Thus, we proposed a three-dimensional(3D) fully convolutional network for segmentation in the retinal OCT images.RESULTS: The algorithm was evaluated according to segmentation accuracy, Kappa coefficient, and F1 score. For the 3D fully convolutional network proposed in this paper, the overall segmentation accuracy rate is 99.56%, Kappa coefficient is 98.47%, and F1 score of retinal fluid is 95.50%. CONCLUSION: The OCT image segmentation algorithm based on deep learning is primarily founded on the 2D convolutional network. The 3D network architecture proposed in this paper reduces the influence of category imbalance, realizes end-to-end segmentation of volume images, and achieves optimal segmentation results. The segmentation maps are practically the same as the manual annotations of doctors, and can provide doctors with more accurate diagnostic data.

基于仿真数据驱动的激光钻进气体喷嘴结构优化

激光钻进采用气体作为循环介质进行清孔,合理的气体流动特性是高效清孔的保障,气体喷嘴是影响气体流动特性的直接因素,其结构设计不合理会严重影响激光钻进的效率。针对激光钻进实验平台中的气体喷嘴,构建喷嘴基本型态,对影响气体清孔效率的喷嘴结构尺寸进行分析,制定仿真方案,通过Fluent模拟气体流场,对清孔效果进行分析,采用神经网络分析喷嘴结构及仿真结果,训练神经网络模型,得出最佳清孔效率时的喷嘴结构参数并进行验证,为喷嘴结构设计提供参考。

PARAMETERS INVERSION OF FLUID-SATURATED POROUS MEDIA BASED ON NEURAL NETWORKS

The multi- layers feedforward neural network is used for inversion ofmaterial constants of fluid-saturated porous media. The direct analysis of fluid-saturated porousmedia is carried out with the boundary element method. The dynamic displacement responses obtainedfrom direct analysis for prescribed material parameters constitute the sample sets training neuralnetwork. By virtue of the effective L-M training algorithm and the Tikhonov regularization method aswell as the GCV method for an appropriate selection of regu-larization parameter, the inversemapping from dynamic displacement responses to material constants is performed. Numerical examplesdemonstrate the validity of the neural network method.

汽轮发电机带有交替径向风道的转子流体与传热耦合分析

针对汽轮发电机带有交替径向通风道的转子发热冷却问题,以一台350 MW水氢氢冷汽轮发电机为研究对象,依据流体力学和传热学的基本理论,首先建立计及旋转的电机全域通风网络模型,采用逐次迭代法计算得到各支路流量和节点压力。其次,建立了带有交替径向风道的发电机转子流体-传热三维物理模型和数学模型,给出了基本假设和相应的边界条件,同时将通风网络计算得到的风速和压力作为转子求解域的耦合边界,采用有限体积法进行求解,计算结果与实测值吻合。然后分析了交替径向风道内流量分配和氢气流动情况,研究了转子内部氢气温度分布和槽楔出风口风温变化规律,探明了转子绕组和铁心轴向温度分布特性,讨论了副槽入口流量和槽楔出口直径对转子流体和温度的影响。得出副槽入口流量应控制在0.1~0.16 m^(3)/s范围内,且选择较小的槽楔出口直径,可以提高通风系统的效率与风量分配均匀性,降低转子轴向热不平衡。

编织知识网络提升《化工原理》学习效果

化工原理是化工类专业的基础课程,是学生在理科基础上建立工程思想的关键课程。该门课程的学习存在知识点多且杂、难度大、不好学等问题,文章提出编织单元知识网络的方法,帮助学生提高学习效率,拓展自学能力。让学生在掌握课程知识的同时,也获得提高效率的学习方法。

鲁南地区流体监测台网优化与监测效能评估

结合鲁南地区地质构造背景,通过自主研发的高精度水位水温实时监测仪器,对地下流体监测站点软硬件设施进行升级改造,建设鲁南地区地下流体数字化监测台网。同时,构建基于层次分析法的鲁南地区地下流体监测效能评价体系,利用模糊综合评价法对各观测站点进行量化的效能评估,为进一步提升该地区地下流体监测台网监测能力提供参考意见。

面向国产超算平台的通用能源管网仿真计算模型

为实现城市能源管网仿真软件的自主可控,基于国产异构高性能计算机“嵩山”超级计算平台,提出一种通用的城市能源管网仿真计算模型。通过优化管网中“非管”组件模型,提高了计算模型对国产异构并行计算机系统的适配性;将不同管网组件的计算过程进行封装,弱化了网络组件在仿真计算过程的依赖性,提升模型在工程实现上的可并行性。供水、燃气和热力三种场景的并行仿真实验,证明了计算模型在解决城市能源供给网络的仿真计算上具有一定的普适性;通过管网实测数据与仿真模型中模拟数据对比结果表明仿真管网压力的误差率在4%以下,其温度的误差率低于2%,同时也说明了提出的管网仿真计算模型在国产超算平台上具有良好的计算通用性。

Intelligent Recognition Method of Insufficient Fluid Supply of Oil Well Based on Convolutional Neural Network

Traditional methods for judging the degree of insufficient fluid supply in oil wells have low efficiency and limited accuracy. To address this problem, a method for intelligently identifying the degree of insufficient fluid supply in oil wells based on convolutional neural networks is proposed in this paper. Firstly, 5000 indicator diagrams with insufficient liquid supply were collected from the oilfield site, and a sample set was established after preprocessing;then based on the AlexNet model, combined with the characteristics of the indicator diagram, a convolutional neural network model including 4 layers of convolutional layers, 3 layers of down-pooling layers and 2 layers of fully connected layers is established. The backpropagation, ReLu activation function and dropout regularization method are used to complete the training of the convolutional neural network;finally, the performance of the convolutional neural network under different iteration times and network structure is compared, and the super parameter optimization of the model is completed. It has laid a good foundation for realizing the self-adaptive and intelligent matching of oil well production parameters and formation fluid supply conditions. It has certain application prospects. The results show that the accuracy of training and verification of the method exceeds 98%, which can meet the actual application requirements on site.

A new matrix-based mathematical model for determining unidirectional circuits in a ventilation network

The occurrence of local circulating ventilation can be caused by many factors, such as the airflow reversion during mine fire,the improper arrangement of local fan or underground fan station and the man-made error input of raw data before network solving. Once circulating ventilations occur,the corresponding branches in the ventilation network corresponding to the relevant airways in ventilation system form circuits,and all the direc- tions of the branches in the circuits are identical,which is the unidirectional problem in ventilation network.Based on the properties of node adjacent matrix,a serial of mathe- matical computation to node adjacent matrix were performed,and a mathematical model for determining unidirectional circuits based on node adjacent matrix was put forward.

局部内冷空气系统流体网络分析与数字化计算方法

燃气轮机内冷空气系统结构、气路流动与运行工况极为复杂,造成了内冷空气系统流体计算难度很大,传统一维流体网络法的计算精度已经不能满足当代发动机分析与设计的需求,急需开发一种面向燃气轮机内冷空气系统的高精度、快速的系统级别的流体计算方法。本文面向某船用燃气轮机高压转子内冷空气系统的卸荷腔,创新性地利用一维-三维联合仿真计算方法,实现了快速求解流体网络节点信息的同时解析关键区域流动传热细节。研究结果表明,与传统一维流体网络法相比,一维-三维耦合计算可以通过边界数据传递与耦合计算提高整体计算精度,与传统CFD计算方法相比,一维-三维耦合计算可以加快计算收敛速度,以卸荷腔高压涡轮腔室为例,一维-三维联合仿真计算方法可以节约计算收敛时间约83%,同时三维流体计算域可以获得关键区域的流场细节,对系统级别计算下掌握关键区域流动特性具有重要意义。

Application of Stochastic Fracture Network with Numerical Fluid Flow Simulations to Groundwater Flow Modeling in Fractured Rocks

The continuum approach in fluid flow modeling is generally applied to porous geological media, but has limited applicability to fractured rocks. With the presence of a discrete fracture network relatively sparsely distributed in the matrix, it may be difficult or erroneous to use a porous medium fluid flow model with continuum assumptions to describe the fluid flow in fractured rocks at small or even large field scales. A discrete fracture fluid flow approach incorporating a stochastic fracture network with numerical fluid flow simulations could have the capability of capturing fluid flow behaviors such as inhomogeneity and anisotropy while reflecting the changes of hydraulic features at different scales. Moreover, this approach can be implemented to estimate the size of the representative elementary volume (REV) in order to find out the scales at which a porous medium flow model could be applied, and then to determine the hydraulic conductivity tensor for fractured rocks. The following topics are focused on in this study: (a) conceptual discrete fracture fluid flow modeling incorporating a stochastic fracture network with numerical flow simulations; (b) estimation of REV and hydraulic conductivity tensor for fractured rocks utilizing a stochastic fracture network with numerical fluid flow simulations; (c) investigation of the effect of fracture orientation and density on the hydraulic conductivity and REV by implementing a stochastic fracture network with numerical fluid flow simulations, and (d) fluid flow conceptual models accounting for major and minor fractures in the 2 D or 3 D flow fields incorporating a stochastic fracture network with numerical fluid flow simulations.

基于卷积神经网络的气动热预测方法

严重的气动加热现象会威胁飞行器的飞行安全,因此在飞行器设计期间需要对其气动热环境进行预测以辅助热防护设计,气动热的预测速度直接影响了飞行器的设计效率。为了探索气动热的快速预测方法以缩短高超声速飞行器的设计周期,本文基于卷积神经网络建立了数据驱动的气动热预测模型。首先,为了实现不同外形飞行器的表面热流预测,提出了一种能够用于卷积神经网络的三维外形几何表达方法。然后,基于该方法分别采用编码器-解码器架构和U-Net架构建立了两种神经网络模型,实现了气动热的快速预测。最后,选取钝锥、钝双锥、升力体和双椭球4类高超声速飞行器典型外形作为研究对象,采用CFD数值模拟方法构建了气动热数据集,在不同的气动热数据集上对建立的模型进行了训练和测试。计算验证结果表明,两种模型针对简单外形预测精度良好,但当外形变复杂时,U-Net模型对外形的感知能力更强,具有更高的预测精度。与其他数据驱动的方法相比,U-Net模型具有更强的学习能力,能够在较少的训练样本下达到相对较高的预测精度。另一方面,由于该方法采用了大量卷积神经网络结构,因此具有更高的建模效率。

基于电网络理论和计算流体力学的血流仿真

人体的上肢动脉包括肱动脉和桡动脉,两段动脉也经常被作为测得人体血压及脉搏波的关键部位,反映出了丰富的生理病理信息,对此进行研究对心血管疾病的诊断具有着重要意义。对比结合了基于弹性腔的电网络模型法和双向流固耦合计算流体力学模型法,分别建立了在相同结构和力学参数下的两种仿真模型,通过两种方法在不同尺度下的优势互补,对人体在健康及动脉粥样硬化状态下的上肢动脉段的血流动力学问题进行了研究。一方面通过电网络模型对动脉脉搏波进行了动态仿真,发现在动脉粥样硬化下脉搏波的主波波峰出现了明显的升高和前滞现象,另一方面通过双向流固耦合计算流体力学法分析了血液流场的局部特性,从而为粥样硬化等心血管疾病的形成机理及预测提供参考。

航天器流体回路泵智能故障诊断方法

开展了航天器流体回路泵智能故障诊断方法研究,将BP神经网络、粒子群优化(PSO)-BP神经网络、遗传算法(GA)-BP神经网络和模糊神经网络4种模型应用于故障诊断,以温度、流量、出口压力、转速4类状态参数作为神经网络输入,状态标志作为输出进行训练,通过均方误差、相关系数、模型训练评分对模型训练效果评价,从而实现模型优选,完成故障诊断。利用在轨遥测数据进行应用验证,结果表明可以准确识别流体回路泵的正常和叶轮卡死泵功能丧失两种在轨实际状态类型,且模糊神经网络相对其他3种神经网络具有更好的诊断效果。