Experimental Study on Jet Blast at Laboratory Scale

The flow field of 3D （three-dimensional） wall-jet is investigated. Jet-blast from airplane is simulated by wall-jet setup using a sonic nozzle at a laboratory scale. Farfield velocity and fluctuation distributions are measured by using X-type hot wire anemometer at four measurement planes. As a result, the flow properties of streamwise component are consistent with data which are obtained in previous researches. The secondary flow is also measured on each measurement plane. Reynolds stresses, v＇v＇ and w＇ w＇, are analyzed from the fluctuation of the secondary flow. The law of similarity is observed in the dimensionless distributions of mean velocity and fluctuation. However, the distributions in nearer field （i.e., in the measurement plane at X/D = 100） tend to disobey the similarity law, especially in the cases of fluctuation. It seems that jet-blast is not fully developed by reaching X/D = 100. The experimental results are compared with computational results which are obtained by CFD （computational fluid dynamics） with SST （shear-stress transport） turbulence model. And it is shown that the results by the simulation with SST turbulence model do not follow the similarity law. The present database of the Reynolds stresses is critically important for development of a new turbulence model of RANS （reynolds-averaged navier-atokes） simulations on wall-jet.

Partial phase flow rate measurements for stratified oil-water flow in horizontal wells

To accurately measure and evaluate the oil-water production profile of horizontal wells, a dynamic measurement experiment of oil-water two-phase flow in horizontal wells and numerical simulation were combined to establish a method for measuring the partial phase flow rate of oil-water two-phase stratified flow in horizontal wells. An experimental work was performed in horizontal oil-water two-phase flow simulation well using combination production logging tool including mini-capacitance sensor and mini-spinner. The combination tool provides a recording of holdup and velocity profiles at five different heights of the borehole cross-section. The effect of total flow rate and water-cut on the response of spinner and capacitive sensor at five measured positions were investigated. The capacitance water holdup interpolation imaging algorithm was used to determine the local fluid property and oil-water interface height, and the measured local fluid speed was combined with the numerical simulation result to establish an optimal calculation model for obtaining the partial phase flow rate of the oil-water two-phase stratified flow in the horizontal well. The calculated flow rates of five measured points are basically consistent with the experimental data, the total flow rate and water holdup from calculation are in agreement with the set values in the experiment too, suggesting that the method has high accuracy.

Comparison of three different CFD methods for dense fluidized beds and validation by a cold flow experiment

This work focuses on a comparison between three different numerical CFD methods, namely Euler-Euler, Euler-Lagrange-stochastic, and Euler-Lagrange-deterministic, to treat a dense spouted bed, A simple cold flow experiment was used to investigate the hydrodynamics of a gas-solid flow in a three dimensional lab-scale spouted bed, In this context, two different air mass flow rates, 0,005 and 0.006 kg/s, were applied during fluidization. The experimental bed behaviour was recorded with a high-speed camera to validate the numerical predictions in terms of bubble size, bed expansion rate, and particle velocities at different reactor heights. The numerical setup was kept similar between all three modelling approaches, At both gas mass flow rates all three approaches are able to capture the overall bed expansion. However, at higher gas mass flow rates, discrepancies between experiment and simulation increase for the Euler-Euler and Euler-Lagrange-stochastic models. The Euler-Lagrange deterministic model most accurately predicts the flow pattern at both mass flow rates. The main reasons for discrepancies between simulation and experiment result from modelling of the collision and friction forces.

Advances on numerical and experimental investigation of ship roll damping

This paper presents recent developments towards efficient and reliable methods for roll damping estimation based on numerical simulations as well as model tests using the harmonic excited roll motion(HERM)technique.A newly designed automatic roll damping estimation procedure shows the advantage of a just-in-time post processing of experimental measurement results.Real-time analysis of the measured roll damping values permits a considerable shortening of the test times.Thus,a large number of investigations can be carried out with relatively manageable effort in order to determine the roll damping behavior of different keel configurations or at operating conditions,e.g.,different sized keels or Froude numbers.In addition,HERM measurement method is applied to investigate the memory effect.For this purpose,different excitation schemes are introduced and the results are analyzed.Moreover,a study of the scale effect on the roll damping properties is conducted,in which experimental and numerical investigations are performed for two scales of a ship model.Furthermore,a method is developed that significantly reduces the effort of Reynolds average Navier-Stokes(RANS)-based simulations of roll motion.The reduction of simulation time is achieved by introducing an artificial damping.The obtained results show that the developed method is very well applicable for numerical as well as in experimental investigations.During the model tests using HERM technique,the model is free and the rudder is used to keep the straight-ahead course.The analysis of the numerical and experimental results shows that the influence of the rudder induced force and moment during HERM tests is not negligible and the contribution of the rudder must be taken into account by estimating the roll damping.Finally,a new concept is developed to investigate the parametric roll behavior of ships,which allows neglecting the consideration of the complex modelling of free surface waves in the simulations.During the RANS computations,a potential-based method is applied to compute the variation of restoring terms due the roll motion.

通过融合物理神经网络重构稀疏或不完整数据流场的实用方法

高分辨率流场重构被普遍认为是实验流体力学领域的一项艰巨任务,因为测量数据在时间和空间上通常是稀疏或不完整的.具体而言,由于实验设备或测量技术的限制,某些关键区域的数据无法测量.本文提出了一种基于融合物理神经网络(PINN)的不完美数据重建流场的实用方法,该网络将已知数据与物理原理相结合.通过圆柱体的尾流作为测试算例.研究了两种不完美数据训练集,一种是不同稀疏度的速度数据,另一种是不同区域缺失的速度数据.为了加速训练收敛,本文采用了余弦退火算法以提高PINN的计算效率.计算结果表明,该方法不仅可以高精度地重建真实的速度场,而且即使在数据稀疏度达到1%或核心流动区域数据被截断的情况下,也可以精确地预测压力场.这项研究提供了令人鼓舞的结论,即PINN可以作为实验流体力学的有潜力的数据同化方法.