大型波流水槽整流区间长度的优化研究.pdf

Investigation of the Optimal Interaction Zone Length fora Large Wave-current FlumeYAO Hai-yuan1,2,ZANG Zhi-peng1,3,FANG Zhuo2,TIAN Ying-hui4,CHENG Ning5(1.State Key Laboratory of Hydraulic Engineering Simulation and Safety,Tianjin University,Tianjin 300072,China;2.Transport Planning and Research Institute,Ministry of Transport,Beijing 100028,China;3.Key Laboratory ofEarthquake Engineering Simulation and Seismic Resilience of China Earthquake Administration,Tianjin University,Tianjin 300350,China;4.Melbourne School of Engineering,The University of Melbourne,Victoria 3010,Australia;5.China Harbour Engineering Company Limited,Beijing 100027,China)Abstract:In this study,a numerical wave-current flume is established based on OpenFOAM softwareaccording to the real configuration of a physical flume.The current inlet is set in front of the wave paddle,which results in the formation of an interaction zone downstream of the current inlet.The aim ofthis study is to investigate the effects of the main influential parameters on the length of the interactionzone.Thus,the optimal configuration of wave-current flumes can be determined to obtain the minimum length of the interaction zone,thereby maximizing the effective test section.The vertical profilesof the time-mean averaged horizontal velocity at different locations are calculated.The length of theinteraction zone is determined according to deviations of the time-mean velocity from the stable valueof less than 10%.Based on the numerical results,the optimal values of the incident angle,locationand width of the current inlet are determined.The effects of water depth,wave period,wave height,and current inlet velocity on the length of the interaction zone are also investigated.Key words:numerical wave-current flume;interaction zone;configuration parameter;optimal designCLC number:TV139.2Document code:Adoi:10.3969/j.issn.1007-7294.2023.06.0050 IntroductionPhysical modeling is one of the most commonly-used methods to study hydrodynamic issues incoastal and offshore engineering1.A large-scale earthquake engineering simulation facility hasbeen installed in Tianjin.In the facility,a large shaking table set up in a wave-current flume isused to simulate the responses of coastal structures under the combined effects of earthquakes andmarine environmental loads.In a physical flume,there is usually an interaction zone when the current goes into the flume and then interferes with waves.After the interaction zone,the wave-current第27卷第6期船舶力学Vol.27 No.62023年6月Journal of Ship MechanicsJun.2023Article ID：1007-7294（2023）06-0835-13Received date:2022-12-15Foundation item:Supported by the China National Key R&D Program(2021YFB2600700);the National NaturalScience Foundation of China(51979192;51579232)and the Program of China CommunicationsConstruction Company Limited(2018-ZJKJ-01)Biography:YAO Hai-yuan(1988-),male,Ph.D.candidate;ZANG Zhi-peng(1981-),male,associate professor,corresponding auhtor,E-mail:;FANG Zhuo(1982-),female,senior engineer;TIAN Ying-hui(1974-),male,professor;CHENG Ning(1986-),male,senior engineer.field tends to be stable,i.e.,the effective test section.Thus,it is necessary to investigate the interaction process between waves and currents in the physical flume to obtain the minimum length of theinteraction zone to maximize the effective test section.For this reason,the configuration of thewave-current flume should be optimized,and the influences of the main parameters on the length ofthe interaction zone should be investigated to meet the desired functions.At present,there are many studies on numerical wave flumes.Xiao et al(2009)2established anumerical wave trough using a meshless method to simulate the propagation process of two-dimensional nonlinear irregular waves from deep to shallow waters.Zhao et al(2010)3developed a numerical flume test based on the nonperiodic high-order spectral method by using an additional potential technique and predicted the generation and propagation of nonlinear waves with different Ursellnumbers.Ning et al(2010)4proposed a completely nonlinear numerical flume model to simulatewave-flow mixing by using the time-domain higher-order boundary element method and used themixed Euler-Lagrangian method to track the instantaneous fluid surface.Higuera et al(2013)5established a numerical wave flume based on OpenFOAM and detailed validation of the new boundary conditions against laboratory experimental data,covering an ample range of different coastal engineering processes.Schmitt and Elsaesser(2015)6summarized and reviewed four different wavegenerator methods and discussed their limitations,requirements for grid and preprocessing,andcomplexities.Jiang et al(2017)7studied fluid resonance in a moonpool formed by two identicalrectangular hulls during in-phase heaving motion by employing a two-dimensional numerical waveflume based on the OpenFOAM package.Gholamipoor and Ghiasi(2020)8proposed a Hermite-type radial point interpolation method,which was used to simulate the meshless numerical flume ofcompletely nonlinear water wave propagation.Hermite-type interpolation,radial basis function andpolynomial function were used to construct shape functions to obtain better stability.Bai et al(2020)9established two-line source wave generators,a piston-type and a flap-type configuration,for use in shallow water,which were applied in OpenFOAM to simulate free surface wave trackingin a numerical flume.There are also a few numerical studies on the wave-current flume.Park et al(2001)10established a numerical wave flume with a finite-difference scheme and a marker-and-cell(MAC)method to investigate the characteristics of nonlinear wave motions and their interactions with a stationary three-dimensional body in the presence of steady uniform currents.Zhang et al(2013)11developed a numerical model to simulate the phenomenon of wave-current interactions,and the dynamicresponse of a porous seabed to combined wave-current loadings was studied.Jia et al(2017)12realized the coupling of wave flow conditions with FLUENT software and applied it to the numericalsimulation of the wave flow force of submarine pipelines.He et al(2018)13used the smoothed particle hydrodynamics(SPH)method to build a numerical wave-current tank(NWCT).The uniformcurrent field was generated by simultaneously imposing directional velocities and hydrostatic pressures in both inflow and outflow regions set below the NWCT.It turned out to be very effective forweakening the undesired oscillatory flow at the beginning stage of current generation.Wang et al(2018)14presented a fully nonlinear numerical method to model the 3D interactions between wavesand varying currents on a large scale using a phase-resolved formulation free of assumptions of836船舶力学第27卷第6期wave steepness or the order of wave-current interactions.The above studies mainly focus on the numerical generation methods and wave-current generation methods or their applications in coastal and ocean structures.These works are different fromwave generation and wave-current generation techniques in a physical flume.In the design of aphysical flume in the large-scale facility in this study,the focuses are the effect of the configurationparameters on the stability of the flow field and the space utilization of the flume,which has seldombeen considered in the previous studies.However,the optimization of a physical flume is also dependent on the selected numerical wave model,which is established entirely coincident with thephysical flume.Based on the numerical results,the effects of the configuration and hydrodynamicparameters on the stability of the wave-current field can be obtained and used when building a realflume.1 Governing equations and model verificationThe numerical wave-current flume is established based on OpenFOAM,an open source program for fluid computational dynamics,which is governed by the Reynolds average Navier-Stokes(RANS)equations,where the momentum equation is as follows:t()ui+xi()uiuj=-pxi+xi()uixi-uiuj+Si(1)The continuous equation is as follows:uixi=0(2)where uiis the velocity component,p is the pressure,g is acceleration due to gravity,is the fluiddensity,ijis the shear stress and Siis the source term.The RNG k-turbulence model is used tosolve the above equations.The free surface is tracked by the VOF method15.The details of the numerical implements in OpenFOAM can be found in many references5,7,9.To be consistent with the physical flume,waves are generated by a wave paddle at the left endof the flume,which is simulated by the moving boundary method.A damping zone is set at the otherend of the flume.Currents are generated from the flow inlet on the bottom in front of the wave paddle,and a flow outlet is set at the end of the flume,as shown in Fig.1.The same flow discharge isgiven at both flow inlets and outlets.This configuration is the same as current generation from watercirculation forced by a pump in a physical flume.The total length of the wave-current flume is denoted as L.The flow is pumped into the flume from the bottom at an angle denoted as.The widthFig.1 Sketch of the numerical wave-current flume第6期YAO Hai-yuan et al:Investigation of the Optimal Interaction 837of the current inlet is denoted as b,and the distance of the current inlet to the balance positions ofthe wave paddle is denoted as c.The water depth is d,the wave period is T and the wave height isH.The velocities at the current inlet and outlet are denoted as U.OpenFOAM is a well-developed CFD model that has been verified by simulating waves fromvarious aspects5,7,16.Here,the wave-only and wave-current cases are further simulated and compared with the measured data in experiments by Umeyama(2011)17.In his study,the velocity profiles at different instants were measured with particleimage velocimetry(PIV)and electromagnetic velocitymetry(EVM).The current inlet is 1.2 m in front of thewave paddle,and the location of measurement is 14.0m away from the wave paddle.The water depth is set to0.3 m.The wave period is 1.0 s,and the wave height is0.036 m.Then,a current with a velocity of 0.08 m/s issuperimposed on the waves.Fig.2 shows the simulatedand measured phase-averaged wave surface elevationsfor the wave-only and wave-current conditions.Thesimulated results coincide well with the measured results.Then,the simulated velocity profiles are compared with the measured values through bothPIV and EVM,as shown in Fig.3.The simulated velocity profiles coincide well with the measured(a)Wave-only(b)Wave-currentFig.3 Instantaneous horizontal-velocity profilesFig.2 Phase-averaged surface elevationsfor both wave-alone and wave-current cases838船舶力学第27卷第6期results,which verifies that the current numerical model can simulate the interactions betweenwaves and current well and can be used for the optimization of the physical flume.2 Descriptions of the wave-current flumeDue to the large width of the present flume,which can reach 90 m,a vertical 2D full-size model is adopted in this study.As shown in Fig.1,the total length of the flume is 70 m.The wave paddleis set at the left end of the flume(x=0),and the damping zone is set at the right end with a length of6.0 m(x=64-70 m).The current direction is the same as that of wave propagation in the flume.After the wave passes by the current inlet,an interaction zone forms downstream of the current inletwith a length of lc,where the current interacts violently with waves.After that,the flow field in theflume tends to be stable,i.e.,during the effective test section.The shaking table is located in the effective test section.According to the design requirements of the shaking table,the length of the interaction zone must be smaller than 9.0 m(i.e.,lc9.0 m)so that the effective test section has a sufficient length for arranging the shaking table and structures.Here,the interaction zone is defined asdeviations of the time-mean horizontal velocity from the stable velocity in the effective test sectionless than 10%.Thus,the aim of this study is to determine the optimal configuration parameter forthe minimum value of lc.The designed environmental parameters of the physical flume are as follows:the water depth is3.0 m,the maximum current velocity in the effective test section is 0.4 m/s,the wave period T=1.8 sand the wave height H=0.3 m.These parameters correspond to the typical designed environmentalconditions for offshore engineering,i.e.,a water depth of 60 m,a wave height of 6.0 m,a wave period of 8.0 s,and a current velocity of 1.8 m/s,based on a model scale of 1:20 according to theFroude similarity.To achieve the minimum length of the interaction zone,a series of configuration parameters ofthe flume are tested,i.e.,=15-60,b=2.0-4.0 m,c=2.0-3.0 m.The ranges of the environmentalparameters tested in the study are d=1.5-3.0 m,T=1.3-2.1 s,H=0.2-0.5 m,and U=0.28-0.80 m/s.The detailed combinations of the above parameters are listed in Tab.1.There are a total of 16 casessimulated in this study.Tab.1 Test conditions for the simulation of wave-current flumeTest serial number12345678Incidenceangle offlow/()1530456030303030Distance fromthe wave-makingplate b/m2.02.02.02.03.04.02.02.0Flow inletwidth c/m2.02.02.02.02.02.02.53.0Depth ofwater d/m3.03.03.03.03.03.03.03.0Wave heightH/m0.30.30.30.30.30.30.30.3Wave periodT/s1.81.81.81.81.81.81.81.8Inlet velocityU/(ms-1)0.600.600.600.600.600.600.480.40第6期YAO Hai-yuan et al:Investigation of the Optimal Interaction 839Tab.1(Continued)Test serial number910111213141516Incidenceangle offlow/()3030303030303030Distance fromthe wave-makingplate b/m2.02.02.02.02.02.02.02.0Flow inletwidth c/m2.02.02.02.02.02.02.02.0Depth ofwater d/m1.51.53.03.03.03.03.03.0Wave heightH/m0.30.20.20.50.30.30.30.3Wave periodT/s1.81.31.81.81.52.11.81.8Inlet velocityU/(ms-1)0.600.280.600.600.600.600.400.80In the numerical simulation,quadrilateral structured mesh systems are adopted,which aredenser near the free surface and the bottom wall.Before the formal simulation,a mesh dependencestudy is carried out,and the detailed results are not given here due to the limitation of the content.Then,the minimum mesh size of x=0.05 m and z=0.005 m is adopted near the free surface andbottom wall for formal simulations,while the mesh size in the middle area of the evaluation domaincan reach up to x=0.1 m and z=0.02 m.The total number of meshes is approximately 300 000.Aworkstation with 16 cores is used for parallel computing.3 Influences of the main parameters on the length of the interaction zone3.1 Effect of incident angle of current inletThe incident angle of the current inlet is one of the most important parameters in flume designand has an important influence on the interaction between waves and water flow.Four incident angles of=15,30,45 and 60 are considered.Other parameters are set as b=2.0 m,c=2.0 m,d=3.0 m,T=1.8 s,H=0.3 m and U=0.6 m/s(i.e.,the depth-mean velocity in the test section is 0.4 m/s).The meshes of the four configurations with different current inlet inc