ANSYS-DYNA教程-第12章.ppt

《ANSYS-DYNA教程-第12章.ppt》由会员分享，可在线阅读，更多相关《ANSYS-DYNA教程-第12章.ppt（38页珍藏版）》请在三一办公上搜索。

1、第 12 章流 体 分 析,主要内容,LS-DYNA 流体分析功能Lagrangian,Eulerian及 ALE物质描述 求解流体控制方程多物质 Euler算法流体/结构耦合算法流体分析相关的关键字,LS-DYNA 流体求解能力,Explicit Eulerian 欧拉算法Explicit ALE 算法Multi-material 算法中一单元可同时容纳三种不同物质流体结构耦合算法可模拟流体、结构和炸药,Yongs interface(杨氏界面)重建Semi-implicit 求解器各种 ALE-mesh 控制技术,未来功能:,Lagrangian、Eulerian 及ALE物质运动描述,基

2、本概念物质坐标系（material coordinate或Lagrange space）:建立在物质点上，如果物质运动或变形，跟随。空间坐标系(spatial coordinate或Euler spaces)：建立在假象的永恒空间，永远不动。参考构形(reference geometery)、初始构形、当前构形：用于描述任意物质点在时空间中的运动（包括质点位置、物质边界）,Lagrangian 描述是指参考构形（网格）为物质坐标属性，即网格的节点固定在相应的物质点上，随物质的变形，则网格变形。,bird impacting rigid wall,特点：单元具有Lagrange属性，在单元变形的

3、过程中积分点始终是同一物质点，因此，对于变形历史相关的本构描述材料（结构材料）有优势。物质边界永远是由单元的SEGMENT构成，物质边界清晰，接触界面清晰。时间步长与变形程度密切相关。如果物质的变形很大，则网格变形同样很大，可能导致计算时间很长，精度下降或程序CRASH。,Inlet,Outlet,Eulerian 描述是指网格具有空间坐标属性.网格节点不动，物质在网格间运动、变形。,特点：单元采用Euler描述，节点或积分点的物质随时间推进而改变，跟踪任何物质点变形历史的工作将非常复杂。通常用于描述流体，或结构材料在高应变率、大变形条件下的流体表现。可模拟任意程度的变形 时间步长与物质变形关

4、系不大 物质边界不清晰,Arbitrary Lagrangian-Eulerian描述是指网格独立于物质坐标和空间坐标，即网格节点可随物质运动而运动，但同时，网格点与物质点可以分离。,节点的位置由算法决定，网格因此保持良好的形态。,ALE,Lagrangian,极度变形的网格,ALE单元的特点：综合Euler和Lagrange描述，即网格可以运动，物质在网格中输运。单物质时，物质边界由ALE网格SEGMENT表征，边界清晰；多物质边界不清晰。时间步长与变形程度相关任一积分步包括Lagrange 和 Euler步 在Lagrange步，网格跟随物质点运动而运动 在Euler步自动优化网格形态 适

5、用LS-DYNA中1-27号材料,流体控制方程,流体的控制方程包括质量、动量及能量守恒方程，守恒方程是建立在物质点速度v和网格运动速度 w基础上。显然，当v=w时，方程演变为？当w=0时，方程演变为？,动量,质量,能量,LS-DYNA 如何求解控制方程？,采用 算子分裂（Operator split）技术，一个时间步做如下两个工作:,2.Euler步，移动节点到新构形，在网格间输运物质。,1.Lagrangian 步，让节点随物质流动,*由此可见，ALE物质描述是算子分裂技术的理论基础。,算子分裂 技术,执行 Lagrangian 步，让网格跟随物质运动,让网格回到初始位置，执行Euler步，

6、将物质输运参量映射到新网格,移动节点的方案,Eulerian 算法:所有节点移动到初始构形,ALE 算法:1.网格优化（涉及到网格形态优化算法）2.*Prescribed motion of nodes following user defined load curves v9603.*Rigid body translation of mesh following mass flow or motion of three user specified nodes v960,*Methods 2 and 3 are only interesting in combination with a

7、multi-material formulation,*由此可见，Euler是ALE的特殊形式,物质输运的实现,移动节点后，速度、压力、应力应变以及其他的 history variables(密度、体积分数等)必须由原构形映射到新构形,在此操作中可采用不同算法，LS-DYNA 有一阶及二阶精度算法,输运算法的计算精度与网格尺寸有关。,一阶精度输运算法 Donor Cell,Donor Cell 算法是一个简单、强大的物质输运方法。基本上，要求节点在每个时间步只发生小位移。,Elements beforenode repositioning,Element centered history va

8、riable,Center element after node repositioning,二阶精度输运算法 Van Leer,van Leer 为二阶精度算法，需要更多的CPU时间，但更加精确。,estimated linear distribution,element afternode repositioning,这种算法首先构造假象的物质量在单元中的线性分布，然后，基于这个函数用积分的形式获得参量在新构形的值。,泰勒杆问题的塑性应变分布,Lagrangian,1st order ALE,2nd order ALE,一阶、二阶精度输运算法的结果比较,多物质 Euler算法,LS-DYN

9、A 容许任意 ALE 或Eulerian 同时容纳最多三种物质，通称为多物质Euler算法。这种算法对于多种材料大变形等复杂分析十分有效。,air,bird,wall,mixed element,100%air,100%bird,流体/结构耦合算法,运动约束方法（constraint based method）罚函数方法（penalty based algorithm）,constraint based method 同时修改结构及流体的速度，强迫两者运动协调。这种方式保证动量守衡，不保证动能守衡。,耦合前,耦合后,penalty based algorithm 跟踪流体/结构的相对位移，节点

10、力与相对位移成正比。这种方法动能守衡，但相对运动约束法不稳定。,o,o,耦合前,Lagrangian 单元,流体单元,耦合后,反力正比于移动距离,o,=物质点或节点,采用多物质Euler方法分析楔形物体高速冲击水面的过程，流固耦合采用基于运动约束法的(type 2)方式。,Eulerian 或 ALE 分析相关的关键字,*CONTROL_ALE(选择物质输运算法)DCT NADV METH AFAC BFAC CFAC DFAC START END AAFAC VFACT VLIMIT EBC,This is the main card for Eulerian and ALE-simulat

11、ions and it is compulsoryInteresting entries:NADV-Number of Lagrangian steps to be taken between each advection step(only for classical mesh smoothing)METH-Advection method,use 2(van Leer)or 4(Donor cell).Methods 1 and 3 are obsoleteAFAC etc.-ALE mesh smoothing parameters.AFAC=-1 turns off mesh smoo

12、thing.DFAC=1 and the rest=0 is normally the best choice.START,END-Birth and death time for mesh smoothingEBC-Automatic Eulerian BCs.A simple way to constrain nodes on the surface in all directions or in the,to the surface,normal direction.,*SECTION_SOLID_ALE（选择单元描述及Euler单元类型）SECID ELFORM AETAFAC BFA

13、C CFAC DFAC START END AAFAC,This is where one chooses element formulation(Eulerian,ALE or multi-material)Interesting entries:ELFORM-5=one material ALE 6=one material Eulerian 7=ambient(an ambient element keeps its initial set of history variables through out the whole simulation).Used for pressure b

14、cs 11=multi-material ALE 12=single material and voidAET-Type of ambient element(not well maintained in the code),*ALE_MULTI-MATERIAL_GROUP_OPTION（多物质Euler算法）PSID/PID,This is where one tells the code which multi-material parts that are using the same material.These parts are grouped together.It is an

15、 keyword command that shouldnt be necessary but it is.Interesting entries:PSID/PID-Part or part set id,defining a group,*INITIAL_VOID_PART（单物质+Void分析）PID,Material type 12 is single material and void.It means that the elements dont have to be completely filled with material(0%-100%).The rest of the v

16、olume is vacuum.This keyword allows the user to specify parts which initially dont contain any material(voided).Interesting entries:PID-Defines part which initially is void.,*CONSTRAINED_LAGRANGE_IN_SOLID（选择流固耦合算法）SLAVE MASTER SSTYP MSTYP NQUAD CTYPE DIREC MCOUPSTART END PFAC FRIC FRCMIN,Fluid-struc

17、ture coupling keywordInteresting entries:NQUAD-The default is to directly couple the motion of the Lagrangian nodes to the fluid flow.There will be leaking if the the Lagrangian slave side has a coarser mesh than the fluid.NQUAD0 creates a denser grid of coupling points to avoid leakage.(only active

18、 for CTYPE=2 and if the slave side is defined as shells or as a set of segments.)CTYPE-Coupling type.2=constraint coupling and 4=penalty coupling.Types 1 and 3 are not very good.DIREC-Chose coupling in all directions or in normal direction only(Active for CTYPE=4).,*CONSTRAINED_LAGRANGE_IN_SOLIDSLAV

19、E MASTER SSTYP MSTYP NQUAD CTYPE DIREC MCOUPSTART END PFAC FRIC FRCMIN,Fluid-structure coupling keywordcont.interesting entries:MCOUP-This flag allows the coupling to be active only for the highest density material in a multi-material mesh.(only for CTYPE=4)START,END-Birth and death time for couplin

20、g.PFAC-Penalty factor for CTYPE=4.FRIC-Coefficient of friction for CTYPE=4 when coupling in normal direction only.FRCMIN-Minimum volume fraction of highest density material for the coupling to be active(default=0.5).,*ALE_REFERENCE_SYSTEM_GROUP（网格控制技术）v960SID STYPE PRTYPE EXPAN,Defines a part set,se

21、gment set or node set to follow a specific prescribed ALE reference system motion.SID-Part,part set,segment set or node set ID.STYPE-0=Part set 1=Node set 2=Segment setPRTYPE-1=Following switch list 0=Eulerian 1=Lagrangian 2=Normal ALE mesh smoothing 3=Prescribed motion following load curves or 3 no

22、des 3=Automatic mesh motion following average mass weighted flow velocity in ALE mesh EXPAN-0=Mesh expansion allowed(for PRTYPE3 only)1=Rigid body translation of mesh,*ALE_REFERENCE_SYSTEM_CURVE v960CID1 CID2 CID3 CID4 CID5 CID6 CID7 CID8CID9 CID10 CID11 CID12,List of load curves defining prescribed

23、 ALE-mesh motion(PRTYPE=3).,*ALE_REFERENCE_SYSTEM_NODE v960NID1 NID2 NID3,List with three nodes defining a prescribed ALE-mesh motion(PRTYPE=3).The mesh is given a rigid body translation/rotation following a coordinate system defined by the three specified nodes.The origin of the coordinate system i

24、s located at node NID1.,*ALE_REFERENCE_SYSTEM_SWITCH v960TIME1TIME2TIME3TIME4TIME5TIME6TIME7TYPE1 TYPE2 TYPE3 TYPE4 TYPE5 TYPE6 TYPE7 TYPE8,The first row contains a list of time states where the prescribed reference system type shifts(PRTYPE=-1).The second row contains the switch list with the different reference system types(see*ALE_REFERENCE_SYSTEM_GROUP).,ADVANCED TRAINING andWORKSHOP,