声振论坛

 找回密码
 我要加入

QQ登录

只需一步,快速开始

查看: 2929|回复: 5

[转帖]接触单元常问的问题及解答

[复制链接]
发表于 2005-8-18 17:15 | 显示全部楼层 |阅读模式

马上注册,结交更多好友,享用更多功能,让你轻松玩转社区。

您需要 登录 才可以下载或查看,没有账号?我要加入

x
这是从XANSYS上转下来的, 挺好, 值得一读: <BR><BR><BR><STRONG> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% <BR> F. A. Qs on CONTACT - ANALYSIS <BR> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% <BR>  <BR> Terminology:  <BR>  <BR> STS - Surface to Surface Contact <BR> NTS - Nodes to Surface Contact <BR> NTN - Node to Node Contact <BR> FKN or Kn - Normal Contact Stiffness  <BR> E - Youngs Modulus  <BR> ICONT - Initial Contact value <BR> PMIN &amp; PMAX - Minimum and Maximum initial Penetration range <BR> PINB - Pinball region Radius <BR> FTOLN - Penetration tolerance <BR> Mu - Friction Coefficient <BR>  <BR> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% <BR> FAQ# 1. On what factor can a contact analysis be judged as "OKAY"? <BR> ======================================= <BR> Answer: Penetration. <BR>  <BR> In physical reality, penetration between 2 contacting bodies never occurs. This <BR> is a mathematical creation so as to activate a contact stiffness between two <BR> bodies. Keeping penetration to a minimum is a best way to simulate a contact <BR> analysis. For acheiving this, the contact Stiffness may be specified as high as <BR> possible as long as a converged FE solution is possible.  <BR>  <BR> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% <BR> FAQ# 2. What affects the accuracy and convergence of a contact analysis? <BR> ======================================= <BR> (FKN) Stiffness used for the Contact is a major factor affecting Convergence of <BR> Contact problems. Using Higher values of Stiffness may diverge the solution. <BR> Its advised to start with a small enough value and get a converged solution. <BR> Re-solve with increased value of stiffness so long as the problem stops <BR> converging.  <BR>  <BR> Trial-and-Error + Experience will help you predict the Contact Stiffness needed <BR> to solve a Contact Analysis.  <BR>  <BR> Another recommended practise is to run the analysis with an initial small <BR> contact stiffness and then slowly increase the stiffness over a series of load <BR> steps. This will ramp the contact stiffness value from the initially considered <BR> Stiffness, thus improving convergence behavior. The results for the last <BR> loadstep would have been analysed with more stiffness and hence would be more <BR> accurate than previous results.  <BR>  <BR> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% <BR> FAQ# 3. What is Hertz Stiffness? How is it calculated? <BR> ======================================= <BR> Hertz Stiffness provides an approximate value for the Penalty Stiffness value.  <BR> Kh = l*E <BR> where, Kh --&gt; Hertz Stiffness <BR> l --&gt; Size/edge length of the element at the contact surface <BR> E --&gt; Youngs Mod of Contact Surface <BR>  <BR> For 2D models, the thickness of element is considered for l <BR>  <BR> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% <BR> FAQ# 4. Simulating contacts between 2 Rubber material bodies. How can I <BR> evaluate E for the rubber so that I can roughly estimate the contact Stiffness? <BR> ======================================= <BR> Consider E for the rubber material portion as:  <BR> E = 6(a+b) where a and b are the first two Mooney-Rivlin contants <BR>  <BR> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% <BR> FAQ# 5. While using NTN contact elements between two (3D) surfaces, is there <BR> any recommended practise for meshing at those two surfaces? <BR> ======================================= <BR> Meshing between the two sides were varying along X.  <BR>  <BR> Surface 1 <BR> - - - - - - - - --- --- --- --- --- --- --- Y <BR> $ $ $ $ $ $ $ $ $ $ $ $ $ $ $ | <BR> - - - - - - - - --- --- --- --- --- --- --- |___ X <BR> finer coarser <BR> Surface 2 <BR>  <BR> $ sign indicates the Node-To-Node Contact Element between the Nodes of Surface <BR> 1 and 2 respectively. If you observe the mesh variation along X direction, it <BR> has varied from Finer to coarser.  <BR>  <BR> In such situations, use of a uniform Kn value for all the N-T-N contact <BR> elements could lead to un-even contact pressures. It is advisable to specify <BR> different values of Kn. If this step was not taken care of, then the finely <BR> meshed portion will experience more stiffness compared to the coarsely meshed <BR> portion. <BR>  <BR> To avoid these issues, its recommended practise to use the same edge length <BR> between the two surfaces 1 &amp; 2. </STRONG><BR>

本帖被以下淘专辑推荐:

回复
分享到:

使用道具 举报

 楼主| 发表于 2005-8-18 17:16 | 显示全部楼层
<STRONG>%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% <BR> FAQ# 6. What are the guidelines to estimate FKN for a contact Analysis. <BR> ======================================= <BR> For bulky contact, where the two are solid bodies  <BR> FKN = 0.1 to 10 ---&gt; Factor of E <BR>  <BR> For contact between two slender bodies ( thin / bending dominated structures)  <BR> FKN = 0.01 to 1 ---&gt; Factor of E <BR>  <BR> Its always advised to start with a lower value of FKN and proceed later with <BR> higher values. The default setting in Ansys is 1, irrespective of what kind of <BR> bodies are participating are getting in Contact. <BR>  <BR> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% <BR> FAQ# 7. When two different material (say Steel and Aluminum) are in Contact, <BR> which material's E should be considered for evaluating FKN <BR> ======================================= <BR> E of the softer material be always considered for estimating FKN. In this case <BR> prefer E of Aluminium.  <BR>  <BR> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% <BR> FAQ# 8. How to specify Sticking and Sliding conditions for Contact between two <BR> bodies. <BR> ======================================= <BR> 'Mu' - Friction coefficient and TAUMAX are the two inputs user needs to <BR> specify.  <BR> If Ft &gt;= Mu*Fn where Fn --&gt; Normal Force &amp; Ft --&gt; Tangential Force, the <BR> bodies slide relative to each opther and the two contacting bodies experince a <BR> Shear force of TAUMAX.  <BR>  <BR> Sticking happens when Ft &lt; Mu*Fn <BR>  <BR> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% <BR> FAQ# 9. How to estimate TAUMAX? <BR> ======================================= <BR> TAUMAX = (Von Mises Yield Stress/1.732) <BR> This is an empirical formula <BR>  <BR> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% <BR> FAQ# 10. I have Higher order elements (with mid-side nodes) on the two <BR> contacting surfaces &amp; I want to analyse the problem using NTN Contact. Can I <BR> create NTN Contact between the mid-side Nodes? <BR> ======================================= <BR> Due to the uneven nature of the kinematically consistent reaction forces at the <BR> nodes of midside-noded elements, NTN contact should NOT be applied to mid-side <BR> nodes. <BR>  <BR> In effect, usage of NTN implies that only lower order elements be used for <BR> underlying mesh of contacting parts. </STRONG><BR>
 楼主| 发表于 2005-8-18 17:16 | 显示全部楼层
<STRONG>%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% <BR> FAQ# 11. While editing the real constants of a Contact element, like FKN can <BR> -ve values be entered? <BR> ======================================= <BR> Answer: In the Real Constant menu, a negative value indicates an absolute <BR> value; and a positive is considered as a factor. <BR>  <BR> So if FKN was specified as 0.1, its considered as 0.1*E <BR> Else, if FKN was specified as -2e10 then the value of FKN for the analysis is <BR> -2e10 <BR>  <BR> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% <BR> FAQ# 12. How does FTOLN affect convergence &amp; Accuracy of results? <BR> ======================================= <BR> After FKN, FTOLN is the next parameter that affects both convergence &amp; <BR> accuracy. Its usually known as Penetration Tolerance(TOLN) <BR>  <BR> This is a factor of the depth of the underlying element. If the underlying <BR> elements depth was [h], and if FTOLN is specified as 0.1 (default value), a <BR> penetration of (0.1*h) is allowed. If this value was exceeded during the <BR> solution, then solution does not converge. Its aborted. <BR>  <BR> In effect, this value can be treated as the Maximum allowable pentration. <BR> Setting too small a value can also affect in the form of excessive iterations <BR> or non-convergence. <BR>  <BR> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% <BR> FAQ# 13. How and where is contact detected? <BR> ======================================= <BR> For STS elements, only the Target elements can penetrate into the Contact <BR> Elements and NOT vice-versa. Gauss-Integration points on the Contact Surface <BR> act as the Contact detection points. This is the default setting and is <BR> recommended for most cases. <BR>  <BR> The Newton-Cotes/Lobatto nodal integration scheme, uses the Nodes of the <BR> Contact elements as Contact Detection points.  <BR>  <BR> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% <BR> FAQ# 14. There are several Contact Surface behaviour options on the Contact <BR> Wizard. What does each one of them do/describe? <BR> ======================================= <BR> STANDARD: Normal contact closing and opening behavior, with normal <BR> sticking/sliding friction behavior. <BR>  <BR> ROUGH: Normal contact closing and opening behavior, but no sliding can occur. <BR> The surfaces are assumed to be so rough that there is infinite Friction and <BR> there can be no relative Sliding.  <BR>  <BR> BONDED: Target and contact surfaces are assumed to get GLUED once contact is <BR> established.  <BR>  <BR> BONDED CONTACT (always): Any contact detection points initially inside the <BR> pinball region or that come into contact are bonded for the remainder of the <BR> analysis. This contact typically can be used to "ADD" two meshes in a Assembly <BR> analysis where two parts have different meshes. A linear Static analysis can <BR> also be solved with this contact. Though ansys would prompt for nonlinear <BR> analysis(due to presence of contact elements) a single iteration is enough <BR>  <BR> BONDED CONTACT (initial contact): Bonds surfaces ONLY in initial contact, <BR> initially open surfaces will remain open. <BR>  <BR> NO SEPERATION: Target and contact surfaces are tied once contact is <BR> established (sliding is permitted). <BR>  <BR> NO SEPERATION (always): Any contact detection points initially inside the <BR> pinball region or that come into contact are tied in the normal direction <BR> (sliding is permitted). <BR>  <BR> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% <BR> FAQ# 15. How should I simulate interference? Between two parts there exists <BR> both interference as well as Gap between their surfaces? Using STS Contact <BR> element, what factors need to be adjusted/set ? <BR> ======================================= <BR> In most of the FE models, the exact interfernce cannot be modeled with all <BR> that accuracy. Hence there exists a parameter called CNOF which can be <BR> tweaked, either to set a interfernce or a Gap.  <BR>  <BR> A positive value of CNOF is considered as INTERFERENCE, while a negative value <BR> of CNOF is considered as GAP. Using CNOF will either move the Contact Surface <BR> inwards/outwards for the Analysis. The Target surface will not be moved. So, if <BR> you are modeling contact btw two flexible bodies... please note that the <BR> surface designated as Target will not be offset.  <BR>  <BR> If the model contains initial geometric penetration+CNOF, it is recommended to <BR> set the initial interference option to "include with ramped effects". If this <BR> setting was not set, there is every chance that due to initial penetrations, <BR> the contact forces will be stepped (not ramped) to a huge value and there is <BR> every possibility for a un-converged solution. <BR>  <BR> If both GAP and Interference exists in the model, different CNOF values will <BR> have to be specified between the various surfaces of the Parts. <BR>  <BR> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% <BR> FAQ# 16. RIGID body motion error. I have the FE model with well defined <BR> constraints. Yet I see this error in the very beginning of the Solution. How <BR> can I avoid this error? <BR> ======================================= <BR> A rigid body motion can occur just before contact is established between two <BR> parts. <BR>  <BR> You may use a displacement so as to initiate contact in the first load step. <BR> In the second load step delete this displacement, and apply the reaction force <BR> obtained from previous load step. This will switch your problem from <BR> displacement controlled problem to a force controlled problem. In the <BR> subsequent load step(3) continue with the load history. <BR>  <BR> Weak springs that are connected to Ground can be used to control Rigid body <BR> motions. The Spring stiffness should be very small compared to the stiffness of <BR> the system. Though, the spring may deform... its of no concern.  <BR>  <BR> The above two techniques involve some experimentation from the user. Also, <BR> they cannot be used in all kinds of problems. They are restricted.  <BR>  <BR> Ansys provides 3 parameters to control RIGID BODY MOTION <BR>  <BR> 1. ICONT 2. PMIN-PMAX and 3. CNOF <BR>  <BR> 1. ICONT: This is a real-constant value, which specified a zone/band around the <BR> surface of Target Elements. Any contact points lying within this Zone, are <BR> shifted to the target surface.  <BR> Ansys uses a small default value of ICONT if not specified by the user. This <BR> value should not be set to ZERO. If set to Zero, Ansys uses a default value. To <BR> turn it OFF specify a a very small value liek 1e-20. <BR>  <BR> If there is a Gap of say 0.25mm between the Target and the contact surface &amp; <BR> this happens to be the first place where contact happens, then specify ICONT as <BR> 10% more than this Gap. ICONT = 0.275mm <BR>  <BR> Please note that ICONT is not any factor, its a constant value unlike other <BR> parameters <BR>  <BR> 2. PMIN-PMAX: Use real constants PMIN and PMAX to specify an initial allowable <BR> penetration range. The Target surface shall be moved into the Contact Surface <BR> and made to lie within PMIN-PMAX. This is acheived with a few interations <BR> before the load history comes into effect. By using this feature, the gaps <BR> lying between Contact and Target Surfaces will be converted into a initial <BR> closed contact. <BR> If any of the nodes on the Target surface had constraints, then that <BR> particular node will not be moved in the constrained direction. <BR>  <BR> 3. CNOF: [ see the previous FAQ. ] <BR>  <BR> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% <BR> FAQ# 17. In a Flexible-Flexible body contact, Why should the surface with finer <BR> Mesh be designated as Contact Surface, and the other surface (with coarse mesh) <BR> be designated as Target? <BR> ======================================= <BR> The Finer mesh will have more contact detection points. This will help the <BR> analysis solve better. Please note that elements/nodes on Target surface do not <BR> detect contact. <BR>  <BR> In a flexible-flexible body contact, the choice of the target and contact <BR> surface can cause a different amount of penetration and thus affect the <BR> solution accuracy. Choose that surface which tends to move towards the other as <BR> the Contact Surface. (Assuming that both surfaces have same mesh density &amp; have <BR> relatively small difference in their structural stiffness) <BR>  <BR> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% <BR> FAQ# 18. Guidelines to choose Target/Contact Surface <BR> ======================================= <BR> 1. If a convex surface comes into contact with a flat or concave surface, the <BR> flat or concave surface should be the target surface. <BR> 2. If one surface has a coarse mesh and the other a fine mesh, the surface <BR> with the coarse mesh should be the target surface. <BR> 3. If one surface is stiffer than the other, the stiffer surface should be the <BR> target surface. <BR> 4. If one surface is higher order and the other is lower order, the lower order <BR> surface should be the target surface. <BR> 5. If one surface is larger than the other, the larger surface should be the <BR> target surface <BR>  <BR> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% <BR> FAQ# 19. What is Pinball Region ? What does it affect? What is its purpose in <BR> Contact Analysis? <BR> ======================================= <BR> Pinball region (PIN<IMG src="http://okok.org/ut/images/smiles/smile_blackeye.gif" align=absMiddle border=0> affects the contact status determination. This Pinball <BR> Region is a circle (2D) or sphere(3D) around the Contact element. This Region <BR> helps figure out how "far" and "near" regions around the Contact element.  <BR>  <BR> If l ==&gt; depth of underlying element, in Rigid-to-FLex Contact, the Pinball <BR> region is 4*l. For flex-flex contact, th Pinball Region is 2*l. <BR>  <BR> The position and motion of a contact element relative to its associated target <BR> surface determines the contact element status. ANSYS monitors each contact <BR> element and assigns a status:  <BR>  <BR> STAT = 0 Open far-field contact <BR>  <BR> STAT = 1 Open near-field contact <BR>  <BR> STAT = 2 Sliding contact <BR>  <BR> STAT = 3 Sticking contact <BR>  <BR> A contact element is considered to be in near-field contact when its contact <BR> element enters a pinball region, which is centered on the integration point of <BR> the contact element.  <BR>  <BR> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% <BR> FAQ# 20. What type of Contact is best suited to simulate Interference fit kind <BR> of problems.  <BR> ======================================= <BR> NTN type of Contact is best suited to simulate Interference Fit kind of <BR> problems. Most interference problems have negligible relative sliding <BR> deformation. <BR>  <BR> NTN Contact is the least expensive in terms of solution times when compared to <BR> STS and NTS type of contact. Also, convergence issues are lesser. <BR>  <BR> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% <BR> FAQ# 21. For NTN elements, how should the axis of NTN element be oriented? (I-J <BR> orientation of element) <BR> ======================================= <BR> Typically the I-J direction should be perpendicular to the contact surfaces. If <BR> the angles are not perpendicular undesired tangential responses will be <BR> generated between the Contact surfaces. <BR>  <BR> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%</STRONG>
发表于 2006-6-22 10:48 | 显示全部楼层
嗯不错啊 <BR>订上去!!!hoho!!
发表于 2006-9-2 16:46 | 显示全部楼层
也支持一下1
发表于 2006-9-2 17:21 | 显示全部楼层
怎么这么多符号啊,排版乱了点吧,不过还是谢谢了。
您需要登录后才可以回帖 登录 | 我要加入

本版积分规则

QQ|小黑屋|Archiver|手机版|联系我们|声振论坛

GMT+8, 2024-11-28 21:48 , Processed in 0.073001 second(s), 20 queries , Gzip On.

Powered by Discuz! X3.4

Copyright © 2001-2021, Tencent Cloud.

快速回复 返回顶部 返回列表