裂纹分析中的分网格问题--张小杰.doc

裂纹分析中的分网格问题 Simwe会员 Zxjcad整理 小楼一夜听春雨分析2维或3维裂纹问题，求应力强度因子和J积分，都要把裂纹尖端的单元划成奇异单元，如图。前面的帖子都没有明确说明这样的网格在cae中如何划分，如果要写inp的话，太麻烦了。ansys有这样的功能，不知道abaqus怎么样，我用的是6.4  YogayogaIf you only want to get K and J integral, you really dont need that. Normal elements can do. 小楼一夜听春雨： 请教，就是用常规单元，（不管是三角形还是四边形单元），只要足够细，得到裂尖附近的应力场，然后怎么计算应力强度因子K和j积分呢？还有，用奇异单元能够计算出裂纹扩展方向q，用常规单元能么？ Yogayoga Abaqus use area integral to get J integral. I assume it convert J to K for elastic case. No singularity is needed. Not sure which method you take to give propagation direction. Some LEFM approaches do need you give a crack tip K field. A fine mesh can give you approximation. But I really dont think they can give proper predition all the time. 小楼一夜听春雨： 用常规单元得到裂尖附近的应力，应变场，J积分如何计算？自己写程序算么？我得数学不太好；我是做疲劳裂纹方面的，J积分的应用是有限制的，必须满足简单加载条件，在循环载荷作用下J积分还有意义么？怎样才能模拟裂纹萌生、扩展、最后破坏的整个过程呢？ Yogayoga Abaqus can do this for you, just check contour integral.If you are looking for fatigue crack propagation, perhaps you can try Paris law. You only need FEM to help calculate K at different crack length if you can not get it analytically. The other thing you can try is to use damage cohesive zone model. Its kind of computational intensive. Other damage type approaches can also help. You do need testing to calibrate parameters in any of these approaches. 小楼一夜听春雨： 在FEM模型中，用两条重合的线或者面表示裂纹，用常规单元，直接在inp文件中*contour integral就可以得到J积分的值么？Paris law适用于长裂纹、小应变情况下的高周疲劳，计算相同疲劳载荷下不同裂纹长度的delta K值，然后计算寿命，是这个样子的把？da/dN如何通过实验数据calibrate？我做的是低周疲劳裂纹，在裂尖位置会产生很大的塑性变形，LEFM已经不适用了，考虑delta J么？damage cohesive zone model是什么model，没接触过，哪本书或文献记载的，我想看看。 Yogayoga Abaqus only need to know your crack tip position. You need to makesure that you give large enough contour to make it converge. Since it use area (2D) or volume (3D) integral, it even doesnt care if you have crack or how many crack tip you have inside contour. You need to make sureyou got what you want. If you are dealing with large scale yielding situation, I really dont think J in