IMPROVEMENT OF FINITE ELEMENT MESH QUALITY BY USING GEOMETRICAL QUALITY MEASURES AND OPTIMIZATION

多边形网格 有限元法 形状优化 度量(数据仓库) 数学优化 计算机科学 反向 联轴节(管道) 数学 几何学 机械工程 结构工程 工程类 数据挖掘 计算机图形学(图像)
作者
Marko Kegl,Boštjan Harl,Dejan Dinevski
出处
期刊:Proceedings of International Structural Engineering and Construction [ISEC Press]
卷期号:2 (1)
标识
DOI:10.14455/isec.res.2015.213
摘要

This paper discusses possible procedures to improve finite element meshes in order to enable an accurate and stable numerical analysis processes. In general, the process of mesh improvement has to address three main aspects: mesh untangling (removal or fix of inverted finite elements), improvement of element shape, and making the element sizes more uniform. This paper focusses on the last two aspects: improving shape and size uniformity. This problem is addressed from purely geometrical aspect and by engaging optimization methods; thus, stress/strain related finite element quality measures are not considered. Two various element quality measures are discussed with emphasis on their coupling with an adequate optimization procedure. These two measures are the inverse mean ratio measure and the size measure. The first one addresses the shape and size of finite elements, but concentrates more on the shape. The later one, on the other hand, addresses only finite element size. Since mesh improvement tasks are typical multi-objective optimization problems, the paper also addresses briefly the procedure how to transform the multi-objective problem into a usual single-objective one that can be solved by employing standard optimization techniques. In this work a gradient-based approximation method was employed to do the optimization. The discussed theory is numerically tested on simply deformed meshes.

科研通智能强力驱动
Strongly Powered by AbleSci AI
科研通是完全免费的文献互助平台,具备全网最快的应助速度,最高的求助完成率。 对每一个文献求助,科研通都将尽心尽力,给求助人一个满意的交代。
实时播报
刚刚
1秒前
科研菜鱼完成签到,获得积分10
1秒前
2秒前
2秒前
tiant014完成签到,获得积分10
2秒前
2秒前
3秒前
华仔应助zhouyan采纳,获得10
3秒前
无花果应助vv采纳,获得10
5秒前
真白白鸭完成签到,获得积分10
5秒前
珍珍发布了新的文献求助10
5秒前
6秒前
7秒前
wanci应助顺利的奇异果采纳,获得30
7秒前
shenlu完成签到,获得积分10
7秒前
量子星尘发布了新的文献求助10
8秒前
8秒前
丘比特应助不开心我的采纳,获得30
8秒前
研友_VZG7GZ应助Linco采纳,获得10
9秒前
猪猪hero发布了新的文献求助10
9秒前
z104发布了新的文献求助10
10秒前
QZR应助chen采纳,获得60
10秒前
11秒前
重要问丝完成签到 ,获得积分10
11秒前
12秒前
13秒前
小二郎应助正直帆布鞋采纳,获得10
13秒前
bbd完成签到,获得积分10
13秒前
彭于晏应助pax采纳,获得10
14秒前
zhaoaotao完成签到,获得积分10
14秒前
hyPang发布了新的文献求助10
15秒前
fairy完成签到,获得积分10
15秒前
15秒前
15秒前
15秒前
魂断红颜发布了新的文献求助10
16秒前
17秒前
17秒前
杨榆藤完成签到,获得积分10
18秒前
高分求助中
(应助此贴封号)【重要!!请各用户(尤其是新用户)详细阅读】【科研通的精品贴汇总】 10000
Basic And Clinical Science Course 2025-2026 3000
人脑智能与人工智能 1000
花の香りの秘密―遺伝子情報から機能性まで 800
Process Plant Design for Chemical Engineers 400
Principles of Plasma Discharges and Materials Processing, 3rd Edition 400
Signals, Systems, and Signal Processing 400
热门求助领域 (近24小时)
化学 材料科学 生物 医学 工程类 计算机科学 有机化学 物理 生物化学 纳米技术 复合材料 内科学 化学工程 人工智能 催化作用 遗传学 数学 基因 量子力学 物理化学
热门帖子
关注 科研通微信公众号,转发送积分 5613711
求助须知:如何正确求助?哪些是违规求助? 4698799
关于积分的说明 14899078
捐赠科研通 4737011
什么是DOI,文献DOI怎么找? 2547125
邀请新用户注册赠送积分活动 1511067
关于科研通互助平台的介绍 1473605