meshless method programme for moving least square approximation

资 源 简 介

meshless method programme for moving least square approximation

详 情 说 明

Meshless methods have gained popularity in computational mechanics due to their flexibility in handling complex geometries without relying on a predefined mesh. One such technique is the Moving Least Squares (MLS) approximation, which is widely used for constructing smooth, high-order interpolations in a mesh-free framework.

The Moving Least Squares method approximates a function by locally fitting a weighted least squares polynomial to scattered data points. Unlike traditional finite element methods, MLS does not require explicit connectivity between nodes, making it particularly useful for problems involving large deformations, crack propagation, or adaptive refinement.

A typical implementation involves: Selecting a set of nodes or particles where field variables are known. Defining a weight function (often Gaussian or spline-based) that determines the influence of neighboring nodes on the local approximation. Solving a weighted least squares problem to determine the coefficients of a polynomial basis, ensuring smoothness and consistency.

The key advantages of MLS include its ability to achieve high accuracy with irregular node distributions and its smooth derivatives, which are beneficial for solving partial differential equations. However, computational cost can be higher than mesh-based methods due to the need for repeated least squares fitting at each evaluation point.

Applications of MLS in meshless methods span across fracture mechanics, fluid-structure interaction, and shape optimization, where traditional meshing becomes cumbersome. Future developments may focus on optimization techniques to reduce computational overhead while maintaining accuracy.