Skip to Main Content
 

Global Search Box

 
 
 
 

Files

File List


Di_Thesis.pdf (23.55 MB)

ETD Abstract Container

Abstract Header

Physically Based Simulation of Various Fabrics with Multi-Level Modeling

Cao, Di
http://rave.ohiolink.edu/etdc/view?acc_num=osu1492519984871715

Abstract Details

2017, Doctor of Philosophy, Ohio State University, Computer Science and Engineering.
Cloth simulation has been studied in computer graphics since the early 1980’s. It has been widely applied as special effects in feature-length films as well as animated movies, virtual fitting, costume design, etc. Textile fibers, yarns and fabric structure are considered to be interdependent in the production of fabric and together they decide the internal mechanical properties. Unfortunately, the currently popular mass-spring model describes the relationship between knot and yarn as a simple linear mass-spring system. Although simple to code and fast to execute, it is neither stable nor an accurate mathematical approximation to the real-world cloth deformations, even though it partially represents the cloth knot-yarn structure. The other main approach to cloth modeling is the finite element method. This approach models the whole piece of cloth as a continuous sheet, discretized into many small elements (often, triangles). However, the flat sheet structure only considers the shape and it ignores all yarn-level details (yarn twist, yarn slip, woven structure, etc.), which limit its accuracy even when a better data-driven stiffness model is used. Our new model is inspired by King et al. [55] and generated by modifying the existing FEM. Unlike the existing FEM, our model incorporates the knot-yarn details by paying attention to microscopic (fibers), mesoscopic (yarns) and macroscopic (fabric structure) aspects of cloth construction. On a macroscopic level, the unit element is a second order nine-node quadrilateral rather than a linear three-node typical triangle of the FEM. A linear combination of functions (shape functions) of each node is used to interpolate the deformation gradient of any point within the element to solve the differential equations governing the cloth’s motion. The strain-stress (stiffness) model used in the FEM approaches is replaced by a discrete physical stress tensor model describing the yarn-yarn crossover structure. The stress is calculated analytically from the microscopic forces such as the yarn stretch and the mesoscopic forces including yarn shear, uncrimping, and locking due to the yarn structure. We then apply the Gaussian quadrature rule to do the interpolation and homogenize over the element surface to get the force on each of the nine nodes. We present a new fabric system with better nonlinearity and anisotropy, faster computation and more realistic simulation. Results show that our new cloth model has advantages over the existing models particularly with regard to producing more natural looking wrinkles and folds.
Richard Parent (Advisor)
Huamin Wang (Committee Member)
Matthew Lewis (Committee Member)
151 p.
Computer Engineering; Computer Science
physically based cloth simulation; multi-level modeling; implicit integration;

Recommended Citations

Citations

  • Cao, D. (2017). Physically Based Simulation of Various Fabrics with Multi-Level Modeling [Doctoral dissertation, Ohio State University]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=osu1492519984871715

    APA Style (7th edition)

  • Cao, Di. Physically Based Simulation of Various Fabrics with Multi-Level Modeling. 2017. Ohio State University, Doctoral dissertation. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=osu1492519984871715.

    MLA Style (8th edition)

  • Cao, Di. "Physically Based Simulation of Various Fabrics with Multi-Level Modeling." Doctoral dissertation, Ohio State University, 2017. http://rave.ohiolink.edu/etdc/view?acc_num=osu1492519984871715

    Chicago Manual of Style (17th edition)

osu1492519984871715
572
© 2017, some rights reserved.Creative Commons License Physically Based Simulation of Various Fabrics with Multi-Level Modeling by Di Cao is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License. Based on a work at etd.ohiolink.edu.
This open access ETD is published by The Ohio State University and OhioLINK.