Elsevier

Computers & Graphics

Volume 37, Issue 7, November 2013, Pages 775-786
Computers & Graphics

Technical Section
A unified smoke control method based on signed distance field

https://doi.org/10.1016/j.cag.2013.05.001Get rights and content

Highlights

  • A unified control framework, which integrates path control, shape control, and mixed control of both.

  • Two new control forces.

  • An adaptive strategy for the divergence field adjustment used in the shape control.

  • A hybrid vortex particle scheme to enhance turbulent flow details.

Abstract

Smoke control involves shape and path. A unified framework able to deal with both of them will enable animators to manipulate the shape and path of smoke animation more effectively. In this paper, we develop such a unified framework. With our approach, path control, shape control, and mixed control of both can be handled satisfactorily in the same framework. In order to develop this framework, we use a signed distance field to provide three control forces: path control force, boundary control force, and shape control force based on medial axis point clouds. The path control force makes the smoke move along the appointed route, the boundary control force keeps the smoke moving through specified regions only, and the shape control force enables the smoke to form various given shapes. The boundary control force and the shape control force are two novel control forces developed in this paper. To make the smoke form the target shape more accurately, we develop an adaptive strategy to adjust the divergence field. We also employ a new hybrid vortex particle scheme to enhance the turbulence flow details. The examples given in this paper indicate that our proposed framework is advantageous over the existing shape control approaches and path control algorithms, and a naive combination of the two.

Introduction

The special effects industry has witnessed a greater emphasis on the use of physically based fluid animation to reproduce realistic fluid effects. Besides realism, the ability to control the fluid behavior is also very important and challenging. Smoke control is an important topic of fluid control. It has drawn the attention of many researchers, and some control algorithms for special effects simulation have been developed. These algorithms can be roughly classified into two groups: path control and shape control. Path control algorithms enable smoke to follow given paths, and shape control methods make smoke form the target shapes.

Since smoke control involves both shape and path, a unified framework able to tackle both of them will enable animators to manipulate the shape and path of smoke animation more effectively. Such a unified framework has not been developed, and the work carried out in this paper indicates that it cannot be achieved by a simple combination of the existing shape control methods and path control algorithms, even with the modifications given in Section 4 of this paper.

In order to address this issue, we propose a unified control algorithm to integrate shape control, path control, and mixed control into the same framework. Our control algorithm translates 3D surface geometry models and space curves representing paths into a signed distance field. Through the signed distance field, we provide two novel control forces: boundary control force and shape control force based on medial axis point clouds. The boundary control force restricts the smoke to the appointed regions, and the shape control force is used to drive the smoke into given shapes. In addition, we use the path control force presented by Kim et al. [1]. In order to improve the accuracy of shape control, we developed an adaptive strategy for divergence field adjustment. By combining the vortex particle method [2] and the Langevin particle method [3] together, we design a hybrid vortex particle scheme to enhance the turbulent flow details. This hybrid vortex particle can freely switch between two identities of vortex particles and Langevin particles depending on its spatial location.

The contributions of our work include: (a) a unified control framework, which integrates path control, shape control, and mixed control of both, (b) two new control forces, i.e., the boundary force restricting the smoke to appointed regions and the shape control force making the smoke form target shapes, (c) an adaptive strategy for the divergence field adjustment used in the shape control, (d) a hybrid vortex particle scheme to enhance turbulent flow details.

Our approach gives a solution to the problem of mixed control of shape and path which has not been addressed by the existing approaches. With our proposed approach, the shape and path of smoke animation can be controlled more effectively.

The rest of the paper is organized as follows. Section 2 provides a brief overview of previous related work. In Section 3, the adopted algorithm is elaborated. Section 4 presents the experimental results. Finally in Section 5, the conclusion of the present work is drawn and a proposal for future work is given. Table 1

Section snippets

Related work

In 1997, Foster and Metaxas [4] introduced embedded controllers which enable animators to control fluid movement. Based on this algorithm, Foster and Fedkiw [5] proposed one modified algorithm, in which 3D parametric space curves are sampled to generate oriented points, and the velocity of these local points is further modified to control fluid movement. Three years later, Rasmussen et al. [6] presented a control algorithm based on particles, which were set accordingly to generate either hard

Algorithm

Developed from Eulerian approaches, our control algorithm consists of two parts, pre-computation and simulation loop. First, we introduce the basic fluid solver in Section 3.1. Next, we investigate the precomputation in Section 3.2 with a main focus on the calculations of the signed distance field and the scaling parameter, and the determination of the direction of our control forces. Finally, we discuss the simulation loop in Section 3.3.

Results and discussions

We have implemented our algorithm on a PC with Intel Core i5 CPU 3.20 GHz and 4 GB RAM. The PBRT library [26] was employed to render the volume data with material density and the NURBS++ library [27] was used to manipulate NURBS curves.

Fig. 8 shows four frames of our path control for the knot model. The resolution of the simulation grid is 128×128×128. We set the path radius to 0.03. The whole precomputation process lasts 201 s. A hybrid vortex particle is sampled every two steps at the smoke

Conclusions

In this paper, we have presented a novel unified technique to control the dynamics of shape-constrained smoke based on signed distance fields without direct manipulation of simulation parameters. In order to develop the unified technique, we have proposed two novel control forces: the boundary control force and the shape control force based on medial axis point clouds. A signed distance field has been used to provide these two control forces and the path control force. The combined application

Acknowledgments

This work was supported by Zhejiang Provincial Natural Science Foundation of China (Grant no. Z1110154), the National Key Basic Research Foundation of China (Grant no. 2009-CB320801), the China 863 program (Grant no. 2012AA011503), the National Natural Science Foundation of China (Grant nos. 61272298, 60933007), and the Major Science and Technology Innovation Team (Grant no. 2010R50040).

References (27)

  • S. Liu et al.

    Ellipsoidal-blob approximation of 3d models and its applications

    Comput Graph

    (2007)
  • Kim Y, Machiraju R, Thompson D. Path-based control of smoke simulations. In: Proceedings of the 2006 ACM...
  • A. Selle et al.

    A vortex particle method for smoke, water and explosions

    ACM Trans Graph

    (2005)
  • F. Chen et al.

    Langevin particlea self-adaptive lagrangian primitive for flow simulation enhancement

    Comput Graph Forum

    (2011)
  • Foster N, Metaxas D. Controlling fluid animation. In: Proceedings of the 1997 Conference on Computer Graphics...
  • Foster N, Fedkiw R. Practical animation of liquids. In: Proceedings of the 28th annual conference on Computer graphics...
  • Rasmussen N, Enright D, Nguyen D, Marino S, Sumner N, Geiger W, et al. Directable photorealistic liquids. In:...
  • Pighin F, Cohen JM, Shah M. Modeling and editing flows using advected radial basis functions. In: Proceedings of the...
  • Schpok J, Dwyer W, Ebert DS. Modeling and animating gases with simulation features. In: Proceedings of the 2005 ACM...
  • Angelidis A, Neyret F. Simulation of smoke based on vortex filament primitives. In: Proceedings of the 2005 ACM...
  • Angelidis A, Neyret F, Singh K, Nowrouzezahrai D. A controllable, fast and stable basis for vortex based smoke...
  • S. Weißssmann et al.

    Filament-based smoke with vortex shedding and variational reconnection

    ACM Trans Graph

    (2010)
  • A. Treuille et al.

    Keyframe control of smoke simulations

    ACM Trans Graph

    (2003)
  • Cited by (15)

    • A stable tensor-based method for controlled fluid simulations

      2019, Applied Mathematics and Computation
      Citation Excerpt :

      Here, we are interested in controlling fluid flow throughout the entire simulation. Most methods related to this problem use external forces to restrict fluid flow to specific paths [10–12]. There are also methods based on deformation of the underlying grid [13,14], adapting the velocity fields according to the deformed grid.

    • Target-driven cloud evolution using position-based fluids

      2020, Computer Animation and Virtual Worlds
    • Physically-based and Data-driven Fluid Simulation Research

      2020, Ruan Jian Xue Bao/Journal of Software
    View all citing articles on Scopus
    View full text