Fluid animation generation method and device based on deep learning and SPH framework

A deep learning and generation device technology, applied in the field of fluid simulation, can solve problems such as computational efficiency constraints, achieve high-precision detail performance, and improve computational efficiency.

Inactive Publication Date: 2018-10-30
TIANJIN UNIV
View PDF4 Cites 9 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

In these improved SPH methods, the calculation efficiency of the pressure item is a big shackle, and effectively improving the calculation efficiency of the pressure item is of great significance to the improvement of the overall execution efficiency of the algorithm, and because large-scale fluid scenes may need to calculate With tens of millions or even hundreds of millions of particles, an effective algorithm acceleration strategy needs to be proposed urgently

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Fluid animation generation method and device based on deep learning and SPH framework
  • Fluid animation generation method and device based on deep learning and SPH framework
  • Fluid animation generation method and device based on deep learning and SPH framework

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0041] In order to achieve the above purpose, the embodiment of the present invention proposes a fluid animation generation method based on deep learning and SPH framework. The method is based on the SPH fluid simulation framework and a standard neural network model for fluid simulation and data training, including the following steps :

[0042] 101: Define the data file of the fluid simulation scene, the definition content includes: fluid parameters, and boundary conditions, etc.;

[0043] Among them, the definition of the fluid simulation scene in step 101, the specific steps are as follows:

[0044] Create data files for fluid simulation scenarios. All fluid data and scene data are defined through external data files. The definitions include but are not limited to: fluid parameters, fluid position and scale, fluid boundary conditions, flow field position and scale, etc.

[0045] The foregoing specific operations are set according to requirements in practical applications, ...

Embodiment 2

[0077] The scheme in embodiment 1 is further introduced below in conjunction with specific calculation formulas and accompanying drawings, see the following description for details:

[0078] 201: Define a scene data file;

[0079] Among them, the definition content includes but not limited to: fluid parameters, fluid position and scale, flow field boundary data, etc. Among them, the fluid parameters include viscosity coefficient, surface tension factor, particle radius and so on.

[0080] 202: Load data from the scene file, and initialize the SPH framework;

[0081] After parsing the data file into fluid data, first, the fluid block is sampled according to the particle radius to generate fluid particles, and the flow field data and boundary conditions are parsed into boundary particles. Then, the data structure of particle neighborhood search is constructed to accelerate the search of adjacent particles. The hash grid structure is used to divide the fluid space into three-di...

Embodiment 3

[0137] The embodiment of the present invention provides a data-driven SPH fluid animation generation device, which corresponds to the generation methods in Embodiments 1 and 2, see Figure 4 , the generator consists of:

[0138] The fluid scene data initialization unit is used to import and initialize the flow field data;

[0139] Among them, the scene data is defined externally, and the defined data includes but not limited to: fluid parameters, fluid position and scale, fluid boundary conditions, flow field position and scale, and the specific operation steps can be found in Examples 1 and 2, the embodiment of the present invention I won't go into details on this.

[0140] A fluid simulation data generating unit, used for obtaining a fluid simulation data set to be trained;

[0141] In the generating unit, according to the solution of the pressure item in an iterative process of the fluid simulation, relevant simulation data before and after the calculation of the pressure...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a fluid animation generation method and device of deep learning and an SPH framework. The method includes: constructing a deep-learning model, and carrying out training on fluid simulation data to generate a data drive item based on the neural network; constructing a neural network model according to features of a data set, setting relevant hyperparameters, then reading thedata set for preprocessing, and carrying out training of the network model; importing the trained neural network model into the SPH fluid simulation framework to replace a pressure item in a fluid simulation step with the same as the data drive item; importing the data drive item into low-precision fluid scene data, and then carrying out SPH fluid simulation calculation; and using a Marching Cubes algorithm to construct a fluid surface model, extracting a fluid surface mesh, outputting the same, storing a fluid mesh of each frame as a data file, and using the same for offline rendering. High-detail showing of fluid under a low-precision SPH simulation scene is realized, and calculation efficiency of a large-scale fluid simulation scene is improved.

Description

technical field [0001] The present invention relates to the field of fluid simulation in computer graphics, and at the same time includes relevant research content in the field of deep learning, especially relates to the SPH (smooth particle dynamics) method in the Lagrangian fluid simulation scheme, which uses the combination of SPH and deep neural network The method of high-precision, high-efficiency fluid simulation research. Background technique [0002] Fluid simulation has always been an important research topic in the field of computer graphics. In order to achieve high realism in fluid animation, a large number of physics-based fluid simulation schemes have been studied and applied. Among them, the Lagrangian method and the Euler method are widely used. Research fluid simulation methods. Compared with the grid-based Euler method, the Lagrangian method has many advantages in detail performance, among which, the SPH algorithm is the most widely studied object in the L...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06T13/20G06T17/20
CPCG06T13/20G06T17/205G06T2207/20081
Inventor 应翔仇强于瑞国喻梅王建荣于健
Owner TIANJIN UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap