Parallel Fabric Piece Simulation via Common Frontier Computation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current computer-aided design and simulation systems face challenges in efficiently simulating complex garments in real-time, as traditional methods either lack scalability or fail to provide realistic physical behavior, especially when dealing with massive data volumes and complex fabric interactions.
Innovation Solution
A computer-implemented method that utilizes parallel solvers to simulate fabric pieces, maintaining their assembly by computing a common frontier, allowing for real-time simulation by leveraging the natural partitioning of garments based on sewing patterns and enabling dynamic load balancing across multiple computers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional parallel computing methods (FEM) are used to divide the garment into smaller domains, then computational scalability is improved, but real-time simulation capability deteriorates due to iterative coupling procedures requiring hours to simulate few seconds of real phenomena
Solution Approach 1:
The garment is divided into multiple fabric pieces that can be simulated in parallel, with each piece handled by a dedicated solver. This segmentation allows computational scalability while maintaining real-time performance by eliminating the need for iterative coupling between domains.
Solution Approach 2:
A common frontier computation mechanism acts as an intermediary between parallel solvers, enabling them to exchange boundary information and maintain coherence without requiring iterative coupling procedures. This mediator allows direct communication between solvers, achieving both scalability and real-time performance.
2Productivity
If Zoning approach from MMOGs is used to partition the virtual universe into zones running in parallel, then massiveness is handled successfully, but physical realism deteriorates because objects are teleported between zones with minimal state continuity and no zone to zone interaction
Solution Approach 1:
The common frontier computation serves as an intermediary that enables continuous physical interaction between fabric pieces simulated in parallel, replacing the teleportation mechanism of Zoning with a physically realistic boundary interaction model that maintains state continuity across piece boundaries.
Solution Approach 2:
The system implements feedback mechanisms where solvers exchange boundary state information and adjust their computations accordingly, ensuring physical realism is maintained across piece boundaries while allowing parallel simulation of large numbers of fabric pieces.
3Reliability
If a single solver simulates the entire garment assembly, then physical coherence is maintained, but computational efficiency deteriorates due to the massive data volume and complexity
Solution Approach 1:
The garment assembly is segmented into multiple fabric pieces, each simulated by a dedicated solver. This segmentation improves computational efficiency by distributing the computational load while maintaining physical coherence through the common frontier computation that ensures consistent boundary conditions between pieces.
4Productivity
If fabric pieces are divided into separate domains for parallel simulation, then computational scalability is improved, but accuracy and coherence on the frontiers deteriorates, requiring iterative coupling procedures that take hours
Solution Approach 1:
The common frontier computation acts as an intermediary that directly communicates boundary state information between parallel solvers, eliminating the need for iterative coupling procedures. This mediator maintains accuracy and coherence on frontiers while preserving computational scalability through parallel execution.
Data Source
Figure 1~3
Figure 4~7
Figure 8~11
AI summary
It is proposed a computer-implemented method for simulating an assembly of fabric pieces. The method comprises providing a plurality of parallel solvers, a first solver simulating at least one first fabric piece and a second solver simulating at least one second fabric piece, and maintaining the first and the second fabric pieces tight by computing a common frontier by the first and second solvers.