Litcius/Paper detail

New insights into error accumulation due to biased particle distribution in semi-implicit particle methods

Guangtao Duan, Takuya Matsunaga, Seiichi Koshizuka, Akira Yamaguchi, Mikio Sakai

2021Computer Methods in Applied Mechanics and Engineering31 citationsDOIOpen Access PDF

Abstract

This study investigates the instability issue at a free surface when the consistent schemes based on variable differences are applied in semi-implicit particle methods. A semi-analytical error-analysis method is proposed to clarify how the incomplete/biased neighbor support triggers error accumulation and instability. Specifically, the discretization models are decomposed into the center-variable components (CVCs) and neighbor-variable components (NVCs). The influence of different components on error accumulation is analyzed theoretically and numerically. Based on the error analysis, new indices are proposed to evaluate the risk of error accumulation due to the biased neighbor support. Then, novel free-surface-detection conditions are proposed from the indices by detecting the particles prone to error accumulation as free surface particles. Numerical examples demonstrated that the proposed conditions detected fewer free-surface particles but produced more stable simulations compared to the existing conditions.

Topics & Concepts

DiscretizationVariable (mathematics)Particle (ecology)Surface (topology)Stability (learning theory)InstabilityError analysisComputer scienceApplied mathematicsMathematicsMechanicsStatistical physicsMathematical analysisPhysicsGeometryOceanographyGeologyMachine learningFluid Dynamics Simulations and InteractionsLattice Boltzmann Simulation StudiesFluid Dynamics and Heat Transfer
New insights into error accumulation due to biased particle distribution in semi-implicit particle methods | Litcius