Introducing a Generative Adversarial Network Model for Lagrangian Trajectory Simulation

TL;DR

This paper presents a GAN-based model for simulating 3D Lagrangian particle trajectories in recirculation zones of flames, combining stochastic recurrent and convolutional neural networks for realistic motion generation.

Contribution

It introduces a novel GAN architecture specifically designed for Lagrangian trajectory simulation in turbulent flow environments.

Findings

The GAN accurately reproduces the statistical properties of real trajectories.

The model generalizes well to unseen flow conditions.

Benchmarking shows high fidelity in trajectory simulation.

Abstract

We introduce a generative adversarial network (GAN) model to simulate the 3-dimensional Lagrangian motion of particles trapped in the recirculation zone of a buoyancy-opposed flame. The GAN model comprises a stochastic recurrent neural network, serving as a generator, and a convoluted neural network, serving as a discriminator. Adversarial training was performed to the point where the best-trained discriminator failed to distinguish the ground truth from the trajectory produced by the best-trained generator. The model performance was then benchmarked against a statistical analysis performed on both the simulated trajectories and the ground truth, with regard to the accuracy and generalization criteria.