Numerical Methods for Simulating Star Formation

Teyssier, Romain; Commerçon, B.

doi:10.3389/fspas.2019.00051

Cited by 27 publications

(25 citation statements)

References 373 publications

(532 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, even with today's most powerful computers, we are still far from being able to self-consistently simulate the formation of molecular clouds in a cosmological context. Simulating the multiphase interstellar medium (ISM) still remain challenging even in "zoom-in" simulations, so that it is necessary to make use of subgrid modules to model unresolved physical processes, such as the formation of molecular clouds, winds from dying stars, and supernovae (Teyssier & Commerçon 2019). Progress in understanding the global baryon cycle will have impact beyond understanding the evolution of star formation and galaxies.…”

Section: Simulating Baryonic Processes In Cosmological Contextmentioning

confidence: 99%

The Cosmic Baryon and Metal Cycles

Péroux

Howk

2020

Annu. Rev. Astron. Astrophys.

266

200

View full text Add to dashboard Cite

Characterizing the relationship between stars, gas, and metals in galaxies is a critical component of understanding the cosmic baryon cycle. We compile contemporary censuses of the baryons in collapsed structures and their chemical make-up and dust content. We show the following: ▪ The HI mass density of the Universe is well determined to redshifts z≈5 and shows minor evolution with time. New observations of molecular hydrogen reveal its evolution mirrors that of the global star-formation rate density, implying a universal cosmic molecular gas depletion timescale. The low-redshift decline of the star-formation history is thus driven by the lack of molecular gas supply due to a drop in net accretion rate related to the decreased growth of dark matter halos. ▪ The metal mass density in cold gas ( T≲104 K) contains virtually all the metals produced by stars for z≳2.5. At lower redshifts, the contributors to the total amount of metals are more diverse; at z < 1, most of the observed metals are bound in stars. Overall, there is little evidence for a “missing metals problem” in modern censuses. ▪ We characterize the dust content of neutral gas over cosmic time, finding the dust-to-gas and dust-to-metals ratios fall with decreasing metallicity. We calculate the cosmological dust mass density in the neutral gas up to z≈5. There is good agreement between multiple tracers of the dust content of the Universe. Expected final online publication date for the Annual Review of Astronomy and Astrophysics, Volume 58 is August 18, 2020. Please see http://www.annualreviews.org/page/journal/pubdates for revised estimates.

show abstract

Section: Simulating Baryonic Processes In Cosmological Contextmentioning

confidence: 99%

The Cosmic Baryon and Metal Cycles

Péroux

Howk

2020

Annu. Rev. Astron. Astrophys.

266

200

View full text Add to dashboard Cite

show abstract

“…Beuther et al 2019;Maud et al 2019). In parallel, heavy numerical developments are undertaken in order to accurately describe the physics of star formation in numerical models (see Teyssier & Commerçon 2019, for a recent review).…”

Section: Introductionmentioning

confidence: 99%

Discs and outflows in the early phases of massive star formation: Influence of magnetic fields and ambipolar diffusion

Commerçon

González

Mignon-Risse

et al. 2022

A&A

View full text Add to dashboard Cite

Context. Massive star formation remains one of the most challenging problems in astrophysics, as illustrated by the fundamental issues of the radiative pressure barrier and the initial fragmentation. The wide variety of physical processes involved, in particular the protostellar radiative feedback, increase the complexity of massive star formation in comparison with its low-mass counterpart. Aims. We aim to study the details of mass accretion and ejection in the vicinity of massive star forming cores using high-resolution (5 au) three-dimensional numerical simulations. We investigated the mechanisms at the origin of outflows (radiative force versus magnetic acceleration). We characterised the properties of the disc forming around massive protostars depending on the physics included: hydrodynamics, magnetic fields, and ambipolar diffusion. Methods. We used state-of-the-art three-dimensional adaptive-mesh-refinement models of massive dense core collapse, which integrate the equations of (resistive) grey radiation magnetohydrodynamics, and include sink particle evolution. For the first time, we include both protostellar radiative feedback via pre-main-sequence evolutionary tracks and magnetic ambipolar diffusion. To determine the role of magnetic fields and ambipolar diffusion play in the formation of outflows and discs, we studied three different cases: a purely hydrodynamical run, a magnetised simulation under the ideal approximation (perfect coupling), and a calculation with ambipolar diffusion (resistive case). In the most micro-physically complex model (resistive MHD), we also investigated the effect the initial amplitude of both magnetic field and solid body rotation have on the final properties of the massive protostellar system. We used simple criteria to identify the outflow and disc material and follow their evolution as the central star accretes mass up to 20 M⊙ in most of our models. The radiative, magnetic, and hydrodynamical properties of the outflows and discs are quantitatively measured and cross-compared between models. Results. Massive stars form in all our models, together with outflows and discs. The outflow is completely different when magnetic fields are introduced, so magneto-centrifugal processes are the main driver of the outflow up to stellar masses of 20 M⊙. Then, the disc properties heavily depend on the physics included. In particular, the disc formed in the ideal and resistive runs show opposite properties in terms of plasma beta; that is, the ratio of thermal-to-magnetic pressures and of magnetic field topology. While the disc in the ideal case is dominated by the magnetic pressure and the toroidal magnetic fields, the one formed in the resistive runs is dominated by the thermal pressure and essentially has a vertical magnetic field in the inner regions (R < 100−200 au). Conclusions. We find that magnetic processes dominate the early evolution of massive protostellar systems (M⋆ < 20 M⊙) and shapes the accretion and ejection as well as the disc formation. Ambipolar diffusion is mainly at work at disc scales and regulates its properties. We predict magnetic field’s topology within the disc and outflows, as well as disc masses and radii to be compared with observations. Lastly, our finding for the outflow and disc properties are reminiscent of the low-mass star formation framework, suggesting that accretion and ejection in young massive and low-mass protostars are regulated by the same physical processes in the early stages.

show abstract

“…There has been a lot of research works on numerical methods for solving conservation laws whose solutions may contain shocks and contact discontinuities, such as the Godunov scheme [1], MUSCL scheme [2,3], ENO [4,5] and WENO [6,7] schemes, hierarchical reconstruction [8,9,10], and many others. Numerical techniques are important for studying high-speed aerodynamic flows which plays a substantial role in aircraft designs, combustion problems and astronomy physics [11,12,13,14]. However, the development of machine learning techniques for solving hyperbolic conservation laws are still in early stage.…”

Section: Introductionmentioning

confidence: 99%

Neural Networks with Inputs Based on Domain of Dependence and A Converging Sequence for Solving Conservation Laws, Part I: 1D Riemann Problems

Huang¹,

Liu²,

Yang³

2021

Preprint

View full text Add to dashboard Cite

Recent research works for solving partial differential equations (PDEs) with deep neural networks (DNNs) have demonstrated that spatiotemporal function approximators defined by auto-differentiation are effective for approximating nonlinear problems, e.g. the Burger's equation, heat conduction equations, Allen-Cahn and other reaction-diffusion equations, and Navier-Stokes equation. Meanwhile, researchers apply automatic differentiation in physics-informed neural network (PINN) to solve nonlinear hyperbolic systems based on conservation laws with highly discontinuous transition, such as Riemann problem, by inverse problem formulation in data-driven approach. However, it remains a challenge for forward methods using DNNs without knowing part of the solution to resolve discontinuities in nonlinear conservation laws. In this study, we incorporate 1st order numerical schemes into DNNs to set up the loss functional approximator instead of auto-differentiation from traditional deep learning framework, e.g. TensorFlow package, which improves the effectiveness of capturing discontinuities in Riemann problems. In particular, the 2-Coarse-Grid neural network (2CGNN) and 2-Diffusion-Coefficient neural network (2DCNN) are introduced in this work. We use 2 solutions of a conservation law from a converging sequence, computed from a low-cost numerical scheme, and in a domain of dependence of a space-time grid point as the input for a neural network to predict its high-fidelity solution at the grid point. Despite smeared input solutions, they output sharp approximations to solutions containing shocks and contacts and are efficient to use once trained.

show abstract

Numerical Methods for Simulating Star Formation

Cited by 27 publications

References 373 publications

The Cosmic Baryon and Metal Cycles

The Cosmic Baryon and Metal Cycles

Discs and outflows in the early phases of massive star formation: Influence of magnetic fields and ambipolar diffusion

Neural Networks with Inputs Based on Domain of Dependence and A Converging Sequence for Solving Conservation Laws, Part I: 1D Riemann Problems

Contact Info

Product

Resources

About