Search NASA⌕ Search

SEARCH · Search NASA

Results for “Direct Numerical Simulation”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,171 records · Page 65
All sourcesNASA NTRSDOE OSTI
NASA NTRS2010

Parameterizing Coefficients of a POD-Based Dynamical System

A method of parameterizing the coefficients of a dynamical system based of a proper orthogonal decomposition (POD) representing the flow dynamics of a viscous fluid has been introduced. (A brief description of POD is presented in the immediately preceding article.) The present parameterization method is intended to enable construction of the dynamical system to accurately represent the temporal evolution of the flow dynamics over a range of Reynolds numbers. The need for this or a similar method arises as follows: A procedure that includes direct numerical simulation followed by POD, followed by Galerkin projection to a dynamical system has been proven to enable representation of flow dynamics by a low-dimensional model at the Reynolds number of the simulation. However, a more difficult task is to obtain models that are valid over a range of Reynolds numbers. Extrapolation of low-dimensional models by use of straightforward Reynolds-number-based parameter continuation has proven to be inadequate for successful prediction of flows. A key part of the problem of constructing a dynamical system to accurately represent the temporal evolution of the flow dynamics over a range of Reynolds numbers is the problem of understanding and providing for the variation of the coefficients of the dynamical system with the Reynolds number. Prior methods do not enable capture of temporal dynamics over ranges of Reynolds numbers in low-dimensional models, and are not even satisfactory when large numbers of modes are used. The basic idea of the present method is to solve the problem through a suitable parameterization of the coefficients of the dynamical system. The parameterization computations involve utilization of the transfer of kinetic energy between modes as a function of Reynolds number. The thus-parameterized dynamical system accurately predicts the flow dynamics and is applicable to a range of flow problems in the dynamical regime around the Hopf bifurcation. Parameter-continuation software can be used on the parameterized dynamical system to derive a bifurcation diagram that accurately predicts the temporal flow behavior.

Kalb, Virginia L.↗
NASA NTRS2009

Numerical and Physical Modeling of the Response of Resonator Liners to Intense Sound and Grazing Flow

Two significant advances have been made in the application of computational aeroacoustics methodology to acoustic liner technology. The first is that temperature effects for discrete sound are not the same as for broadband noise. For discrete sound, the normalized resistance appears to be insensitive to temperature except at high SPL. However, reactance is lower, significantly lower in absolute value, at high temperature. The second is the numerical investigation the acoustic performance of a liner by direct numerical simulation. Liner impedance is affected by the non-uniformity of the incident sound waves. This identifies the importance of pressure gradient. Preliminary design one and two-dimensional impedance models were developed to design sound absorbing liners in the presence of intense sound and grazing flow. The two-dimensional model offers the potential to empirically determine incident sound pressure face-plate distance from resonator orifices. This represents an important initial step in improving our understanding of how to effectively use the Dean Two-Microphone impedance measurement method.

Hersh, Alan S.↗
NASA NTRS2004

Part 1 of a Computational Study of a Drop-Laden Mixing Layer

This first of three reports on a computational study of a drop-laden temporal mixing layer presents the results of direct numerical simulations (DNS) of well-resolved flow fields and the derivation of the large-eddy simulation (LES) equations that would govern the larger scales of a turbulent flow field. The mixing layer consisted of two counterflowing gas streams, one of which was initially laden with evaporating liquid drops. The gas phase was composed of two perfect gas species, the carrier gas and the vapor emanating from the drops, and was computed in an Eulerian reference frame, whereas each drop was tracked individually in a Lagrangian manner. The flow perturbations that were initially imposed on the layer caused mixing and eventual transition to turbulence. The DNS database obtained included transitional states for layers with various liquid mass loadings. For the DNS, the gas-phase equations were the compressible Navier-Stokes equations for conservation of momentum and additional conservation equations for total energy and species mass. These equations included source terms representing the effect of the drops on the mass, momentum, and energy of the gas phase. From the DNS equations, the expression for the irreversible entropy production (dissipation) was derived and used to determine the dissipation due to the source terms. The LES equations were derived by spatially filtering the DNS set and the magnitudes of the terms were computed at transitional states, leading to a hierarchy of terms to guide simplification of the LES equations. It was concluded that effort should be devoted to the accurate modeling of both the subgridscale fluxes and the filtered source terms, which were the dominant unclosed terms appearing in the LES equations.

Okong'o, Nora A.↗
NASA NTRS2004

Helicity in Supercritical O2/H2 and C7H16/N2 Mixing Layers

This report describes a study of databases produced by direct numerical simulation of mixing layers developing between opposing flows of two fluids under supercritical conditions, the purpose of the study being to elucidate chemical-species-specific aspects of turbulence, with emphasis on helicity. The simulations were performed for two different fluid pairs -- O2/H2 and C7H16/N2 -- at similar values of reduced pressure.

Okongo, Nora↗
NASA NTRS2003

Irreversible Entropy Production in Two-Phase Mixing Layers

This report presents a study of dissipation (irreversible production of entropy) in three-dimensional, temporal mixing layers laden with evaporating liquid drops. The purpose of the study is to examine the effects of evaporating drops on the development of turbulent features in flows. Direct numerical simulations were performed to analyze transitional states of three mixing layers: one without drops, and two that included drops at different initial mass loadings. Without drops, the dissipation is essentially due to viscous effects. It was found that in the presence of drops, the largest contribution to dissipation was made by heating and evaporation of the drops, and that at large length scales, this contribution is positive (signifying that the drops reduce turbulence), while at small scales, this contribution is negative (the drops increase turbulence). The second largest contribution to dissipation was found to be associated with the chemical potential, which leads to an increase in turbulence at large scales and a decrease in turbulence at small scales. The next smaller contribution was found to be that of viscosity. The fact that viscosity effects are only third in order of magnitude in the dissipation is in sharp contrast to the situation for the mixing layer without the drops. The next smaller contribution - that of the drag and momentum of the vapor from the drops - was found to be negative at lower mass loading but to become positive at higher mass loading.

Okongo, Nora↗
NASA NTRS2003

Update on Area Production in Mixing of Supercritical Fluids

The focus of this research is on supercritical C7H16/N2 and O2/H2 mixing layers undergoing transitions to turbulence. The C7H16/N2 system serves as a simplified model of hydrocarbon/air systems in gas-turbine and diesel engines; the O2/H2 system is representative of liquid rocket engines. One goal of this research is to identify ways of controlling area production to increase disintegration of fluids and enhance combustion in such engines. As used in this research, "area production" signifies the fractional rate of change of surface area oriented perpendicular to the mass-fraction gradient of a mixing layer. In the study, a database of transitional states obtained from direct numerical simulations of the aforementioned mixing layers was analyzed to investigate global layer characteristics, phenomena in regions of high density-gradient magnitude (HDGM), irreversible entropy production and its relationship to the HDGM regions, and mechanisms leading to area production.

Okongo, Nora↗
NASA NTRS2003

Noncircular Cross Sections Could Enhance Mixing in Sprays

A computational study has shown that by injecting drops in jets of gas having square, elliptical, triangular, or other noncircular injection cross sections, it should be possible to increase (relative to comparable situations having circular cross section) the entrainment and dispersion of liquid drops. This finding has practical significance for a variety of applications in which it is desirable to increase dispersion of drops. For example, in chemical-process sprays, increased dispersion leads to increases in chemical- reaction rates; in diesel engines, increasing the dispersion of drops of sprayed fuel reduces the production of soot; and in household and paint sprays, increasing the dispersion of drops makes it possible to cover larger surfaces. It has been known for some years that single-phase fluid jets that enter flow fields through noncircular inlets entrain more fluid than do comparable jets entering through circular inlets. The computational study reported here was directed in part toward determining whether and how this superior mixing characteristic of noncircular single phase jets translates to a similar benefit in cases of two-phase jets (that is, sprays). The study involved direct numerical simulations of single- and two-phase free jets with circular, elliptical, rectangular, square, and triangular inlet cross sections. The two-phase jets consisted of gas laden with liquid drops randomly injected at the inlets. To address the more interesting case of evaporating drops, the carrier gas in the jets was specified to be initially unvitiated by the vapor of the liquid chemical species and the initial temperature of the drops was chosen to be smaller than that of the gas. The mathematical model used in the study was constructed from the conservation equations for the two-phase flow and included complete couplings of mass, momentum, and energy based on thermodynamically self-consistent specification of the enthalpy, internal energy, and latent heat of vaporization of the vapor.

Bellan, Josette↗
NASA NTRS2012

Effects of Riblets on Skin Friction in High-Speed Turbulent Boundary Layers

Direct numerical simulations of spatially developing turbulent boundary layers over riblets are conducted to examine the effects of riblets on skin friction at supersonic speeds. Zero-pressure gradient boundary layers with an adiabatic wall, a Mach number of M1 = 2.5, and a Reynolds number based on momentum thickness of Re = 1720 are considered. Simulations are conducted for boundary-layer flows over a clean surface and symmetric V- groove riblets with nominal spacings of 20 and 40 wall units. The DNS results confirm the few existing experimental observations and show that a drag reduction of approximately 7% is achieved for riblets with proper spacing. The influence of riblets on turbulence statistics is analyzed in detail with an emphasis on identifying the differences, if any, between the drag reduction mechanisms for incompressible and high-speed boundary layers.

Duan, Lian↗
NASA NTRS2012

Numerical Study of Pressure Fluctuations due to High-Speed Turbulent Boundary Layers

Direct numerical simulations (DNS) are used to examine the pressure fluctuations generated by fully developed turbulence in supersonic turbulent boundary layers with an emphasis on both pressure fluctuations at the wall and the acoustic fluctuations radiated into the freestream. The wall and freestream pressure fields are first analyzed for a zero pressure gradient boundary layer with Mach 2.5 and Reynolds number based on momentum thickness of approximately 2835. The single and multi-point statistics reported include the wall pressure fluctuation intensities, frequency spectra, space-time correlations, and convection velocities. Single and multi-point statistics of surface pressure fluctuations show good agreement with measured data and previously published simulations of turbulent boundary layers under similar flow conditions. Spectral analysis shows that the acoustic fluctuations outside the boundary layer region have much lower energy content within the high-frequency region. The space-time correlations reflect the convective nature of the pressure field both at the wall and in the freestream, which is characterized by the downstream propagation of pressure-carrying eddies. Relative to those at the wall, the pressure-carrying eddies associated with the freestream signal are larger and convect at a significantly lower speed. The preliminary DNS results of a Mach 6 boundary layer show that the pressure rms in the freestream region is significantly higher than that of the lower Mach number case.

Duan, Lian↗
NASA NTRS2013

Roughness Induced Transition in a Supersonic Boundary Layer

Direct numerical simulation is used to investigate the transition induced by threedimensional isolated roughness elements in a supersonic boundary layer at a free stream Mach number of 3.5. Simulations are performed for two different configurations: one is a square planform roughness and the other is a diamond planform roughness. The mean-flow calculations show that the roughness induces counter rotating streamwise vortices downstream of the roughness. These vortices persist for a long distance downstream and lift the low momentum fluid from the near wall region and place it near the outer part of the boundary layer. This forms highly inflectional boundary layer profiles. These observations agree with recent experimental observations. The receptivity calculations showed that the amplitudes of the mass-flux fluctuations near the neutral point for the diamond shape roughness are the same as the amplitude of the acoustic disturbances. They are three times smaller for the square shape roughness.

Balakumar, Ponnampalam↗
NASA NTRS2013

Understanding How Kurtosis Is Transferred from Input Acceleration to Stress Response and Its Influence on Fatigue Llife

High cycle fatigue of metals typically occurs through long term exposure to time varying loads which, although modest in amplitude, give rise to microscopic cracks that can ultimately propagate to failure. The fatigue life of a component is primarily dependent on the stress amplitude response at critical failure locations. For most vibration tests, it is common to assume a Gaussian distribution of both the input acceleration and stress response. In real life, however, it is common to experience non-Gaussian acceleration input, and this can cause the response to be non-Gaussian. Examples of non-Gaussian loads include road irregularities such as potholes in the automotive world or turbulent boundary layer pressure fluctuations for the aerospace sector or more generally wind, wave or high amplitude acoustic loads. The paper first reviews some of the methods used to generate non-Gaussian excitation signals with a given power spectral density and kurtosis. The kurtosis of the response is examined once the signal is passed through a linear time invariant system. Finally an algorithm is presented that determines the output kurtosis based upon the input kurtosis, the input power spectral density and the frequency response function of the system. The algorithm is validated using numerical simulations. Direct applications of these results include improved fatigue life estimations and a method to accelerate shaker tests by generating high kurtosis, non-Gaussian drive signals.

Kihm, Frederic↗
NASA NTRS2013

Comparative Study on High-Order Positivity-preserving WENO Schemes

In gas dynamics and magnetohydrodynamics flows, physically, the density and the pressure p should both be positive. In a standard conservative numerical scheme, however, the computed internal energy is obtained by subtracting the kinetic energy from the total energy, resulting in a computed p that may be negative. Examples are problems in which the dominant energy is kinetic. Negative may often emerge in computing blast waves. In such situations the computed eigenvalues of the Jacobian will become imaginary. Consequently, the initial value problem for the linearized system will be ill posed. This explains why failure of preserving positivity of density or pressure may cause blow-ups of the numerical algorithm. The adhoc methods in numerical strategy which modify the computed negative density and/or the computed negative pressure to be positive are neither a conservative cure nor a stable solution. Conservative positivity-preserving schemes are more appropriate for such flow problems. The ideas of Zhang & Shu (2012) and Hu et al. (2012) precisely address the aforementioned issue. Zhang & Shu constructed a new conservative positivity-preserving procedure to preserve positive density and pressure for high-order WENO schemes by the Lax-Friedrichs flux (WENO/LLF). In general, WENO/LLF is too dissipative for flows such as turbulence with strong shocks computed in direct numerical simulations (DNS) and large eddy simulations (LES). The new conservative positivity-preserving procedure proposed in Hu et al. (2012) can be used with any high-order shock-capturing scheme, including high-order WENO schemes using the Roe's flux (WENO/Roe). The goal of this study is to compare the results obtained by non-positivity-preserving methods with the recently developed positivity-preserving schemes for representative test cases. In particular the more difficult 3D Noh and Sedov problems are considered. These test cases are chosen because of the negative pressure/density most often exhibited by standard high-order shock-capturing schemes. The simulation of a hypersonic nonequilibrium viscous shock tube that is related to the NASA Electric Arc Shock Tube (EAST) is also included. EAST is a high-temperature and high Mach number viscous nonequilibrium flow consisting of 13 species. In addition, as most common shock-capturing schemes have been developed for problems without source terms, when applied to problems with nonlinear and/or sti source terms these methods can result in spurious solutions, even when solving a conservative system of equations with a conservative scheme. This kind of behavior can be observed even for a scalar case (LeVeque & Yee 1990) as well as for the case consisting of two species and one reaction (Wang et al. 2012). For further information concerning this issue see (LeVeque & Yee 1990; Griffiths et al. 1992; Lafon & Yee 1996; Yee et al. 2012). This EAST example indicated that standard high-order shock-capturing methods exhibit instability of density/pressure in addition to grid-dependent discontinuity locations with insufficient grid points. The evaluation of these test cases is based on the stability of the numerical schemes together with the accuracy of the obtained solutions.

WENO↗
NASA NTRS2014

Turbulence Modeling Verification and Validation

Computational fluid dynamics (CFD) software that solves the Reynolds-averaged Navier-Stokes (RANS) equations has been in routine use for more than a quarter of a century. It is currently employed not only for basic research in fluid dynamics, but also for the analysis and design processes in many industries worldwide, including aerospace, automotive, power generation, chemical manufacturing, polymer processing, and petroleum exploration. A key feature of RANS CFD is the turbulence model. Because the RANS equations are unclosed, a model is necessary to describe the effects of the turbulence on the mean flow, through the Reynolds stress terms. The turbulence model is one of the largest sources of uncertainty in RANS CFD, and most models are known to be flawed in one way or another. Alternative methods such as direct numerical simulations (DNS) and large eddy simulations (LES) rely less on modeling and hence include more physics than RANS. In DNS all turbulent scales are resolved, and in LES the large scales are resolved and the effects of the smallest turbulence scales are modeled. However, both DNS and LES are too expensive for most routine industrial usage on today's computers. Hybrid RANS-LES, which blends RANS near walls with LES away from walls, helps to moderate the cost while still retaining some of the scale-resolving capability of LES, but for some applications it can still be too expensive. Even considering its associated uncertainties, RANS turbulence modeling has proved to be very useful for a wide variety of applications. For example, in the aerospace field, many RANS models are considered to be reliable for computing attached flows. However, existing turbulence models are known to be inaccurate for many flows involving separation. Research has been ongoing for decades in an attempt to improve turbulence models for separated and other nonequilibrium flows. When developing or improving turbulence models, both verification and validation are important steps in the process. Verification insures that the CFD code is solving the equations as intended (no errors in the implementation). This is typically done either through the method of manufactured solutions (MMS) or through careful step-by-step comparisons with other verified codes. After the verification step is concluded, validation is performed to document the ability of the turbulence model to represent different types of flow physics. Validation can involve a large number of test case comparisons with experiments, theory, or DNS. Organized workshops have proved to be valuable resources for the turbulence modeling community in its pursuit of turbulence modeling verification and validation. Workshop contributors using different CFD codes run the same cases, often according to strict guidelines, and compare results. Through these comparisons, it is often possible to (1) identify codes that have likely implementation errors, and (2) gain insight into the capabilities and shortcomings of different turbulence models to predict the flow physics associated with particular types of flows. These are valuable lessons because they help bring consistency to CFD codes by encouraging the correction of faulty programming and facilitating the adoption of better models. They also sometimes point to specific areas needed for improvement in the models. In this paper, several recent workshops are summarized primarily from the point of view of turbulence modeling verification and validation. Furthermore, the NASA Langley Turbulence Modeling Resource website is described. The purpose of this site is to provide a central location where RANS turbulence models are documented, and test cases, grids, and data are provided. The goal of this paper is to provide an abbreviated survey of turbulence modeling verification and validation efforts, summarize some of the outcomes, and give some ideas for future endeavors in this area.

Rumsey, Christopher L.↗
NASA NTRS2014

Comparing Experiment and Computation of Hypersonic Laminar Boundary Layers with Isolated Roughness

Streamwise velocity profile behavior in a hypersonic laminar boundary layer in the presence of an isolated roughness element is presented for an edge Mach number of 8.2. Two different roughness element types are considered: a 2-mm tall, 4-mm diameter cylinder, and a 2-mm radius hemisphere. Measurements of the streamwise velocity behavior using nitric oxide (NO) planar laser-induced fluorescence (PLIF) molecular tagging velocimetry (MTV) have been performed on a 20-degree wedge model. The top surface of this model acts as a flat-plate and is oriented at 5 degrees with respect to the freestream flow. Computations using direct numerical simulation (DNS) of these flows have been performed and are compared to the measured velocity profiles. Particular attention is given to the characteristics of velocity profiles immediately upstream and downstream of the roughness elements. In these regions, the streamwise flow can experience strong deceleration or acceleration. An analysis in which experimentally measured MTV profile displacements are compared with DNS particle displacements is performed to determine if the assumption of constant velocity over the duration of the MTV measurement is valid. This assumption is typically made when reporting MTV-measured velocity profiles, and may result in significant errors when comparing MTV measurements to computations in regions with strong deceleration or acceleration. The DNS computations with the cylindrical roughness element presented in this paper were performed with and without air injection from a rectangular slot upstream of the cylinder. This was done to determine the extent to which gas seeding in the MTV measurements perturbs the boundary layer flowfield.

Bathel, Brett F.↗
NASA NTRS2014

Comparitive Study of High-Order Positivity-Preserving WENO Schemes

In gas dynamics and magnetohydrodynamics flows, physically, the density ρ and the pressure p should both be positive. In a standard conservative numerical scheme, however, the computed internal energy is The ideas of Zhang & Shu (2012) and Hu et al. (2012) precisely address the aforementioned issue. Zhang & Shu constructed a new conservative positivity-preserving procedure to preserve positive density and pressure for high-order Weighted Essentially Non-Oscillatory (WENO) schemes by the Lax-Friedrichs flux (WENO/LLF). In general, WENO/LLF is obtained by subtracting the kinetic energy from the total energy, resulting in a computed p that may be negative. Examples are problems in which the dominant energy is kinetic. Negative ρ may often emerge in computing blast waves. In such situations the computed eigenvalues of the Jacobian will become imaginary. Consequently, the initial value problem for the linearized system will be ill posed. This explains why failure of preserving positivity of density or pressure may cause blow-ups of the numerical algorithm. The adhoc methods in numerical strategy which modify the computed negative density and/or the computed negative pressure to be positive are neither a conservative cure nor a stable solution. Conservative positivity-preserving schemes are more appropriate for such flow problems. too dissipative for flows such as turbulence with strong shocks computed in direct numerical simulations (DNS) and large eddy simulations (LES). The new conservative positivity-preserving procedure proposed in Hu et al. (2012) can be used with any high-order shock-capturing scheme, including high-order WENO schemes using the Roe's flux (WENO/Roe). The goal of this study is to compare the results obtained by non-positivity-preserving methods with the recently developed positivity-preserving schemes for representative test cases. In particular the more di cult 3D Noh and Sedov problems are considered. These test cases are chosen because of the negative pressure/density most often exhibited by standard high-order shock-capturing schemes. The simulation of a hypersonic nonequilibrium viscous shock tube that is related to the NASA Electric Arc Shock Tube (EAST) is also included. EAST is a high-temperature and high Mach number viscous nonequilibrium ow consisting of 13 species. In addition, as most common shock-capturing schemes have been developed for problems without source terms, when applied to problems with nonlinear and/or sti source terms these methods can result in spurious solutions, even when solving a conservative system of equations with a conservative scheme. This kind of behavior can be observed even for a scalar case as well as for the case consisting of two species and one reaction.. This EAST example indicated that standard high-order shock-capturing methods exhibit instability of density/pressure in addition to grid-dependent discontinuity locations with insufficient grid points. The evaluation of these test cases is based on the stability of the numerical schemes together with the accuracy of the obtained solutions.

Schemes↗
NASA NTRS2014

DNS of Flows over Periodic Hills using a Discontinuous-Galerkin Spectral-Element Method

Direct numerical simulation (DNS) of turbulent compressible flows is performed using a higher-order space-time discontinuous-Galerkin finite-element method. The numerical scheme is validated by performing DNS of the evolution of the Taylor-Green vortex and turbulent flow in a channel. The higher-order method is shown to provide increased accuracy relative to low-order methods at a given number of degrees of freedom. The turbulent flow over a periodic array of hills in a channel is simulated at Reynolds number 10,595 using an 8th-order scheme in space and a 4th-order scheme in time. These results are validated against previous large eddy simulation (LES) results. A preliminary analysis provides insight into how these detailed simulations can be used to improve Reynoldsaveraged Navier-Stokes (RANS) modeling

DNS↗
NASA NTRS2014

DNS of Laminar-Turbulent Transition in Swept-Wing Boundary Layers

Direct numerical simulation (DNS) is performed to examine laminar to turbulent transition due to high-frequency secondary instability of stationary crossflow vortices in a subsonic swept-wing boundary layer for a realistic natural-laminar-flow airfoil configuration. The secondary instability is introduced via inflow forcing and the mode selected for forcing corresponds to the most amplified secondary instability mode that, in this case, derives a majority of its growth from energy production mechanisms associated with the wall-normal shear of the stationary basic state. An inlet boundary condition is carefully designed to allow for accurate injection of instability wave modes and minimize acoustic reflections at numerical boundaries. Nonlinear parabolized stability equation (PSE) predictions compare well with the DNS in terms of modal amplitudes and modal shape during the strongly nonlinear phase of the secondary instability mode. During the transition process, the skin friction coefficient rises rather rapidly and the wall-shear distribution shows a sawtooth pattern that is analogous to the previously documented surface flow visualizations of transition due to stationary crossflow instability. Fully turbulent features are observed in the downstream region of the flow.

Duan, L.↗
NASA NTRS2012

Radiative Feedback from Primordial Protostars and Final Mass of the First Stars

In this contribution, we review our efforts toward understanding the typical mass-scale of primordial stars. Our direct numerical simulations show that, in both of Population III.1 and III.2 cases, strong UV stellar radiative feedback terminatesmass accretion onto a protostar.AnHII region formed around the protostar very dynamically expands throughout the gas accreting envelope, which cuts off the gas supply to a circumstellar disk. The disk is exposed to the stellar UV radiation and loses its mass by photoevaporation. The derived final masses are 43 Stellar Mass and 17 Stellar Mass in our fiducial Population III.1 and III.2 cases. Much more massive stars should form in other exceptional conditions. In atomic-cooling halos where H2 molecules are dissociated, for instance, a protostar grows via very rapid mass accretion with the rates M* approx. 0.1 - 1 Stellar Mass/yr. Our newstellar evolution calculations show that the protostar significantly inflates and never contracts to reach the ZAMS stage in this case. Such the "supergiant protostars" have very low UV luminosity, which results in weak radiative feedback against the accretion flow. In the early universe, supermassive stars formed through this process might provide massive seeds of supermassive black holes.

star formation↗