Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “computational costs”

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 163 records · Page 9
All sourcesNASA NTRSDOE OSTI
DOE OSTI2026

Predicting Atomistic Transitions with Transformers

Accurate knowledge of the atomistic transition pathways in materials and material surfaces is crucial for many material science problems. However, conventional simulation techniques used to find these transitions are extremely computationally intensive. Even with large-scale, accelerated material simulations, the computational cost constrains the applicable domain in practice. Machine learning models, with the potential to learn the complex emergent behaviors governing atomistic transitions as a fast surrogate model, have great promise to predict transitions with a vastly reduced computational cost. Here, we demonstrate how transformers can be trained to predict atomistic transitions in nano-clusters. We show how we evaluate physical validity of the predictions and how a multitude of additional, different microstates can be generated by slightly varying the data provided to the model.

36 MATERIALS SCIENCE↗
DOE OSTI2025

Embedded random phase approximation for magnetic systems: H 2 dissociative adsorption on Fe(110)

The random phase approximation (RPA), a method for treating electron correlation, has been shown to be superior to standard density functional theory (DFT) approximations in numerous cases. However, the RPA’s computational cost is substantially higher than that of DFT, particularly restricting its application to extended surfaces. The recently introduced embedded RPA (emb-RPA) approach [Wei et al., J. Chem. Phys. 159(19), 194108 (2023)] reduces this computational cost by approximately two orders of magnitude. While previous applications of emb-RPA focused on non-spin-polarized systems, here we extend the approach to ferromagnetic ones. Unlike other embedded correlated wavefunction methods, such as embedded complete active space self-consistent field theory, emb-RPA is advantageous for spin-polarized systems because the RPA is compatible with unrestricted DFT solutions, which are eigenfunctions of the spin angular momentum operator S z but not the total spin-squared operator S 2 . By applying emb-RPA with specific magnetization constraints, we achieved a speedup of two to three orders of magnitude (one order when accounting for the one-time embedding potential optimization cost) with only small errors (∼50 meV) compared to full periodic RPA. Moreover, emb-RPA significantly reduces the over-binding errors of DFT approximations. In conclusion, we anticipate that the acceleration enabled by the spin-polarized emb-RPA approach will broaden the applicability of RPA to magnetic materials.

Density functional theory↗
DOE OSTI2021

Comparison of Radiation Models for a Turbulent Piloted Methane/Air Jet Flame: A Frozen-Field Study

Numerical modeling of radiative transfer in nongray reacting media is a challenging problem in computational science and engineering. The choice of radiation models is important for accurate and efficient high-fidelity combustion simulations. Different applications usually involve different degrees of complexity, so there is yet no consensus in the community. In this paper, the performance of different radiative transfer equation (RTE) solvers and spectral models for a turbulent piloted methane/air jet flame are studied. The flame is scaled from the Sandia Flame D with a Reynolds number of 22,400. Three classes of RTE solvers, namely the discrete ordinates method, spherical harmonics method, and Monte Carlo method, are examined. The spectral models include the Planck-mean model, the full-spectrum k-distribution (FSK) method, and the line-by-line (LBL) calculation. The performances of different radiation models in terms of accuracy and computational cost are benchmarked. The results have shown that both RTE solvers and spectral models are critical in the prediction of radiative heat source terms for this jet flame. The trade-offs between the accuracy, the computational cost, and the implementation difficulty are discussed in detail. The results can be used as a reference for radiation model selection in combustor simulations.

David, Chloe↗
NASA NTRS1997

Simplifying silicon burning: Application of quasi-equilibrium to (alpha) network nucleosynthesis

While the need for accurate calculation of nucleosynthesis and the resulting rate of thermonuclear energy release within hydrodynamic models of stars and supernovae is clear, the computational expense of these nucleosynthesis calculations often force a compromise in accuracy to reduce the computational cost. To redress this trade-off of accuracy for speed, the authors present an improved nuclear network which takes advantage of quasi- equilibrium in order to reduce the number of independent nuclei, and hence the computational cost of nucleosynthesis, without significant reduction in accuracy. In this paper they will discuss the first application of this method, the further reduction in size of the minimal alpha network. The resultant QSE- reduced alpha network is twice as fast as the conventional alpha network it replaces and requires the tracking of half as many abundance variables, while accurately estimating the rate of energy generation. Such reduction in cost is particularly necessary for future generation of multi-dimensional models for supernovae.

Hix, W. R.↗
DOE OSTI2024

RandONets: Shallow networks with random projections for learning linear and nonlinear operators

Deep neural networks have been extensively used for the solution of both the forward and the inverse problem for dynamical systems. However, their implementation necessitates optimizing a high-dimensional space of parameters and hyperparameters. This fact, along with the requirement of substantial computational resources, pose a barrier to achieving high numerical accuracy, but also interpretability. Here, to address the above challenges, we present Random Projection-based Operator Networks (RandONets): shallow networks with random projections and tailor-made numerical analysis methods that learn accurately and fast linear and nonlinear operators. Building on previous works, we prove that RandOnets are universal approximators of linear and nonlinear operators. Due to their simplicity, RandONets provide a one-step transformation of the input space, facilitating interpretability. For the evaluation of their performance, we focus on operators of PDEs. We show, that RandONets outperform by several orders of magnitude, both in terms of numerical approximation accuracy and computational cost, the “vanilla” DeepONets. Hence, we believe that our method will trigger further developments in the field of scientific machine learning, for the development of new ‘’light”schemes that will provide high accuracy while reducing dramatically the computational cost. A MATLAB toolbox for RandONets, including demos, is available on GitHub at https://github.com/GianlucaFabiani/RandONets.

Interpretable machine learning↗
DOE OSTI2024

Physics-assisted Latent Space Dynamics Learning for Stiff Collisional-radiative Models

Collisional-radiative (CR) models describe the atomic processes in a plasma by tracking the population density in the ground and excited states for each charge state of the atom/ion. These models predict important plasma properties such as charge state distributions and radiative emissivity and opacity. Accurate descriptions of the CR balance of the plasma are essential in fusion whole device modeling, especially when significant impurities are introduced into the plasmas. In an integrated fusion plasma and CR simulation, a CR model, which is a high-dimensional stiff ODE, needs to be solved on each grid point in the configuration space, and can overwhelm the overall computational cost. In this work, we propose a machine-learning-based method that discovers a latent space and learns its corresponding latent dynamics, which can capture the essential physics to make accurate predictions at much lower online computational cost. The proposed approach is physics-assisted, due to its combination of a physical latent space and a data-driven latent space. It has been demonstrated that the proposed architecture can predict both the full-order CR dynamics and a physical quantity of interest accurately.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
NASA NTRS2001

A Rapid Aerodynamic Design Procedure Based on Artificial Neural Networks

An aerodynamic design procedure that uses neural networks to model the functional behavior of the objective function in design space has been developed. This method incorporates several improvements to an earlier method that employed a strategy called parameter-based partitioning of the design space in order to reduce the computational costs associated with design optimization. As with the earlier method, the current method uses a sequence of response surfaces to traverse the design space in search of the optimal solution. The new method yields significant reductions in computational costs by using composite response surfaces with better generalization capabilities and by exploiting synergies between the optimization method and the simulation codes used to generate the training data. These reductions in design optimization costs are demonstrated for a turbine airfoil design study where a generic shape is evolved into an optimal airfoil.

Rai, Man Mohan↗
DOE OSTI2020

Neural-based time series forecasting of loss of coolant accidents in nuclear power plants

During the last few years, deep learning in neural networks has demonstrated impressive successes in the areas of computer vision, speech and image recognition, text generation, and many others. However, sensitive engineering areas such as nuclear engineering benefited less from these efficient techniques. In this work, deep learning expert systems are utilized to model and predict time series progression of a design-basis nuclear accident, featuring a loss of coolant accident. Two major findings are accomplished in this work. First, the ability to train expert systems with high accuracy, which could help nuclear power plant operators to figure out plant responses during the accident. Second, building fast, efficient, and accurate deep models to simulate nuclear phenomena, which could be valuable to nuclear computational science. In this work, large amount of time series data is obtained from simulation tools by simulating different conditions of the base-case/nominal accident scenario. Four critical outputs/responses are monitored during the accident (e.g. temperature, pressure, break flow rate, water level). Two approaches are adopted in this work. The first approach is to use feedforward deep neural networks (DNN) to fit all time steps and outputs in a single model. The second approach is to use long short-term memory (LSTM) to fit all time steps together for each reactor response separately. Both DNN and LSTM demonstrate very good performance in predicting the test and base-case scenarios, with accuracy as low as 92% and as high as 99%, where these test scenarios are unknown to the expert systems and are not included in the model training. In addition, both approaches demonstrate a significant reduction in computational costs, as the deep expert system is able to accurately predict the accident 100,000 times faster than the original simulation tool. Given sufficient data, the methodology adopted in this study demonstrates that DNN/LSTM expert systems can be used as a decision support system to model advanced time series phenomena within nuclear power plants with high accuracy and negligible computational costs.

42 ENGINEERING↗
DOE OSTI2020

Newton versus the machine: solving the chaotic three-body problem using deep neural networks

ABSTRACT Since its formulation by Sir Isaac Newton, the problem of solving the equations of motion for three bodies under their own gravitational force has remained practically unsolved. Currently, the solution for a given initialization can only be found by performing laborious iterative calculations that have unpredictable and potentially infinite computational cost, due to the system’s chaotic nature. We show that an ensemble of converged solutions for the planar chaotic three-body problem obtained using an arbitrarily precise numerical integrator can be used to train a deep artificial neural network (ANN) that, over a bounded time interval, provides accurate solutions at a fixed computational cost and up to 100 million times faster than the numerical integrator. In addition, we demonstrate the importance of training an ANN using converged solutions from an arbitrary precise integrator, relative to solutions computed by a conventional fixed precision integrator, which can introduce errors in the training data, due to numerical round-off and time discretization, that are learned by the ANN. Our results provide evidence that, for computationally challenging regions of phase space, a trained ANN can replace existing numerical solvers, enabling fast and scalable simulations of many-body systems to shed light on outstanding phenomena such as the formation of black hole binary systems or the origin of the core collapse in dense star clusters.

Breen, Philip G.↗
DOE OSTI2023

Tensor-based flow reconstruction from optimally located sensor measurements

Reconstructing high-resolution flow fields from sparse measurements is a major challenge in fluid dynamics. Existing methods often vectorize the flow by stacking different spatial directions on top of each other, hence confounding the information encoded in different dimensions. Here, we introduce a tensor-based sensor placement and flow reconstruction method which retains and exploits the inherent multidimensionality of the flow. We derive estimates for the flow reconstruction error, storage requirements and computational cost of our method. We show, with examples, that our tensor-based method is significantly more accurate than similar vectorized methods. Furthermore, the variance of the error is smaller when using our tensor-based method. While the computational cost of our method is comparable to similar vectorized methods, it reduces the storage cost by several orders of magnitude. The reduced storage cost becomes even more pronounced as the dimension of the flow increases. We demonstrate the efficacy of our method on three examples: a chaotic Kolmogorov flow, in situ and satellite measurements of the global sea surface temperature and three-dimensional unsteady simulated flow around a marine research vessel.

Mechanics↗
DOE OSTI2021

Modeling laser-driven ion acceleration with deep learning

Developments in machine learning promise to ameliorate some of the challenges of modeling complex physical systems through neural-network-based surrogate models. High-intensity, short-pulse lasers can be used to accelerate ions to mega-electronvolt energies, but to model such interactions requires computationally expensive techniques such as particle-in-cell simulations. Multilayer neural networks allow one to take a relatively sparse ensemble of simulations and generate a surrogate model that can be used to rapidly search the parameter space of interest. In this study, we created an ensemble of over 1,000 simulations modeling laser-driven ion acceleration and developed a surrogate to study the resulting parameter space. A neural-network-based approach allows for rapid feature discovery not possible for traditional parameter scans given the computational cost. A notable observation made during this study was the dependence of ion energy on the pre-plasma gradient length scale. While this methodology harbors great promise for ion acceleration, it has ready application to all topics in which large-scale parameter scans are restricted by significant computational cost or relatively large, but sparse, domains.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
NASA NTRS2018

Hybrid Reduced Order Model for N2-N2 Interactions for Application to Dissociation and Energy Transfer Processes

Recent work in the aerothermodynamics community has focused on the development of reduced-order models for thermo-chemical non-equilibrium which avoid the restrictive assumptions of multi-temperature models and the prohibitive cost associated with State-to-State (StS) models. In the present work, this is accomplished by lumping energy states together and assuming groups of states are roughly in equilibrium. As a result, the non-equilibrium behavior of a gas can be captured at a reduced computational cost from a full StS simulation. In this work, we present a hybrid grouping model for studying energy transfer and dissociation in a mixture of nitrogen molecules due to N2-N2 reactions. This is accomplished by making use of a grouping strategy informed by data from the N2-N StS kinetic data. However, due to the massive computational cost associated with constructing StS data for the N2-N2 system, the kinetic data for the hybrid grouping model are calculated using the quasi-classical trajectory (QCT) method by sampling states for trajectory within the groups. This general framework is called the Maximum-Entropy Quasi-Classical Trajectory (ME-QCT) method. The primary challenge associated with this method is that rates for reverse grouped reactions cannot be obtained through detailed balance at a group level, due to the variation of group internal temperatures. To construct the full model for N2-N2 grouped kinetics using the ME-QCT method, detailed balance is invoked at the microscopic level, allowing for the calculation of the full kinetic data from QCT. Results will be presented using the full ME-QCT model for the N2-N2 system in an isothermal and isochoric reactor simulation. In addition, simple CFD test cases for a one-dimensional standing shock and for a quasi-one-dimensional nozzle will be used for demonstration of the ME-QCT method. This method allows for the calculation of non-equilibrium behavior for the N2-N2 system without the prohibitive cost of a full StS simulation. Moreover, it enables the construction of a unified model for the dissociating and recombining non-equilibrium flows.

nonequilibrium modeling↗
DOE OSTI2022

Hybrid RANS-LES of the Atmospheric Boundary Layer for Wind Farm Simulations: Preprint

Wind farm simulations often do not accurately represent wake-atmospheric boundary layer (ABL) interactions, blade boundary layer (BL) dynamics, and turbine-turbine interactions. In this work, we use Active Model Split (AMS), a new hybrid Reynolds-Averaged Navier Stokes (RANS)-large eddy simulation (LES) model, which is well suited to capture these effects because the model can (i) accurately simulate the ABL with the Coriolis effect, (ii) is accurate in adverse pressure gradients such as those near wind turbine blades, and (iii) has sufficiently low computational cost to simulate multiple turbines while resolving the blade BL. For simplicity and consistency we develop AMS to be used throughout the domain rather than in a zonal method. We implement our work in the massively parallel flow solver, Nalu-Wind, so that our model can access the compute resources needed for blade-resolved simulations of multiple wind turbines. To accomplish these aims, we modify the baseline AMS by changing the RANS contribution to SST k - omega with a length scale limiter, adding the Coriolis effect, and developing an appropriate wall treatment. We show that AMS of the ABL with the Coriolis effect matches LES reference results better than those obtained with RANS. We describe our plans to add buoyancy effects and wind turbines to our AMS simulations.

atmospheric boundary layer↗
DOE OSTI2023

Unifying principles for catalytic hydrotreating processes (Final Technical Report)

This project builds on the hypothesis that the hydrotreating processes for the removal of oxygen and sulfur are fundamentally similar at the atomic-scale and existing knowledge from the treatment of petroleum derived feedstock can be leveraged for the design of novel catalysts for the upgrade of bio-oil. We tested this hypothesis by comparing computed potential energy diagrams for hydrodesulfurization (HDS) of thiophene over MoS 2 with hydrodeoxygenation (HDO) of furan over MoO 3 and concluded that certain aspects, such as catalyst promotion with transition metals, are valid strategies for both reactions. On the other hand, we also noticed significant differences in the mechanism for hydrogen (H 2 ) activation, which requires sites with metallic character. While MoS 2 is known to have metallic edge states that can catalyze H 2 dissociation, this elementary step is prohibitively slow on defect-free oxides. Only in the presence of vacancies or by creating metal/oxide interfaces can efficient H 2 activation sites during HDO be formed. The need for bifunctional catalyst when it comes to efficient and selective HDO or dehydrogenation reactions was further corroborated in joint experimental and theoretical studies of the Guerbet reaction for the coupling of biomass derived oxygenates over PdCu alloys, nitrate reduction over In-promoted Pd nanoparticles, and ethylene dehydroaromatization over Ga-exchanged ZSM-5 zeolites. All of these catalytic systems have in common that catalytic sites with distinct functional requirements are needed to create a working catalyst. Detailed computational studies were carried out for HDO of m -cresol and phenol on Ru-modified TiO 2 surfaces, which allowed us to attribute catalytic activity to the metal/oxide interface. A surprising finding was that heterolytic cleavage of the H-H bond across the Ru/TiO 2 interface was critically important, despite lower barriers for homolytic H 2 activation on Ru metal. The explanation lies in the high barriers for hydrogen spillover from Ru to TiO 2 , which becomes unnecessary in the heterolytic activation pathway. Moreover, we also reported that proton transfer steps between metal and oxide sites are mediated by weakly adsorbed surface water. During attempts to develop and validate a kinetic Monte Carlo (kMC) model for HDO reactions at the Ru/TiO 2 interface, it became clear that lateral interactions are paramount to describe realistic surface chemistry and without these interactions, the reduction and hydroxylation behavior from our simulations was inconsistent with reported experiments. To assess the importance of lateral interactions in popular computational catalyst design strategies relying on the identification of reactivity descriptors, which can be used along with Brønsted–Evans–Polanyi (BEP) and scaling relations as input to a microkinetic model (MKM) to make predictions for activity or selectivity trends, we compared predicted trends with those obtained from descriptor-based kMC models. We critically evaluated the benefits of kMC over MKM in terms of trend predictions and computational cost when using only a small set of input parameters. After confirming that in the absence of lateral interactions the kMC and MKM approaches yield identical trends and mechanistic information, we observed substantial differences between the two kinetic models when lateral interactions were introduced. The mean-field implementation applies coverage corrections directly to the descriptors, causing an artificial overprediction of the activity of strongly binding metals. In contrast, the cluster expansion in kMC implementation can differentiate among the highly active metals but it is very sensitive to the set of included interaction parameters. Considering that computational screening relies on a minimal set of descriptors, for which MKM makes reasonable trend predictions at a ca. three orders of magnitude lower computational cost than kMC, we concluded that the MKM approach does provide an overall better entry point for computational catalyst design. Overall, this project has led to 11 peer-reviewed publications, and their scientific is impact is well illustrated by their combined 702 citations.

08 HYDROGEN↗
DOE OSTI2023

Subspace recursive Fermi-operator expansion strategies for large-scale DFT eigenvalue problems on HPC architectures

Quantum mechanical calculations for material modeling using Kohn–Sham density functional theory (DFT) involve the solution of a nonlinear eigenvalue problem for N smallest eigenvector-eigenvalue pairs, with N proportional to the number of electrons in the material system. Here, these calculations are computationally demanding and have asymptotic cubic scaling complexity with the number of electrons. Large-scale matrix eigenvalue problems arising from the discretization of the Kohn–Sham DFT equations employing a systematically convergent basis traditionally rely on iterative orthogonal projection methods, which are shown to be computationally efficient and scalable on massively parallel computing architectures. However, as the size of the material system increases, these methods are known to incur dominant computational costs through the Rayleigh–Ritz projection step of the discretized Kohn–Sham Hamiltonian matrix and the subsequent subspace diagonalization of the projected matrix. This work explores the potential of polynomial expansion approaches based on recursive Fermi-operator expansion as an alternative to the subspace diagonalization of the projected Hamiltonian matrix to reduce the computational cost. Subsequently, we perform a detailed comparison of various recursive polynomial expansion approaches to the traditional approach of explicit diagonalization on both multi-node central processing unit and graphics processing unit architectures and assess their relative performance in terms of accuracy, computational efficiency, scaling behavior, and energy efficiency.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

Predicting Partial Atomic Charges in Metal–Organic Frameworks: An Extension to Ionic MOFs

Molecular simulation is an invaluable tool to predict and understand the usage of metal–organic frameworks (MOFs) for gas storage and separation applications. Accurate partial atomic charges, commonly obtained from density functional theory (DFT) calculations, are often required to model the electrostatic interactions between the MOF and adsorbates, especially when the adsorbates have dipole or quadrupole moments, such as water and CO 2 . Machine learning (ML) models have been previously employed to predict partial charges and avoid the computational cost associated with DFT calculations. However, previous ML models suffer from small training data sets, which limit their scope of application. In this work, we introduce two novel machine learning models, PACMOF2-neutral and PACMOF2-ionic, aimed at predicting the density-derived electrostatic and chemical (DDEC6) partial atomic charges for both neutral and ionic MOFs. These models not only yield DFT-level accuracy at a fraction of the computational cost but also demonstrate a remarkable improvement in prediction of adsorption, as validated with grand canonical Monte Carlo simulations. Furthermore, the robustness and fast computational time of the PACMOF2 models, along with their transferability to other porous materials such as covalent organic frameworks and zeolites, underscores their potential in high-throughput screening of MOFs for diverse applications.

36 MATERIALS SCIENCE↗
DOE OSTI2025

A Comparative Study of Physics‐Informed and Data‐Driven Neural Networks for Compound Flood Simulation at River‐Ocean Interfaces: A Case Study of Hurricane Irene

Simulating compound flooding (CF) at the river-ocean interface within large-scale Earth System Models (ESMs) presents significant challenges due to complex interactions between river discharge, storm surge, and tides. This study assesses the comparative advantages of physics-informed and data-driven machine learning (ML) approaches for enhancing local ESM performance. We systematically compare data-driven neural network models (i.e., CNNs, U-Net, Long Short-Term Memory (LSTM), Gated Recurrent Unit), and physics-informed neural network (PINN) models, including vanilla PINN and a finite-difference-based PINN (FD-PINN). Specifically, FD-PINN is introduced to enhance computational efficiency, accelerating vanilla PINNs by ∼6.5 times while improving accuracy. To enhance data-driven model training, a new data-generation approach is developed to sample historical fluvial and coastal flood events, which ensures a robust data set for extreme event prediction. The models are evaluated using a realistic one-dimensional river domain extracted from an ESM's river mesh and the Hurricane Irene event as an independent test case. Results show that FD-PINN achieves accurate predictions with significantly reduced computational costs relative to vanilla PINNs. Among data-driven models, the best overall performance is achieved by a CNN-LSTM hybrid, which balances accuracy and efficiency. While a fully connected CNN (CNN-FC) provides the best accuracy, it incurs high computational cost. Architectures lacking strong temporal modeling tend to underperform on unseen events. These findings highlight the importance of sequence-aware designs for robust generalization. This study reveals the trade-offs between physics-informed and data-driven models and proposes an adaptive hybrid framework for integrating ML into ESMs to enhance local flood simulations.

Earth Systems Modeling↗
DOE OSTI2025

Multidimensional Modeling of Mixture Formation in a Hydrogen-Fueled Heavy-Duty Optical Engine With Direct Injection

Hydrogen (H 2 ), as a carbon-free fuel, is considered as one of the most promising solutions to reduce the carbon footprint of hard-to-decarbonize energy and transportation sectors. As such, hydrogen-fueled internal combustion engines (H 2 ICEs) have recently been receiving increasing attention, particularly in applications such as on-road/off-road heavy-duty transport and combined heat and power. The direct injection (DI) of gaseous hydrogen into the combustion chamber offers great potential for achieving high power density and high engine efficiency, while mitigating the risk of backfire and reducing pre-ignition. However, the numerical simulation of H 2 DI system remains a formidable challenge associated with the high computational cost of reproducing compressible supersonic flow and shocks in narrow injector passages and in near-nozzle regions. In general, there is a lack of well-established and validated practices for the modeling of high-pressure H 2 DI in large-bore engines. Here, to this end, this study focuses on computational fluid dynamics (CFD) modeling of the mixture formation process in a heavy-duty optical engine employing a medium-pressure H 2 DI system. Both large eddy simulations (LES) and Reynolds Averaged Navier–Stokes (RANS) simulations are performed and evaluated against optical data. Gaseous hydrogen is injected into the combustion chamber via a centrally located outward opening hollow-cone injector at a pressure of 40 bar. Simulations are carried out for two injection timings, namely, −120 and −60 °CA. The numerical predictions for H 2 distribution in different horizontal and vertical planes during the compression stroke are systematically compared against optical data obtained through planar laser-induced fluorescence (PLIF) measurements. Overall, the LES approach using the Dynamic Structure model is found to have good predictive capabilities for the early jet penetration in terms of length and shape, as well as the later H 2 distributions. However, the unsteady RANS approach with the renormalization group $k - ϵ$ model, which is widely used by industry to model heavy-duty ICEs, significantly underpredicts the H 2 mixing, even at similar mesh resolution to that used in LES. These results indicate that there is a need for the improvement of mixing submodels within the RANS approach when applied to H 2 DI simulations.

LES↗