Conference PaperPDF Available

Visualization of Jet Impingement and Ignition in a Piston-cylinder Chamber

Authors:
Visualization of Jet Impingement and Ignition in a
Piston-cylinder Chamber
Nicholas Brunhart-Lupo
nicholas.brunhart-lupo@nrel.gov
National Renewable Energy Laboratory
Golden, Colorado, USA
Shashank Yellapantula
shashank.yellapantula@nrel.gov
National Renewable Energy Laboratory
Golden, Colorado, USA
Kenny Gruchalla
kenny.gruchalla@nrel.gov
National Renewable Energy Laboratory
Golden, Colorado, USA
Ray Grout
ray.grout@nrel.gov
National Renewable Energy Laboratory
Golden, Colorado, USA
Figure 1.
Three frames of temperature iso-volumes showing the impingement of one of the fuel jets and the ame stabilization
in regions of high residence time. Using advanced rendering techniques (i.e., emission and reection) provided a better
understanding of the relationship between the ow and the chamber geometry.
Abstract
The transportation sector accounts for almost a third of the
United States’ energy consumption. We are developing pre-
dictive simulations of the complex in-cylinder processes of
internal combustion engines to improve performance and
decrease pollutant emissions. We provide a rst look at the
jet impingement and ignition inside a piston-cylinder cham-
ber using volume rendering. Through the visualization, we
constructed a complete 3D picture of the jet impingement
on the piston-bowl walls leading to the formation of pockets
of auto-ignition of the fuel-air mixture. These auto-ignition
pockets subsequently lead to ame stabilization, releasing
the chemical energy for conversion to useful work. A de-
tailed physical picture of ame stabilization, as provided by
Permission to make digital or hard copies of all or part of this work for
personal or classroom use is granted without fee provided that copies are not
made or distributed for prot or commercial advantage and that copies bear
this notice and the full citation on the rst page. Copyrights for components
of this work owned by others than ACM must be honored. Abstracting with
credit is permitted. To copy otherwise, or republish, to post on servers or to
redistribute to lists, requires prior specic permission and/or a fee. Request
permissions from permissions@acm.org.
e-Energy ’21, June 28–July 2, 2021, Virtual Event, Italy
©2021 Association for Computing Machinery.
ACM ISBN 978-1-4503-8333-2/21/06. . . $15.00
hps://doi.org/10.1145/3447555.3466595
the visualization techniques presented in this study, help in
the design of next-generation internal combustion engines.
CCS Concepts Human-centered computing Sci-
entic visualization.
Keywords
combustion, scientic visualization, high per-
formance computing
ACM Reference Format:
Nicholas Brunhart-Lupo, Shashank Yellapantula, Kenny Gruchalla,
and Ray Grout. 2021. Visualization of Jet Impingement and Ignition
in a Piston-cylinder Chamber. In The Twelfth ACM International
Conference on Future Energy Systems (e-Energy ’21), June 28–July
2, 2021, Virtual Event, Italy. ACM, New York, NY, USA, 7 pages.
hps://doi.org/10.1145/3447555.3466595
1 Introduction
In 2019, the United States transportation sector accounted for
28% of the U.S. energy use and 28.5% of the greenhouse gas
emissions, with 91% of that energy originating from internal
combustion engines [
8
]. The inside of an internal combustion
engine is a highly complex and tumultuous environment. As
high performance computing moves into the exascale era,
we can begin to numerically resolve more of the turbulence
length and time scales in high-delity combustion simula-
tions, allowing us to study in detail the turbulence-chemistry
e-Energy ’21, June 28–July 2, 2021, Virtual Event, Italy Brunhart-Lupo, et al.
interaction and develop combustion devices for increased
eciency and decreased emissions.
We are using a variety of volume rendering tools and tech-
niques to gain a qualitative understanding of the interaction
between chemical kinetics and fuel-air mixture preparation
inside an advanced compression ignition automotive engine.
In compression ignition engines, the initiation of combustion
occurs through auto-ignition of the fuel when it is mixed
with hot compressed air in the piston cylinder. This combus-
tion of fuel releases the chemical energy stored in fuel bonds
to generate useful work. Auto-ignition of fuel-air mixture
is highly dependent on the ratio of fuel and air in the mix-
ture, also known as fuel-air equivalence ratio,
𝜙
. The latest
technologies rely on compression ignition where mixture
preparation is critical, and the interaction between the fuel
injection and the combustion-chamber geometry establishes
the ow features that drive the mixture preparation.
In the current study, we performed a high-delity simula-
tion of fuel jet injection and impingement on the piston bowl,
as observed in production engine hardware. The main objec-
tive of this study was to investigate mixing, auto-ignition,
and subsequent combustion in a realistic setting. We used a
variety of volume rendering tools and techniques, as outlined
in the paper, to get a rst-look qualitative understanding of
how the geometry of the combustion chamber impacts ow
features and thus inuencing mixing and combustion.
2 Description of the Simulation
Production and research engines use pistons with varying
geometry, ranging from at to complex re-entrant bowls
that concentrate and enhance the ow structures. We used
a piston bowl conguration resembling a production tur-
bocharged diesel engine for this study (see Figure 3). In com-
pression ignition engines, fuel is typically added near the
top of the compression stroke through a multi-hole injector,
with the details of the injector design specic to the fuels and
operating mode under study. For this study, we inject gas-
phase fuel through seven discrete jets into high-temperature
air as a generic conguration.
We initialized the velocity eld within the cylinder by
constructing a velocity eld from a homogeneous isotropic
turbulence (HIT) simulation. This background turbulence
eld is instrumental in introducing variations in the trajec-
tory of each jet. We inject seven evenly spaced gaseous
CH4
jets arranged on a circle with a diameter of
0.025 m
. The
jets were injected at an angle of
30°
. A parabolic prole, as
a function of radius, was imparted to the injection velocity.
The simulation parameters are summarized in the appendix
(Table 1).
The reacting simulation is performed using PeleC [
9
], a
code currently under development as a part of the DOE
Oce of Science’s Exascale computing project. PeleC is a
reacting ow code, solving a fully compressible version of
the Navier-Stokes equation with chemical reactions. PeleC
uses the adaptive mesh renement library AMReX to adapt
the simulation grid on the y to eciently resolve localized
features. PeleC uses an embedded boundary methodology to
simulate complex geometries without the need for a meshing
step. In the current study, cells were tagged for renement
based on a temperature gradient criteria of 𝑇>75.0 K.
3 Visualization
We are primarily interested in understanding the interac-
tion of the ow with the geometry of the piston bowl (i.e.,
the jet impingement in the piston-cylinder chamber) and
how it impacts the mixing between fuel and air. However,
visualizing that interaction is dicult without the geometry
occluding the ow or the ow occluding the geometry. We
found that we were able to get the best qualitative sense of
the jet impingement through a combination of illumination
and reection. We produced this illumination-based visual-
ization of the simulation using Blender [
17
], with simulation
data exported through a YT Project [18] workow.
3.1 Cinematic Rendering
Standard scientic visualization tooling, such as Paraview [
1
],
and YT provide ecient volume rendering pipelines, with
a number of features that can be used to tease out various
aspects of the data. In the case of YT, for example, transfer
functions for the volume renderer are multidimensional, and
thus gives great control and option to the visualization scien-
tist [
10
]. However, these tools do not provide more advanced
rendering capabilities, like scattering, global illumination,
and reections, which can be useful in highlighting data
characteristics [
2
,
14
]. While these advanced rendering tech-
niques are gaining usage in medical imaging [
6
,
7
,
15
], cine-
matic scientic volume rendering has largely been restricted
to visualization for communication [
3
] not analysis. The gen-
eral lack of support for advanced rendering in these parallel
visualization tools is likely due to the inherent diculties of
the path-tracing problem in distributed contexts [16].
Blender’s rendering capabilities are geared for cinematic
video production, not scientic visualization. Blender has
limited data mapping and ltering capabilities and cannot
be easily used in a distributed manner (i.e., split frame ren-
dering). However, it does provide many advanced render-
ing features. The volumetric path provides great control
over volume rendering, providing physically-based volume
absorption, scattering, and emission, with arbitrary user-
written functions to control each, even separately, if desired.
It also provides strong animation control, which is more
easily available and simpler to use than in other packages.
To compare approaches, consider the two frames pre-
sented in Figure 2. On the left is a frame as rendered from
Paraview’s built-in volume renderer, and the right is a frame
Visualization of Jet Impingement and Ignition in a Piston-cylinder Chamber e-Energy ’21, June 28–July 2, 2021, Virtual Event, Italy
(a) Paraview (b) Blender
Figure 2.
Comparison between Paraview and Blender products. Reections are seen in Figure 2b, with more denition to the
volumetric surface as provided by scattering.
as rendered in Blender. Figure 2b shows the eects of scatter-
ing and absorption, which provides a solidity to the volume.
Note that the underlying cylinder exhibits reection, provid-
ing a sense of rearward-facing surfaces and an indication of
depth. We have also applied other eects to the cylinder to
highlight the geometry while also limiting obfuscation of
volume structure behind the surface, such as giving back-
faces lesser opacity and the hard geometry a rough glass
material with emission to ensure visibility.
This is not an issue limited to Paraview and YT. For ex-
ample, when using OSPRay, a CPU-only ray-tracer built on
top of Intel’s Embree [
19
], some shadowing is apparent un-
der light sources; however, the volume itself cannot emit
light, and there are no reections
1
. Another option, also inte-
grated into Paraview, is NVIDIA’s IndeX volume rendering
software development kit[
13
]. This kit provides a richer set
of features than Paraview’s built-in renderer, such as edge
detection and gradient lighting options, but again, is aimed
at distributed rendering and does not support emission or
reective geometry.
The primary challenge to using Blender is that it does
not provide a simple facility to directly import volume data
into its Cycles renderer
2
. The rendering code is intended
to be used in conjunction with cinematic eects, such as
smoke and re simulation that are generated from within
the package itself. To circumvent this limitation, we sepa-
rate the volume into slices in the
𝑧
-axis and pack them in a
standard 2D image. Combined with an unpacking routine
implemented in a shader, this provides a usable 3D volume
sampling.
3.2 Data Workow
The purpose of the workow is to transform the adaptive
mesh renement data into a rectilinear grid format that is
amenable to Blender.
1As of Paraview 5.7.
2
At the time of this project, experimental OpenVDB [
12
] exists in some
builds but is immature.
Using the YT Project library, we ingested the data and
resampled it into a smoothed covering grid. We sampled the
grids at the minimum resolution of the AMR grid. Special
care must be taken when interpolating between dierent
AMR renements to avoid block artifacts.
After resampling, we exported the data as an HDF5 dataset
(one per time-step), along with an XMDF metadata descriptor.
We then explored the data using Paraview to nd illustrative
datasets, eective colormaps, data bounds, and to check for
artifacts.
Once we selected a subset of the data, we exported these
through a second set of python scripts. These scripts took
a requested dataset from an HDF5 le, normalized it, given
provided bounds, and packed the volume into a single 16-bit
greyscale PNG image (see Appendix Fig. 4). We accomplished
the packing by tiling slices of the data in the
𝑧
-dimension in a
grid, with care taken to not exceed CUDA texture dimension
limits of 65536
×
32768. Thus, for our volume with dimensions
of 512
×
512
×
192, we can construct an image of dimensions
1536 ×32768, with 𝑧-slices tiled 3×64.
The volume shader in Blender contains a mapper that can
look up the correct tile given the
𝑥
,
𝑦
, and
𝑧
generated mesh
coordinates (see Appendix Fig. 5).
3.3 Render
We produced video frames with Blender’s Cycles renderer.
Blender provides a ‘Principled Volume’ shader which com-
bines several previously disparate shading elements into a
single node that can provide volume density, scattering, and
emission eects. Appendix Figure 5 shows the volume shader
used in drawing the volume.
We rendered the individual frames at 1080p on an 8-node
GPU cluster. The nodes are equipped with two Intel Xeon
E5-2640v4 CPUs and an Nvidia K80. Approximate rendering
time was 8 to 16 minutes per frame. High render times re-
sulted from increased sampling (576 antialiased) and small
volume stepping size to both reduce noise and banding ar-
tifacts. Using structure-aware denoising was only partially
eective; volumes take on a ‘paintbrush’ look with features
e-Energy ’21, June 28–July 2, 2021, Virtual Event, Italy Brunhart-Lupo, et al.
smeared in regions where the background is visible through
the volume, or at the interface where colors shift dramatically
and are intermixed with noise.
4 Discussion
Advanced visualization techniques described in the previous
section proved to be critical in understanding key physics
related to fuel jet mixing, impingement, and subsequently
combustion in the high-delity open piston-cylinder simu-
lation. As mentioned in the Sec. 2, the simulation setup is
purely axisymmetric in terms of boundary conditions, but
due to the background turbulent ow eld, each of the fuel
jets follow a dierent trajectory with varying levels of mix-
ing with hot air in the chamber. Due to high jet velocity
leading to a low residence time of fuel-air pockets, no auto-
ignition is observed in the initial times. At later times, we can
observe jet impingement on the piston bowl walls leading
to a disintegration of fuel jets and the formation of high res-
idence time pockets with a high propensity of ignition. We
can see these pockets auto-ignite rst releasing heat and in-
creasing the temperature of the mixture. Subsequently, these
high-temperature regions mix with fresh un-burnt fuel-air
mixture leading to the formation of more high-temperature
regions, which results in stabilizing a ame in the jet periph-
ery. Appendix Fig. 6 exhibits selected frames from the video
showing the evolution of a jet.
One of the key contributions of the visualization tech-
niques was in improving the physical and the geometrical
interactions of the role played by jet impingement on auto-
ignition and subsequent ame stabilization. Cylinder mor-
phology tends to cause a high dispersion of regions that
feature a propensity of auto-ignition of the fuel-air mixture.
This high spatial variation in location of these auto-ignition
kernels makes the problem of characterizing their behavior
challenging. The location of these auto-ignition kernels plays
an important role in determining the transient evolution of
these kernels into a stable ame or extinguished kernels.
For instance, the cylinder walls are typically cold, and heat
transfer to walls could extinguish an auto-ignition kernel
situated adjacent to the piston-cylinder walls. Through the
advanced techniques used here, it was less challenging to see
the actual impact interface of the ame on the cup geometry
(as seen in Figure 2). Reection helped us judge distances
between features, as well as helped ght the occlusion of
these auto-ignition kernels. Physically accurate volume ren-
dering provided a stronger intuition of the ame density
and structure– while primarily used for cinematic render-
ing, Blender also allowed us to explore the mappings of data
to density, scattering, and emission by using the available
interactive progressive rendering mode.
The use of Blender presents some limitations; rstly in
space. Data must be translated to the sliced format for con-
sumption, which adds to disk costs. However, as we use a
non-oating image format to store slices, the additional disk
cost is not overly burdensome from quantization. Next, there
is a limitation in drawing frames. Cinematic rendering of
any kind greatly increases the time-to-frame for any visu-
alization, due to path tracing, etc. Splitting a single frame
across multiple nodes is not practical, so single nodes must
be capable of loading the whole dataset and drawing. This
is less of an issue of memory as we are using compact im-
ages, however as data sizes increase, a single image will no
longer be able to hold all slices
3
. At times banding artifacts
can also be seen in the output. This is a Blender specic
issue; the volume pipeline does not permit randomization in
step-size/starting point. To combat this, we use very small
step sizes at the cost of an increased render time.
We note that this analysis lends itself to volume rendering,
as it allows researchers the ability to see, for example, areas
of high temperature that would otherwise be occluded by low
temperature structures. Other techniques, like ow visualiza-
tion, are also useful to properly analyze this data, however,
there is a disconnect in the visualization tool-sets. Visual-
ization tools don’t support cinematic volume rendering and
these ow techniques can be laborious to integrate into cin-
ematic packages. Therefore, cinematic rendering should be
(and is) used in composition with other tools.
4.1 Conclusion
In this study, we utilized advanced visualization techniques
to analyze high delity simulation of discrete gaseous fuel
jets injection in a production piston bowl chamber. Volume
rendering of simulation data provided important insights
about regions where auto-ignition occurred, subsequently
leading to ame stabilization. The role of piston-bowl geom-
etry on jet impingement and the formation of high residence
pockets of the fuel-air mixture was also highlighted through
novel visualization strategies. The video is available at [5].
In terms of visualization, we have traded scalability and
interactivity for advanced lighting modalities. Naturally, the
workows presented here will become intractable as these
combustion simulations scale. For that reason, we continue
to investigate and experiment with OpenVDB[
12
] support in
Blender as a pipeline for delivering high-delity volumes to
the renderer. The computational and communication issues
in scaling path-tracing to distributed environments remain
daunting; however, we believe that this work shows they are
worth addressing to bring these useful features to scientic
visualization tooling.
3Due to CUDA/image reader library limitations
Visualization of Jet Impingement and Ignition in a Piston-cylinder Chamber e-Energy ’21, June 28–July 2, 2021, Virtual Event, Italy
Acknowledgments
This work was authored by the National Renewable Energy
Laboratory, operated by Alliance for Sustainable Energy,
LLC, for the U.S. Department of Energy (DOE) under Con-
tract No. DE-AC36-08GO28308. This research was supported
by the Exascale Computing Project (17-SC-20-SC), a joint
project of the U.S. Department of Energy’s Oce of Science
and National Nuclear Security Administration, responsible
for delivering a capable exascale ecosystem, including soft-
ware, applications, and hardware technology, to support
the nation’s exascale computing imperative. The views ex-
pressed herein do not necessarily represent the views of the
DOE or the U.S. Government.
References
[1]
Utkarsh Ayachit. 2015. The ParaView Guide: A Parallel Visualization
Application. Kitware, Inc.
[2]
Dirk Bartz, Douglas Cunningham, Jan Fischer, and Christian Wallraven.
2008. The Role of Perception for Computer Graphics. In Eurographics
2008 - State of the Art Reports, Theoharis Theoharis and Philip Dutre
(Eds.). The Eurographics Association. hps://doi.org/10.2312/egst.
20081045
[3]
Kalina Borkiewicz, AJ Christensen, Helen-Nicole Kostis, Greg Shirah,
and Ryan Wyatt. 2019. Cinematic Scientic Visualization: The Art
of Communicating Science. In ACM SIGGRAPH 2019 Courses (SIG-
GRAPH ’19). Association for Computing Machinery, New York, NY,
USA, Article 5, 273 pages. hps://doi.org/10.1145/3305366.3328056
[4]
Cynthia Brewer and Mark Harrower. 2019. ColorBrewer2. hp://
colorbrewer2.org.
[5]
Nicholas Brunhart-Lupo, Shashank Yellapantula, Kenny Gruchalla,
and Ray Grout. 2021. Visualization of Jet Impingement and Ignition in
a Piston-cylinder Chamber (Video). hps://youtu.be/75CSgrxL-Bk.
[6]
L.C. Chu, P.T. Johnson, and E.K. Fishman. 2018. Cinematic rendering
of pancreatic neoplasms: preliminary observations and opportunities.
Abdominal Radiology 43 (2018), 3009–3015. hps://doi.org/10.1007/
s00261-018- 1559-3
[7]
E. Dappa, K. Higashigaito, J. Fornaro, S. Leschka, S. Wildermuth, and
H. Alkadhi. 2016. Cinematic rendering – an alternative to volume
rendering for 3D computed tomography imaging. Insights Imaging 7
(2016), 849–856. hps://doi.org/10.1007/s13244-016-0518-1
[8]
Stacy C Davis and Robert G Boundy. 2019. Transportation Energy
Data Book: Edition 39. Oak Ridge National Laboratory (2019), 455.
hps://doi.org/10.2172/1767864
[9]
R. Grout. 2019. AMReX-Combustion/PeleC. hps://github.com/
AMReX-Combustion/PeleC.
[10]
Kenny Gruchalla, Mark Rast, Elizabeth Bradley, John Clyne, and Pablo
Mininni. 2009. Visualization-Driven Structural and Statistical Analysis
of Turbulent Flows. In Advances in Intelligent Data Analysis VIII, Niall
Adams, Céline Robardet, Arno Siebes, and Jean-François Boulicaut
(Eds.). Lecture Notes in Computer Science, Vol. 5772. Springer Berlin /
Heidelberg, 321–332. hp://dx.doi.org/10.1007/978-3- 642-03915- 7_28
10.1007/978-3-642-03915-7_28.
[11]
J. D. Hunter. 2007. Matplotlib: A 2D graphics environment. Computing
in Science & Engineering 9, 3 (2007), 90–95. hps://doi.org/10.1109/
MCSE.2007.55
[12]
Ken Museth. 2013. VDB: High-Resolution Sparse Volumes with
Dynamic Topology. ACM Trans. Graph. 32 (07 2013), 27:1–27:22.
hps://doi.org/10.1145/2487228.2487235
[13] NVIDIA. 2019. IndeX. hps://developer.nvidia.com/index.
[14]
Bernhard Preim, Alexandra Baer, Douglas Cunningham, Tobias Isen-
berg, and Timo Ropinski. 2016. A Survey of Perceptually Motivated
3D Visualization of Medical Image Data. Comput. Graph. Forum 35, 3
(June 2016), 501–525. hps://doi.org/10.1111/cgf.12927
[15]
S.P. Rowe, P.T. Johnson, and E.K. Fishman. 2018. Initial experience
with cinematic rendering for chest cardiovascular imaging. The British
Journal of Radiology 91 (Feb 2018). Issue 1082. hps://doi.org/10.1259/
bjr.20170558
[16]
Min Shih, Silvio Rizzi, Joseph Insley, Thomas Uram, Venkatram Vish-
wanath, Mark Hereld, Michael E. Papka, and Kwan-Liu Ma. 2016.
Parallel distributed, GPU-accelerated, advanced lighting calculations
for large-scale volume visualization. In 2016 IEEE 6th Symposium
on Large Data Analysis and Visualization (LDAV). 47–55. hps:
//doi.org/10.1109/LDAV.2016.7874309
[17]
The Blender Foundation. 2019. Blender: A 3D Modelling and Rendering
Package. hp://www.blender.org.
[18]
M. J. Turk, B. D. Smith, J. S. Oishi, S. Skory, S. W. Skillman, T. Abel, and
M. L. Norman. 2011. yt: A Multi-code Analysis Toolkit for Astrophysi-
cal Simulation Data. The Astrophysical Journal Supplement Series 192,
Article 9 (Jan. 2011), 9 pages. hps://doi.org/10.1088/0067-0049/192/1/9
arXiv:astro-ph.IM/1011.3514
[19]
Ingo Wald, Sven Woop, Carsten Benthin, Gregory S. Johnson, and
Manfred Ernst. 2014. Embree: A Kernel Framework for Ecient CPU
Ray Tracing. ACM Trans. Graph. 33, 4, Article 143 (July 2014), 8 pages.
hps://doi.org/10.1145/2601097.2601199
A Additional Figures & Tables
Table 1.
Demonstration calculation conguration details
and resolution requirements
Domain size
𝐿𝑥=𝐿𝑦104 mm
𝐿𝑧39 mm
Initial Conditions
Chamber Air temperature, 𝑇𝑐1400 K
Chamber Air Pressure, 𝑃𝑐101 325.0 Pa
Initial velocity uctuation, 𝑢2.5 m s1
Boundary Conditions
Jet inlet velocity, 𝑈𝑗35 m s1
Jet temperature, 𝑇𝑗350 K
Injection hole diameter, 𝑑𝑗0.0016 m
Mesh Resolution
Base mesh, 𝐵0.1 mm
Base mesh, 𝑁𝑋 ×𝑁 𝑌 ×𝑁 𝑌 (256 ×256 ×96)
Levels of renement 1
E. resolution at nest level
203 ×106m
(512 ×512 ×192)
B Blender Conguration
Although there are a number of rendering packages, Blender
was chosen as it is an open-source solution for cinematic
rendering.
The solution presented here is not limited to this dataset,
and can be easily adapted to other simulation products by
generating the required slice-image-packs, and directing
Blender to those slices. As with other visualization tools,
e-Energy ’21, June 28–July 2, 2021, Virtual Event, Italy Brunhart-Lupo, et al.
Figure 3. Schematic of the simulation conguration.
Figure 4.
An example subset of a volume slice pack as provided to Blender. Each
𝑧
slice of the volume is tiled to a dierent
portion of the image.
the user can then select color mappings, tweak density map-
pings, etc.
Blender uses a binary le format, making communica-
tion of settings somewhat dicult. However, the core of the
volume rendering settings are captured in Figure 5.
We show temperature on the ‘GnBu’ colormap[
4
] with
values from
400 K
to
2400 K
. The quantity
𝑌𝐶𝐻4
, representing
the mass fraction of fuel in the mixture, is plotted using the
‘inferno’ colormap from Matplotlib[
11
] with a range (as a
fraction of volume) of 0 % to 80 %.
Visualization of Jet Impingement and Ignition in a Piston-cylinder Chamber e-Energy ’21, June 28–July 2, 2021, Virtual Event, Italy
Figure 5.
Volume rendering shader used. Volume coordinates are provided automatically and remapped so as to acquire
the volume sample from a tile-pack texture. The sample is then mapped to a color, an emissive value, and a density, which
parameterizes the volume shader. Note that while the image node uses linear ltering, the volume sampling process is cubic.
(a) Initial state (b) Fuel injection (c) Jet formation
(d) Plume contact (e) Plume impact spreading (f) Auto-ignition regions apparent
Figure 6.
Selected frames of temperature iso-volumes showing the evolution of the jets, their impingement, and the ame
stabilization. High temperature regions in subgure
f
indicate auto-ignition of the fuel and onset of ame stabilization. The
emission and reection (see
d-f
) help us better understand the relationship between the piston-chamber geometry and the jets.
ResearchGate has not been able to resolve any citations for this publication.
Conference Paper
Full-text available
The Advanced Visualization Lab (AVL) is part of the the National Center for Supercomputing Applications (NCSA) at the University of Illinois at Urbana-Champaign. The AVL is led by Professor Donna Cox, who coined the term "Renaissance Team", with the belief that bringing together specialists of diverse backgrounds creates a team that is greater than the sum of its parts, and members of the AVL team reflect that in our interdisciplinarity. We specialize in creating high-quality cinematic scientific visualizations of supercomputer simulations for public outreach.
Article
Full-text available
Pancreatic cancer is the third most common cause of cancer death and CT is the most commonly used modality for the initial evaluation of suspected pancreatic cancer. Post-processing of CT data into 2D multiplanar and 3D reconstructions has been shown to improve tumor visualization and assessment of tumor resectability compared to axial slices, and is considered the standard of care. Cinematic rendering is a new 3D-rendering technique that produces photorealistic images, and it has the potential to more accurately depict anatomic detail compared to traditional 3D reconstruction techniques. The purpose of this article is to describe the potential application of CR to imaging of pancreatic neoplasms. CR has the potential to improve visualization of subtle pancreatic neoplasms, differentiation of solid and cystic pancreatic neoplasms, assessment of local tumor extension and vascular invasion, and visualization of metastatic disease.
Article
Full-text available
Volume rendering (VR) represents today’s standard three-dimensional (3-D) image post-processing technique, and often is used to visualize complex anatomical information. Recently, a novel 3-D technique for post-processing of computed tomography (CT) image data has been introduced, which is called cinematic rendering (CR). The objective of this review is to illustrate the image appearance and potential value of CR in comparison with conventional VR in a number of various applications and different anatomical regions. Similar to VR, CR best visualizes high density and high contrast structures such as bones and contrast-enhanced vessels, but at the same time provides a more natural and photo-realistic illumination of the rendered data. Further research will be necessary for determining possible advantages of CR over conventional VR and over two-dimensional (2-D) image post-processing for CT image data. Teaching Points • Cinematic rendering is a novel post-processing technique for 3D visualization of CT image data. • Compared to volume rendering, CR results in a more photo-realistic representation of anatomy. • Similar to volume rendering, CR provides best image quality of high density structures.
Article
Full-text available
This survey provides an overview of perceptually motivated techniques for the visualization of medical image data, including physics-based lighting techniques as well as illustrative rendering that incorporate spatial depth and shape cues. Additionally, we discuss evaluations that were conducted in order to study the perceptual effects of these visualization techniques as compared to conventional techniques. These evaluations assessed depth and shape perception with depth judgment, orientation matching, and related tasks. This overview of existing techniques and their evaluation serves as a basis for defining the evaluation process of medical visualizations and to discuss a research agenda.
Conference Paper
Full-text available
We describe Embree, an open source ray tracing framework for x86 CPUs. Embree is explicitly designed to achieve high performance in professional rendering environments in which complex geometry and incoherent ray distributions are common. Embree consists of a set of low-level kernels that maximize utilization of modern CPU architectures, and an API which enables these kernels to be used in existing renderers with minimal programmer effort. In this paper, we describe the design goals and software architecture of Embree, and show that for secondary rays in particular, the performance of Embree is competitive with (and often higher than) existing state-of-the-art methods on CPUs and GPUs.
Article
Full-text available
The analysis of complex multiphysics astrophysical simulations presents a unique and rapidly growing set of challenges: reproducibility, parallelization, and vast increases in data size and complexity chief among them. In order to meet these challenges, and in order to open up new avenues for collaboration between users of multiple simulation platforms, we present yt (available at http://yt.enzotools.org/) an open source, community-developed astrophysical analysis and visualization toolkit. Analysis and visualization with yt are oriented around physically relevant quantities rather than quantities native to astrophysical simulation codes. While originally designed for handling Enzo's structure adaptive mesh refinement data, yt has been extended to work with several different simulation methods and simulation codes including Orion, RAMSES, and FLASH. We report on its methods for reading, handling, and visualizing data, including projections, multivariate volume rendering, multi-dimensional histograms, halo finding, light cone generation, and topologically connected isocontour identification. Furthermore, we discuss the underlying algorithms yt uses for processing and visualizing data, and its mechanisms for parallelization of analysis tasks.
Article
Three-dimensional (3D) computed tomography (CT) with volume rendering (VR) and maximum intensity projection (MIP) is an integral component of CT interpretation for diagnosis and treatment planning of vascular pathology. Cinematic rendering (CR is a new rendering algorithm that incorporates a more advanced lighting model than that used for VR, to create photorealistic 3D CT images. The post-processing tool is not yet widely available and its ultimate clinical utility has yet to be assessed in well-designed studies. Nonetheless, the striking level of detail and enhanced depiction of 3D anatomic relationships holds potential for improvements in diagnosis, interventional or operative planning and patient centered care. In this pictorial essay, a series of complex cardiovascular cases are presented to demonstrate the enhanced display capabilities associated with cinematic rendering and the utility to guide patient management and understanding are discussed.
Article
We have developed a novel hierarchical data structure for the efficient representation of sparse, time-varying volumetric data discretized on a 3D grid. Our “VDB”, so named because it is a Volumetric, Dynamic grid that shares several characteristics with B+trees, exploits spatial coherency of time-varying data to separately and compactly encode data values and grid topology. VDB models a virtually infinite 3D index space that allows for cache-coherent and fast data access into sparse volumes of high resolution. It imposes no topology restrictions on the sparsity of the volumetric data, and it supports fast (average O(1)) random access patterns when the data are inserted, retrieved, or deleted. This is in contrast to most existing sparse volumetric data structures, which assume either static or manifold topology and require specific data access patterns to compensate for slow random access. Since the VDB data structure is fundamentally hierarchical, it also facilitates adaptive grid sampling, and the inherent acceleration structure leads to fast algorithms that are well-suited for simulations. As such, VDB has proven useful for several applications that call for large, sparse, animated volumes, for example, level set dynamics and cloud modeling. In this article, we showcase some of these algorithms and compare VDB with existing, state-of-the-art data structures.