Built site for gh-pages

.nojekyll

@@ -1 +1 @@
-c47ebe3d
+5cc97586

GahlotLi2024SEG/paper.html

@@ -277,7 +277,7 @@ <h1 class="title">A Digital Twin for Geological Carbon Storage with Controlled I
 
 <section id="abstract" class="level2 unnumbered">
 <h2 class="unnumbered anchored" data-anchor-id="abstract">Abstract</h2>
-<p><em>We present an uncertainty-aware Digital Twin (DT) for geologic carbon storage (GCS), capable of handling multimodal time-lapse data and controlling CO<sub>2</sub> injectivity to mitigate reservoir fracturing risks. In GCS, DT represents virtual replicas of subsurface systems that incorporates real-time data and advanced generative Artificial Intelligence (genAI) techniques, including neural posterior density estimation via simulation-based inference and sequential Bayesian inference. These methods enable the effective monitoring and control of CO<sub>2</sub> storage projects, addressing challenges such as subsurface complexity, operational optimization, and risk mitigation. By integrating diverse monitoring data, e.g., geophysical well observations and imaged seismic, DT can bridge the gaps between seemingly distinct fields like geophysics and reservoir engineering. In addition, the recent advancements in genAI also facilitate DT with principled uncertainty quantification. Through recursive training and inference, DT utilizes simulated current state samples, e.g., CO<sub>2</sub> saturation, paired with corresponding geophysical field observations to train its neural networks and enable posterior sampling upon receiving new field data. However, it lacks decision-making and control capabilities, which is necessary for full DT functionality. This study aims to demonstrate how DT can inform decision-making processes to prevent risks such as cap rock fracturing during CO<sub>2</sub> storage operations.</em></p>
+<p><em>We present an uncertainty-aware Digital Twin (DT) for geologic carbon storage (GCS), capable of handling multimodal time-lapse data and controlling CO<sub>2</sub> injectivity to mitigate reservoir fracturing risks. In GCS, DT represents virtual replicas of subsurface systems that incorporate real-time data and advanced generative Artificial Intelligence (genAI) techniques, including neural posterior density estimation via simulation-based inference and sequential Bayesian inference. These methods enable the effective monitoring and control of CO<sub>2</sub> storage projects, addressing challenges such as subsurface complexity, operational optimization, and risk mitigation. By integrating diverse monitoring data, e.g., geophysical well observations and imaged seismic, DT can bridge the gaps between seemingly distinct fields like geophysics and reservoir engineering. In addition, the recent advancements in genAI also facilitate DT with principled uncertainty quantification. Through recursive training and inference, DT utilizes simulated current state samples, e.g., CO<sub>2</sub> saturation, paired with corresponding geophysical field observations to train its neural networks and enable posterior sampling upon receiving new field data. However, it lacks decision-making and control capabilities, which is necessary for full DT functionality. This study aims to demonstrate how DT can inform decision-making processes to prevent risks such as cap rock fracturing during CO<sub>2</sub> storage operations.</em></p>
 </section>
 <section id="introduction" class="level2">
 <h2 class="anchored" data-anchor-id="introduction">Introduction</h2>
@@ -297,7 +297,7 @@ <h2 class="anchored" data-anchor-id="conclusion-and-discussion">Conclusion and d
 <div id="fig-flow" class="quarto-figure quarto-figure-center quarto-float anchored">
 <figure class="quarto-float quarto-float-fig figure">
 <div aria-describedby="fig-flow-caption-0ceaefa1-69ba-4598-a22c-09a6ac19f8ca">
-<a href="./figs/Figure1.png" class="lightbox" data-glightbox="description: .lightbox-desc-1" data-gallery="quarto-lightbox-gallery-1"><img src="./figs/Figure1.png" class="img-fluid figure-img" style="width:90.0%"></a>
+<a href="./figs/Figure1.png" class="lightbox" data-gallery="quarto-lightbox-gallery-1" data-glightbox="description: .lightbox-desc-1"><img src="./figs/Figure1.png" class="img-fluid figure-img" style="width:90.0%"></a>
 </div>
 <figcaption class="quarto-float-caption-bottom quarto-float-caption quarto-float-fig" id="fig-flow-caption-0ceaefa1-69ba-4598-a22c-09a6ac19f8ca">
 Figure&nbsp;1: Pressure state at time-step <span class="math inline">\(k=3\)</span> (second row) and a comparative analysis of pressure outputs (first and third) from the digital twin at time-step <span class="math inline">\(k=4\)</span>
@@ -307,7 +307,7 @@ <h2 class="anchored" data-anchor-id="conclusion-and-discussion">Conclusion and d
 <div id="fig-fracture" class="quarto-figure quarto-figure-center quarto-float anchored">
 <figure class="quarto-float quarto-float-fig figure">
 <div aria-describedby="fig-fracture-caption-0ceaefa1-69ba-4598-a22c-09a6ac19f8ca">
-<a href="./figs/Figure2.png" class="lightbox" data-glightbox="description: .lightbox-desc-2" data-gallery="quarto-lightbox-gallery-2"><img src="./figs/Figure2.png" class="img-fluid figure-img" style="width:90.0%"></a>
+<a href="./figs/Figure2.png" class="lightbox" data-gallery="quarto-lightbox-gallery-2" data-glightbox="description: .lightbox-desc-2"><img src="./figs/Figure2.png" class="img-fluid figure-img" style="width:90.0%"></a>
 </div>
 <figcaption class="quarto-float-caption-bottom quarto-float-caption quarto-float-fig" id="fig-fracture-caption-0ceaefa1-69ba-4598-a22c-09a6ac19f8ca">
 Figure&nbsp;2: Injection rate samples and KDE curve(a), CDF curve of fracture probability versus injection rate with 95% confidence interval(b)
@@ -927,7 +927,7 @@ <h2 class="anchored" data-anchor-id="references">References</h2>
     </div>
   </div>
 </footer>
-<script>var lightboxQuarto = GLightbox({"selector":".lightbox","openEffect":"zoom","loop":false,"closeEffect":"zoom","descPosition":"bottom"});
+<script>var lightboxQuarto = GLightbox({"closeEffect":"zoom","openEffect":"zoom","selector":".lightbox","loop":false,"descPosition":"bottom"});
 window.onload = () => {
   lightboxQuarto.on('slide_before_load', (data) => {
     const { slideIndex, slideNode, slideConfig, player, trigger } = data;

Rex2024SEG/paper.html

@@ -288,7 +288,7 @@ <h2 class="anchored" data-anchor-id="methodology">Methodology</h2>
 <div id="fig-unsup" class="quarto-figure quarto-figure-center quarto-float anchored">
 <figure class="quarto-float quarto-float-fig figure">
 <div aria-describedby="fig-unsup-caption-0ceaefa1-69ba-4598-a22c-09a6ac19f8ca">
-<a href="figures/workflow.001.jpeg" class="lightbox" data-glightbox="description: .lightbox-desc-1" data-gallery="quarto-lightbox-gallery-1"><img src="figures/workflow.001.jpeg" class="img-fluid figure-img" style="width:90.0%"></a>
+<a href="figures/workflow.001.jpeg" class="lightbox" data-gallery="quarto-lightbox-gallery-1" data-glightbox="description: .lightbox-desc-1"><img src="figures/workflow.001.jpeg" class="img-fluid figure-img" style="width:90.0%"></a>
 </div>
 <figcaption class="quarto-float-caption-bottom quarto-float-caption quarto-float-fig" id="fig-unsup-caption-0ceaefa1-69ba-4598-a22c-09a6ac19f8ca">
 Figure&nbsp;1: Our initial velocity model (a) and it’s CIG (b), Set of Smoothed velocity models used as training inputs (c), Migrated CIGs from smoothed velocity models as training targets (d), Surrogate that learns the mapping between initial velocity models, their CIGs, and target models, to produce the corresponding CIGs for the target velocity models (e)
@@ -298,8 +298,8 @@ <h2 class="anchored" data-anchor-id="methodology">Methodology</h2>
 <div id="fig-sup" class="quarto-figure quarto-figure-center quarto-float anchored">
 <figure class="quarto-float quarto-float-fig figure">
 <div aria-describedby="fig-sup-caption-0ceaefa1-69ba-4598-a22c-09a6ac19f8ca">
-<p><a href="figures/valid_sample=8.png" class="lightbox" data-glightbox="description: .lightbox-desc-2" data-gallery="quarto-lightbox-gallery-2"><img src="figures/valid_sample=8.png" class="img-fluid figure-img"></a></p>
-<p><a href="figures/valid_sample=6.png" class="lightbox" data-glightbox="description: .lightbox-desc-3" data-gallery="quarto-lightbox-gallery-3"><img src="figures/valid_sample=6.png" class="img-fluid figure-img"></a></p>
+<p><a href="figures/valid_sample=8.png" class="lightbox" data-gallery="quarto-lightbox-gallery-2" data-glightbox="description: .lightbox-desc-2"><img src="figures/valid_sample=8.png" class="img-fluid figure-img"></a></p>
+<p><a href="figures/valid_sample=6.png" class="lightbox" data-gallery="quarto-lightbox-gallery-3" data-glightbox="description: .lightbox-desc-3"><img src="figures/valid_sample=6.png" class="img-fluid figure-img"></a></p>
 </div>
 <figcaption class="quarto-float-caption-bottom quarto-float-caption quarto-float-fig" id="fig-sup-caption-0ceaefa1-69ba-4598-a22c-09a6ac19f8ca">
 Figure&nbsp;2: Column 1 represents the target CIG of the smoothed velocity models. Column 2 represents the predicted CIG from our FNO. The last column shows the amplified absolute difference between the first 2 columns
@@ -916,7 +916,7 @@ <h3 class="anchored" data-anchor-id="references">References</h3>
     </div>
   </div>
 </footer>
-<script>var lightboxQuarto = GLightbox({"loop":false,"closeEffect":"zoom","descPosition":"bottom","selector":".lightbox","openEffect":"zoom"});
+<script>var lightboxQuarto = GLightbox({"selector":".lightbox","descPosition":"bottom","closeEffect":"zoom","loop":false,"openEffect":"zoom"});
 window.onload = () => {
   lightboxQuarto.on('slide_before_load', (data) => {
     const { slideIndex, slideNode, slideConfig, player, trigger } = data;

erdinc2024SEG/abstract.html

@@ -330,7 +330,7 @@ <h2 class="anchored" data-anchor-id="methodology">Methodology</h2>
 <div id="fig-sup" class="quarto-figure quarto-figure-center quarto-float anchored">
 <figure class="quarto-float quarto-float-fig figure">
 <div aria-describedby="fig-sup-caption-0ceaefa1-69ba-4598-a22c-09a6ac19f8ca">
-<a href="figures/IMAGE2024_figure1_new.png" class="lightbox" data-gallery="quarto-lightbox-gallery-1" data-glightbox="description: .lightbox-desc-1"><img src="figures/IMAGE2024_figure1_new.png" class="img-fluid figure-img"></a>
+<a href="figures/IMAGE2024_figure1_new.png" class="lightbox" data-glightbox="description: .lightbox-desc-1" data-gallery="quarto-lightbox-gallery-1"><img src="figures/IMAGE2024_figure1_new.png" class="img-fluid figure-img"></a>
 </div>
 <figcaption class="quarto-float-caption-bottom quarto-float-caption quarto-float-fig" id="fig-sup-caption-0ceaefa1-69ba-4598-a22c-09a6ac19f8ca">
 Figure&nbsp;1: Training-loss formulation: field observations (RTMs and wells) along with column-wise masks (<span class="math inline">\(\mathbf{A}\)</span>) are sampled initially. Conditioned on these masks, further sub-sampling masks (<span class="math inline">\(\mathbf{\widetilde{A}}\)</span>) are generated. The neural denoising model <span class="math inline">\(h_\theta\)</span>, informed by the sub-sampling masks, noisy sub-sampled well information and RTMs, is trained to reconstruct the complete, noise-free velocity model.
@@ -344,7 +344,7 @@ <h2 class="anchored" data-anchor-id="results">Results</h2>
 <div id="fig-unsup" class="quarto-figure quarto-figure-center quarto-float anchored">
 <figure class="quarto-float quarto-float-fig figure">
 <div aria-describedby="fig-unsup-caption-0ceaefa1-69ba-4598-a22c-09a6ac19f8ca">
-<a href="figures/IMAGE2024_figure2_new.png" class="lightbox" data-gallery="quarto-lightbox-gallery-2" data-glightbox="description: .lightbox-desc-2"><img src="figures/IMAGE2024_figure2_new.png" class="img-fluid figure-img"></a>
+<a href="figures/IMAGE2024_figure2_new.png" class="lightbox" data-glightbox="description: .lightbox-desc-2" data-gallery="quarto-lightbox-gallery-2"><img src="figures/IMAGE2024_figure2_new.png" class="img-fluid figure-img"></a>
 </div>
 <figcaption class="quarto-float-caption-bottom quarto-float-caption quarto-float-fig" id="fig-unsup-caption-0ceaefa1-69ba-4598-a22c-09a6ac19f8ca">
 Figure&nbsp;2: The denoising generative model is tested on two previously unseen seismic images. For each case, we fed the corresponding RTM into the network and generated velocity samples (200 samples for each example). Subsequently, we computed both the posterior mean and variance. To assess the accuracy of our model, we included the SSIM between the ground-truth to the posterior means
@@ -969,7 +969,7 @@ <h2 class="anchored" data-anchor-id="references">References</h2>
     </div>
   </div>
 </footer>
-<script>var lightboxQuarto = GLightbox({"closeEffect":"zoom","loop":false,"descPosition":"bottom","openEffect":"zoom","selector":".lightbox"});
+<script>var lightboxQuarto = GLightbox({"selector":".lightbox","descPosition":"bottom","closeEffect":"zoom","loop":false,"openEffect":"zoom"});
 window.onload = () => {
   lightboxQuarto.on('slide_before_load', (data) => {
     const { slideIndex, slideNode, slideConfig, player, trigger } = data;

search.json

@@ -4,7 +4,7 @@
     "href": "GahlotLi2024SEG/paper.html",
     "title": "A Digital Twin for Geological Carbon Storage with Controlled Injectivity",
     "section": "",
-    "text": "We present an uncertainty-aware Digital Twin (DT) for geologic carbon storage (GCS), capable of handling multimodal time-lapse data and controlling CO2 injectivity to mitigate reservoir fracturing risks. In GCS, DT represents virtual replicas of subsurface systems that incorporates real-time data and advanced generative Artificial Intelligence (genAI) techniques, including neural posterior density estimation via simulation-based inference and sequential Bayesian inference. These methods enable the effective monitoring and control of CO2 storage projects, addressing challenges such as subsurface complexity, operational optimization, and risk mitigation. By integrating diverse monitoring data, e.g., geophysical well observations and imaged seismic, DT can bridge the gaps between seemingly distinct fields like geophysics and reservoir engineering. In addition, the recent advancements in genAI also facilitate DT with principled uncertainty quantification. Through recursive training and inference, DT utilizes simulated current state samples, e.g., CO2 saturation, paired with corresponding geophysical field observations to train its neural networks and enable posterior sampling upon receiving new field data. However, it lacks decision-making and control capabilities, which is necessary for full DT functionality. This study aims to demonstrate how DT can inform decision-making processes to prevent risks such as cap rock fracturing during CO2 storage operations.",
+    "text": "We present an uncertainty-aware Digital Twin (DT) for geologic carbon storage (GCS), capable of handling multimodal time-lapse data and controlling CO2 injectivity to mitigate reservoir fracturing risks. In GCS, DT represents virtual replicas of subsurface systems that incorporate real-time data and advanced generative Artificial Intelligence (genAI) techniques, including neural posterior density estimation via simulation-based inference and sequential Bayesian inference. These methods enable the effective monitoring and control of CO2 storage projects, addressing challenges such as subsurface complexity, operational optimization, and risk mitigation. By integrating diverse monitoring data, e.g., geophysical well observations and imaged seismic, DT can bridge the gaps between seemingly distinct fields like geophysics and reservoir engineering. In addition, the recent advancements in genAI also facilitate DT with principled uncertainty quantification. Through recursive training and inference, DT utilizes simulated current state samples, e.g., CO2 saturation, paired with corresponding geophysical field observations to train its neural networks and enable posterior sampling upon receiving new field data. However, it lacks decision-making and control capabilities, which is necessary for full DT functionality. This study aims to demonstrate how DT can inform decision-making processes to prevent risks such as cap rock fracturing during CO2 storage operations.",
     "crumbs": [
       "Home",
       "Abstracts",
@@ -16,7 +16,7 @@
     "href": "GahlotLi2024SEG/paper.html#abstract",
     "title": "A Digital Twin for Geological Carbon Storage with Controlled Injectivity",
     "section": "",
-    "text": "We present an uncertainty-aware Digital Twin (DT) for geologic carbon storage (GCS), capable of handling multimodal time-lapse data and controlling CO2 injectivity to mitigate reservoir fracturing risks. In GCS, DT represents virtual replicas of subsurface systems that incorporates real-time data and advanced generative Artificial Intelligence (genAI) techniques, including neural posterior density estimation via simulation-based inference and sequential Bayesian inference. These methods enable the effective monitoring and control of CO2 storage projects, addressing challenges such as subsurface complexity, operational optimization, and risk mitigation. By integrating diverse monitoring data, e.g., geophysical well observations and imaged seismic, DT can bridge the gaps between seemingly distinct fields like geophysics and reservoir engineering. In addition, the recent advancements in genAI also facilitate DT with principled uncertainty quantification. Through recursive training and inference, DT utilizes simulated current state samples, e.g., CO2 saturation, paired with corresponding geophysical field observations to train its neural networks and enable posterior sampling upon receiving new field data. However, it lacks decision-making and control capabilities, which is necessary for full DT functionality. This study aims to demonstrate how DT can inform decision-making processes to prevent risks such as cap rock fracturing during CO2 storage operations.",
+    "text": "We present an uncertainty-aware Digital Twin (DT) for geologic carbon storage (GCS), capable of handling multimodal time-lapse data and controlling CO2 injectivity to mitigate reservoir fracturing risks. In GCS, DT represents virtual replicas of subsurface systems that incorporate real-time data and advanced generative Artificial Intelligence (genAI) techniques, including neural posterior density estimation via simulation-based inference and sequential Bayesian inference. These methods enable the effective monitoring and control of CO2 storage projects, addressing challenges such as subsurface complexity, operational optimization, and risk mitigation. By integrating diverse monitoring data, e.g., geophysical well observations and imaged seismic, DT can bridge the gaps between seemingly distinct fields like geophysics and reservoir engineering. In addition, the recent advancements in genAI also facilitate DT with principled uncertainty quantification. Through recursive training and inference, DT utilizes simulated current state samples, e.g., CO2 saturation, paired with corresponding geophysical field observations to train its neural networks and enable posterior sampling upon receiving new field data. However, it lacks decision-making and control capabilities, which is necessary for full DT functionality. This study aims to demonstrate how DT can inform decision-making processes to prevent risks such as cap rock fracturing during CO2 storage operations.",
     "crumbs": [
       "Home",
       "Abstracts",