Final edits to the paper

paper/paper.bib

@@ -100,59 +100,6 @@ @misc{interactive_Jupyter_widgets
     howpublished={\url{https://github.com/jupyter-widgets/ipywidgets}}, 
     year={2010}
 }
-%%%%%%%%%%%%%%%%%%%%%%
-%
-%%%%%%%%%%%%%%%%%%%%%%
-@article{Pearson:2017,
-    url = {http://adsabs.harvard.edu/abs/2017arXiv170304627P},
-    Archiveprefix = {arXiv},
-    Author = {{Pearson}, S. and {Price-Whelan}, A.~M. and {Johnston}, K.~V.},
-    Eprint = {1703.04627},
-    Journal = {ArXiv e-prints},
-    Keywords = {Astrophysics - Astrophysics of Galaxies},
-    Month = mar,
-    Title = {{Gaps in Globular Cluster Streams: Pal 5 and the Galactic Bar}},
-    Year = 2017
-}
-
-@book{Binney:2008,
-    url = {http://adsabs.harvard.edu/abs/2008gady.book.....B},
-    Author = {{Binney}, J. and {Tremaine}, S.},
-    Booktitle = {Galactic Dynamics: Second Edition, by James Binney and Scott Tremaine.~ISBN 978-0-691-13026-2 (HB).~Published by Princeton University Press, Princeton, NJ USA, 2008.},
-    Publisher = {Princeton University Press},
-    Title = {{Galactic Dynamics: Second Edition}},
-    Year = 2008
-}
-
-@article{gaia,
-    author = {{Gaia Collaboration}},
-    title = "{The Gaia mission}",
-    journal = {Astronomy and Astrophysics},
-    archivePrefix = "arXiv",
-    eprint = {1609.04153},
-    primaryClass = "astro-ph.IM",
-    keywords = {space vehicles: instruments, Galaxy: structure, astrometry, parallaxes, proper motions, telescopes},
-    year = 2016,
-    month = nov,
-    volume = 595,
-    doi = {10.1051/0004-6361/201629272},
-    url = {http://adsabs.harvard.edu/abs/2016A%26A...595A...1G},
-}
-
-@article{astropy,
-    author = {{Astropy Collaboration}},
-    title = "{Astropy: A community Python package for astronomy}",
-    journal = {Astronomy and Astrophysics},
-    archivePrefix = "arXiv",
-    eprint = {1307.6212},
-    primaryClass = "astro-ph.IM",
-    keywords = {methods: data analysis, methods: miscellaneous, virtual observatory tools},
-    year = 2013,
-    month = oct,
-    volume = 558,
-    doi = {10.1051/0004-6361/201322068},
-    url = {http://adsabs.harvard.edu/abs/2013A%26A...558A..33A}
-}
 
 
 @article{Ragauskas2014,
@@ -195,5 +142,6 @@ @article{begum2024review
   author={Begum, Yasmin Ara and Kumari, Sheetal and Jain, Shailendra Kumar and Garg, Manoj Chandra},
   journal={Environmental Science: Advances},
   year={2024},
-  publisher={Royal Society of Chemistry}
+  publisher={Royal Society of Chemistry},
+  doi={10.1039/D4VA00109E}
 }

paper/paper.md

@@ -39,18 +39,18 @@ bibliography: paper.bib
 `Virtual Engineering` (`VE`) is a Python software framework designed to accelerate the research and development of engineering processes that are fundamentally defined by multiple unit operations executed in series. 
 VE supports a wide variety of different multi-physics models and joins them to simulate an entire end-to-end process. To automate the execution of this model sequence, `VE` provides (i) a robust method to communicate between models, 
 (ii) a high-level, user-friendly interface to set model parameters and enable optimization, and (iii) an overall model-agnostic approach that allows new computational units to be swapped in and out of workflows. 
-<!-- Although the `VE` approach was developed to support the process of the low-temperature conversion of biomass to fuel, we have designed each component to easily accommodate new domains and unit models. -->
-Although the VE framework currently supports the biochemical conversion of biomass-to-fuel, we have designed each component to easily accommodate new domains and unit models.
+Although the `VE` framework was developed to support the biochemical conversion of biomass to fuel, we have designed each component to easily accommodate new domains and unit models.
+<!-- Although the VE framework currently supports the biochemical conversion of biomass to fuel, we have designed each component to easily accommodate new domains and unit models. -->
 
 # Statement of need
 
 Many industrial and manufacturing operations consist of a sequence of discrete processing steps, including physicochemical transformations, to produce a final product. Often, optimizing the performance of each individual step—such as yield or energy efficiency—does not lead to the overall best outcome. Therefore, it is essential to ensure connectivity between each step and optimize the whole process. This process optimization can depend not only on operating parameters for each step but also on the choice and order of these steps.
 
-Numerical simulations that can support the analysis and optimization of such systems frequently require linking multiple individual models together—each associated with different steps in the overall operational process—so that the outputs of one model can inform the inputs of the next. These models can span multiple levels of physical fidelity and computational costs. Virtual Engineering (`VE`) is a Python package that enables the creation of this type of model sequence. 
+Numerical simulations that can support the analysis and optimization of such systems frequently require linking multiple individual models together—each associated with different steps in the overall operational process—so that the outputs of one model can inform the inputs of the next. These models can span multiple levels of physical fidelity and computational costs. `Virtual Engineering` (`VE`) is a Python package that enables the creation of this type of model sequence. 
 
-`VE` was originally developed to support the simulation and optimization of the biochemical conversion of lignocellulosic biomass to fuels. This bioconversion process was modeled by linking previously developed computational models as each of three important unit operations (Figure \ref{fig:VE_diagram}):  (i) the _pretreatment_ of the feedstock to make cellulose more accessible [@sitaraman_multiphysics_2015], (ii) an _enzymatic hydrolysis_ step to digest lingnocellulose into sugars [@sitaraman_coupled_2019;@lischeske2019two], and (iii) a _bioconversion_ step to convert sugars into products in a bioreactor [@rahimi_computational_2018].  Finally, the capital and operating costs of the process, and the product's (in this case, ethanol) subsequent minimum feasible selling price through discounted cash flow analysis, were calculated using an Aspen Plus process simulation with techno-economic analysis (TEA) [@humbird2011process]. 
+`VE` was originally developed to support the simulation and optimization of the biochemical conversion of lignocellulosic biomass to fuels. This bioconversion process was modeled by linking previously developed computational models as three important unit operations (Figure \ref{fig:VE_diagram}):  (i) the _pretreatment_ of the feedstock to make cellulose more accessible [@sitaraman_multiphysics_2015], (ii) an _enzymatic hydrolysis_ step to digest lingnocellulose into sugars [@sitaraman_coupled_2019;@lischeske2019two], and (iii) a _bioconversion_ step to convert sugars into products in a bioreactor [@rahimi_computational_2018].  Finally, the capital and operating costs of the process, and the product's (in this case, ethanol) subsequent minimum feasible selling price through discounted cash flow analysis, were calculated using an Aspen Plus process simulation with techno-economic analysis (TEA) [@humbird2011process]. 
 
-![An example of the end-to-end processes for converting biomass to fuel at low temperature defined and automated within `VE` by connecting different unit models.\label{fig:VE_diagram}](figs/VE_diagram.pdf){ width=100% }
+![An example of the end-to-end process for biochemical convertion of biomass to fuel defined and automated within `VE` by connecting different unit models.\label{fig:VE_diagram}](figs/VE_diagram.pdf){ width=100% }
 
 ![Example of `ipywidgets` element defining the choices of EH unit model and its input parameters.\label{fig:controls}](figs/eh_widgets.png){ width=100% }
 
@@ -71,6 +71,8 @@ The `VE` package provides an optimization capability that can be easily accessed
 
 To demonstrate these capabilities of the `VE` package, we optimize the oxygen uptake rate (OUR) that determine fuel yields in the bioconversion process (Figure \ref{fig:VE_diagram}), by solving the box-constrained optimization problem with two controls summarized in Table \ref{tab:opt_problem}. It is important to note that the objective function, OUR, is an output of the bioreactor (BR) model, while acid loading and enzyme loading are independent inputs to the pretreatment (PT) and enzymatic hydrolysis (EH) models, respectively, which indirectly affect overall fuel yields. 
 
+
+
 : Optimization problem with objective and control variables in different unit operations.\label{tab:opt_problem}
 
 |            |                 | Lower bound| Upper bound| Units |
@@ -90,13 +92,13 @@ We solve this optimization problem for two different modeled feedstocks: switchg
 
 To accelerate the optimization process, we developed surrogate models for the computationally expensive unit operations of EH and BR, utilizing Gaussian process regression and leveraging dimension reduction and active importance sampling. As a result, one can obtain predictions of EH and BR outputs within seconds.
 
-Figure \ref{fig:opt_results} illustrates the optimization process for these two different feedstocks. The contour plot of OUR is obtained by sweeping through the parameter space and serves as the background to visualize the behavior of the objective function. This contour plot verifies that our optimization algorithm follows the gradients as expected and converges to a reasonable final solution.
+Figure \ref{fig:opt_results} illustrates the optimization process for these two different feedstocks. The contour plots of OUR are obtained by sweeping through the parameter space and serves as the background to visualize the behavior of the objective function. These contour plots verify that our optimization algorithm follows the gradients as expected and converges to a reasonable final solution.
 
 To demonstrate the robustness of the optimization algorithm, we used different initial values for the control variables in the switchgrass and corn stover cases. Table \ref{tab:opt_results} displays the initial and final values of the controls and the objective, as well as the change in the OUR.
 
-![](figs/opt_acid_enz_fs_1_new.png){ width=95% }
+![](figs/opt_acid_enz_fs_1_new.png){ width=93% }
 
-![The result of optimizing the amount of acid loading during PT and enzyme loading during EH to maximize OUR for a modeled switchgrass (top) and corn stover (bottom) feedstock; the triangle indicates the initial control values, the square the optimal control values, and the connecting line the path of the optimization algorithm as it identifies the maximum.\label{fig:opt_results}](figs/opt_acid_enz_fs_2_new.png){ width=95% }
+![The result of optimizing the amount of acid loading during PT and enzyme loading during EH to maximize OUR for a modeled switchgrass (top) and corn stover (bottom) feedstock. The triangles represent the initial control values, the squares indicate the optimized values, and the lines connecting them show the algorithm's paths to the maximum.\label{fig:opt_results}](figs/opt_acid_enz_fs_2_new.png){ width=93% }
 
 : Optimization results\label{tab:opt_results}
 
@@ -116,7 +118,9 @@ One interesting outcome is the difference in the optimal amount of enzyme loadin
 
 The `VE` package was used to quantify the effects of modifying the enzyme loading and total enzymatic hydrolysis processing time on the minimum fuel selling price and the optimization features were used to identify the porosity of the initial feedstock necessary to maximize the OUR. These results, plus a live demonstration of the `VE` notebook usage, were presented at the 2021 American Institute of Chemical Engineers (AIChE) Annual Meeting [@young_aiche_2021]. A more comprehensive analaysis---including details on the surrogate modeling methodology, model validation against experimental results, and multi-variable optimization outcomes to maximize OUR---was presented at the 2023 Bioenergy Technologies Office (BETO) peer review [@young_beto_2023].
 
-Finally, we reiterate that, while developed for the specific biomass conversion process discussed above, the `VE` framework is comprised of generalizable components that can readily extend to model a wide array of industrial and manufacturing workflows. An immediate example for future extension is on thermochemical conversion pathways such as biomass/plastics/solid-waste pyrolysis or gasification that involves similar chemically reacting flows except in a gas-solid environment. `VE` framework also provides a pathway to combine biochemical and thermochemical approaches and provide overall optimal process conditions, in line with current research directions on hybrid integrated approaches [@begum2024review]. As another example, we may consider the processing of iron ore for steelmaking. This industrial process generally includes multiple stages of grinding, separation, treatment, and some form of pelletizing. Each of the processes can be performed in different ways---e.g., grinding may be performed with bar or ball grinders or using high-pressure grinding rollers that all have different pros and cons as well as different operational parameters. The `VE` framework could be used to study different grinding methods within the context of a broader iron ore processing operation to optimize different objectives, such as yield, cost, or energy usage. Similar design process questions can arise in applications such as pharmaceutical production, microelectronics fabrication, agricultural processing, and more.
+Finally, we reiterate that, while developed for the specific biomass conversion process discussed above, the `VE` framework is comprised of generalizable components that can readily extend to model a wide array of industrial and manufacturing workflows. An immediate example for future extension is on thermochemical conversion pathways such as biomass/plastics/solid-waste pyrolysis or gasification that involves similar chemically reacting flows except in a gas-solid environment. `VE` framework also provides a pathway to combine biochemical and thermochemical approaches and provide overall optimal process conditions, in line with current research directions on hybrid integrated approaches [@begum2024review]. 
+
+As another example, we may consider the processing of iron ore for steelmaking. This industrial process generally includes multiple stages of grinding, separation, treatment, and some form of pelletizing. Each of the processes can be performed in different ways---e.g., grinding may be performed with bar or ball grinders or using high-pressure grinding rollers that all have different pros and cons as well as different operational parameters. The `VE` framework could be used to study different grinding methods within the context of a broader iron ore processing operation to optimize different objectives, such as yield, cost, or energy usage. Similar design process questions can arise in applications such as pharmaceutical production, microelectronics fabrication, agricultural processing, and more.
 
 <!-- # Example of optimizing the design of bioreactor