Enabling immersive engagement in energy system models with deep learning

Bugbee, Bruce; Bush, Brian; Gruchalla, Kenny; Potter, Kristin; Brunhart-Lupo, Nicholas; Krishnan, Venkat

doi:10.1002/sam.11419

Cited by 9 publications

(4 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We also employ tools for web-browserbased 2D and immersive 3D exploration of the highdimensional features of BSM output (Figures 8 and 9). 31,32 The 3D tools pack more information (scenarios, data points, and dimensions) into the visualization than the 2D tools typically do, and the hands-on, immersive creation and exploration of simulation results accelerate discovery. These visualization-oriented analysis tools support several workflows.…”

Section: Discussionmentioning

confidence: 99%

Simulation process and data flow for a large system dynamics model

et al. 2022

View full text Add to dashboard Cite

This paper documents the workflow and supporting technologies that a large system dynamics model, the biomass scenario model, employs to streamline the data preparation, simulation, quality control, and analysis process at the National Renewable Energy Laboratory. The workflow centers on automation of routine aspects of the flow of data between data stores, simulations, and visualizations. It enforces quality checks on data, reproducibility of computations, and traceability of results, while maintaining complete archives of modeling and analysis artifacts. The resulting frictionless simulation/analysis environment supports large-scale sensitivity analysis, interactive creation of ensembles of simulations, and rapid visualization-based exploration of simulation results.

show abstract

Section: Discussionmentioning

confidence: 99%

Simulation process and data flow for a large system dynamics model

et al. 2022

View full text Add to dashboard Cite

show abstract

“…The method does not explicitly handle time series: in the examples, we just use the final time step of the model for the output variables. More sophisticated methods like tensor decomposition (Bugbee et al, 2019) or functional principal components analysis (Ramsay and Silverman, 2013) would appropriately handle the multivariate time series that comprise the output in simulation models, but the local-sensitivity method would remain the same after such preprocessing was applied.…”

Section: Discussionmentioning

confidence: 99%

Adaptively Sampling via Regional Variance-Based Sensitivities

Bush,

Wendelberger,

Hanes

2021

Preprint

Self Cite

View full text Add to dashboard Cite

Inspired by the well-established variance-based methods for global sensitivity analysis, we develop a local total sensitivity index that decomposes the global total sensitivity conditions by independent variables' values. We employ this local sensitivity index in a new method of experimental design that sequentially and adaptively samples the domain of a multivariate function according to local contributions to the global variance. The method is demonstrated on a nonlinear illustrative example that has a three-dimensional domain and a three-dimensional codomain, but also on a complex, high-dimensional simulation for assessing the industrial viability of the production of bioproducts from biomass.

show abstract

“…190−192 Machine learning models can also be of help here; these models can help guide the user in finding optimal visualization configurations and ease parameter searches and can be used in tandem with immersive visualization. 193 Another approach is to help bring more minds together to attack a single problem. Collaborative visualization, where multiple researchers jointly analyze a data set (in both the quantitative and qualitative), has been an area of research since the early 1990s.…”

Section: ■ Mesoscale Modeling Methodsmentioning

confidence: 99%

“…To reduce complexity (or make the dynamics perceptible), we may have to go beyond the monitor. Immersive visualization and virtual reality have demonstrated benefits for some classes of problems. − Machine learning models can also be of help here; these models can help guide the user in finding optimal visualization configurations and ease parameter searches and can be used in tandem with immersive visualization . Another approach is to help bring more minds together to attack a single problem.…”

Section: Mesoscale Modeling Methodsmentioning

confidence: 99%

Bridging Scales in Bioenergy and Catalysis: A Review of Mesoscale Modeling Applications, Methods, and Future Directions

et al. 2021

View full text Add to dashboard Cite

Between the molecular and reactor scales, which are familiar to the chemical engineering community, lies an intermediate regime, here termed the “mesoscale,” where transport phenomena and reaction kinetics compete on similar time scales. Bioenergy and catalytic processes offer particularly important examples of mesoscale phenomena owing to their multiphase nature and the complex, highly variable porosity characteristic of biomass and many structured catalysts. In this review, we overview applications and methods central to mesoscale modeling as they apply to reaction engineering of biomass conversion and catalytic processing. A brief historical perspective is offered to put recent advances in context. Applications of mesoscale modeling are described, and several specific examples from biomass pyrolysis and catalytic upgrading of bioderived intermediates are highlighted. Methods including reduced order modeling, finite element and finite volume approaches, geometry construction and import, and visualization of simulation results are described; in each category, recent advances, current limitations, and areas for future development are presented. Owing to improved access to high-performance computational resources, advances in algorithm development, and sustained interest in reaction engineering to sustainably meet societal needs, we conclude that a significant upsurge in mesoscale modeling capabilities is on the horizon that will accelerate design, deployment, and optimization of new bioenergy and catalytic technologies.

show abstract

Enabling immersive engagement in energy system models with deep learning

Cited by 9 publications

References 31 publications

Simulation process and data flow for a large system dynamics model

Simulation process and data flow for a large system dynamics model

Adaptively Sampling via Regional Variance-Based Sensitivities

Bridging Scales in Bioenergy and Catalysis: A Review of Mesoscale Modeling Applications, Methods, and Future Directions

Contact Info

Product

Resources

About