FES Funded ProjectsOutputs
| Title |
Category |
Date |
Authors |
| Introduction to Machine Learning: A Practical WorkshopThis full-day workshop to introduce machine learning for materials discovery and analysis was organized and offered by several members of FES.
This workshop provided an overview of machine learning applied by UofA researchers at
Engineering and Chemistry departments. The introduction to the practical application included
data processing and preparation aspects, as well as creating and running machine-learning models
on free demo version software. Participants had an opportunity to learn step-by-step how to
handle the data, use the models, and interpret the results.
This workshop was offered twice, first on August 13, 2018 and second on August 16, 2018.T12-P01 University of Alberta | Activity | 2018-08-13 | Oliynyk, A., Adutwum, L., Mar, A., "Ajay Ganesh ", "Anjana Puliyanda ", "Kaushik Sivaramakrishnan ", Gokul Sai Subraveti, Kasturi Nagesh Pai, Prasad, V. | | Machine Learning and Models: How we find optimal materials for Solar and CCS technologiesFuture Energy Systems hosted its first Interdisciplinary Lunch and Learn, Machine Learning and Models: How we find optimal materials for Solar and CCS technologies. The session brought together two research groups from different themes and faculties that had never previously had an opportunity to collaborate.T02-P02, T12-P01 University of Alberta | Activity | 2018-05-15 | Oliynyk, A., Alex Gzyl, Jan Poehls, Mar, A., Rajendran, A., Gokul Sai Subraveti, Kasturi Nagesh Pai, Prasad, V. | | Real-Time Production Optimization of Steam-Assisted-Gravity-Drainage Reservoirs Using Adaptive and Gain-Scheduled Model-Predictive Control: An Application to a Field ModelJournal article presenting two novel workflows to handle nonlinear reservoir dynamics using model predictive control.T07-C02 University of Alberta | Publication | 2019-02-01 | | | Reduced-order modelling of Pressure-swing adsorption processes for Pre-combustion CO2 captureThe threat of global warming and climate change is a major concern caused by the increase in atmospheric concentrations of greenhouse gases, mainly CO2 . Capturing CO2 from fossil-fuel based power plants is one of the means to mitigate anthropogenic CO2 emissions. Pre-combustion technology using solid adsorbents has emerged as a potential separation technique for capturing CO2 in fossil-fuel based Integrated Gasification Combined Cycle (IGCC) power plants. The design, optimization and integration of these processes is, however, complex because the processes are discrete and cyclic in nature. These processes depend on several operating conditions like times of each step, feed velocity, pressures etc. The current models that describe these processes are computationally expensive, thus, making optimization very challenging. These problems can be tackled by developing reduced-order models that significantly lowers the computational times. Further, incorporating these reduced-order models into system-level models would also simplify the computational complexities.
In the current study, reduced-order models are developed for different pressure-swing adsorption (PSA) processes. Suitable multivariate statistical analysis, such as causal analysis; multivariate regression etc. are performed based on simulation-generated data in order to understand how each input variable would impact the process outputs; and to find the set of input variables that best describe the behavior of process outputs, a key aspect towards developing reduced-order models. The PSA processes are required to meet certain regulatory targets for CO2 capture. Hence, classifier models are also developed to identify the set of operating conditions (input variables) for each process that would meet the required output targets. At the meeting, validation of reduced-order models against the detailed model and the results from statistical analysis will be presented.T02-P02 University of Alberta | Activity | 2018-10-29 | | | Bridging molecular properties to systems level indicators for adsorbent based post-combustion carbon capture using machine learningFossil fuels are an essential backbone of our existing energy infrastructure. Using fossil fuels sustainably would involve capturing and sequestering CO2 from large emitters. Existing capture technologies such as absorption are energy intensive and thus are not readily implemented. Adsorption processes which use solid sorbents have shown promise in capturing CO2 at high purities and recoveries at low energy consumptions.
Recent developments in material synthesis have made it possible to synthesize thousands of novel solid sorbents for adsorbent-based CO2 capture. Sorbent-based processes are cyclic in nature and hence challenging to simulate and optimize. They involve the solution of a stiff set of PDEs, and optimization of these nonlinear processes is a complex and time-consuming task. The key challenge is to identify solid sorbents, based on their thermodynamic properties, that can guarantee performance when used in a large-scale. In this project, thousands of novel adsorbents are filtered for the application of CO2 capture. The process-related data for the simulations of these materials are analyzed using statistical techniques such as principal component analysis (PCA) and appropriate machine learning models to find links between material properties and process performance. This study shows how system-level outputs such as energy consumption, productivity, purity, and recovery of a vacuum swing adsorption unit relate to molecular level material inputs such as isotherm properties. The results show important links that will help both material scientists and process design engineers bridge the existing gaps and to direct efforts to develop solid sorbents that have a higher probability of success.T02-P02 University of Alberta | Activity | 2018-10-31 | Kasturi Nagesh Pai, "Thomas Burns ", Gokul Sai Subraveti, "Sean Collins ", Li, Z., Prasad, V., "Tom Woo ", Rajendran, A. | | Machine Learning-Based Multiobjective Optimization of Pressure Swing AdsorptionT02-P02 University of Alberta | Publication | 2019-01-01 | | | Experimentally validated machine learning frameworks for accelerated prediction of cyclic steady state and optimization of pressure swing adsorption processesT02-P02 University of Alberta | Publication | 2020-01-01 | | | Generalized, Adsorbent-Agnostic, Artificial Neural Network Framework for Rapid Simulation, Optimization, and Adsorbent Screening of Adsorption ProcessesT02-P02 University of Alberta | Publication | 2020-01-01 | | | Practically Achievable Process Performance Limits for Pressure-Vacuum Swing Adsorption-Based Postcombustion CO2 CaptureT02-P02 University of Alberta | Publication | 2021-01-01 | | | Artificial Neural Network-Based Surrogate Models for Rapid Simulation, Optimization of Pressure Swing AdsorptionT02-P02 University of Alberta | Activity | 2020-11-20 | | | Physics-Based Neural Networks for Simulation and Synthesis of Cyclic Adsorption ProcessesT02-P02 University of Alberta | Publication | 2022-03-01 | | | Experimental validation of an adsorbent-agnostic artificial neural network (ANN ) framework for the design and optimization of cyclic adsorption processesT02-P02 University of Alberta | Publication | 2022-06-01 | | | Can a computer "learn" nonlinear chromatography?: Physics-based deep neural networks for simulation and optimization of chromatographic processesT02-P02 University of Alberta | Publication | 2022-06-01 | | | Practically Achievable Process Performance Limits for Pressure-Vacuum Swing Adsorption Based Post-Combustion CO2 CaptureT02-P02 University of Alberta | Publication | 2021-11-09 | | | Hybrid-AI Based Modelling of Pressure Swing AdsorptionT02-P02 University of Alberta | Publication | 2021-11-09 | | | Adsorbent Agnostic Machine-Assisted Adsorption Process Learning and Emulation (MAPLE) FrameworkT02-P02 University of Alberta | Publication | 2021-11-09 | | | How Can (or Why Should) Process Engineering Aid the Screening and Discovery of Solid Sorbents for CO2 Capture?T02-P02 University of Alberta | Publication | 2023-08-22 | |
|
|
