Point Source Capture

AI/ML Approaches for Research and Development Application

Approaches
Use machine learning approaches based on known labels/expectations, usually for
  • Classification (e.g., Logistic, Support Vector Machine)
  • Regression (e.g., Linear, Kernel Ridge)
  • Collect data on diverse PSC systems
  • Explore grid-market interactions using cost models.
  • Discovery of materials and predicting properties of solvents & sorbents
  • Clarify capture levels & net emissions
  • Validate PSC technologies for other applications
  • Identify site qualifications to inform screening.
  • Predict PSC system applicability based on the nature of an industry operation
  • Integrate waste heat to reduce PSC energy use
  • Fully integrate industrial PSC processes to effectively use low-carbon fuels and feedstocks
  • Predict feedstock impacts on product properties
  • Run integrated controls to meet purity requirements
  • Mange diverse flows/content from varied sources
  • Track the quantity & quality of CO2 and flag any issues
  • Predict the water and resource usage to avoid operation disruption
  • Structured numeric data, text, image, maps, sequences, time-series data
  • Reduce Capture Costs from dilute streams
  • Expand capabilities in process modeling, TEA & LCA
  • Reduce and validate PSC performance and cost
  • Scale up technology for low-carbon power, materials & supply chain
  • Adapt power plant PSC technology for industry
  • Reduce process energy intensity and emissions
  • Product low-CO2 product/construction materials
  • Ensure quality and minimum purity of captured CO2
  • Water and resource usage management
Approaches
Use machine learning approaches without known labels/expectations, usually for
  • Clustering
  • Dimension reduction
  • Collect data on diverse PSC systems
  • Optimize emissions in stead and non-steady states
  • Increase process intensification to improve carbon capture efficiency
  • Identify site qualifications to inform screening.
  • Guide operations to cut CO2 and improve the product.
  • Structured or unstructured data
  • Reduce Capture Costs from dilute streams
  • Expand capabilities in prcoess modeling, TEA & LCA
  • Reduce and validate PSC performance and cost
  • Scale up technology for low-carbon power, materials & supply chain
  • Product low-CO2 product/construction materials
Approaches
Use complex neural network architecture for
  • Image and spatial classification
  • Time series regression
  • Natural language processing
  • Explore grid-market interactions using cost models.
  • Discovery of materials and predicting properties of solvents & sorbents
  • Mange diverse flows/content from varied sources.
  • Predict PSC system applicability based on the nature of an industry operation
  • Predict the water and resource usage to avoid operation disruption
  • Structured numeric data, text, image, maps, sequences, time-series data
  • Reduce Capture Costs from dilute streams
  • Reduce and validate PSC performance and cost
  • Adapt power plant PSC technology for industry
  • Reduce process energy intensity and emissions
  • Water and resource usage management
Approaches
Combine more than one machine learning models to handle the tasks
  • Random Forest
  • XGBoost
  • Committee Machine
  • Collect data on diverse PSC systems
  • Fill critical knowledge gaps in the complex interplay among mateirals, processing, properties & performance.
  • Conduct technoeconomic Analyses using cost data from FEED studies
  • Optimize emissions in steady and non-steady states
  • Develop optimization models for PSC control systems.
  • Increase process intensification to improve carbon capture efficiency
  • Validate PSC technologies for other applications
  • Provide a control strategy to optimize overall PSC system operations.
  • Assess real-time data on fuels and feedstocks
  • Predict the water and resource usage to avoid operation disruption
  • Structured numeric data, text, image, maps, sequences, time-series data
  • Reduce Capture Costs from dilute streams
  • Expand capabilities in prcoess modeling, TEA & LCA
  • Reduce and validate PSC performance and cost
  • Reduce process energey intensity and emissions
  • Ensure quality and minimum purity of captured CO2
  • Water and resource usage management
Approaches

Integrate scientific equations and knowledge into machine learning algorithms (usually using Deep Learning), such as Physics-Informed Machine Learning

  • Discovery of materials and predicting properties of solvents & sorbents.
  • Validate PSC technologies for other applications
  • Provide a control strategy to optimize overall PSC system operations
  • Predict the water and resource usage to avoid operation disruption
  • Structured numeric data, text, image, maps, sequences, time-series data
  • Scientific equations, field equations, partial differential equations
  • Reduce Capture Costs from dilute streams
  • Reduce and validate PSC performance and cost
  • Water and resource usage management
Large Language Models for UNSTRUCTURED DATA MINING Large Language Models for GENERATIVE DISCOVERY Transfer Learning Science-based Learning Edge Computing Advance Sensor Reinforcement Learning Operator Learning
Characterize Materials
Develop Structural/Functional Materials
Enable Smart Components w/embedded sensors
Optimize regional Planning and Operations
Modeling/Simulation Bridging Scales
Accelerate Flow Sheet Initialization
Optimize Turbine Operations
Assess/Control H2 Combustion Characteristics