爱荷华大学：基于云的机器学习提高异质原料向可再生天然气的转化率.pdf

1、Craig JustDonald E.Bently Professor in EngineeringIowa Wastewater and Waste to Energy Research Program(IWWERP)AVEVA World Conference 2025Cloud-based Machine Learning Solution to Increase Heterogenous Feedstock Conversion to Renewable Natural GasCivil&Environmental EngineeringIIHR Hydroscience&Engine

2、eringPatent Pending:MACHINE LEARNING-ENABLED OPTIMIZATION OF BIOGAS PRODUCTIONCivil&Environmental EngineeringIIHR Hydroscience&EngineeringThe University of Iowa and industry partners increase profitability of renewable natural gas facilitiesChallengeLack of real-time information decreased efficiency

3、 and increased downtimeLack of trend data limited operator insightLack of a pathway to a digital twin solution limited longer-term value propositionSolutionDeployed AVEVA Connect Data Service,Communities,Advanced Analytics,and Visualization to deliver real-time data and trends to facility operatorsR

4、esultsDelivered a cloud-based,machine learning solution to predict feedstock and digester composition in real-timeThe solution provides a pathway to digital twin implementationCivil&Environmental EngineeringIIHR Hydroscience&EngineeringThe June 2024 issue of Water Environment&Technology provides a d

5、igital twin“maturity spectrum”frameworkLevel 4.Two-Way Data Integration and InteractionFocus:implementing two-way data communication to enable process optimization based on data analysis and model simulationsBenefits:enables closed-loop control systems for automated process adjustments and facilitat

6、es scenario simulations for optimizing operationsWater sector adoption:limited,but pioneering projects existCivil&Environmental EngineeringIIHR Hydroscience&EngineeringThe June 2024 issue of Water Environment&Technology provides a digital twin“maturity spectrum”frameworkLevel 5.Autonomous Operations