首页|New Findings on Machine Learning from Earth and Environmental Sciences Division Summarized (Efficient Prediction of Hydrogen Storage Performance In Depleted Gas Reservoirs Using Machine Learning)
New Findings on Machine Learning from Earth and Environmental Sciences Division Summarized (Efficient Prediction of Hydrogen Storage Performance In Depleted Gas Reservoirs Using Machine Learning)
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
NSTL
By a News Reporter-Staff News Editor at Robotics & Machine Learning DailyNews Daily News - Data detailed on Machine Learning have been presented. According to news reportingfrom Los Alamos, New Mexico, by New sRx journalists, research stated, “Underground hydrogen (H2)storage (UHS) has e merged as a promising technology to facilitate the widespread adoption of fluctu atingrenewable energy sources. However, the current UHS experience primarily fo cuses on salt caverns, withno working examples of storing pure H2 in porous res ervoirs.”Funders for this research include Los Alamos National Laboratory, Laboratory Dir ected Research andDevelopment (LDRD) program.The news correspondents obtained a quote from the research from Earth and Enviro nmental SciencesDivision, “A key challenge in UHS within porous rocks is the un certainty in evaluating storage performancedue to complicated geological and op erational conditions. While physics -based reservoir simulationsare commonly us ed to quantify H2 injection and withdrawal processes during storage cycles, they arecomputationally demanding and unsuitable for providing rapid support to UHS operations. In this study,we develop efficient reduced -order models (ROMs) fo r UHS in depleted natural gas reservoirs using deepneural networks (DNNs) based on comprehensive reservoir simulation data sets. The ROMs can accuratelyforeca st UHS performance metrics (H2 withdrawal efficiency, produced H2 purity, produc ed gas -waterratio) across various geological and operational conditions and ar e over 22000 times faster than physics-based simulations. Then, we employ the R OMs for sensitivity analysis to assess the impact of geologicaland operational parameters on UHS performance and conduct uncertainty quantification to characte rizepotential performance and associated probabilities. Lastly, we present a fi eld case study from the Dakotaformation of the Basin field in the Intermountain -West (I -WEST) region, USA. Based on the ROMs’predictions, Dakota formation i s favorable for UHS due to its high H2 withdrawal efficiency and purity,and low water production risk. By optimizing operational parameters, we can further imp rove the storageperformance in Dakota formation and reduce the uncertainty in U HS performance prediction.”
Los AlamosNew MexicoUnited StatesNorth and Central AmericaCyborgsElementsEmerging TechnologiesGasesHydrogenInorganic ChemicalsMachine LearningEarth and Environmental Sciences Division
NETL
NSTL
