首页|Minjiang University Details Findings in Machine Learning (Unraveling the Effects of Sodium Carbonate On Hydrothermal Liquefaction Through Individual Biomass Model Component and Machine Learning-enabled Prediction)
Minjiang University Details Findings in Machine Learning (Unraveling the Effects of Sodium Carbonate On Hydrothermal Liquefaction Through Individual Biomass Model Component and Machine Learning-enabled Prediction)
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
Elsevier
Current study results on Machine Learning have been published. According to news reporting originating in Fuzhou, People’s Republic of China, by NewsRx journalists, research stated, “Despite sodium carbonate (Na2CO3) being commonly utilized as a catalyst in actual biomass hydrothermal liquefaction (HTL), its impact on individual biomass components hasn’t been well-examined. This study thus delves into the role of Na2CO3 in HTL of biomass model components (carbohydrate, lignin, protein, and lipid) at varying conditions.” Funders for this research include Natural Science Foundation of Fujian Province, Fashu Research Foundation, Minjiang University, National Sciences and Engineering Research Council Discovery, Canada. The news reporters obtained a quote from the research from Minjiang University, “Na2CO3 at 5 wt% amplified carbohydrate degradation into biocrude and aqueous-gaseous products (AG), resonating with previous work on carbohydrate-rich feedstocks. While Na2CO3 has a marginal effect on lignin HTL, it negatively influences lipid HTL. Here, 5 wt% and 13.5 wt% Na2CO3 decrease the biocrude yield from 95.6% to less than 10%, simultaneously increasing the AG yield to approximately 90%. This is presumably due to the interaction of lipid decomposition intermediates (fatty acids) with sodium cations, resulting in water-soluble soap. For protein HTL, a low Na2CO3 concentration (5 wt%) has no significant impact on product formation, but excessive Na2CO3 (27 wt%) converts a considerable portion of biocrude into AG. In addition to these explorations that provided insights for actual biomass HTL, we also developed a machine learning model to adequately predict HTL biocrude yield by taking Na2CO3 effect into consideration. The Adaboost machine learning model displayed the most satisfactory prediction performance (training and testing R2: 0.96, 0.8) among the three investigated machine learning models. The feature importance analysis reveals lipid content and Na2CO3 concentration as pivotal factors over HTL process conditions and other biochemical components.”
FuzhouPeople’s Republic of ChinaAsiaAlkaliesAnionsCarbonatesCarbonic AcidCyborgsEmerging TechnologiesMachine LearningMinjiang University
NETL
NSTL
Elsevier
