TY - JOUR
T1 - Efficient Catalytic Conversion of Waste Peanut Shells into Liquid Biofuel
T2 - An Artificial Intelligence Approach
AU - Li, Pan
AU - Du, Zeji
AU - Chang, Chun
AU - Zhao, Shiqiang
AU - Xu, Guizhuan
AU - Xu, Chunbao Charles
PY - 2020/2/20
Y1 - 2020/2/20
N2 - Artificial intelligence approach can be used to solve complicated process problems. A hybrid methodology comprising artificial neural network (ANN) and genetic algorithms (GA) was utilized to model and optimize the methanolysis process of waste peanut shells. Acid catalytic methanolysis of waste peanut shells into liquid biofuel-methyl levulinate (ML) was investigated. The combination of sulfuric acid with extremely low concentration and Al2(SO4)3 was identified as the efficient mixed acid catalytic system. The ML yield under optimal conditions optimized by response surface methodology (RSM) was 16.49%, while the ML yield optimized by ANN-GA was 17.61%. The results showed that ANN-GA had a higher optimizing ability than the RSM model. Meanwhile, the methanolysis kinetics provided insights into the reaction routes for ML production. Moreover, Al2(SO4)3 can be recycled and reused five times without much decrease of the ML yield. This study suggested that waste peanut shells can be used as potential raw materials for ML production and ANN-GA can serve as a powerful tool for biofuel processing technology. © 2020 American Chemical Society.
AB - Artificial intelligence approach can be used to solve complicated process problems. A hybrid methodology comprising artificial neural network (ANN) and genetic algorithms (GA) was utilized to model and optimize the methanolysis process of waste peanut shells. Acid catalytic methanolysis of waste peanut shells into liquid biofuel-methyl levulinate (ML) was investigated. The combination of sulfuric acid with extremely low concentration and Al2(SO4)3 was identified as the efficient mixed acid catalytic system. The ML yield under optimal conditions optimized by response surface methodology (RSM) was 16.49%, while the ML yield optimized by ANN-GA was 17.61%. The results showed that ANN-GA had a higher optimizing ability than the RSM model. Meanwhile, the methanolysis kinetics provided insights into the reaction routes for ML production. Moreover, Al2(SO4)3 can be recycled and reused five times without much decrease of the ML yield. This study suggested that waste peanut shells can be used as potential raw materials for ML production and ANN-GA can serve as a powerful tool for biofuel processing technology. © 2020 American Chemical Society.
UR - http://www.scopus.com/inward/record.url?scp=85080905040&partnerID=8YFLogxK
UR - https://www.scopus.com/record/pubmetrics.uri?eid=2-s2.0-85080905040&origin=recordpage
U2 - 10.1021/acs.energyfuels.9b03433
DO - 10.1021/acs.energyfuels.9b03433
M3 - RGC 21 - Publication in refereed journal
SN - 0887-0624
VL - 34
SP - 1791
EP - 1801
JO - Energy and Fuels
JF - Energy and Fuels
IS - 2
ER -
