Bioethanol production optimization through machine learning algorithm approach: biomass characteristics, saccharification, and fermentation conditions for enzymatic hydrolysis

In this study, the optimized decision-making system (OD-MS) algorithm in machine learning for optimizing the enzymatic hydrolysis saccharification, fermentation conditions, and its yield was studied. Two hundred fifty datasets were collected as training data from various studies and experimental res...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Biomass conversion and biorefinery Ročník 13; číslo 8; s. 7287 - 7299
Hlavní autoři: Vinitha, Nithianantharaj, Vasudevan, Jaikumar, Gopinath, Kannappan Panchamoorthy
Médium: Journal Article
Jazyk:angličtina
Vydáno: Berlin/Heidelberg Springer Berlin Heidelberg 01.06.2023
Témata:
Biotechnology
Energy
Original Article
Renewable and Green Energy
Glucose
Support vector machine
Ethanol yield
Optimized decision-making system algorithm
Machine learning algorithm
ISSN:2190-6815, 2190-6823
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:In this study, the optimized decision-making system (OD-MS) algorithm in machine learning for optimizing the enzymatic hydrolysis saccharification, fermentation conditions, and its yield was studied. Two hundred fifty datasets were collected as training data from various studies and experimental results. Initially, the data of various biomass and product conditions were collected, and their correlation coefficient values were determined using the Pearson correlation matrix. Test datasets were analyzed in an optimized decision-making system to predict the data value, and the process was repeated until the desired data was achieved. 3-D surface analysis was performed to determine the product yield ranges based on biomass characteristics and process conditions. Maximum glucose yield (> 50 g/L) and ethanol yield (> 40 g/L) were achieved with the increase in cellulose (> 73%), S-temp (55–60 °C), S-pH (7–9), S-shaking speed (180–200 rpm), F-pH (4.5), F-time (< 40 h), and F-shaking speed (120–150 rpm) and decrease in hemicellulose (< 10%), lignin (< 10%), S-time (< 20 h), and F-time (20 h). Weighted rank was assigned to biomass characteristics and process conditions to find the optimum parameter conditions using correlation values for obtaining a better yield. Two-step validation was done to find the accuracy of biomass characteristics and process conditions. A 95% of accuracy was found while comparing the actual dataset with the data predicted using the OD-MS algorithm for various biomass characteristics and process conditions. The R 2 value of this model was found to be 0.9762.
ISSN:2190-6815
2190-6823
DOI:10.1007/s13399-022-03163-z