Disentangled Representation Aided Physics-Informed Neural Network for Predicting Syngas Compositions of Biomass Gasification DOI
Shaojun Ren, Shiliang Wu,

Qihang Weng

et al.

Energy & Fuels, Journal Year: 2024, Volume and Issue: 38(3), P. 2033 - 2045

Published: Jan. 9, 2024

Machine learning (ML) has been extensively studied and applied in the biomass gasification field currently. However, insufficient experimental data tends to cause a mismatch between ML model physical mechanism, particularly for feedstocks that do not appear training set, becoming significant challenge creating credible models gasification. Therefore, this study proposes disentangled representation-aided physics-informed neural network method, briefly called DR-PINN, predict syngas components. First, DR-PINN extracts latent variables represent feedstock properties through representation generates synthetic samples gasification-related variable space cover full range of types. Then, employs inequality constraints embed priori monotonic relationships into loss function. Finally, are simultaneously considered process realize synergy complementarity actual information existing knowledge using an evolutionary algorithm. As result, shows good prediction performance (the within set: R2 ≈ 0.96, root-mean-square error (RMSE) 1.7; outside 0.81, RMSE 3). Moreover, even with can strictly abide by prior relationships, consistency degree equal 1. Overall, proposed demonstrates superior generalization interpretability compared other methods, such as RF, GBR, SVM, ANN, PINN.

Language: Английский

Catalytic pyrolysis of corncob with Ni/CaO dual functional catalysts for hydrogen-rich gas DOI
Hongyu Liu, Yuting Tang, Xiaoqian Ma

et al.

Journal of the Taiwan Institute of Chemical Engineers, Journal Year: 2023, Volume and Issue: 150, P. 105059 - 105059

Published: July 22, 2023

Language: Английский

Citations

13
Continuous biohydrogen and volatile fatty acids production from cheese whey in a tubular biofilm reactor: Substrate flow rate variations and microbial dynamics DOI Creative Commons
Omprakash Sarkar, Ulrika Rova, Paul Christakopoulos

et al.

International Journal of Hydrogen Energy, Journal Year: 2024, Volume and Issue: 59, P. 1305 - 1316

Published: Feb. 15, 2024

The present study demonstrates the influence of substrate flow rate on continuous biohydrogen and volatile fatty acids production from acidogenic fermentation cheese whey in a tubular biofilm reactor. Three bioreactors with varied (2 mL/min, 5 8 mL/min) were examined for 75 days. At mL/min rate, evolution was higher (3.88 mL H2/h), while its conversion efficiency lower compared to 2 rate. formation ammonium also influenced by rates. slightly at (12.74 ± 2.42 gCOD/L) (18.09 2.01 while, decreasing (11.85 0.78 gCOD/L). Substrate significantly affected pattern composition showing acetic acid, butyric propionic acid 4.72 1.46 gCOD/L 10.41 0.91 (5 1.78 0.13 mL/min). Continuous input maintained pH reactor due replacement fresh substrate, thereby controlling feedback inhibition boosting metabolite production. Hydrogen-producing Firmicutes confirmed pivotal role microbial community's significant contribution converting waste bioenergy. Overall, results support use operation mode large-scale However, ensure efficacy system using or wastewater, low rates are recommended.

Language: Английский

Citations

5
Brewery spent grain valorization through fermentation: Targeting biohydrogen, carboxylic acids and methane production DOI Creative Commons

Jacobo Pérez-Barragán,

Cristina Martínez-Fraile, Raúl Muñoz

et al.

Process Safety and Environmental Protection, Journal Year: 2024, Volume and Issue: 191, P. 206 - 217

Published: Aug. 30, 2024

This study investigated three different fermentation approaches to explore the potential for producing biohydrogen, carboxylic acids, and methane from hydrolysates of thermally dilute acid pretreated brewer's spent grains (BSG). Initially, research focused on maximizing volumetric hydrogen production rate (HPR) in continuous dark (DF) BSG by varying hydraulic retention time (HRT). The highest HPR reported date 5.9 NL/L-d was achieved at 6 h HRT, with a Clostridium-dominated microbial community. effect operational pH (4, 5, 6, 7) acidogenic then investigated. A peak concentration 17.3 g CODequiv./L recorded an associated productivity 900.5 ± 13.1 mg CODequiv./L-h degree acidification 68.3 %. Lactic bacteria such as Limosilactobacillus Lactobacillus were dominant 4–5, while Weissella, Enterococcus, Lachnoclostridium appeared 7. Finally, this evaluated biochemical DF broth unfermented found high yields 659 517 NmL CH4/g-VSadded, respectively, both within one week. Overall, results showed that can be low-cost feedstock bioenergy valuable bio-based chemicals circular economy.

Language: Английский

Citations

5
Comprehensive Review of Carbon Capture and Storage Integration in Hydrogen Production: Opportunities, Challenges, and Future Perspectives DOI Creative Commons
Seyed Mehdi Alizadeh, Yasin Khalili, Mohammad Ahmadi

et al.

Energies, Journal Year: 2024, Volume and Issue: 17(21), P. 5330 - 5330

Published: Oct. 26, 2024

The growing emphasis on renewable energy highlights hydrogen’s potential as a clean carrier. However, traditional hydrogen production methods contribute significantly to carbon emissions. This review examines the integration of capture and storage (CCS) technologies with processes, focusing their ability mitigate It evaluates various techniques, including steam methane reforming, electrolysis, biomass gasification, discusses how CCS can enhance environmental sustainability. Key challenges, such economic, technical, regulatory obstacles, are analyzed. Case studies future trends offer insights into feasibility CCS–hydrogen integration, providing pathways for reducing greenhouse gases facilitating transition.

Language: Английский

Citations

5
Disentangled Representation Aided Physics-Informed Neural Network for Predicting Syngas Compositions of Biomass Gasification DOI
Shaojun Ren, Shiliang Wu,

Qihang Weng

et al.

Energy & Fuels, Journal Year: 2024, Volume and Issue: 38(3), P. 2033 - 2045

Published: Jan. 9, 2024

Machine learning (ML) has been extensively studied and applied in the biomass gasification field currently. However, insufficient experimental data tends to cause a mismatch between ML model physical mechanism, particularly for feedstocks that do not appear training set, becoming significant challenge creating credible models gasification. Therefore, this study proposes disentangled representation-aided physics-informed neural network method, briefly called DR-PINN, predict syngas components. First, DR-PINN extracts latent variables represent feedstock properties through representation generates synthetic samples gasification-related variable space cover full range of types. Then, employs inequality constraints embed priori monotonic relationships into loss function. Finally, are simultaneously considered process realize synergy complementarity actual information existing knowledge using an evolutionary algorithm. As result, shows good prediction performance (the within set: R2 ≈ 0.96, root-mean-square error (RMSE) 1.7; outside 0.81, RMSE 3). Moreover, even with can strictly abide by prior relationships, consistency degree equal 1. Overall, proposed demonstrates superior generalization interpretability compared other methods, such as RF, GBR, SVM, ANN, PINN.

Language: Английский

Citations

4