OPTIMIZATION OF BIOGAS REFORMING PROCESSES USING HYBRID DEEP LEARNING ALGORITHMS: PREDICTION OF OUTPUT PARAMETERS WITH CNN-LSTM MODEL

Yıl 2024, Cilt: 11 Sayı: 23, 301 - 316, 31.08.2024

Saadin Oyucu , Münür Sacit Herdem

https://doi.org/10.54365/adyumbd.1488710

Öz

This study examines the application of a hybrid deep learning algorithm, combining Convolutional Neural Networks (CNN) with Long Short-Term Memory (LSTM), for predicting various output parameters in biogas reforming processes. The objective is to develop predictive models that enhance the management of these processes. The CNN-LSTM model was selected for its proficiency in capturing long-term dependencies and complex features in time-series data, and it was benchmarked against other models, including Support Vector Regression (SVR). This research evaluates crucial outputs of biogas reforming, such as the methane conversion rate, the hydrogen-to-carbon monoxide ratio, and the synthesis gas composition. The effectiveness of the CNN-LSTM model was assessed using RMSE, MAE, and MAPE metrics. After different training epochs, the RMSE for CONMET (%) was recorded at 0.1905, MAE at 0.1311, and MAPE at 0.0036, demonstrating the model's high accuracy in prediction. This study marks a significant advancement in incorporating machine learning techniques into optimizing and controlling biogas reforming processes for industrial applications. The success of the CNN-LSTM model, particularly in managing complex biochemical processes, underscores the potential of deep learning techniques. Future efforts will explore the model's application across different biogas plants and aim to refine optimization parameters further.

Anahtar Kelimeler

Biogas Reforming, CNN-LSTM, SVR, Sustainable Energy

HİBRİT DERİN ÖĞRENME ALGORİTMALARI KULLANILARAK BİYOGAZ REFORM SÜREÇLERİNİN OPTİMİZASYONU: CNN-LSTM MODELİ İLE ÇIKTI PARAMETRELERİNİN TAHMİNİ

Öz

Bu çalışma, biyogaz reform süreçlerinde çeşitli çıktı parametrelerini tahmin etmek için Evrişimli Sinir Ağları (CNN: Convolutional Neural Networks) ve Uzun Kısa Süreli Bellek (LSTM: Long Short-Term Memory) algoritmalarını birleştiren hibrit bir derin öğrenme modelinin uygulamasını incelemektedir. Çalışmanın amacı, bu süreçlerin yönetimini iyileştiren tahmine dayalı modeller geliştirmektir. CNN-LSTM modeli, zaman serisi verilerindeki uzun vadeli bağımlılıkları ve karmaşık özellikleri yakalama konusundaki yeterliliği nedeniyle seçilmiş ve Destek Vektör Regresyonu (SVR: Support Vector Regression) gibi diğer modellerle karşılaştırılmıştır. Araştırma metan dönüşüm oranı, hidrojen-karbon monoksit oranı ve sentez gazı bileşimi gibi biyogaz reformunun önemli çıktılarını değerlendirmektedir. Modelin etkinliği RMSE, MAE ve MAPE ölçümleri kullanılarak değerlendirilmiştir. Farklı eğitim dönemleri sonrasında, metan dönüşüm oranı için RMSE 0,1905, MAE 0,1311 ve MAPE 0,0036 olarak kaydedilmiştir. Elde edilen sonuçlar, modelin tahmin başarısındaki yüksek doğruluğu ortaya koymaktadır. Bu çalışma, makine öğrenimi tekniklerinin endüstriyel uygulamalarda biyogaz reform süreçlerinin optimize edilmesi ve kontrol edilmesine katkı sağlayabileceğini göstermektedir. CNN-LSTM modelinin özellikle karmaşık biyokimyasal süreçleri yönetmedeki başarısı, derin öğrenme tekniklerinin potansiyelini vurgulamaktadır. Gelecekteki çalışmalar, modelin farklı biyogaz tesislerinde uygulanmasını ve optimizasyon parametrelerinin daha da iyileştirilmesini amaçlayacaktır.

Anahtar Kelimeler

Biyogaz Reformlama, CNN-LSTM, SVR, Sürdürülebilir Enerji

Kaynakça

• Kougias PG, Angelidaki I. Biogas and its opportunities - A review Keywords. Front. Environ. Sci. 2018; 12(June):1-22.
• Abanades S. A conceptual review of sustainable electrical power generation from biogas. Energy Sci. Eng. 2022; 10(2):630-655, doi: 10.1002/ese3.1030.
• Phan TS. Hydrogen production from biogas : Process optimization using ASPEN Plus. International Journal of Hydrogen Energy; 47(100): 42027-42039. 2022. HAL Id : hal-03563223.
• Vita A, Italiano C, Previtali D, Fabiano C, Palella A, Freni F, Bozzano G, Pino L, Manenti F. Methanol synthesis from biogas: A thermodynamic analysis. Renew. Energy 2018; 118: 673-684, doi: 10.1016/j.renene.2017.11.029.
• da Silva Pinto RL, Vieira AC, Scarpetta A, Marques FS, Jorge RMM, Bail A, Jorge LMM, Corazza ML, Ramos LP. An overview on the production of synthetic fuels from biogas. Bioresour. Technol. Reports, 2022; 18(1): 101104. doi: https://doi.org/10.1016/j.biteb.2022.101104.
• Minutillo M, Perna A, Sorce A. Green hydrogen production plants via biogas steam and autothermal reforming processes: energy and exergy analyses. Appl. Energy, 2020; 277: 115452. doi: https://doi.org/10.1016/j.apenergy.2020.115452.
• Chein RY, Chen WH, Chyuan Ong H, Loke Show P, Singh Y. Analysis of methanol synthesis using CO2 hydrogenation and syngas produced from biogas-based reforming processes. Chem. Eng. J. 2021; 426:130835. doi: https://doi.org/10.1016/j.cej.2021.130835.
• Marchese M, Giglio E, Santarelli M, Lanzini A. Energy performance of Power-to-Liquid applications integrating biogas upgrading, reverse water gas shift, solid oxide electrolysis and Fischer-Tropsch technologies. Energy Convers. Manag. 2020; 6: 100041. doi: https://doi.org/10.1016/j.ecmx.2020.100041.
• Al-Wahaibi A, Osman AI, Al-Muhtaseb AAH, Alqaisi O, Baawain M, Fawzy S, Rooney DW. Techno-economic evaluation of biogas production from food waste via anaerobic digestion. Sci. Rep. 2020; 10(1):1-16. doi: 10.1038/s41598-020-72897-5.
• Roy PS, Song J, Kim K, Park CS, Raju ASK.NCO2 conversion to syngas through the steam-biogas reforming process. J. CO2 Util. 2018; 25:275-282. doi: https://doi.org/10.1016/j.jcou.2018.04.013.
• Zhao X, Joseph B, Kuhn J, Ozcan S. Biogas Reforming to Syngas: A Review. iScience, 2020; 23(5): 101082. doi: https://doi.org/10.1016/j.isci.2020.101082.
• Lee J, Hong S, Cho H, Lyu B, Kim M, Kim J, Moon I. Machine learning-based energy optimization for on-site SMR hydrogen production. Energy Convers. Manag. 2021; 244: 114438. doi: https://doi.org/10.1016/j.enconman.2021.114438.
• Byun M, Lee H, Choe C, Cheon S, Lim H. Machine learning based predictive model for methanol steam reforming with technical, environmental, and economic perspectives. Chem. Eng. J. 2021; 426: 131639. doi: https://doi.org/10.1016/j.cej.2021.131639.
• Abdelkareem MA, Soudan B, Mahmoud MS, Sayed ET, AlMallahi MN, Inayat A, Olabi AG. Progress of artificial neural networks applications in hydrogen production. Chem. Eng. Res. Des. 2022; 182: 66-86. doi: https://doi.org/10.1016/j.cherd.2022.03.030.
• Irie K, Tüske Z, Alkhouli T, Schlüter R, Ney H. LSTM, GRU, highway and a bit of attention: An empirical overview for language modeling in speech recognition. Annu. Conf. Int. Speech Commun. Assoc. 08-12-Sept, pp. 3519–3523, 2016, doi: 10.21437/Interspeech.2016-491.
• Chen Y, Fu Z. Multi-Step Ahead Forecasting of the Energy Consumed by the Residential and Commercial Sectors in the United States Based on a Hybrid CNN-BiLSTM Model. Sustain. 2023;15(3):1-15. doi: 10.3390/su15031895.
• Chicco D, Warrens MJ, Jurman G, The coefficient of determination R-squared is more informative than SMAPE, MAE, MAPE, MSE and RMSE in regression analysis evaluation. PeerJ Comput. Sci. 2021; 7:1-24. doi: 10.7717/PEERJ-CS.623.
• Barupal DK, Fiehn O. Generating the blood exposome database using a comprehensive text mining and database fusion approach. Environ. Health Perspect. 2019; 127(9): 2825-2830. doi: 10.1289/EHP4713.
• Khair U, Fahmi H, Al Hakim S, Rahim R. Forecasting Error Calculation with Mean Absolute Deviation and Mean Absolute Percentage Error. J. Phys. Conf. Ser. 2017; 930(1):1-9. doi: 10.1088/1742-6596/930/1/012002.
• De Myttenaere A, Golden B, Le Grand B, Rossi F. Mean Absolute Percentage Error for regression models. Neurocomputing 2016; 192: 38-48. doi: 10.1016/j.neucom.2015.12.114.