Prediction of product distribution and bio-oil heating value of biomass fast pyrolysis

• Published in: Chemical engineering and processing Vol. 130; pp. 36 - 42
• Main Authors: Chen, Xing, Zhang, Huiyan, Song, Yao, Xiao, Rui
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.08.2018
• Subjects: Artificial neural network; Bio-oil heating value; Biomass fast pyrolysis; Product distribution; Support vector machine; Biomass fast pyrolysis; Artificial neural network; Support vector machine; Bio-oil heating value; Product distribution
• ISSN: 0255-2701; 1873-3204
• DOI: 10.1016/j.cep.2018.05.018

[Display omitted] •Product distribution and bio-oil heating value from biomass pyrolysis can be predicted using artificial intelligence models.•Nearly 200 correlated samples were collected to ensure the applicability of the models and describe biomass pyrolysis.•The link between pyrolytic products and the biomass feedstock and operating conditions was established by deep data mining. Prediction models for product distribution and bio-oil heating value of biomass pyrolysis, were established in this paper using some artificial intelligence algorithms, i.e. artificial neural network (ANN) and support vector machine (SVM). Correlated samples about biomass pyrolysis were collected as data set. The modeling results showed that both ANN and SVM models can estimate the pyrolytic product yield and bio-oil heating value successfully. In all cases, the ANN model accorded well with the experimental data in the training set while SVMs performed better in the prediction set, which indicated that SVM model has a better predicting performance in biomass pyrolysis. The established prediction model can be used as a good reference of the modeling study of biomass fast pyrolysis.