Biochemical methane potential prediction of plant biomasses: Comparing chemical composition versus near infrared methods and linear versus non-linear models. (January 2015)

• Record Type: Journal Article
• Authors: Godin, Bruno; Mayer, Frédéric; Agneessens, Richard; Gerin, Patrick; Dardenne, Pierre; Delfosse, Philippe; Delcarte, Jérôme
• Is Part Of: Bioresource technology. Volume 175(2015)
• Page Start: 382
• Page End: 390
• Publication Date: 2015-01
• Subjects: asl above sea level -- BMP biochemical methane potential -- C.-V. cross-validation -- CV coefficient of variation -- DM dry matter -- eDOM enzymatically digestible organic matter -- MedRE median standard residual error of prediction -- MLR multiple linear regression -- n number of samples -- PLS partial least square -- NIR near infrared -- R2Med coefficient of determination of prediction based on median variables -- RPDMed ratio of the median standard deviation of the variable to MedRE -- SD mean standard deviation -- SDMed median standard deviation -- SEL standard error of laboratory -- Val. validation -- VS organic matter (volatile solids) -- VST Van Soest; Anaerobic digestion -- Biogas -- Multivariate data analysis -- Chemometrics -- Prediction; Biomass -- Periodicals; Biomass energy -- Periodicals; Bioremediation -- Periodicals; Agricultural wastes -- Periodicals; Factory and trade waste -- Periodicals; Organic wastes -- Periodicals; Bioénergie -- Périodiques; Déchets agricoles -- Périodiques; Déchets industriels -- Périodiques; Déchets organiques -- Périodiques; Déchets (Combustible) -- Périodiques; 662.88
• DOI: 10.1016/j.biortech.2014.10.115
• Languages: English
• ISSNs: 0960-8524

Abstract

Highlights: Predictions based on the NIR spectrum were most reliable to estimate the BMP. NIR predictions of the BMP made by local models were reliable and quantitative. Non-linear models gave more reliable predictions than linear models. Biomass presentation form did not influence the model's prediction performances. Abstract: The reliability of different models to predict the biochemical methane potential (BMP) of various plant biomasses using a multispecies dataset was compared. The most reliable prediction models of the BMP were those based on the near infrared (NIR) spectrum compared to those based on the chemical composition. The NIR predictions of local (specific regression and non-linear) models were able to estimate quantitatively, rapidly, cheaply and easily the BMP. Such a model could be further used for biomethanation plant management and optimization. The predictions of non-linear models were more reliable compared to those of linear models. The presentation form (green-dried, silage-dried and silage-wet form) of biomasses to the NIR spectrometer did not influence the performances of the NIR prediction models. The accuracy of the BMP method should be improved to enhance further the BMP prediction models.