Biomass classification by cluster analysis based on compositional data

Solid biomass waste has found various applications in the field of biorefining, which aims to obtain added value from organic industrial byproducts which would otherwise be discarded. The choice of conversion pathways best suited for a given type of biomass is dependent on its composition, since different biorefining techniques benefit from having more or less of certain components. This study aims to compile three datasets of biomass compositional data (NREL LAPs compositional data of fruit waste, proximate analysis, and ultimate analysis) and apply clustering analysis in order to classify the data into distinct groupings. The average composition of the groups can then be studied and suitable biorefining pathways can be proposed for each group. The results obtained demonstrate how these unsupervised learning methods give rise to clusters that make sense in terms of biomass categories. The groups obtained for both the NREL and proximate analysis compositional data each comprise types of biomass with high amounts of an individual component (hemicellulose, lignin+protein, cellulose, extractives, and ash for NREL LAP data; ash, volatile matter, and fixed carbon for proximate analysis data), while ultimate analysis data was grouped mainly based on its carbon to oxygen ratio. Various conversion pathways were then proposed for each group based on their average composition and current research of biorefining techniques

• Autores: Toro Delgado, Juan Andrés
• Tipo de recurso: Trabajo de grado de pregrado
• Fecha de publicación: 2020
• Institución: Universidad de los Andes
• Repositorio: Séneca: repositorio Uniandes
• Idioma: eng
• OAI Identifier: oai:repositorio.uniandes.edu.co:1992/49059
• Acceso en línea: http://hdl.handle.net/1992/49059
• Palabra clave: Biomasa; Ingeniería bioquímica; Microbiología; Análisis cluster; Aprovechamiento de residuos; Ingeniería
• Rights: openAccess
• License: http://creativecommons.org/licenses/by-nc-nd/4.0/