Exposing and understanding synergistic effects in co-pyrolysis of biomass and plastic waste via machine learning. (February 2023)

• Record Type: Journal Article
• Authors: Prasertpong, Prapaporn; Onsree, Thossaporn; Khuenkaeo, Nattawut; Tippayawong, Nakorn; Lauterbach, Jochen
• Is Part Of: Bioresource technology. Volume 369(2023)
• Publication Date: 2023-02
• Subjects: AshB Ash content of biomass -- AshP Ash content of plastic -- BOY Bio-oil yields -- B:P Biomass-to-plastic ratio -- CB Carbon content of biomass -- CP Carbon content of plastic -- ClB Chlorine content of biomass -- ClP Chlorine content of plastic -- FCB Fixed carbon content of biomass -- FCP Fixed carbon content of plastic -- GF Gas conditioning flow rate -- HB Hydrogen content of biomass -- HP Hydrogen content of plastic -- HR Heating rate -- ML Machine learning -- MCB Moisture content of biomass -- MCP Moisture content of plastic -- MVB Volatile matter content of biomass -- MVP Volatile matter content of plastic -- NB Nitrogen content of biomass -- NP Nitrogen content of plastic -- OB Oxygen content of biomass -- OP Oxygen content of plastic -- SB Sulfur content of biomass -- SM Sample mass loading -- SMOTER Synthetic minority over-sampling technique for regression -- SP Sulfur content of plastic -- SS Sample size -- SE Synergistic effects -- SMOGN Gaussian noise -- SHAP Shapley additive explanation -- XGB XGBoost -- T Reaction temperature -- t Reaction time -- M Initial mass of the sample on a dry basis -- mliquid Mass of the total liquid fraction -- xi, j Value of zero and variance of unity for a feature j at a datapoint i -- x∼i, j Transformed data of a feature j at a datapoint i from the original data -- xˇi, j Original data of a feature j at a datapoint i -- x¯j Mean of transformed data of a feature j -- yBOY Bio-oil yieldss -- yBOY, B BOY value from pyrolysis experiment of the biomass -- yBOY, P BOY value from pyrolysis experiment of the plastic -- ySE Synergistic effectss -- σj Variance of transformed data of a feature j -- λ transformation parameter -- ω Biomass-to-plastic ratioss -- yBOY, individual BOY value from an individual pyrolysis experiment of the biomass and plastic -- yBOY, co-pyrolysis BOY value from the co-pyrolysis experiment of a biomass-and-plastic mixture; AI -- Bioenergy -- Mixed biomass -- Waste-to-energy -- Regression; 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.2022.128419
• Languages: English
• ISSNs: 0960-8524

Abstract

Graphical abstract: Highlights: Machine learning prediction of oil yields and synergistic effects of input variables from co-pyrolysis of biomass and plastic waste was conducted. XGBoost gave prediction accuracy with R 2 of 0.86–0.90 using 26 input features. With Shapley additive explanation, temperature and biomass-to-plastic ratio identified as the top two important features. Abstract: During co-pyrolysis of biomass with plastic waste, bio-oil yields (BOY) could be either induced or reduced significantly via synergistic effects (SE). However, investigating/ interpreting the SE and BOY in multidimensional domains is complicated and limited. This work applied XGBoost machine-learning and Shapley additive explanation (SHAP) to develop interpretable/ explainable models for predicting BOY and SE from co-pyrolysis of biomass and plastic waste using 26 input features. Imbalanced training datasets were improved by synthetic minority over-sampling technique. The prediction accuracy of XGBoost models was nearly 0.90 R 2 for BOY while greater than 0.85 R 2 for SE. By SHAP, individual impact and interaction of input features on the XGBoost models can be achieved. Although reaction temperature and biomass-to-plastic ratio were the top two important features, overall contributions of feedstock characteristics were more than 60 % in the system of co-pyrolysis. The finding provides a better understanding of co-pyrolysis and a way of further improvements.