Parametric study and global sensitivity analysis for co-pyrolysis of rape straw and waste tire via variance-based decomposition

Author:  Xu, L., Jiang, Y., Qiu, R.

Journal: Bioresource Technology

DOI:  10.1016/j.biortech.2017.09.141

Keywords: Co-pyrolysis, Rape straw, Data-driven modelling, Global sensitivity analysis, response-surface methodology, municipal solid-waste, thermal-decomposition, process optimization, bayesian-approach, solar pyrolysis, tyre pyrolysis

Abstract: 

In present study, co-pyrolysis behavior of rape straw, waste tire and their various blends were investigated. TGFTIR indicated that co-pyrolysis was characterized by a four-step reaction, and H2O, C-H, O-H, CO2 and C=O groups were the main products evolved during the process. Additionally, using BBD-based experimental results, best-fit multiple regression models with high R-2-pred values (94.10% for mass loss and 95.37% for reaction heat), which correlated explanatory variables with the responses, were presented. The derived models were analyzed by ANOVA at 95% confidence interval, F-test, lack-of-fit test and residues normal probability plots implied the models described well the experimental data. Finally, the model uncertainties as well as the interactive effect of these parameters were studied, the total-, first- and second-order sensitivity indices of operating factors were proposed using Sobol' variance decomposition. To the authors' knowledge, this is the first time global parameter sensitivity analysis has been performed in (co-) pyrolysis literature.