Probabilistic uncertainty analysis based on Monte Carlo simulations of co-combustion of hazelnut hull and coal blends: Data-driven modeling and response surface optimization

Abstract

The aim of present study is to investigate the thermogravimetric behaviour of the co-combustion of hazelnut hull (HH) and coal blends using three approaches: multi non-linear regression (MNLR) modeling based on Box–Behnken design (BBD) (1), optimization based on response surface methodology (RSM) (2), and probabilistic uncertainty analysis based on Monte Carlo simulation as a function of blend ratio, heating rate, and temperature (3). The response variable was predicted by the best-fit MNLR model with a predicted regression coefficient (R2pred) of 99.5%. Blend ratio of 90/10 (HH to coal, %wt), temperature of 405°C, and heating rate of 44°Cmin−1 were determined as RSM-optimized conditions with a mass loss of 87.4%. The validation experiments with three replications were performed for justifying the predicted-mass loss percentage and 87.5%±0.2 of mass loss were obtained under RSM-optimized conditions. The probabilistic uncertainty analysis were performed by using Monte Carlo simulations.