Unlocking waste potential: A neural network approach to forecasting sustainable acetaldehyde production from ethanol upcycling in biomass waste gasification

Abstract

The upcycling of biomass waste gasification process for ethanol and acetaldehyde production has been augmented by integrating secondary and tertiary processes. A simulation model using Aspen Plus, developed from experimental investigations to assess the viability of this proposed system. Furthermore, this model has been utilized to create an Artificial Neural Network (ANN) prediction model. Process sustainability analysis demonstrated an energy efficiency of 64 % while economic viability up to 80 % process efficiency with an Internal Rate of Return (IRR) of 6 % and a payback period of 2107 days. An optimization strategy and an artificial neural network (ANN)-based predictive model have been developed. Optimization results revealed that a gasifier temperature of around 600 °C, a gasifying agent ratio of 2.0, and an acetaldehyde reactor temperature of 300 °C yield better process outcomes and revenue. The ANN model exhibited robust performance with coefficient of determination (R2) values ranging from 0.92 to 0.98 for acetaldehyde, hydrogen, and total revenue. Mean absolute error (MAE) and mean absolute percentage error (MAPE) fell within the range of 0.03–0.1 and 0.1–1.12 %, respectively. Consequently, given the favorable sustainability and predictive model performance, this study can be used for similar endeavors.