Prediction of product yields using fusion model from Co-pyrolysis of biomass and coal

Prediction of MSW pyrolysis products based on a deep artificial neural network

Pyrolysis is a promising method for recovering resources and energy products from municipal solid waste (MSW). Predicting MSW pyrolysis products is crucial for establishing an efficient pyrolysis system for resource recovery. In this study, a database was established based on MySQL to record relevant information on MSW pyrolysis, which includes the MSW ultimate analysis results, proximate analysis results, parameters of pyrolysis operation and yields of pyrolysis products, etc. Based on the database and with help of a deep artificial neural network (ANN) which contains 10 hidden layers, a prediction model was successfully established to predict the yield of char, liquid and gas products from MSW pyrolysis. The results showed that the coefficients of determination for predicting the yields of char, liquid and gas from the MSW pyrolysis are 0.841, 0.84, and 0.85, respectively; these values demonstrate an accuracy comparable to that achieved for product prediction from single biomass, indicating a successful model performance. The results also show that ash content and temperature are the most important input factors influencing the outputs, namely, yields of char, liquid and gas. The results of this study can help to achieve a more efficient design of the pyrolysis system and improve the recovery of the desired pyrolysis products.

The effect of torrefaction temperature and catalyst loading in Microwave-Assisted in-situ catalytic Co-Pyrolysis of torrefied biomass and plastic wastes

In the current study, the effect of torrefaction temperatures (125-175 °C) and catalyst quantity (5-15 g) on co-pyrolysis of torrefied sawdust (TSD) and polystyrene (PS) are investigated to obtain value-added products. The role of torrefaction in co-pyrolysis of TSD: PS was analyzed to understand the product yields, synergy, and energy consumption . As the torrefaction temperature increases, oil yield (48.3-59.6 wt%) and char yield (24.3-29 wt%) increase while gas yield (27.4-11.4 wt%) decreases. Catalytic co-pyrolysis showed a significant level of synergy when compared to non-catalytic co-pyrolysis. For the conversion (%), a positive synergy maximum (-2.6) exists at a torrefaction temperature of 175 °C and 15 g of KOH catalyst. To develop the model, polynomial regression-based machine learning was used to predict pyrolysisproduct yields and energy usage variables. The developed models showed significant prediction accuracy (R2 > 0.98), suggesting the experimental values and the predicted values matched well.