Optimal Variable Combinations of Co-pelletization of Eucalyptus Sawdust and Natural Forest Sawdust

Abstract

Eucalyptus sawdust is a residue from fast-growing forest processing. Utilization of Eucalyptus sawdust as fuel by co-pelletization with natural forest sawdust could solve problems of waste disposal and material supply limitation of pellet production. This work targeted at providing a useful reference for improving pellet properties and optimizing variable combinations. Experiments were arranged using response surface methodology with a central composite design and carried out using a uniaxial piston-cylinder densification apparatus. Based on the analysis of variance of effects of variables on responses, the optimal models were all selected as quadratic and expressed in the form of regression equations. The differences between adjusted R 2 and predicted R 2 are all within 0.2 and the adequate precision is all greater than four illustrating the reliability of the established models. The optimal variable combinations were obtained according to the desired response goals and validation experiments were conducted. Errors between response predicted and actual values were calculated and applied to improve the model accuracy. After modification, final errors were reduced to less than 9% and could be accepted by pellet production. Experimental results also demonstrated that pellet properties and higher heating value were all increased by co-pelletization and optimization. This will be very beneficial for pellet application. Furthermore, these results could provide the required data for designing a suitable machine of pellet production.