Modeling, simulation and optimization of biomass diffusers

Author(s)

Larissa de Souza Noel Simas Barbosa

• Document type: Doctoral Thesis
• Press: Piracicaba.
• Institution: Universidade de São Paulo (USP). Escola Superior de Agricultura Luiz de Queiroz (ESALA/BC)
• Defense date: 2019-04-05

Examining board members

Paulo Seleghim Junior; Helcio Rangel Barreto Orlande; Gonçalo Amarante Guimarães Pereira

• Advisor: Paulo Seleghim Junior; Patricia Aparecida Santiago Monteiro

Abstract

Diffusers can be used to extract sucrose from sugarcane and sugarbeet, to extract tannin from black acacia bark and to extract oil from oleaginous plants. Although they operate simply from a mechanical and chemical point of view, problems that occur in their operation can compromise the extraction efficiency. The modeling, simulation and optimization of biomass diffusers is essential to better understand such problems and propose optimized strategies for their project and operation. In this context, the main objective of the proposed thesis is to model and simulate the operation of a ten- stage biomass diffuser to later optimize it mono and multi-objectively considering the possibility of connection among all extraction stages. In the case of the mono-objective optimization, in which a fixed liquid imbibition flow rate is considered, the main objective of the optimization is to maximize the concentration of solute in the liquid that leaves the diffuser from random changes in the connectivity coefficients of the diffuser. For the multi-objective optimization, the liquid imbibition flow rate entering the diffuser is considered variable. Therefore, the optimization of the system requires the maximization of two objectives simultaneously (the maximization of the solute concentration in the liquid and minimization of the final solute concentration in the biomass) from random changes in the diffuser connectivity coefficients. In addition, since the minimization of operational problems is related to the monitoring of liquid level height in the fiber bed, different types of signals for level measurement were tested on an experimental extraction stage. From the simulation and optimization results, we can conclude that new connections among the stages of the diffuser can optimize the extraction process which runs counter to common sense that connections must be sequential. In the case of the mono-objective optimization, 82 to 89% of the fluid is directed to the next stage along the extraction stages (except for stage 6, in which this value is 100%). The recirculation of the fluid to the same stage is more pronounced in the final stages (stages 1 to 4), with the optimized diffuser having 14 to 18% of fluid recirculation for the same stage in these stages while for the remaining stages the recirculation varies between 0 and 11%. For the initial stages (stages 5, 7, 8 and 9), fluid recirculation to the previous stage is more prominent, varying from 6 to 16% while for the remaining stages it is only 0 to 2%. Stage 6, in turn, is the only one that connects only to the next stage. For the multi-objective optimization, similar results were found, with the probability of 85 to 100% of the fluid being circulated to the next stage being between 93 and 99%. Regarding recirculation to the same stage, the probability of 0 to 15% of the fluid being recirculated varies from 58 to 79% for stages 1 to 4, from 49 to 55% for the medium stages 5, 6 and 7 and from 60 to 90% for the last three stages (stages 8, 9 and 10). For the connection to the immediately preceding stage, the probability that 0 to 15% of the fluid being directed to the previous stage is greater in stages 1, 2, 8 and 9, ranging from 31 to 38%. For liquid level measurement, two different types of signals, electrical conductivity and infrared radiation were tested. The tests were first carried out only with water and then with sugarcane bagasse. Conductivity meters were stable and repeatable when tested with water. In the tests performed with sugarcane bagasse, the conductivity signal was insufficient for level measurement. For water-tested infrared meters, they have shown some instability symptoms and variability. When tested with cane, the infrared meters exhibit instability and distinct variances that depend on the position of the meter in the diffuser and, consequently, bed compactation. It can be then inferred that an increase in the signal variance can be also an indicative of increased bed compactation and, consequently, decreased bed permeability. (AU)

• FAPESP's process: 14/08402-3 - Modeling of extraction process of sucrose in industrial diffusers: studies with sugar cane and energy cane
• Grantee: Larissa de Souza Noel Simas Barbosa
• Support Opportunities: Scholarships in Brazil - Doctorate (Direct)