Avaliação da viabilidade do uso de redes neurais artificiais para o desenvolvimento de um softsensor de biomassa de microalgas em fotobiorreator

• Autor(a) principal: Palma, Guilherme Meneghetti
• Data de Publicação: 2022
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional da UFSCAR
• Texto Completo: https://repositorio.ufscar.br/handle/ufscar/15895
• topic: Cultivo; Interpolação; Otimização; Cultivation; Interpolation; Optimization; Airlift; Scenedesmus obliquus; ENGENHARIAS::ENGENHARIA QUIMICA::PROCESSOS INDUSTRIAIS DE ENGENHARIA QUIMICA

Microalgae are natural sources of biomass, unsaturated fatty acids, carotenoids, xanthophylls, vitamins, proteins, minerals and enzymes; compounds of great commercial importance. Furthermore, since biofuels have become a product of great interest to industry, especially in the energy sector, the production of microalgae has come to be widely explored by scientific institutions and private organizations. In addition to being one of the most photosynthetic efficient organisms, it requires a very small cultivation space when produced in bioreactors. Thus, the optimization of the microalgae production process in photobioreactors is of great economic interest. Currently, the measurement of biomass values and cell concentration inside reactors is one of the major limitations for the optimization and control of the process, after all, it occurs in a non-automated way by cell counting and dry mass determination methods. Therefore, the objective of this project was to evaluate the feasibility of using artificial neural networks to create a biomass instantaneous inference softsensor inside a bioreactor from light intensity data obtained by red, green and blue light sensors connected to a microcontroller. The study was carried out by cultivating the microalgae Scenedesmus obliquus in modified BG-11 culture medium in a 6L Airlift-type photobioreactor with illumination provided from a white LED panel. The content presented in this study showed artificial neural networks with MSE = 0,0278 [mg/L]2, R2=0,93 and provides substantial information for the accomplishment of the control and optimization of microalgal production inside bioreactors based only on light intensity information.