Development platform of analytical sensors, based on artificial intelligence, applied to industrial biotechnological processes: monitoring of alcoholic fermentation

• Grant number: 19/09080-3
• Support Opportunities: Research Grants - Innovative Research in Small Business - PIPE
• Start date: February 01, 2021
• End date: January 31, 2022
• Field of knowledge: Engineering - Chemical Engineering - Chemical Process Industries
• Principal Investigator: Moisés Alves
• Grantee: Moisés Alves
• Company: Bioprocess Improvement Consultoria e Pesquisas em Bioprocessos Ltda
• CNAE: Fabricação de malte, cervejas e chopes ; Fabricação de álcool ; Fabricação de biocombustíveis, exceto álcool
• City: Campinas
• Pesquisadores principais: Daniel Ibraim Pires Atala
• Associated researchers: Elmer Alberto Ccopa Rivera
• Associated scholarship(s): 21/00667-1 - Development platform of analytical sensors, based on artificial intelligence, applied to industrial biotechnological processes: monitoring of alcoholic fermentation, BP.PIPE

Abstract

Currently, bioprocesses play a key role in the industry. A bottleneck in monitoring and controlling bioprocesses is often caused by the lack of reliable sensors, particularly for biotechnological variables. A solution to this problem can be found through the design of software sensors, algorithms used to infer variables of difficult measure from simple and low-cost measures. The objective of this project is the development and experimental evaluation of robust software sensor that allows real-time inference of cell, substrate and product concentrations from secondary measures such as pH, temperature, turbidity and CO2 flow for alcoholic fermentation processes on basis the process of batch fermentation. For that, it will be perfomed the elaboration of a database of the variables of the batch fermentation process by means of ten bench tests at different temperatures. LabVIEW® software will be used for data collection of secondary variables and chromatographic analysis to collect the main data. The database and artificial intelligence techniques through artificial neural network (RNA) models will be used for the construction of software sensors that describe the kinetics of the process to meet changing in the fermentation conditions. RNAs will be characterized by three main aspects; architecture, training algorithm and activation function. Finally, the validation of the sensor software will be carried out by means of new tests of the fermentation process, comparing the prediction of the virtual sensors with the results obtained from conventional chromatographic techniques. It is intended, therefore, to obtain an efficient monitoring of the alcoholic fermentation process in real time that allows to make adjustments in the conditions of operation, control and to keep the process operating in the ideal conditions. (AU)