2021 Volume 19 Pages 32-38
The fermentation process is commonly employed for brewed beverages and foods e.g. amino acids, vitamins, enzymes, and antibiotics. The fundamental technology was, however, already established between the 1960s and 1970s and very few signs of progress have been made since then. The biggest reason is, such complex biological process requires precise and sensitive control which is even difficult for skilled engineers. In this paper, several experiments to evaluate the effectiveness of the bioproduction management system with artificial intelligence (AI) are described. A set of detailed-describable sensor data from cultivation batch integrates into convolutional data, which allows the AI to optimize bioproduction with appropriately formulated process control variables. The demonstration involves experimental microbial production with 5L culture tanks. The system achieved a certain degree of improvement in yield for productivity factors through prediction and control instruction by the convolutional data and the AI. We are confident with the result and the AI would become an innovator in emerging bioeconomy markets through inheriting tacit knowledge of experts and beyond them.